Bi.

lTISH LIBRARY
LENDING DIVISION

2 6 OCT:?77

RODYNA I(] 83 4 0
BY

DANIEL BERNOULLI

&
HYDRA ULICS
BY

JOHANN BERNOULL~

TRANSLATED FROM THE LATIN BY

THOMAS CARlvfODY and HELMUT KOBUS
PREFACE BY
HUNTER ROUSE
Under the Auspices of the
Iowa Institute of Hydraulic Research

k. DOVER PCBLICATIONS, INC .

NEW YORK
 Copyright© 1968 by Dover Publications, Inc.
All rights reserved under Pan American and
International Copyright Conventions.

Published in Canada by General Publishing Company, Ltd.,

30 Lesmill Road. Don :'.\1ills, Toronto, Ontario.
Published in the United Kingdom by Constable and Company, Ltd.,
10 Orange Street, London W.C. 2.

This volume, first published by Dover Publications, Inc., in 1968,

contains new English translations by Thomas Carmody (presently
Associate Professor of Civil Engineering, University of Arizona, Tucson)
and Helmut Kobus presently Gruppenleiter der Gruppe Wasserbau,
\'ersuchsanstalt flir Wasserbau und Schiffbau, Berlin) of:
H;·drod;namica, by Daniel Bernoulli, as published by Johann Reinhold
Dulsecker at Strassburg in 1738;
Hydraulica, by Johann Bernqulli, as published by Marc-Michel
Bousquet et Cie. at Lausanne and Geneva in 1743.
This volume also contains a Preface to the English Translations by
Hunter Rouse, Dean, College of Engineering, The University of Iowa,
Iowa City.

library of Congress Catalog Card Number: 68-u668

Manufactured in the United States of America

DOVER PUBLICATIONS, INC.

180 Varick Street
New York, N.Y. 10014
 CONTENTS
Page
PREFACE TO THE ENGLISH
TRANSLATION Vll

BY HUNTER RousE

HYDRODYNAMICS I

BY DANIEL BERNOULLI

HYDRAULICS 35 1
BY j OHANN BERNOULLI

INDEX TO BOTH TREATISES

."'!

I
""
 CELSISSIME At~E SERENISSIME
PRINCE PS,
DOMINE GRATIOSISSIME.

On aufus fuHfem Sereniffimo Nomini

Tuo lrydrodynamtcam hanc in-
fcribere, nifi ilia Academia: Scien.
tiarum, fub um bone T uo Petropoli
florentis, confilio & fubfidiis a me
confcripta fui.ffet. Novimus quan-
tum Tibi , Sereniffime Princeps ,
Magnanime Academix Protecror, pofl: Augufiam illam
orbis borealis Palladem, debeamus, idque cum toto orbe
literato, qui pra:clara fibi porro ab Academia, amrenis
benevolentia:: Tua:: radiis collufirata, pollicetur, pia &
immortali recolemus memoria. Florebit in a:ternitatis
facrario apud Rufficam gentem Tuorum in illam merito-
rum

(Reproduction of first page of dedication to

Daniel Bernoulli's H ydrodynamics.
See page xix for translation .)
 PREFACE
TO THE ENGLISH TRANSLATIONS

In the belief that students attaining the doctoral level should know
something about the background of their profession, the writer began
in 1960 to offer a graduate course at the University of Iowa on the
history of hydraulics. Instead of attending lectures, every stud_ent
was expected to read the Institute book on the subject [r], select a
lesser-known investigator from the past in each of his three required
doctoral languages (English, of course, included), and submit original
monographs summarizing the respective lives and works. If he so
preferred, a student could prepare instead one monograph on a single
subject, such as pipe resistance, the roots of which would carry him
into several source languages. Furthermore, a student whose lan-
guage background was sufficiently broad to need no further exercise
could concentrate on a single national literature, like Russian
hydraulics.
The lack of Latin in the writer's background-which had proved
particularly troublesome when he was seeking to digest early treatises
on hydraulics-led him to suggest to two students with foreign-
language upbringing (the one, Italian; the other, German) who were
able as well to read Latin that they compare the works on fluid
motion of the two Bernoullis, Johann and Daniel. The idea soon
grew to the point of involving the complete translation of the two
present books, in part as the regular course requirement, in part as
salaried employment with the Iowa Institute of Hydraulic Research,
and in no small part as a labor of love. To the American-born
Thomas Carmody (who really began the undertaking) fell the task of
actual translation; but to the German-born Helmut Kobus (who had
studied far more Latin) fell that of checking meticulously every word
and thought and of preparing the manuscript for the printer after his
fellow translator had left for another university. Kobus, in turn,
finally left the country, and it remained to Carmody (and the writer)
to read proof.
 Vlll PREFACE TO THE ENGLISH TRANSLATIONS

Why the Bernoullis' works should have been singled out for trans-
lation seems at first thought rather obvious, if only because of the fre-
quency with which the name Bernoulli is on a hydraulician's lips.
But it is only Daniel to whom one is making reference, and the word is
gradually spreading that' the theorem bearing his name is nowhere to
be found in his habitually cited Hydrodynamica [2]. Not until the last
few years has mention of either the work Hydraulica [3] or its author
Johann Bernoulli appeared in fluids literature with any frequency
whatever, and this almost exclusively in the writings of C. Truesdell
[4]. It is Truesdell's thesis that, whereas Daniel has received too
much credit for the formulation of the Bernoulli theorem, Johann
has received too little. Readers who have not studied Latin, and
who may never have the chance of seeing the original works, can now
judge the matter for themselves. They can also marvel at the many
familiar concepts which Daniel did originate and for which he has
received almost no credit at all.
To understand the rather curious relationship between Johann
Bernoulli and his son Daniel, one must know something of the family
itself [5]. Basel had become a university town in 1460, a center of
early printing that attracted such Renaissance writers as Erasmus and
Paracelsus, a refuge for Huguenots during the Reformation of 1530,
and finally by the seventeenth century a very literate city of strong
family ties. To this city, in 1622, came a Huguenot from Antwerp by
the name of Bernoulli. He established himself as a merchant and
raised sons who also became merchants. One of these fathered a
dozen children, of whom four lived: a mathematician, an artist,
another mathematician, and at last a merchant. The oldest was
Jakob Bernoulli (1654-1705), who became professor of mathematics
at the university and finally rector. The second mathematician was
Johann Bernoulli (1667-1748), who was trained by his brother,
worked with the French mathematician L'Hopital at Paris, taught
mathematics for ten years in Holland, and then succeeded his brother
as professor at Basel. It was he who had as sons Nikolaus, Daniel,
and Johann II.
Though Jakob was an extremely able mathematician in his own
right, his great contribution to the present history was the education
of his younger brother, who in turn taught his own sons. Unfortu-
nately, friction developed and steadily increased between the brothers,
and their early collaboration eventually changed to rivalry. Jo-
hann's bitterness was increased by L'Hopital's publication in his own
name of various discoveries communicated to him in Johann's many
letters-not to mention an entire course of instruction which he had

-~-- - . . . . _;_ ..
 HUNTER ROUSE lX

given him in Paris-but he nonetheless became upon Newton's death

the foremost mathematician in the world. Because of his close
association with Leibniz (it was Johann who first applied the new
calculus-and, in fact, introduced the word "integral"), he sided
eloquently with him against Newton in the fluxion-calculus contro-
versy, to the chagrin of the Royal Society.
Daniel Bernoulli (1700-1782) was born at Groningen near the
middle of his father Johann's ten-year Dutch professorship. He
studied under his father at Basel after the latter returned to fill the
chair left vacant by Jakob's death. In his twenties Daniel spent
seven or eight years as professor of mathematics at St. Petersburg, a
period darkened only by the early death there of his older brother
Nikolaus, also a mathematician. Daniel won or shared, in the
course of his life, ten prizes awarded by the Paris Academie des
Sciences for the solution of designated problems. The first of these,
received at the age of twenty-four, involved the design of a. clepsydra
for the exact measurement of time at sea. The third, for a paper on
tides, was shared with Euler and Maclaurin. One which he divided
with his father dealt with the inclination of the planetary orbits, and
another, shared with Johann II, was on the best form of anchors.
Still another had to do with the nature and cause of ocean currents.
In each of these he displayed considerable mathematical ability, to be
sure, but above all a keen physical perception and the ingenuity to
produce a solution regardless of the method used [1].
Daniel's Hydrodynamica was begun in 1729, during his sojourn in
Russia, and an uncompleted manuscript of it was left at St. Peters-
burg when he returned to Basel four years later. In the course of its
revision and completion, he wrote for permission to dedicate it to the
Empress of Russia, as an acknowledgment of his debt to that country.
When the book was finally published in Germany in 1 738, he re-
quested that the Russian manuscript be destroyed, but it is still pre-
served in the files of the Soviet Academy of Science (as remarked in
the Preface of a Russian translation of the German edition that
recently appeared l6] ).
A younger colleague of Daniel's, Leonhard Euler ( 1707- 1783), had
also studied mathematics under Johann and, largely as the result of
Daniel's influence, was also invited to St. Petersburg by Catherine I.
There he became professor of mathematics when Daniel returned to
Basel. Euler was eventually to surpass the great Johann Bernoulli as a
mathematician (fifty pages were finally required in his eulogy to list
merely the titles of his writings), and hence it is hardly surprising that
not only Daniel but also Johann was to defer to his judgment at even
 X PREFACE TO THE ENGLISH TRANSLATIONS

so early a date. Daniel wrote him near the end of 1734 that he had
arranged with Dulsecker in Strasbourg to publish his new work, but
not for three years could he report that it was nearly finished. Early
in 1738 he sent several copies to St. Petersburg, but by the end of the
year Euler wrote that they had not yet arrived. On March 7 of the
following year Daniel again protested the total lack of news as to
their fate [ 7].
On the very same day Daniel's father notified Euler that he was
sending him the first part of his own manuscript Hydraulica. Now
except for a brief criticism of Newton's cataract hypothesis in 1716,
there is no record that Johann had written anything whatever on the
subject of hydraulics until some months after his son's treatise was off
the press, when he stated in a letter to Euler that he was preparing a
manuscript on hydraulics which was already well along. Never-
theless, he indicated in the first part that it was written in 1732, a full
year ahead of the Russian version of his son's! The second part fol-
lowed the first to St. Petersburg in 1740, and the two were eventually
published, as Johann had requested, in the Memoirs of the Imperial
Academy of Science for 1737 and 1738 (which were printed, respec-
tively, in 1744 and 1747). They actually first appeared in his col-
lected works, published in Switzerland in 1743.
As surprising as Johann's obvious attempt at seeming to predate his
son in publication was Euler's delay in acknowledging Daniel's book
until such time as he could do the same for Johann's. In fact, of the
two letters of acknowledgment and praise, written on the same day,
that to Johann was far more flattering. A subsequent letter that he
wrote Johann about the book was quoted in part in the foreword to the
Swiss version of Hydraulica as translated herein (seep. 347). Johann's
reason for using only the first paragraph of Euler's letter is evident
from the following translation of the second:
Truly, regarding the force by which vessels are driven backward,
I certainly do not have the least doubt concerning that very method
you use for determining this; but when, for pipes attached hori-
zontally to a vessel, you find that the pressure driving the vessel
backward is different from that which agrees with the hypothesis
of Your son, that force as it is determined by Your Illustrious Son
seems to me certainly to be more suited to the truth than Yours;
may I have said this without offense to You. Indeed, from the
formula which you present for retroaction in this case it follows that
the retroaction can be indefinitely great, even if the orifice is very
small and the motion very slow, and the expression given by Your
Son does not contain this inconsistency; but I am convinced that if
you will deem it worthy to subject this part to examination once

- -- . . . . . _;_ ..
 HUNTER ROUSE Xl

more, Your theory will agree most perfectly with Your Son's idea;
indeed, I suspect that fractions have to be inverted, and, with this
done, it will agree most perfectly with the truth and with Your
Son's expression.

Apparently Daniel first saw his father's treatise when it appeared in

1743 in the collected works, for he thereafter wrote Euler:

I beg your Excellence to tell me in sincere friendship and confi-

dence your opinion of my father's Opera, particularly of the last
volume. I for my part have reason of the highest degree to com-
plain about it: The new mechanics problems stem mostly from me,
and my father had even seen my solutions before he solved them in
his own manner, and nevertheless I am not acknowledged with
even a word, which I find the more annoying as my solution is not
yet published. My first solution of rotation around an instan-
taneous center, found from the nature of the least inertia, he has
questioned and also contemned for a long time, and finally he pub-
lished it as his own. But since by a miraculous hazard I obtained
a page from his manuscript in which this, his pretended solution,
was written, and I complained about it through my brother, he
barely let me pass as a second inventor. The matter is roughly the
same with the remaining new problems in mechanics. Of my
entire Hydrodynamics, of which indeed I in truth need not credit one
iota to my father, I am robbed all of a sudden, and therefore in one
hour I lose the fruits of a work of ten years. All propositions are
taken from my Hydrodynamics; nevertheless, my father calls his
writings Hydraulics, now first discovered, anno r732, since my Hydro-
dynamics was printed only in 1 738. Meanwhile my father has got-
ten everything from me, except that he thought of a different
general method to determine the increment of velocity, which in-
vention consists of some few pages. What my father does not claim
completely for himself he contemns, and finally, as the height ofmy
misfortune, he inserts the letter of your Excellence in which you,
too, diminish my inventions in a field of which I am fully the first,
even the only, author and which I claim to have exhausted com-
pletely. Your Excellence says that I have determined the pressure
of fluids flowing through a conduit in no other way than for the
steady state, whereas I show immediately on page 259 toward
a - vv
the bottom that generally the pressure i s - - - ; and what, on the
2C
other hand, has my father done in this important new field? The in-
vention of the argument comes from me; the idea to consider the
conduit as cut at the point where the pressure is required comes
from me; that one should require the acceleration of the last par-
ticle at the first instant of interruption is my idea; finally, that from
this very acceleration, hindered either partly or completely, the
pressure of the elemental volume determines the pressure of the
 Xll PREFACE TO THE ENGLISH TRANSLATIONS

water in the conduit-this also comes from me; and my father has
done absolutely nothing else than determine the velocity in his own
manner and by repeated reasoning, which is his only invention in
the entire work. The argument about the reaction of fluids my
father does not yet understand today; nevertheless, he refutes me in
the corollary on page 488 [page 336 of present translation; see also
foregoing paragraph from Euler's letter]. All of this is still the
least about which I can complain. In the beginning it seemed
almost unbearable to me; but finally I took everything with resig-
nation; yet I also developed disgust and contempt for my previous
studies, so that I would rather have learned the shoemaker's trade
than mathematics. Also, I have no longer been able to persuade
myself since then to work out anything mathematical. :My entire
remaining pleasure is to work some projects on the blackboard now
and then for future oblivion. I could not accept with a clear con-
science the call to Berlin, even if the King should give me the honor
to send me one, and I beg you therefore not to think of me any
more with respect to this matter. However, I am strongly obliged
to your Excellence for your kind services; your most valuable friend-
ship presents me with an innermost and true pleasure, and I esteem
such [friendship] much higher in itself than in the profit which
could arise to me from it. I could not abstain from complaining to
your Excellence, as my best friend, seeing that the occasion might
well arise that you vindicate me of the unjust suspicion of plagia-
rism without doing wrong to my father, and also bring it about
that the truth, as far as the controversial points between my father
and me are concerned, does not suffer any injury. It does not
seem proper to me to defend myself.

Far from being guilty of plagiarism, Daniel had based his treatise on
material that was not only original but lasting in its interest. The
reader will find in the following pages of translation the initial
appearance of topics that are still prominent in the literature even
today-from the kinetic theory of gases to the principle of jet
propulsion. Daniel was also the first to connect manometers to
piezometric openings in the walls of vessels, to consider the establish-
ment with time of flow in a long conduit, and to attempt to predict
conduit pressure in terms of the velocity. However, his deriva-
tion of what has come to be known as the Bernoulli theorem will
hardly satisfy any reader but the most casual. There is no doubt
that Daniel understood the theorem in its two-term (velocity head and
piezometric head) form. However, the simplicity of the relationship
in comparison with the cumbersomeness of his analysis leads one to
suspect that-despite his claim of invariably reasoning first and ex-
perimenting thereafter-he actually knew the answer in advance.
The artifice of cutting the conduit and relating the pressure to the
assumed acceleration seems forced at best.

---
-
. . -·· ~
 HUNTER ROUSE Xlll

According to Truesdell [8], most of Daniel's difficulty lay in his

imperfect understanding of fluid pressure. Straub, on the contrary,
points out that a paper published by Daniel in the St. Petersburg
Commentarii of I 729 had already contained "an essentially correct
formula for the so-called hydrodynamic pressure" [ 7J. In any event,
though his Hydrodynamica divided a moving fluid for convenience into a
series of slices normal to the direction of motion, in no case was action
stated to occur between them. Pressure was treated rather as a con-
dition existing at the conduit wall which would produce a jet (or
manometric column) equal in height to the piezometric head if the
wall were pierced. Now Johann also assumed the fluid to move in
normal slices, but the concept of pressure as a mutual interaction at
their surface of contact was essential in his analysis. Moreover, in
order to avoid the anomaly of a discontinuity in pressure and velocity
at an abrupt conduit contraction or expansion, he imagined the
actual flow section to change gradually before the contraction or
after the expansion, in a manner reminiscent of ewton's cataract.
He called this transition "gurges," which Truesdell insists should be
translated as "eddy," despite the fact that Johann referred to the
flow within the imaginary throat-like passage rather than to that
around it. In fact, far from visualizing circulatory motion in the
zone of separation, he specifically considered it to be occupied by
stagnant fluid: "And, accordingly, there is formed along the in-
definitely small length HG something like a throat, IFGH, contracting
from the wide into the narrow, through which the liquid must pass,
the acceleration being continuous but nevertheless augmented gradu-
ally, with a rather small portion of the liquid (which fills the small
space IFD) remaining at perpetual rest" [page 357 of the present
volume]. The word "gurges" has hence invariably been translated
herein as "throat."
Johann was obviously a stride beyond his son in his analysis of the
Bernoulli relationship. Whereas Daniel had treated pressure pri-
marily in terms of height of a manometer column or jet, Johann
visualized it as a force, albeit one acting over the area of the slice as a
whole. It remained to Euler-under the considerable stimulus of
Johann's analysis rather than Daniel's-to originate the concept of
pressure at a point and to incorporate the pressure gradient into his
equations of acceleration. These he finally integrated for specific
conditions [g], thereby first deriving in a rigorous manner what is
now known as the Bernoulli equation. As in the case of many
another aspect of fluid motion for which he has not received proper
credit, it is thus the name of Euler which should be most often upon
the hydraulician's lips.
XlV PREFACE TO THE ENGLISH TRANSLATIONS

Although Daniel's lengthier treatise is much easier to read than

Johann's, each will give the reader some difficulty in accustoming
himself to the mathematical and the physical style that the respective
author used. He must hence recall that the calculus was still a rela-
tively new concept; that the energy relationship was not yet correctly
written and even less well understood than the momentum relation-
ship, with which it was thought by many to be in conflict; and that
the theory of dimensions was wholly in its infancy. So long as the
only dimension was length, little difficulty was encountered, as in
Daniel's interrelationship of velocity and piezometric heads. Kine-
matics was somewhat more complex, as witness his expression of
velocity through either its head or a proportionality with an observed
time of fall; in fact, it was Johann who first introduced the coefficient
of proportionality g. Dynamics was the most confusing, for force had
not yet been universally defined in terms of length, time, and mass
through the Newtonian equation, and their interrelationship was still
arbitrary; even Euler, whose equations were in fact dimensionally
homogeneous, still defined mass and weight as he saw fit.
Except for calling attention to certain of the anomalies or actual
errors, the translators have in large part left the assessment of the
authors' analyses to the reader himself. To be sure, many manner-
isms of the times have been eliminated, such as the excessive use of
abbreviations and symbols like the ampersand and equality sign as
parts of speech, and the mathematical notation has been very slightly
modernized. Translators' insertions are invariably placed within
brackets, and the bracketed numbers refer to the lists of cited works at
the ends of the two books. For the convenience of the reader, each
illustration has been introduced into the text at its first point of ref-
erence. The original figures have been reproduced in all cases.
This Preface would not be complete without reference to a contem-
porary undertaking [JO] begun in the Bernoulli city of Basel some two
decades ago and already assuming tangible form: publication of the
collected writings of the Bernoulli family, in particular the members
referred to herein. The project is under the general editorship of Dr.
Otto Spiess, and the first of well over a dozen projected volumes (the
early letters of Johann) recently appeared under his direct guidance
[ r r]. Another, containing Daniel's correspondence as edited by
Dr. Hans Straub, is now in preparation. The scientific world stands
greatly in the debt of Dr. Spiess and his colleagues for the impetus
that they have given to the tremendous task of making this material
generally available. The writer acknowledges his particular grati-
tude for stimulating conferences with Drs. Spiess and Straub, for

. . .• -
._
 HUNTER ROUSE xv

their willingness to review the foregoing pages, and their kindness in

providing portraits of Daniel and Johann from approximately the
times their books were published. A final touch was Dr. Spiess'
indication, during a visit to his home by the writer, of the impend-
ing appearance of Daniel's Hydrodynamica as translated into his native
tongue [ 1 2].
Iowa City, Iowa HUNTER ROUSE
January, 1967

REFERENCES

[1] RousE, H., and INCE, S., History of Hydraulics, Iowa Institute of
Hydraulic Research, 1957. Dover reprint, 1963.
[2] BERNOULLI, D., Hydrodynamica, sive de viribus et motibus fiuidorum
commentarii, Dulsecker, Strasbourg, I 738.
[3] BERNOULLI, J., Hydraulica nunc primum detecta ac demonstrata directe ex
fundamentis pure mechanicis. Anno 1732. Opera Omnia, Vol. 4,
Bousquet, Lausanne and Geneva, 1743.
[4] TRUESDELL, C., Editor's Introduction to Vol. II 12 of Euler's Opera
Omnia, Fussli, Zurich, 1954.
[5] SPIESS, 0., "Die Basler Mathematiker Bernoulli," Bulletin de
!'Association Suisse des Electriciens, Vol. 43, No. 8, 1952.
[6] BERNOULLI, D., Gidrodinamica iii Zapiski o Silakh i Dvizheniyakh
Zhidkostei, translated by V. S. Gokhman, A. I. Nekrasov, K. K.
Vaumgart, and V. I. Smirnov, lzdatel'stvo Akademii Nauk SSSR, 1959.
[7] STRAUB, H., private correspondence with the writer, 1964-65.
[8] TRUESDELL, C., "Zur Geschichte des Begriffes 'innerer Druck,'"
Physikalische Blatter, Vol. 12, No. 7, 1956.
[9] EULER, L., "Principes generaux de l'etat d'equilibre des fluides";
"Principes generaux du mouvement des fluides"; "Continuation
des recherches sur la theorie du mouvement des fluides"; Histoire de
l' Academic de Berlin, 1953-55.
[IO] TRUESDELL, C., "The New Bernoulli Edition," Isis, Vol. 49, Pt. 1,
No. 155, 1958.
[11] SPIESS, 0., Ed., Der Briefwechsel von Johann Bernoulli, Birkhauser,
Basel, 1955.
[12] FLIERL, K., "Des Daniel Bernoulli, Hydrodynamik," Verojfent-
lichungen des Forschungsinstituts des Deutschen Museums fur die Geschichte
der Naturwissenschaften und der Technik, Reihe C, Nr. 1a, Munich,
1965.
 DANIELIS BERNOULLI Jou. Pit.
MED. PROF. BAS IL.
ACAD. SCIENT. IMPER. PETROPOLITANJE, PRIUS MATHESEOS
SUBLIMIORIS PROF. ORD. NUNC MEMBRI ET PROF. HONOR.

HYDRODYNAMICA,
SIVE
DE VIRIBUS ET MOTIBUS FLUIDORUM
COMMENT ARII.
OPUS ACADEMICUM
AB .AUCTO RE, DUM PETRO POLI AGERET,
CONGESTUM~

~
~
_,--. . .
'
I
...
~l:£iit~x
/

~RGENTORA.Tl,
Sumptibus JOHANNIS l\.EINHOLDI DULSECJCERI,
Anno M D CC XXXVlll.
T.r ri~ Jon. fuNR, D£cKERI, Typographi Bafilienfis.
 BY
DANIEL BERNOULLI, Son of Johann,
MEDICAL PROFESSOR AT BASEL
Formerly Professor Ordinarius of Higher Mathematics
Now a Member and Honorary Professor of the
Imperial Academy of Science of St. Petersburg

HYDRODYNAMICS
OR

Commentaries on Forces and Motions of Fluids.

AN ACADEMIC WORK
Written by the Author While He was Engaged at
St. Petersburg

STRASSBURG,
PUBLISHED BY

JOHANN REINHOLD DULSECKER

Anno MDCCXXXVIII
TYPE BY JOH. HENR. DECKER, BASILIEN TYPOGRAPHER

( Translation of original title page,

reproduced on preceding page.)
 TO THE
MOST ELEVATED
and
MOST SERENE
Prince and Lord

LORD
ERNST
JOHANN
Duke, by the Grace of God, of
CURLANDIA
and
SEMGALLIA
IN LIVONIA

. . .
 Most Elevated and Most Serene
PRINCE
Most Beloved Lord

I would not have dared to dedicate this Hydrodynamics to Your

Most Serene Name if it had not been written by me with the
advice and help of the Academy of Sciences which is flourishing
under your protection at St. Petersburg. We know how much we
owe You, Most Serene Prince, Magnanimous Protector of the
Academy, second to that August Athene of the northern world, and
with pious and undying memory we shall reflect upon this with the
entire learned world, which is inspired henceforth by the excellent
Academy enlightened by the pleasant rays of your benevolence.
Flourishing in a shrine of eternity will be: among the Russian
people, the greatness of Your services to them; among the Curlands,
the memory of the fortunate predictions which divine providence
has destined for them under Your Sceptre; and finally, among
universal peoples, the perpetual admiration of Your most glorious
life. The illustrious destinies of the present times show us how dear to
the gods is the Majesty of the Russian Scepter and the happiness of
Your people. May these [gods] grant the prosperous occurrence of
great deeds; may they grant a great length oflife and Reign to You;
may they grant successors of your blood, rivaling your virtues in a
long series to the end of all time, with the universal world approving.
Thus dedicates,
Most Serene and Most Elevated Prince,
MOST GRACIOUS LORD,
Written at Basel To Your Highness
IO March 1738 A Most Humble & Most Obsequious
Servant
DANIEL BERNOULLI
 PREFACE

F inally our Hydrodynamics is published after all the obstacles

which delayed its printing for almost eight years have been
overcome; perhaps it would never have seen the light if all that
labor had been mine alone. With pleasure indeed I acknowledge
that the principal portions of this work are due to the guidance,
counsel, and support of the Academy of Science of St. Petersburg.
The opportunity for the book arose from the very purpose of the
latter, according to which the first Professors who had gathered to
form it were retained to write a Diatribe on some useful and, as
much as possible, new subject, and they were advised, certainly, that
the Theory of forces and motions of fluids, unless it would be under-
taken with an unwilling Minerva, is neither a useless nor a trite
matter, as anyone will easily concede. In addition, in order to
suppress the Reader's boredom, from the outset I paid attention to a
variety of things, especially in the last five Sections, and I inserted
analytical, physical, and mechanical examples, theoretical as well as
practical, some geometric, nautical, astronomical, and others, the
understanding of which, nevertheless, did not seem to support as
much as to postulate the exposition of the undertaken work. The
calm Reader, understanding these matters, will easily correct
whatever mistakes have escaped [this] hasty person. The intent of
this writing is unique: that I may be useful to the Academy, all
labors of which are organized in such a way as to promote public
convenience and an increase of good literature.

Z....,,_ . . ..,,e. ..
'
 CONTENTS
Page

FIRST CHAPTER
Which Is an Introduction, and which Contains Several
Matters To Be Noted in Advance

SECOND CHAPTER
Which Deals with Standing Fluids and Their Equilibrium,
Either Between One Another or Related to Other Forces 18

THIRD CHAPTER
Concerning the Velocities of Fluids Flowing out of a Vessel
Formed in any Way Whatsoever through any Kind of
Opening Whatever 35

FOUR TH CHAPTER
Concerning the Various Times which can be Expected in the
Effiux of Water 71

FIFTH CHAPTER
Concerning the Motion of Water from Constantly Full Vessels 101

SIXTH CHAPTER
Concerning Fluids not Flowing out but Moving within the
Walls of Vessels I 24

SEVENTH CHAPTER
Concerning the Motion of Water through Submerged Vessels,
where it is Shown by Examples how Significantly Useful is
 CONTENTS

the Principle of the Conservation of Live Forces, even in

those Cases in which Continually some Part of Them is to
be Considered Lost 139

EIGHTH CHAPTER
Concerning the Motion of Homogeneous as well as Hetero-
geneous Fluids through Vessels of Irregular and Abrupt
Shape, where from the Theory of Live Forces, a Part of
Which is Continually Absorbed, are explained Excellently
Singular Phenomena of Fluids driven through Several
Orifices, after General Rules have been Set Forth for
Defining the Motions of Fluids Anywhere 159

NINTH CHAPTER
Concerning the Motion of Fluids that are Pushed forth not by
their own Weight but by an Outside Force, and
particularly concerning Hydraulic Machines and their
Ultimate Grade of Perfection that can be Attained, and
how this could be Perfected further through the Mechanics
of Solids as well as of Fluids 183

TENTH CHAPTER
Concerning Properties and Motions of Elastic Fluids, but
especially of Air 226

ELEVENTH CHAPTER
Concerning Fluids acting in a Vortex, while also Concerning
Those which are Contained in Moving Vessels 275

TWELFTH CHAPTER
Which shows the Statics of Moving Fluids, which I call
H ydraulico-Statics

THIRTEENTH CHAPTER
Concerning the Reaction of Fluids flowing out of Vessels and
the Impetus of the Same, after They have Flowed out, on
the Planes against which They Strike 315
HYDRODYNAMICS
BY

DANIEL BERNOULLI

HYDRODYNAMIC.LE
SECTIO PRIMA.
Qf!_,, intr6itus tfi, 'T.Jlflriaque continet pr,nor11na11.
§. L
 FIRST CHAPTER
Which Is an Introduction) and which Contains Several
Matters To Be Noted in Advance

§1. Since the Theory of Fluids is twofold, of which the one, Hydro-
statics, considering the pressures and various equilibria of stagnant
liquids, and the other, Hydraulics, considering the motion of fluids,
have been treated separately by writers, and since I indeed under-
stood both of them to be interrelated by so close a link that the one is
in very great need of the other, by no means did I hesitate to combine
them, inasmuch as the order of things seemed to require it, and to de-
scribe them both under the common and more general name of
Hydrodynamics. However, although from the most ancient times
the Theory of fluids has been continuously refined, nevertheless it did
not gain very noteworthy additions. Certainly the knowledge of the
ancient Mathematicians was terminated by this, that they understood
the common equilibrium of standing fluids or also of bodies together
with the fluids within which they lie, about which Archimedes wrote.
And since in addition it is self-evident that, where equilibrium does
not exist, motion occurs toward the region of lesser pressure, hence
they were able to contrive various games and hydraulic machines,
serving excellently partly for pleasure and partly for public interests,
in which matter certainly they showed themselves to be very
ingenious. They also perceived, but rather as through a veil,
those motions which are due to the pressure of the air. But they
were clearly ignorant of the true reasons and accurate measures in
matters of Hydraulics, and thus they were merely standing on the
threshold.
§2. The effiux of water from a vessel through a very small orifice
serves excellently for defining the motion of fluids. But although it
was not wholly unknown to Frontinus and others, as some believe, that
the velocity of water flowing out of a vessel or container increases
because of an increased height of water above the point of effiux, it
 2 HYDRODYNAMICS, CHAPTER I

nevertheless must be known that indeed the same Frontinus, in com-

puting small amounts of water, or, rather, the water to be expended,
committed disgraceful and improper errors. Benedetto Castelli first
pondered about the relation between velocities and heights, but he
believed a false law, thinking that the two follow the same proportion.
After this Torricelli finally observed that the velocities increase in the
ratio of the square roots of the heights, which everyone followed. But
even though they were not agreeing on the absolute measure of velo-
city, they nevertheless began experiments by which they estimated
that measure to be defined, from which it is customary to acknowl-
edge especially that which was performed by Guglielmini and
repeated eight times, although it departs, of course, from other ex-
periments performed at that time. However, it is usual that all
[experimentsJ performed under different circumstances differ from
each other, and it is not always safe-as we shall say about many in
the appropriate place-to pass judgment concerning the velocity of
the water according to the quantity of the same flowing in a definite
time through a definite orifice. Thus, when we call to account the
Guglielminian experiment of which we just made mention, it should
be concluded that the velocity [obtained] from the quantity of water
which flowed through a given orifice in a given time was not greater
than that which is due to the fourth part of the height of the surface of
the water above the orifice. And there are other experiments by the
same author which are enumerated in Book 2, Prop. 1, aquarum
fiuentium mensura, by virtue of which the water flowing out can ascend
by its own velocity to two-thirds of that height. Among the works of
Mariotte and others there stand out those [experiments] which decide
for half the height; this diversity of the velocities so estimated not-
withstanding, I persuade myself that the true velocities hardly dif-
fered from each other, that they had been in the usual ratio to the
heights of the water, and that they were everywhere approximately
such as those which are due to the entire height. But those [experi-
ments] which were last mentioned, which at first glance seem to
militate for half the height, by number many among the works of the
authors, without doubt moved Newton, a Man immortal for his
merits, to speak somewhat more boldly about the Theory by which he
had found that water springing vertically upwards from a vessel
through a very small opening can ascend to half the height of the water
standing in the vessel, although he contradicts that assertion in all the
experiments which were conducted concerning these heights directly.
He published the Theory in the first edition of Principia Mathematica
Philosophiae Natura/is and attacked it from the pressure by which the

··.... . --iO •
 SEVERAL MATTERS TO BE NOTED IN ADVANCE 3

water situated in front of the orifice and just about to flow out is
driven into motion. But since the nature of the matter by no means
seems always to permit that the force exciting the water to flow out be
defined a priori, and since, rather, concerning this it is hardly pleasing
to resolve it otherwise than from the phenomena of motion, that is,
a posteriori, which I often found, the thinking relied upon for that
principle must be mistrusted. Hence the Man just praised changed
his statement in the second edition of his Work, and he changed it
back somewhat in the third, affirming that the water rises indeed to
the total height but that the stream which it forms is contracted or
made slender in front of the orifice, thus giving satisfaction to both the
phenomenon of the velocity and that of the quantity flowing out in a
given time, which seemed to contradict each other. But although it
is not to be denied that the contraction of the aqueous filament is the
true reason on account of which the velocity of the water flowing out
cannot be estimated from the quantity, nevertheless I consider that
the Theory is not to be overemphasized, because it is accidental and not
everywhere faithful to itself even while the velocity does not vary, let
alone different reasons such as friction, viscosity of the water, and
other similar things. Thus when the water flows out not through a
simple orifice, but through a small cylindrical pipe, the stream is not
notably contracted, the velocity being preserved, after that has been
excepted which is lost by it because of friction. But if, this notwith-
standing, anyone supposes that the flow of water can be deduced cor-
rectly and wholly from the pressure, I may have asked this, that he
pay attention to the more composite cases, for example to the flow of
water, which Mariotte calls extraordinary, from a vessel which some
diaphragm perforated by an orifice separates into two cavities to be
filled with water, so that the water is forced to flow through the two
orifices. Mariotte speaks about this motion in his excellent Traite du
mouvement des eaux, Part IV, p.m. 442.
§3. Since these things are so, anyone will decide easily for himself
how little hope there is that somehow the Laws of motions for fluids
will be reduced to the rules of pure Geometry without any physical
hypothesis, since certainly on the threshold itself they may have
gotten away from the clearsightedness of this Man superior and in-
comparable in ability; nor do I believe that these things which I am
about to present in this work can endure all mathematical rigor. The
principles of the Theory are physical and are to be accepted, not
without generosity, as approximately true. But yet, after the prin-
ciples have been accepted, all will be geometric, subject to no restric-
tions, and connected to each other by a necessary interrelationship.
 4 HYDRODYNAMICS, CHAPTER I

Nevertheless, I cannot but feel well concerning those physical prin-

ciples with which I became strongly involved, since indeed they led
me by the hand to exposing many new properties concerning both the
equilibrium and the motion of fluids, which, unless the love of the
undertaken work deceives me, will some day promote Hydrody-
namics significantly, if they are refined more than I was allowed [by
circumstances] to do. At this point it may be suitable to admonish-
since to many anything which is new is customarily suspect-that I
conceived the whole Theory in my mind, wrote the treatise, commu-
nicated most ofit privately among friends, even sketched some things
in the presence of our Society, before I undertook any experiment,
lest I should be liable to be deceived by preconceived measures
through a false opinion which is nevertheless approximately satis-
factory to those [measures], and that even at some time Men most
perspicacious in known theorems confessed openly that they cannot
persuade themselves so, nor do they consider that [the theorems] are
about to be confirmed by experiments; and after all that had hap-
pened, at last the experiments were made before friends, and they
agreed with the Theory as much as I myself could barely hope. But
now let us return to that from which we digressed.
§4. After Authors were certain concerning the variety of velocities
caused by changing the heights, they began to consider more com-
posite vessels, namely those furnished with variously inclined and
unequally large ducts. But Frontinus already knew in his own
time the nature of these ducts to some degree, knowing that a quan-
tity of discharge is increased by the slope or depression of a pipe,
that is, of the designated duct which has been attached to a reservoir
or even sometimes to a small river; whence also he ordered that the
pipes be arranged in line, as he says, and put at the same height.
And indeed Frontinus is unjustly accused in this regard by some
that he had no understanding of velocity; however, when he makes
a calculation of all the water received and compares the latter
with that about to be expended, I do not see how he can be excused.
By experience, also, he had been thoroughly informed, which de-
serves to be mentioned, that more water is expended than should be
through a pipe of both the proper size and position to which pipes of
larger size are immediately attached. I will show in the appropriate
place that this is so, and that it had been indicated correctly by
Fabretti, although otherwise very skilled Men indicated that it was
not evident to them, or rather that they were in doubt concerning it.
§5. However, what the ancients observed obscurely and without
true measurements, that at last Mr. Guglielmini grasped in his trea-

.;(l •
I
 SEVERAL MATTERS TO BE NOTED IN ADVANCE 5
tise aquarum fluentium mensura by means of the following more accurate
and more general proposition, saying that the velocity of the water
flowing through an inclined conduit is the same as if it would have flowed from
a vessel through an orifice similar and equal in cross section,just as far below the
surface of the water as the section is below the horizontal [passing] through the
beginning of the canal, which proposition Denis Papin attacked, he him-
self diverging greatly from the truth. But since we are at it, in order
that we may review the principal comments of both Hydrostatics and
Hydraulics, in this place the following remark is also to be listed about
investigating the pressure of fluids from an impetus, namely that the
force of a fluid dashing against a perpendicular plane at a given velocity is
equal to the weight of the cylinder offluid erected above that plane, of which the
altitude is such that from it something movable, by falling freely from rest,
would acquire the velocity of the fluid. With the help of this most useful
Problem one may estimate the force of fluids driving machines or
(which is the nature of the wind) propelling ships, the motion of solid
bodies in resistant media, and many other things. About Hydro-
statics, however, which is particularly concerned with very slender
tubes or capillaries, I say nothing, because thus far it could not be
reduced to the general Laws common to all fluids. Besides, it is
uncertain as to which Author will have first observed the nature of
these small pipes; nevertheless it is agreed that the observation is
recent, because concerning it there is nothing to be seen in books pub-
lished before these last seventy or eighty years.
§6. In addition to those cited, Authors from the times of Galilei
rather celebrated in aquarian matters are Torricelli, Borelli, Viviani,
Pascal, Boyle; and of a more recent age are Varignon, Newton,
Poleni, Hermann, and Jakob and Johann Bernoulli, the discoveries of
whom are found in the Commentaries of the Royal Academy of Science of
Paris, Principia Mathematica Philosophiae Naturalis, the treatise de
Castellis and notes pertaining to Frontinus, Phoronomia, the Acta
Eruditorum of Leipzig, and various other works. But these discoveries
about the curvatures generated from the pressure of a fluid, and others
of this sort, were presented by Geometers because they are easily
reduced to pure Geometry; however, concerning the rest, I pass in
silence over things worthy of all praise.
Since these things have been presented which pertain to the works
of others, I feel that it is reasonable that I advise sincerely, the
opinion of my colleagues also having been considered, whether any
and how many additions to Hydrodynamics can be or must be hoped
for from the former. Briefly therefore, as much as I will be able, I
will indicate the important points of the undertaken work.
6 HYDRODYNAMICS, CHAPTER I

§7. First in order are shown the outstanding Theorems which

pertain to the equilibrium of standing fluids: the custom of the prac-
tice has seemed to me to demand it, although I acknowledge rather
freely that no new propositions have been added by me. Indeed the
method of demonstrating, as far as I know, is original with me, but
since it is easy to compose for oneself innumerable demonstrations,
there is little also in this portion that I claim for myself. In addition,
some phenomena of capillary tubes are reviewed in passing, and
finally on occasion of the pressure which fluids exert on the sides of
vessels, many new and different theorems are added about the shape
of bladders full of liquid, about their capabilities for elevating loads,
about the construction and strength of aqueducts, and about other
associated things.
§8. Thereafter the motion of fluids flowing out of vessels is treated,
and since all who were engaged in this matter up to now will have
considered in their own Theories the unique and most obvious case in
which the orifice is taken as infinitely small with respect to the internal
area of the vessel, ours is recommended not a little for its own breadth,
for it extends itself to the situation of an orifice of any size whatever,
and indeed to vessels of any shape whatever. For although considera-
tion of the internal shape of the vessel is least required when the ori-
fice can be considered as infinitely small, nevertheless without it the
motion of the water cannot be defined when [ the orificeJ is of notable
magnitude. General corollaries are deduced from the Theory which
illustrate splendidly the variable motion of water and the disposition
of the same, and they confirm whatever either experience has shown
or the attributes of the matter indicate manifestly through themselves.
Certainly, when the internal areas ,are, for example, moderately
greater than the area of the orifice, the Theory shows that the error
that follows from the consideration of the orifice as infinitely small is
unnoticeable, and accordingly our several additions will perhaps
seem rather useless. But let me wish that individuals, even if only
those in the future, consider for themselves that-apart from the fact
that I am writing not only for [scholars of] hydraulics but also for
Geometers who are likewise delighted by bare truths-the use of our
meditations is very great in other affairs, which [use] they will under-
stand more when they have considered that the motion begins from
rest and passes through infinite changes before it attains a certain
speed, and that the greatest changes occur often in so short a period of
time that they can in no way be clearly perceived by the senses,
nevertheless that they are to be determined at the individual points
both so that the motion be understood correctly in one's mind and
 SEVERAL MATTERS TO BE NOTED IN ADVANCE 7
because then various Theorems can be deduced. So I noticed (let
this example, for the sake of the momentum of the discussion, be
representative of all) that it cannot occur that the pressure of water
flowing through a conduit at a given velocity be defined along the
sides of it unless those changes, which let me call instantaneous, be
understood correctly in one's mind, no matter how imperceptible
[they are] to the senses. Concerning these things, as it was I who
first thought about them, I thus added a new portion to the Theory of
water with the most pleasing success, which, because it considers both
the motion and the pressure of fluids together, seemed most suitably
called hydraulico-statics. After these specimens of the general Theory
are shown concerning cylindrical vessels, both the simple and those
which are furnished with pipes, there are also determined, especially
in these latter ones, the changes which arise at the beginning of flow
while the given grade of velocity is being reached, and this certainly
under the hypothesis of very wide vessels. But it is to be noted that
these changes are quite perceptible, even if the vessels are of infinite
size, and that they can be demonstrated by experiments, until the
water flowing out of a very wide vessel through a simple orifice im-
mediately at the first instant of time has the entire velocity which it
can attain. The previously mentioned changes depend upon both the
length and the shape of the pipe. Finally also analytical calculations
are included for finding times of a different kind together with the
physical notation pertaining thereto. And finally, as the theory
indicates that it cannot occur that water ascends much beyond the
uppermost surface of a bubbling spring, it is shown at the end of the
section that there does not pertain to our hypotheses the singular
phenomenon which I myself have observed rather frequently and can
imitate at will, and of which mention is made in the History ef the Royal
Academy ef Science ef Paris, for the year I 702, where it is said that it
happens sometimes that water in leaping fountains rises to three or
four times that height which corresponds to the uppermost surface of
the water, but that soon nevertheless the enormous thrusting of water
is depressed to the ordinary height, and afterwards the genuine under-
standing of that phenomenon is conveyed with the true measures
sought from our Theory, and the method of producing that unusual
surge and finally of increasing it at will is indicated.
§9. Further, the Theory is extended to the examination of motions
from constantly full vessels, to which certainly as much water is con-
tinuously supplied as flows out of them. The nature of these consists
most of all in this, that emanating fluids approach more and more that
level of velocity which is due to the full height of the surface of the
8 HYDRODYNAMICS, CHAPTER I

fluid above the orifice, but that they never wholly attain it, except
after an infinite time. Nevertheless the water is shown to converge so
quickly to that velocity that after an unnoticeably short time it very
nearly acquires the full value, except when it is carried through very
long streams or aqueducts and discharged through a great orifice.
Then indeed the accelerations are not so rapid that they cannot be
perceived, which is confirmed by a singular example taken from the
book du mouvement des eaux by Mr. Mariotte. But since the motion
begins from rest and increases forever, formulas are given by the aid
of which either from the time of fl.ow or from the quantity of water
discharged the velocity at individual points of time can be defined,
and vice versa.
§10. Following that, fluids are considered which are being moved
within vessels, where chiefly the reciprocal or oscillatory motions of the
fluids are submitted to measurements and their relations are indicated.
However, Newton gave a similar Theorem for the oscillations of a
fluid in a pipe of uniform cross section ( the two extreme legs of which
are vertical, the intermediate part horizontal), which Theorem my
Father rendered more generally in the Commentaries ef the Imperial
Academy ef Science ef St. Petersburg, Book II, p. 201, for any given in-
clination whatever of the extreme legs with respect to the horizon.
Our Theory explains the entire matter without any restriction, con-
sidering that at individual places the pipes are variable at will with
respect to direction, position, or area. Next it is shown in which
cases it may occur that the different oscillations of a swinging object
are Isochronous, under which conditions the length of a simple
Isochronous pendulum is determined most generally. But in addi-
tion to this type of oscillation certain others are subjected to examina-
tion in the subsequent section, such as those which occur in pipes
immersed in infinite or even confined water, in which there is a need
for singular caution, insofar as all the applied phenomena are respon-
sible for the departure in the calculation; but if the same things have
been neglected, the difference between them becomes as great as it is
between the laws of motion which are valid for perfectly elastic bodies
and those for pliable bodies.
§11. After this I progress to other more composite subjects, consider-
ing certainly the motion of either homogeneous or heterogeneous fluids
which are forced to fl.ow through one or more orifices before they are
discharged into the air, where that rule commonly accepted concern-
ing the surge of the water to the uppermost level of the surface fails
decidedly, with even the ordinary laws of pressure ceasing to hold.
However, of all these things not even a vestige is found among the
 SEVERAL MATTERS TO BE NOTED IN ADVANCE 9

\ :u - of the Authors except that which Mariotte has at the place

ci ed above, Part IV, p.m. 442, du mouvement des eaux, where certainly
he :ho\, that he had been taught by experience that the flow of
\ ~ -er i- retarded; however, at the same time it is clear how far off he
had been from the true Theory of these motions, and this Theory
eems indeed to avoid the influence of almost all the principles- cus-
tomarily applied up to this point in similar cases, so that there is
nothing which further confirms the superiority of our [principles].
Certainly the experiments performed do not allow me to doubt the
truth of these any further. However, these things which have been
considered do not lack their usefulness whenever they can be of great
importance in improving hydraulic machines.
§12. Comments follow about hydraulic machines, about which it is
shown principally that there is some definite termination of perfection
beyond which one may not be able to proceed. But the falling short
of this ultimate level of perfection in many widely accepted machines
is subjected to numerical calculation with added rules or concepts, to
which in constructing new machinery one should pay attention. In
place of an example is mentioned the machine de Marly, * very well
known throughout the world, about which it is shown, if only the de-
scriptions are to be trusted, that it supplies not more than about one
fifty-sixth part of that quantity of water which, the remaining things
being equal, a theoretically perfect machine can supply. Also special
examination is made of a machine most familiar from very ancient
times right up to our age, namely the Cochlea [waterscrew] of Archi-
medes, not unworthy of the attention of Geometers as much because
of the understanding of those things which pertain to pure Geometry
as of those which pertain to Hydraulics.
§13. There follow some specimens of the motion of elastic fluids,
such as air and exploded gunpowder, these things having been set
forth which pertain to the nature of these fluids; but I myself consider
these not differently from physical hypotheses, about which I will
affirm nothing confidently. The Propositions and Problems of this
section are new and selected with the intention that they can give
occasion for illustrating or even solving many physical questions.
Certain things are added about the estimation of live forces innate to
elastic fluids, which at some time probably will be of frequent use in
mechanical practice. Indeed, it is shown that the effect of, let us say,
one pound of inflamed gunpowder can be greater in elevating weights

* [Huge river-driven pumping works, built 1684 in Marly-le-Roi, a suburb of

Paris, to supply the fountains of Versailles-Trans.]
IO HYDRODYNAMICS, CHAPTER I

than that which one hundred very robust men can accomplish by
continuous labor within one day's span.
§14. Further, the circular motion of fluids is treated, and as well
fluids which are standing in vessels having been set in motion; several
other matters are intermixed. But the [statements] which are prof-
fered on circular motion can serve in a certain way in explaining the
phenomena of gravity through vortices; the remaining things may be
applied as far as possible.
§15. The previously discussed Theory of motions is brought right
back again to the equilibrium of fluids, but of moving fluids, of which
the laws have not yet been shown. It is amazing, since motion is
defined elsewhere from pressure, that here by the inverse method the
pressure is sought from the motion by defining it beforehand from the
environment. Nor should I have believed that another way could be
begun safely apart from that which I followed. However, I con-
sidered that the conduit through which the water flows is shortened
in that place and at that point of time which comply with the question;
and afterwards, through our previously mentioned rules, I investi-
gated the acceleration of a particle of water just on the verge of
flowing out. From that acceleration one was able to understand the
pressure upon that aqueous particle, which compression, by the
nature of fluids, is equal to the pressure on the sides of the conduit.
After this pressure has been determined, it is evident what should
happen if the conduit would have been perforated in just this place
and a small pipe were in place of the orifice; indeed it will happed
that the water in it ascends up to a certain standing level in the little
pipe, sustained by the water below flowing through the conduit, so
that here equilibrium is presen-r-betwee._n flowing and standing water:
but under this name I considered ~ ~ i s Theory could be con-
veniently called hydraulico-statics. Further, it may merit being noted
that this same Theory is in turn a basis and a-source of other pre-
viously unknown motions. The Theorems which are presented are
not only new, but also the majority are unexpected, of the truth of all
of which I was not able to convince myself clearly until I had con-
ducted experiments which removed all my doubt. But they have a
significant use whenever the true estimation of the pressure of water
flowing through aqueducts or streams is based upon them, and hence
for deducing the required strengths of pipes. From this also depend
the accurate measures of water to be expended through small water
meters inserted laterally in a stream. In Physiology those things
which pertain to the motion of liquids in an animal body are already
better understood, and there are others.
 SEVERAL MATTERS TO BE NOTED IN ADVANCE I I

§16. Finally, I progress to explaining certain other methods by

which water can cause pressure: namely, while it is flowing out of an
orifice, water thus presses a vessel in the opposite direction no dif-
ferently than a cannon ball pushes back the cannon from which it is
driven out. Many new properties of that repulsion are discovered
which illustrate the nature of pressures splendidly, and their general
laws in mechanics will indicate this matter to those thinking seriously
about it. I performed these investigations because it seemed to me
that they can at some time offer the occasion for discovering some-
thing new for navigation without oars or the help of the wind, about
which matter I may in the proper place convey a little information,
although I know that the origins of all things of this sort in themselves
eem ridiculous to most people. Finally certain Theorems are also
included on the force of water from impulse and back-pressure thence
developed that bodies encounter while moving in fluids.
§17. And certainly these are things which seemed to me to allow
geometric deduction from accepted principles. But since there is
nothing in Theory so rigorously proven that it does not require some
restriction in its application to solid bodies, therefore it is readily
evident that no Theory regarding fluids should be anticipated that
satisfies most fully all measurements ascertained by experience; I
want those who will try to confirm our Theorems by experiments to be
mindful of this matter. Certainly they discover some, but not per-
fect, agreement everywhere, which is either more strict or more lax
depending on the circumstances of things. But whenever I myself
performed some experiment, first of all I pondered until the principles
of the Theory agreed with the proposed case, and thus the experiment
never or very rarely failed me. Indeed, not only was I accustomed
to discern beforehand in which region the difference would be, if it
was to be noticeable, but also how large it would be; thus it is clear
enough in itself, if I judge correctly, that fluids certainly follow the
laws which we assert to be prescribed for them, although they en-
counter everywhere now greater, now lesser obstacles. In addition,
I performed not a few experiments, of which I placed individual ones
at the end of the section to which they pertain; but I was especially
anxious to confirm propositions previously unknown and for the most
par t paradoxical enough. Concerning belief in the experiments,
there is nothing which anyone may doubt, since I performed the im-
portant ones in the presence of Friends after the Theory had been
made public, nevertheless leaving a large portion of the experiments
·hich I conceived in my mind to be performed by others, since it is
not pleasing to go through them individually. After our propositions
12 HYDRODYNAMICS, CHAPTER I

have been read thoroughly, anyone may propose innumerable others

for himself, and I judged that it was not my task to explain all that are
desired by me; nevertheless, I explained some.
§18. But now, finally, the understanding of the principles that we
have mentioned so often is to be rendered. The primary one is the
conservation ef live forces, or, as I say, the equality between actual descent and
potential ascent. I shall make use of this latter term, because that which
the other one signifies finds perhaps a more liberal usage among some
Philosophers, who indeed are inclined to the name vis viva only. I
plan, here and in our work to come, to speak of this matter a little
more fully .
§1 9. After Galilei had shown that a body descending either verti-
cally or on some curved plane acquires the same velocity as long as the
height of fall is the same, which can be shown from the nature of
pressures, Huygens made use of this same proposition, but fortunately
for a more general hypothesis, in bringing out the laws of motion of
elastic bodies resulting from percussion, and, certainly, in stabilizing
the center of oscillation ofa compound pendulum; indeed, he brought
forth this axiom of his own in the following words: If any number ef
weights begin to be moved in some way by the force ef their own gravity, and the
individual weights return to rest ef their own accord, the center ef gravity ef the
same group ef bodies will return to the original height, where by the phrase
in some way he understands either they may strike each other during descent,
or they may press, or the bodies may act on each other in any other way. From
that axiom at once follows the principle of the conservation of live
forces, which Huygens himself showed also, and in which it is assumed:
If any number ef weights begin to be moved in some way by the force ef their
own gravity, the velocities ef the individual weights everywhere will be such that
the products gathered from the squares ef these [velocitiesJ multiplied by their
appropriate masses are proportional to the vertical height through which the
center ef gravity ef the composite ef the bodies descends multiplied by the masses
ef all ef them. It is amazing how much utility this hypothesis may have
in mechanical Philosophy; indeed, my Father, if anyone, noticed this
correctly; he showed it vaguely, but among the foremost, in the
Parisian edition of his Discours sur les Lois du mouvement and in Book II
of the Commentaries ef the Imperial Academy ef Science ef St. Petersburg, and
it is the same that I employed for investigating in fluids the laws of
motion arising from their own gravity; for I set the velocities of the
particles constantly to be such that, after the individual particles were
moved vertically upward to the state of rest, their common center of
gravity ascended to the original height. However, I preferred to
adopt this hypothesis, on account of the reason mentioned above, with
 SEVERAL MATTERS TO BE NOTED IN ADVANCE I3

Huygenian rather than Paternal words, and to mark it with the name
of the equality between actual descent and potential ascent, rather than by that
other of conservation of live forces, which some even yet dislike, chiefly in
England, I know not by what misfortune. Certainly it seems to me
that in the entire Leibnitian doctrine about live forces there is nothing
concerning which not everyone, in his own manner of speaking,
would agree, which, unless I am in error, I showed clearly in the
Commentaries of the Imperial Academy of Science of St. Petersburg, Book I,
p. r3rff., and to which I wished to commission a place here, lest any
of the Readers be offended by these words, and so that he knows that
nothing is accepted by me which is not received in Mechanics by all,
and which does not join by a necessary link with what Galilei already
showed when he established that the increments of velocities follow a
proportion composed of pressures and instants of time.
§20. Although, concerning the remainder, the principle mentioned
above is universal, nevertheless it is not to be treated without circum-
spection, because it often occurs that the motion carries over into
another material. So, for instance, the position of the former is valid
for determining the rules of motions from percussion, if only the bodies
be perfectly elastic; but when they are not so, it is easy to see that a
portion of the live forces, or of the potential ascent, expended in the com-
pression of bodies is not restored to the bodies, but remains impressed
in the certain fine material to which it has transferred. If, neverthe-
less, the matter is correctly considered, whenever the ratio is known
which exists between the portion residual to the bodies and that which
transfers to the fine material, it will appear that that inconvenience
can be obviated easily, and thus the laws of motions can be defined
properly for pliable bodies. Something similar occurs in calculating
the motion of water, where sometimes it is clear that a portion of the
potential ascent is lost continuously; this matter should be taken into
consideration in any case in the calculations to be performed; having
paid proper attention to this, I came to detect many new Theorems
about the flow of water, which [fact] is to be seen in Chapters VI and
VII, and about which I do not yet see whether they could be proven,
much less invented, by any other method.
§21. So, therefore, I bave not used our principle recklessly, and in
this way much presents itself that was previously unknown, not only
about the motion of water but also, as one can see, surprisingly, about
its pressure, which, with no Analysis yet performed, no one will have
easily foreseen or expected. But when it happens that neither can all
of the potential ascent be conserved because of the nature of the situa-
tion, nor can it be foreseen how much may be absorbed, the motion of
14 HYDRODYNAMICS, CHAPTER I

fluids cannot be determined accurately enough, nor do I estimate that

it can be done by any other method. Therefore, let me wish that the
Reader be cautious in deducing corollaries from our Theory which
often, on account of changed conditions, may not accurately agree
with experiments.
§22. From the previous statements it is already sufficiently evident
that it is required from our theory that the velocity of the individual
particles of fluid be defined from an assumed velocity which exists at
some place such as the point of efflux. In the same manner, it was
necessary to add another hypothesis, which is this: After we under-
stand, in the mind of course, that the fluid was divided into layers
perpendicular to the direction of motion, let us consider that the par-
ticles of fluid of the same layer are moved at the same velocity, so that
everywhere the velocity of the fluid is reciprocally proportional to the
corresponding area of the vessel. This hypothesis is familiar, although
it is known further that fluid moves slightly more slowly at the sides
of a vessel but faster in the middle, which happens on account of the
attrition, and that other exceptions as well are frequently to be made.
Nevertheless, a noticeable error can very rarely arise from defects of
this sort.
§23. Let me terminate here these warnings about our hypotheses
with an enumeration of the phenomena which can both illustrate and
confirm somewhat the conservation of live forces in the motion of fluids.
Indeed, many of them occur in this very text, which, however, I will
not treat on account of the calculations they require. However, that
which is observed concerning a drop having fallen into standing water
is trivial and obvious: certainly it creates rings in the surface of the
standing water, and either the larger the drop or the higher the fall,
the more of them it creates, and there is no doubt but that these rings
would propagate themselves without end unless the viscosity of the
fluid and other similar things were [acting] as a hindrance. Also,
whenever it is pleasing to observe another effect from drops of this
sort while many smaller drops are being projected upward from the
surface of the water below, then what pertains here especially appears
constantly: that the higher the drops surge, the fewer they are in
number and the smaller in volume; and when the height of the fall
was two feet, rather often the smaller drops ascended beyond the height
of fall, particularly when the water was passing through a large ori-
fice. Also it is worthwhile to note here what is observed regarding a
particle of water in a narrow conduit which is horizontal and covered
by a perforated cap at the end towards which the water flows. It is
certain that at the instant at which the water reaches right up to the
 SEVERAL MATTERS TO BE NOTED IN ADVANGE I5

cap, a few drops spring forth with a great impetus, and soon the
motion of all the water is established. But anyone could easily be-
lieve that the water next to the orifice continues to be moved at the
proper velocity, and that the remainder of the flow is unchanged;
his, however, would represent the conservation of live forces very
poorly; that violent instantaneous efflux of water, just like an explosion,
represents it splendidly; more about these things elsewhere.
§24. These are the things which I wished to point out in advance
concerning our hypotheses and both their eminence and deficiency.
It remains that I say something about the nature of the fluids toward
which our extensive efforts will certainly be directed, not because I
consider that I have more understanding of them than others, but
because I believe it is a sin to depart from this custom common to all
·,Titers. And first, certainly, all usually agree that internal motion
exists in all fluids whatever, without which indeed no one under-
-rands correctly such fluidity, effervescences of different fluids, dis-
solutions of solids submerged in fluids, evaporations, and infinitely
many other phenomena. Hence a great part of the most solid matter
liquefies with sufficient heat, which impels all things into motion; but
that internal motion causes the particles not to remain adjacent to one
another, but rather to move to and fro, by which it occurs that without
friction they withdraw from a spot upon receiving a very small im-
pulse, which would certainly not happen if the same particles were
placed adjacent to one another as in a pile of sand. So it is easy to
understand that the dust from egg shells held in a pan over a fire is
said to imitate boiling milk. But the greater the heat is, the more
\1.olent is the motion of all the particles, and these are dispersed at a
2'Teater interval from one another; this agrees with the dilatation of
all fluids from added heat, and their contraction from cold, to which
law even water itself, not yet frozen, is subject; but that it is ofa dif-
fere nt nature when it is frozen seems to be deducible from another
cause occurring by chance, namely this: that water supports air
particles in its interstices which thus do not increase the volume of the
·rnter, just as sugar dissolved in water does not increase its volume;
that at the time of impending freezing the motion of the aqueous
particles is lessened; that thus the same particles approach each other
more; and that thus they drive the air particles from their own inter-
stices, which, then less suitably arranged in a different place, can in-
crease the volume, just as sugar not yet dissolved can increase the
·olume of the water with which it is mixed. Hence the reason is
easily deduced as to why ice made of water well purged of air before
freezing becomes not specifically lighter, but rather somewhat
 - -~-
. ~-

HYDRODYNAMICS, CHAPTER I

heavier. But Mariotte instituted excellent experiments about the

true solution of air in water right up to the point of saturation, and he
enumerated these in his Traite du mouvement des eaux. Therefore, there
is an opportunity for doubt that fluids (as I said) are frozen when the
intestine motion ceases or is greatly diminished, when, namely, the
particles collapse on each other and become adjacent, and at the same
time expel the heterogeneous particles from the interstices, if any
linger there; nevertheless, the hardness of frozen bodies is hence not
understood more clearly, since it seems that with that motion stop-
ping, a body is formed with a nature halfway between a fluid and a
solid, unless something else occurs, and it should be compared with a
pile of sand; but whatever occurs to the matter, lest certainly I follow
conjectures, in the meantime it will be pleasing to assume: that as
many particles as you please gravitate to each other, or, in order that
I use the word familiar to the English, attract each other, and that the
attraction increases significantly when the particles approach one
another; that it is of different strength in different bodies, for example
less in oil than in water, the ice of which is harder; that fluids, the
particles of which either attract each other more strongly or are
moving in a slower motion, freeze more quickly and more easily.
Then it would be pleasing to conjecture that water impregnated with
sugar or salt freezes more slowly, because the particles of sugar or salt
placed between the aqueous particles diminish the attraction of the
latter, so that these cannot be joined and frozen solid unless the
heterogeneous particles are driven off; and certainly in all fluids
which are impregnated with heterogeneous particles, at the time of
freezing a certain expulsion or secretion or precipitation of portions
from the pores occurs. There are an infinite number of other pheno-
mena of either solid or fluid bodies which agree altogether wonder-
fully with the principle of mutual gravitation, such that it is a pity that
the principle itself is established so far above the human mind that I
consider that there is no one who can understand it in any way.
§25. At last let one admit to having been warned that this treatise is
considered by me as Physical rather than Mathematical, and that I
did not conduct this plan to obtain a Geometric method in hypotheses,
definitions, and other devices to be set down in advance beyond
measure, and I follow everywhere the order and discourse of Geo-
meters who customarily start from the beginning, having explained
the propositions, and that I treat all in such an order that the individ-
ual things are deduced properly from the first premises, and they
leave nothing unproven behind them, however much this may already
have been proven by so many others. I did not have any trouble in
 SEVERAL MATTERS TO BE NOTED IN ADVANCE I 7
understanding those things which have been handed down by others,
whether they were definitions and axioms, or even theorems, never-
theless I did not omit their proofs, which are new, and, finally, even
in the first section are placed the proofs of the Theorems demonstrated
at random by others; and since certain terms occur that are neither
explained nor used by others, I shall give their definitions in the text
itself. I shall propose the remainder sometimes in the form of Propo-
sitions, Theorems, Problems, Corollaries, and Scholia, according to
the custom of the Geometers, and sometimes I shall give it in a con-
tinuously explained discourse.
One thing remains about which I wish the Reader to be especially
warned in advance: that I was not able to apply to this work that
diligence or attention which I should have, and which I myself
desired. And therefore I have no doubt but that some errors will have
crept in while I was doing the calculations, which I hope no one will
employ wrongly; others which met my eye while I lightly read over
the treatise I myself corrected; nevertheless I am convinced that still
others remain even yet.
 SECOND CHAPTER
Which Deals with Standing Fluids and Their
Equilibrium, Either Between One Another
or Related to Other Forces

THEOREM I

§1. The surface of a standing fluid is para}lel to the horizon.

PROOF. Let the vessel ABCD (Fig. 1) contain the fluid EBCF, the

FIGURE I

surface EGF of which, if it can be done, is made not parallel to the

horizon. Let an elemental volume in the rather elevated position a
be considered, which by its own gravity is driven vertically downward
by a force represented by ac. Let this force be resolved into the two
components ad and ab, one perpendicular to the surface and the other
tangent to it. But since there is nothing present which resists this
latter force, this cannot but spread its own effect and therefore draw
the elemental volume itself toward E, which would be contrary to the
hypothesis of stagnation, or of the permanent state. Therefore, it is
necessary that the tangential force ab be everywhere zero, which does
not occur otherwise when the entire surface is parallel to the horizon.
Q.E.D.
 STANDING FLUIDS AND THEIR EQUILIBRIUM

§2. COROLLARY. Hence the truth of the general propositions is

understood: that certainly the surface of a fluid, the portions of which
are acted upon by any type of force whatever, always composes itself
so that any volume element whatever, placed on the surface, is drawn
in the direction perpendicular to the surface.

THEOREM 2

§3. A homogeneous fluid contained m two connecting pipes

fo rmed in any way whatever is in a state of equilibrium when both
surfaces are placed at the same level, that is, [when] they maintain
an equal vertical distance from the lowest point of the vessel.
PROOF. Let fluid be contained in the vessel ABC (Fig. 2) composed
of two connecting legs or pipes, and let it be placed in each leg to the

d.
···· ······· ···· 1)·

B
FIGURE 2

same height. I say that this position cannot be changed without

some heavy body betaking itself from a lower position to a higher,
which would be contrary to the nature of weights. For if the surface
E descends to e, and in the other part Dis elevated from D to d, then,
si nce the remaining part of the vessel is full of the same fluid before
a nd after the position has been changed, the effect of all this change is
manifested in this: that the particle Ee will have ascended to Dd.
Besides, the same is evident as well from the first Theorem, since a
pipe in standing water can be assumed to be formed in any way what-
ever, in which certainly the water will maintain the position which it
had before, since it is just the same whether the water enclosed in
the pipe is pressed by t]i.e sides of the pipe or by the surrounding
water.
20 HYDRODYNAMICS, CHAPTER II

ScHoLrnM 1

§4. If in the first proof of the preceding paragraph the total mass at
DBE is considered to have exchanged its position with the position
dBe, it is easily shown that the center of gravity of the entire mass has
ascended to a higher position, which is no less absurd. But since in
our proof there is no particle at Ee which will not ascend after the
position has been changed, I considered that the proof will be more
precise and more clear if there be no consideration of the center of
gravity.

ScHOLIUM 2

§5. We have some individual phenomena concerning capillary

tubes. For instance, water ascends above the level in a rather narrow
tube, the other extremity of which is submerged in the water, while
Mercury does not reach that level. Truly, since I considered this
very carefully at one time, I came to the same conclusion more or
less which my uncle, Jakob Bernoulli, of blessed memory, once gave
in his dissertatio de gravitate aetheris, namely that the water in a rather
narrow tube ascends there beyond that level because the number of
aereo-aethereal particles at the base of the column which lies above
the water in the tube is less than the number of the particles at a
similar base beyond the tube. This is certainly understood from the
following: after the globules have been placed next to each other on a
horizontal table, if a circle is made with a compass, some particles are
necessarily excluded because they cannot be divided; but the pres-
sures of the aereo-aethereal columns (of which one base is in the tube,
the other outside the tube) are in proportion to the bases, that is, to
the numbers of particles in the bases; hence, if the number of particles
in the first base equals a, in the second base equals a + b, and the
pressure in the first column equals g, the pressure in the other column
will be a + b g, thus the difference of the pressures equals~ g, to which
a a
the height of the water above the level must be equated. In order
that these things be understood more clearly, it will have to be con-
sidered that g is proprotional to the square of the diameter which
corresponds to the surface of the fluid contained in the tube, and that
a is also proportional to the same square because of the extreme
smallness of the particles, such that the ratio of g to a is to be con-
sidered constant, and accordingly that the height of the water above
the level must follow the proportion of b itself. But, which is in-
trinsically evident, b is in proportion to the periphery of the surface of
 STANDING FLUIDS AND THEIR EQUILIBRIUM 21

the fluid contained in the tube; therefore, the height above the level
will be in proportion to that same periphery, which experience has
confirmed for a long time now. If further we should now consider
different fluids, we will see that the previously mentioned periphery is
the more complicated and accordingly the larger, the greater are the
fluid particles, and since the height of the fluid above the level depends
upon the magnitude of this periphery, we understand why this height
does not follow the inverse ratio of the specific gravity in the same tube.
Thus if the same tube is immersed in spirits of wine and water, the
former will ascend less than the latter, although, nevertheless, the
spirits should ascend more on account of the lesser gravity. But this
indicates, if I have followed the matter correctly, that the pardtles of
the spirits of wine are less [in number] than the particles of water.
Nevertheless, never in my judgment can the ascent above the level in
any fluid be changed into descent, and I should believe that all fluids
are of the same nature in this respect, unless some other reason not yet
considered so far appears in addition, and ifwe should argue accord-
ing to our hypothesis, it should be said that Mercury also would have
ascended above the level if only its particles were not attracted
mutually to each other by a greater force than are the particles of
water. Indeed, to this attraction I attribute all those things which
make Mercury go in a different way. At the end of this section let
me show the experiments which led me to this thinking.

LEMMA

§6. Let the cylindrical pipe ABDC (Fig. 3), the base of which is
perpendicular to the sides of the pipe, be inclined to the horizon in any

FIGURE 3
22 HYDRODYNAMICS, CHAPTER II

way whatever, and let it be considered filled with water right up to

AB. I say that the pressure of all the water on the base CD is equal
to the weight of the aqueous cylinder the base of which is CD and the
height of which is the vertical DE terminated by the horizontal BE.
PROOF. When the shape of the pipe is cylindrical, and in addition
the base is perpendicular to the sides of the pipe, anyone sees that the
action of the fluid on the base is the same as if there were a solid
cylinder of the same weight above the inclined plane; but it is estab-
lished from mechanics that the pressure of the solid cylinder on the
base is that which is defined in the proposition, and therefore the
action of the fluid will be such, if only one does not regard the adhe-
sion of the fluid to the sides of the pipe and also the behavior of the
same with regard to capillary tubes, from which we diverted our
thinking. Q.E.D.

THEOREM 3

§7. Now if, generally, a vessel AHMB (Fig. 4) is formed in any way
whatever and filled with water right up to DE, the pressure of the
A

(
~rj
,s
FIGURE 4

water on the individual particles in the vessel, such as at G or H, will

always be equal to the weight of an aqueous cylinder the base of
which is the surface of that particle and the altitude of which is equal
to the vertical distance of the same particle from the aqueous surface.
PROOF. I. Let there be considered at G the small cylindrical pipe
CG entering the vessel perpendicularly, and let this pipe be under-
stood to be full of the same liquid right up to C, located on the line
 STANDING FLUIDS AND THEIR EQUILIBRIUM 23
ED. If now the vessel is considered to be perforated at G, the fluid
in both places will be in equilibrium (through §3); therefore, the fluid
in the tube CG presses against the interior just as much as the fluid of
the vessel presses against the exterior. But the former pressure
agrees with the proposition (through §6), and therefore the other
[does] also.
II. But if, in place of the point G, another point His assumed so
that a line which enters the vessel perpendicularly at that point lies
within the vessel, then the whole vessel RHSON can be considered
united with the former at H, full of water right up to PO. Thus,
indeed, it appears, if the particle at H which is common to both
vessels is perforated, that the fluid thus will be in equilibrium (§3),
and therefore that the pressure of each at His equal. But the pres-
sure of the fluid in RSN is that which is indicated in the proposition
(according to the first part of this proof), and therefore it is the
pressure of the fluid which is in the vessel AME. Q.E.D.

ScHOLIUM

§8. The equilibria of standing fluids in more composite cases are

easily deduced from these propositions. However, I, being content
with the proofs which I just gave of the fundamental propositions in
hydrostatics, do not wish to follow all of them, for the understanding
of our practice does not require it. But those things which are im-
portant to the pressures of fluids not being at rest surely require a
more profound investigation. Nor yet has the pressure of fluids
flowing through conduits or pipes at a given rate of speed been
properly determined by anyone, although this type of argument may
be very useful in hydraulic matters as well as in many others. But it is
not advisable to deal with these latter things before we have com-
mented on the motion of fluids.
§9. From the preceding an understanding is evident of the powers of
bladders by which immense weights can be supported. Hence even
the force is known which is sustained by the walls of a pipe in which
water is standing. Since it is customarily handled by writers on
hydrostatics, we will now treat this argument, particularly since
many other things are supported by it which we will have to discuss.
At first let there be the bladder onmp (Fig. 5), placed between a
hard floor and the weight B, into which water is poured through the
pipe FRo, the vertical leg of which, for the sake of brevity, we will
make incomparably longer than the diameter of the bladder. The
weight B will not be elevated immediately. But if water is poured
 HYDRODYNAMICS, CHAPTER II

in further right up to F, for instance, finally the weight will become

elevated; however, there will be equilibrium, since the region of con-
tact cd remains with respect to the orifice o just as the weight Bis with
respect to the weight of the aqueous cylinder of height FR standing

Ji'

FIGURE 5
A
FIGURE 6

above the base o. And so the absolute determination of the elevation

depends upon the structure of the bladder; if, for example, it would
have been composed of perfectly flexible filaments admitting no ex-
tension, and if it also had a natural Spherical shape, it is readily
apparent that the regions of contact end and gpe will be equal and
folded, and that the remaining expanded portion will have the shape
of a spherical Zone. And so through Geometry a measure of the
elevation np is deduced, which will be zero as long as the greatest
circle of that bladder will have a smaller ratio to the orifice o than that
which exists between the weight Band the weight of the previously
 STANDING FLUIDS AND THEIR EQUILIBRIUM

mentioned aqueous cylinder, and the whole bladder will not be

unfolded before the height will be infinite, that is, never. But if the
fibers are of a different nature, the situation will be otherwise; many
persons who have discussed the shape of an inflated bladder and who
wanted to apply it to the muscular caverns in the classifications of
animals have not considered this sufficiently. I now wish to treat
this matter a little more fully.
§10. Let there be a bladder DC (Fig. 6) and a weight Phung from
the same, and at the same time [let the bladder beJ attached to the
tube DA, the length of which, in turn, we may consider for our own
benefit to be incomparably greater than the length DC. After these
things have been established, certainly anyone easily ascertains that,
after the vessel and the tube have been filled, the former will become
inflated and will lift the appended weight P. But no one will know
the state of equilibrium and the shape of the bag unless the structure
of the bladder and of its fibers is clearly understood; since these
matters are so, we will examine some individual cases which can
occur rather frequently.
CASE I
§11. Let the bladder be composed of the longitudinal fibers DpC,
DmC, etc., in the form of meridians, concurring uniformly at the points
D and C, or the Poles, perfectly flexible and uniform, the individual
ones of which are connected mutually to the next by minute trans-
verse fibers, which are so lax that they admit sufficient extension under
a minimum or practically null force. Thus any fiber DpC whatever
will be curved in the shape of an elastic, and the whole vessel will
assume the form of a solid which is generated from the revolution of
this curve about the axis DC. If, further, the height AD is infinite,
the elastic DpC becomes a rectangle, and then the maximum breadth
of the bladder is to the length of the axis DC as 2 5 is to 1 1, more or
less, and the length of the arc DpC is to the same axis approximately
as 5 is to 2, so that the vessel will be shortened by three fifths at the
maximum elevation of the weight.

CASE II
§12. If, after the remaining things have been established as before,
the minute transverse filaments no, mp, etc., which are perpendicular
to the longitudinal fibers are resistant to extension, it appears that
the shape of the fiber DopC cannot be determined unless two kinds of
forces are considered to be applied at any one point, one of which acts
perpendicularly to the curve and presses the fiber outward, and the
 HYDRODYNAMICS, CHAPTER II

other is perpendicular to the axis DC of the curve and draws inward.

Thus it is easily seen that an infinite number oflaws for these pressures
can be devised so that the fiber DopC conforms to any given curve
whatever, and thus, for instance, even to a circular one, which shape
is attributed by most Physiologists to fibers which pertain to small
muscular mechanisms. But there is still another way by which the
longitudinal fiber DopC can acquire the shape of a circular arc,
namely, when the transverse fibers np, mp, etc., are completely absent.
Thus, indeed, while the bladder is inflated, an opening is made
between two adjacent longitudinal fibers DopC and DnmC through
which the fluid escapes, but at the same time, since it cannot flow out
fast enough, it extends the fibers and composes them to the circular
shape. And in this case the greatest shortening of the bladder, which
in the first case was¾ of the total length of the uninflated bladder, now
is only approximately 1\ .
§13. It follows from the above that it is difficult to determine cor-
rectly the shape of an inflated bladder to which a weight is appended
since, indeed, there is no one who can know perfectly the nature of
the minute fibers. Nevertheless, here I will transcribe certain
examples, which seem to be especially plausible, from my notes, with-
out proof, which, if anyone desires, he will find in Book III, Commen-
taries ef the Academy ef Science ef St. Petersburg. But first of all I will
give the equation for the curve which is formed from the two kinds of
forces, as I mentioned in the previous paragraph, with these follow-
ing any law whatever.
§14. Thus, let the thread AEG (Fig. 7) be fixed at the two points A
and G. Let the straight line AG be drawn, and let there be two in-
finitely close points D and Eon the thread from which the perpen-

B .....
C-·· ··

FIGURE 7
 STANDING FLUIDS AND THEIR EQUILIBRIUM

diculars DB and EC are drawn to AG; moreover, let the small line
DF be parallel to the line AG. It is known that at the individual
points D or E two forces variable in any way whatever are applied,
one of which is everywhere perpendicular to the curve, the other
everywhere perpendicular to AG; we will set the first one equal to A
at the point D and equal to A + dA at the point E, the other = Cat
the point D, and = C + dC at the point E. Further, let AB = x,
BD = y, AD = s, BC = dx, FE = dy, DE = ds, since the element of
the curve is assumed to be of constant magnitude. The radius of the
Osculating circle at point Dis R, at point Eis R + dR. I say that the
following will be the equation pertaining to the curve : -A dR - R dA
= (R dC dx + 2C dy ds + C dx dR ) ds, or, CR ddx having been substi-
tuted for C dy ds (for R is = ~d!s), one will have -A dR - R dA =

-ARds - RCdx
(R dC dx + CR dds + C dy ds + C dx dR )/ds, or
ds
Jcdy.
§15. It is seen from the preceding equation that, when the forces
which are perpendicular to the curve act alone, AR becomes a con-
stant quantity, because certainly thus C becomes o. Therefore, then,
the radius of the osculating circle everywhere follows an inverse ratio
to the corresponding force. But if the forces perpendicular to the
axis are present alone, then, with the letter A vanishing, there results
- RC
dsdx -- Jc drn.
:.r But t h.1s equation
. can b e integrate
. d an d re d uce d

to the form RC (dx ) 2 = a constant quantity; from this it appears that

the force multiplied by the radius of the osculating circle is everywhere
in inverse proportion to the square of the sine which the ordinate
makes with the curve. Similarly, the canonic equation admits inte-
gration when the forces which are perpendicular to the axis are all
equal to one another or proportional to the element ds of the curve.
So, indeed, after dC has been set = o, one obtains -A dR - R dA =
2n dy ds + n dx dR, by considering n as a constant quantity, with
which, after the equation has been properly treated, there results
f
ny dy + mm dy - ns ds = A dx, where mis the constant arising from
the integration.
If, in addition, the forces normal to the curve are assumed propor-
tional to the ordinates y, the last equation can be reduced further to
this:

_ dx = (2ff- gf) dy; J(2ny + 2mm) 2

- (2ff- ~f·
 HYDRODYNAMICS, CHAPTER II

the constants f and m of which will have to be applied to particular

cases, while n and g depend upon the relationship of the forces at some
particular point; whence, if g = o, the catenary appears, and if
n = o, the elastic curve appears; but generally the equation serves for
determining the curvature of a uniformly heavy cloth over which
fluid is lying. The most simple case in this matter is that when it is
supposed thatf = m = o, for then, indeed, there results

- dx = -,==-==gy=dy==
v1 4nnhh - ggyy'
or, after the integration has been performed, with the addition of the
. d constant, x
require = - J4nnhh 2nh
- - - yy + -, w h'1ch.1s t h e equation
.
gg g
of a semicircle to which certainly the cloth will adjust itself in the
following hypothesis: let a rope of heavy cloth AEG (Fig. 8) be curved

A G-

FIGURE 8

in a semicircle, the diameter AG of which is set level, and let fluid lie
over the rope right up to AG. If the weight of the fluid is equal to the
weight of the rope, I say that a perfectly flexible rope of uniform
thickness will preserve the semicircular shape. But in what manner
it is to be effected that the weights of the rope and the fluid become
equal is well known from the elements of Geometry. Finally, ifit is
stated that the forces A as well as Care everywhere proportional to the
corresponding ordinates y (which hypothesis seems clearly to agree
most closely with the true shape of the bladder in Fig. 6), then again
the canonic equation which contains differentials of the third Order
can be reduced simply to a differential equation, and this should be
solved easily through quadratures. If, indeed, A = my and C = ny,
I say that the nature of the curve ADG in Fig. 7 is expressed by this
equation:

dx = (g 3 + ½myy) dy/\! (1 3 + ½nyy) 2 - (g 3 + ½myy) 2

 STANDING FLUIDS AND THEIR EQUILIBRIUM 29
in which the letters] and g of constant magnitude appear again from
the integrations; but the value of the letter n becomes negative when
the equation is applied to determining the shape of the inflated
bladder.
§16. I did not wish to pursue these things too much, because they
do not pertain very closely to Hydrodynamics. Indeed, I include
nothing about elastic fluids because I arranged to treat the theory of
them separately; but, nevertheless, because it pertains to the pressures
of elastic fluids, the former can easily be deduced and proven from the
nature of simply heavy fluids shown above by assuming that the fluid
is destitute of elasticity and that a cylinder of the same fluid of in-
finite or almost infinite altitude is lying above it; but we will mention
how these things are to be understood in the proper place. Now,
indeed, I continue to that which is customarily sought above all in
aquatic matters, namely, how great the strength of conduits must be
in order to resist the pressure of water, where chiefly conduits are con-
sidered which carry water to fountains, about which I will also say
a few things.
§17. The pressure of water standing in conduits ought to be
properly distinguished from the pressure of flowing water, although
no one, as far as I know, has paid any attention to it up to this time;
hence it is that the rules presented by others are valid only for stand-
ing water even though they use words which can persuade equally
that these apply to flowing water. But in order that the distinction
of either Theory may appear in its own light, I will give a certain
example the proof of which will be evident from what is below. In
place of a reservoir let a very wide vessel ABCD (Fig. g) be full of
water right up to EF, and be connected in its lower part to a hori-
zontal cylindrical pipe MOmo through which it is understood that

FIGURE g
 HYDRODYNAMICS, CHAPTER II

water can flow without impediment. Let the vertical NG be drawn,

terminated by the horizontal EH. After these things have been so
prepared, I say that if the entire orifice Oo is obstructed by one's
finger, the point N is pressed outward according to the total height
NG; that if half the orifice is obstructed, this pressure is diminished by
a fourth part of the original, and if, finally, after the finger is removed
the water is allowed to flow very freely, that all the pressure vanishes,
in the same way that the whole is customarily confused with the part
or even with nothing by the Authors. But I will demonstrate that
the pressure can even be made negative and thus be changed into
suction. But since I cannot treat this before I have treated the whole
theory of flowing water, I shall now consider standing water only,
just as if the entire orifice Oo were closed.
§18. Moreover, it is definite from Mechanics that the walls of the
pipe MOmo ( the diameter of which we will consider to be incom-
parably less than the height NG) are not extended differently than if
they were arranged in the rectangular shape MOmo (Fig. ro), and

..M~---~

m (J

FIGURE IO

if they had a weight P appended which is equal to the weight of

the aqueous prism, the three sides of which are: first, the radius
of the pipe; second, the length of the same; and third, the altitude
of the water above the pipe. From this proposition is known not only
the ratio of the tensions if the al ti tu des of the water or the diameter of the
pipes were different, but also the very measure of the tensions. Thus,
accordingly, if the strength of the pipes is greater than that tension,
there will be no danger of rupture; if otherwise, the pipe will be cer-
tainly ruptured. In addition, experiments of this sort are difficult
and expensive. Therefore, the strength of lead or iron pipes could be
understood more easily if it were known from experiment how much
weight a thread of lead or iron of given thickness can sustain without
danger of rupture. At the end of the section I will add a similar
 STANDING FLUIDS AND THEIR EQUILIBRIUM 31

experiment performed by me to show how from this the strength of

a pipe of given thickness and diameter can be deduced.

EXPERIMENTS WHICH PERTAIN TO CHAPTER II

PERTAINING TO §5
Concerning capillary tubes: Innumerable experiments concerning the
nature of these tubes have been undertaken by many, among whom
Georg Bernhard Bilfinger stands out, who not only collected the im-
portant ones but also added many of his own; see Commentaries of the
Imperial Acaderrry of Science of St. Petersburg, Book 2, p. 233ff.
I. In order that it might appear properly to the eye how contrary in
Character mercury and the rest of the fluids are in this area, I ordered
a glass vessel ABD (Fig. 11) to be made composed of two vertical legs,

D
E·
F
. l .o
I

FIGURE I I

of which the one AB had a diameter of three or four lines and the
other BC of hardly a third part of a line. When the vessel was filled
with any liquid whatever, the surface was higher in the narrower leg
than in the wider, as at D and G; however, mercury alone was more
depressed in the narrower than in the wider, as at F and G.
II. In order to show that mercury differs from the nature of the
other fluids for no other reason than on account of the stronger mutual
attraction of its own particles, I reflected on these experiments: indeed,
I filled a slender pipe with mercury by suction and erected it slowly
from its horizontal position. Accordingly, the mercury, although not
32 HYDRODYNAMICS, CHAPTER II

all ofit, flowed out, and the vertical height of the mercury remaining
in the pipe was consistent with itself in every position. However,
when the mercury is suspended in this way in the pipe, and if then the
extremity of the pipe is brought in touch with the mercury standing
in the vessel, it all flows out directly. The prior Phenomena, unless
I am mistaken, indicate that the same thing occurs with mercury and
the other fluids when there is no opportunity for an attractive force;
but the last phenomenon shows that mercury attracts itself very
strongly.
III. Let there be assumed a cylindrical glass pipe of a diameter of
three or four lines, furnished with a base of delicate Paper or of a very
thin plate of iron prepared and perforated in the middle by a tiny
little orifice, as Fig. 12 shows. Let the pipe ACDB be inclined and

..M ·· ··-· ···· N

o(

d
FIGURE 12

filled entirely with mercury, then erected little by little; what hap-
pened before will occur, and although the pipe is very wide, never-
theless, not all the mercury flows out, but part of it will remain sus-
pended, as for example MCDN, and the smaller is its little orifice o,
the greater will be this amount. When then the base is submerged in
the mercury in some associated vessel just a little bit, so that the sub-
merged part of the pipe is Ca, not only does the mercury not ascend in
the pipe right up to f3 (it having been assumed that Ca = Mf3 of
course) but also almost all of it flows out, until the surface MN
reaches a. Next I submerged the empty pipe ACDB reasonably
deeply in the mercury which was in the other vessel, and nevertheless
nothing of it began to flow from the vessel into the pipe before it had
been submerged to the height CM; and then suddenly it flowed right
up until it had reached a level in each part, namely right up to MN, if
 STANDING FLUIDS AND THEIR EQUILIBRIUM 33
it was submerged up to that point. All these things are deduced
easily from the mutual attraction of the mercurial particles. In
addition, I performed a test to investigate the relationship which
exists between the height MC and the area o of the little orifice; at any
rate, it is probable that that diameter is in a reciprocal ratio to the
diameter pertaining to the little orifice; nevertheless, I was not able to
confirm the idea sufficiently by experiment, sometimes because of the
impurity of the mercury which I used, which caused the height of the
suspended mercury to be not completely consistent with itself when
the orifice had not been varied in repeated experiments, or sometimes
also because it is difficult to measure very small orifices accurately.
Indeed, the orifices must be a minimum, since the height of the sus-
pended mercury is barely six or eight lines when the diameter of the
orifice equals the sixth part of a line; nevertheless, let me tell the
method which I have used. Indeed, by means of copper wires of
different thickness which are used in musical instruments, the very
small diameters of which I found very correctly from the length and
the weight of them, I perforated the little paper CD; but in this way
shreds usually appear around the walls of the orifice which impede the
effiux, and thus it easily happens that the orifice is greater than is the
thickness of the wire.

PERTAINING TO §18

CONCERNING THE STRENGTH OF PIPES. A round copper wire the dia-

meter of which was 2 / 11 of a Paris line, to which successively con-
tinuously greater weights were added, did not break until the weight
exceeded 18 Nuremberg pounds. Then I observed that a very thin
lead plate, which was of rectangular shape, 5/4 of a line wide, 1/ 131
line thick, was broken when to it was appended a weight of three and
a half ounces. From these two observations it followed, with all the
remaining things being equal, that the copper wire is more than 28
times as strong as the lead wire. From the previous experiment it is
also deduced that, if a copper pipe should have a diameter of I foot
and the thickness of the walls were 2 / 11 line, it can sustain water to a
height of 518 feet before it is broken. In this calculation I used 70
pounds for the weight ofa cubic foot of water. But if the same pipe
is of lead, it will sustain water to a height of 18 feet in the light of the
other observation, and it can bear a height of water of 99 feet if the
walls of the pipe are a whole line in thickness. This agrees with what
Mariotte has in his Traite du mouvement des eaux, p. 472, where indeed
he says that a lead pipe, the diameter of which was I foot and the
34 HYDRODYNAMICS, CHAPTER II

thickness of the walls two and a half lines, has supported water with-
out rupture to the height of a hundred feet, and that while he was
observing this he shaved off the sides little by little until at last they
were diminished to a thickness of one line, and that then at last the
force of the water destroyed the pipe.
From the observed strength of copper wire, the strength of cannons
is also determined: let there be, for instance, a cannon, the internal
diameter of the barrel of which is three inches; moreover, the
thickness of the walls not far from the touch-hole, where the force of
the powder is greatest, is customarily more or less equal to the in-
ternal diameter, so that the total diameter is thrice the internal dia-
meter of the barrel. Because, therefore, this thickness is not to be
neglected with respect to the internal diameter of the barrel, we shall
consider all the material concentrated in the middle and thus at a
distance of three inches from the axis of the barrel. This having been
established, the maximum height of water which a cannon can sup-
port not far from the touch-hole will be 1.;- · 12 · 3 · 2 · 518 = 205,128,
which force exceeds the elasticity of natural air by about seven thou-
sand times. But I will show in the following that ignited gunpowder
can exert a force for rupturing any cannon greater indeed than that
which was mentioned, but nevertheless not exceeding it much. But
the cannons obtain the additional strength that they require from
belts or bands which are called plattes bandes et moulures, apart from the
fact that at the very rear of the cannon (d l'endroit de la culasse) the
thickness is greater than that which we assumed. Nevertheless, we
will not be surprised that quite a few cannons are shattered.
 THIRD CHAPTER
Concerning the Velocities of Fluids Flowing out of a
Vessel Formed in any Way Whatever through
any Kind of Opening Whatever

§1. Before we may attempt to define the motion of water developing

from its own gravity, we will look again at what we set forth in the
First Chapter, §§18, 19, 20, 21 and 22, concerning the principles to be
applied to the following matters.
We will recollect, certainly, that the potential ascent of a System, the
individual portions of which are moved at any velocity whatever,
indicates the vertical height to which the center of gravity of that
System reaches if the individual particles, their motion having been
turned upward with the proper velocity, are understood to ascend as
far as they can; and that the actual descent denotes the vertical height
through which the center of gravity descends after the individual
particles have come to rest. Then as well we will be mindful that,
necessarily, the potential ascent is equal to the actual descent when all the
motion remains in the scattered material, and none of it goes over
into unobservable or other type material not pertaining to the system,
and, finally, that the motion of fluids is approximately such that
everywhere the velocity is reciprocally proportional to the corre-
sponding area of the vessel, concerning which we will add certain
other things in the proper place. Now it is fitting to examine the
following proposition.

PROBLEM

§2. If water flows through a conduit formed in any way whatever

and its velocity is known in some place, find the potential ascent of all
the water contained in the conduit.
SOLUTION. Let there be formed in any way whatever the conduit
ST (Figs. 13 and 14), through which part bcfg water flows; it is
 HYDRODYNAMICS, CHAPTER III

assumed, if on the axis ae is taken some point n through which the

plane pm, perpendicular to the axis, passes, that all aqueous particles

,.s

r---------.s
FIGURE 13

existing on that plane will flow at an equal velocity and indeed such
that it is everywhere inversely proportional to the area of the section
pm. Moreover, let the velocity of the water at gJ be such as is due to
the vertical height gs, that is, let the potential ascent of the aqueous

JS.

H
-4

I
V
-- X
1
;o
A
N i

E
FIGURE 14

stratum at gJ be equal to the line qs, and since heights of this sort are in
proportion to the squares of the velocities, it follows that the potential
ascent of the water at pm is equal to the fourth proportional of the
square of the area pm, the square of the area gJ, and the altitude qs,
. (gf)2
mdeed equals (pm) 2 • qs. Thus, with these things having been set
forth, we will assume: that the curve BPG in Fig. 14 is the scale of the
areas of the conduit, so that, with AN = an, NP denotes the area at
pm; hence that the curve HIK is the scale of the potential ascents, so that
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 37

(EG) 2
NI= (NP) 2 ·qs. Now let it be assumed that the individual elements
of the curve HIK have a weight equal to the weight of the correspon-
ding aqueous stratum, and that the center of gravity of that curve
falls at the point L, and let LO be drawn perpendicular to the axis
AE; thus LO will be the desired potential ascent of all the water. From
mechanics, moreover, it follows that if a third curve UX,Z be formed,
2
the ordinate NX of which is everywhere equal to ~~) , LO will be
equal to the fourth proportional of the area AEGB and AE,ZU and the
line qs or EK. Therefore, that which is sought is evident. Q.E.I.
§3. For instance, if there is a conic conduit, the anterior and pos-
terior surfaces gf and be of which have diameters in proportion as m
is to n, the potential ascent of the water will be
3ms
-,------'"----~·qs
n(mm + mn + nn) ·

PROBLEM

§4. Given infinitely small variations, with respect to pos1t10n as

well as to velocity, which correspond to the anterior surface of the
water, find the variations pertaining to the potential ascents throughout
the water.
SOLUTION. Let the area AEGB = M [Fig. 14], the area AE,ZU =

N, qs = v; the potential ascent will be,:. Truly, because the quantity

of water in the conduit is considered constantly the same, the area
AEBG will be invariable, and thus dM = o, so that the differential of
. l ascent is
t h e potentza . s1mp + v dN, an d a 1so d'Ar.
. 1y N dv M b . d firom
Jv 1s o tame

the variation of the position of the water. Therefore, the proposition

is evident. Q.E.I.

ScHOLIUM

§5. These propositions can serve for defining the motion of fluid
moving within vessels, that is, not flowing out, as I shall show in the
proper place; but certainly when the fluid flows out through an ori-
fice, a more appropriate computation is established differently,
namely, as follows.
 HYDRODYNAMICS, CHAPTER III

PROBLEM

§6. Find the difference in potential ascent after a volume element has
flowed out through an orifice.
SOLUTION. Let us consider that water flows out of the vessel aimb
(Fig. 15) formed in any way whatever; let the base im be perforated

- - - - - - - - . ! ' · · · ··· ~

FIGURE 15

by the orifice pl; let the quantity of water remaining in the vessel after
a given quantity of it has already flowed out be cimd; moreover, let
the volume element pnol flow out in an infinitely short instant, with
the surface cd descending to the position ef. Let a section gh be
assumed in the middle of the water, parallel to the surfaces cd and ef
and to the base im itself; and let the velocity of any one of the particles
on gh be such that it can ascend to a height qs or v when the volume
element has not yet flowed out, and to a height qz or v + dv after that
very volume element has flowed out. With all these things having
been set forth, the increment of potential ascent of the water is sought
after the position cimd is replaced with the position eipnolmj, that is,
after the volume element has emerged.
As before, let the curve CGI (Fig. 16) be drawn as the scale of
the areas, where, precisely, CD or EF will represent the area of the
aqueous surface before or after the efflux of the volume element,
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 39

···· -£- ·, Ii ..... U.f . ·- ...... .. ,_,_....

r ·
J [ ..
0 JI
FIGURE 16

GH the proper assumed area, IL the size of the base, PL the size of
the orifice, while the very small connecting parallelogram PNOL
corresponds to the cylindrical volume element pnol. Next, let there
be constructed another curve TRU, the ordinates of which are again
equal to the square of the line GH divided by the corresponding or-
dinate of the curve CGI, to which curve in the same manner is annexed
the small parallelogram LO YX, the side LX of which, certainly, is
equal to the square of the line GH divided by the line PL.
Therefore, now it is apparent that the potential ascent of the water
before the effiux of the volume element is equal to the fourth propor-
tional of the area DCIPL, the area DTUL, and the height qs, and that
the same after the effiux of the volume element is equal to the fourth
proportional of the area FEIPNOL, the area FWUXYOL, and the
height qz:.; moreover, in both analogies the first terms (namely, the
area DCIPL and the area FEIPNOL) are equal to each other; there-
fore, if either one of these areas be indicated by M, the area DTUL by
N, the area FWUXYOL by N + dN, the height qs by v, and qz:. by
v + dv, the increment of potential ascent during the effiux of the
. Ndv + vdN
volume element will be M · Thus, if now it is assumed that
LD = x, FD= -dx, DC= y, HG= m, and PL= n, one will have
mm mm -ydx
D T = - , LX = - , LO = - - (because the area DFEC = area
y n n
mmydx mmdx
LONP ), and hence dN = LOYX - DFWT = - - - + - - ,
nn y
 HYDRODYNAMICS, CHAPTER III

from which now the increment of the potential ascent sought 1s

(x dv _ mm; dx + mm; dx) / M. Q.E.I.

PROBLEM

§7. With these same propositions retained, find the infinitely

small actual descent of the water while the volume element flows
out.
SOLUTION. Since in Fig. 15 the water changes position cdmi for
position efmlonpi, it is evident that in either position the center of
gravity of the portion efmi of the water is in the same place, and
therefore it can be understood that only the small portion cdje (which
equals -y dx, while the total mass of the water equals M) has de-
scended into lonp. Now let the height of the small aqueous particle
cdfe above the volume element lonp be x, and the height of the center
of gravity of the water efmi above the base be b; then the height of
the center of gravity of all the water in the position cdmi above the
base will be b_Y:; · (x - b), and in position efmlonpi the same

height will be ( M ~ dx)b, whence the difference of the heights or

-ydx
the required actual descent equals ------x;r- x, which equation indicates
that the volume element which flowed out is to be multiplied by the
height of the water above the orifice, and the product is to be divided
by the quantity of the water in order to obtain the actual descent which
occurs when the volume element flows out. Q.E.I.

PROBLEM

§8. Determine the motion of a homogeneous fluid flowing out of a

given vessel through a given orifice.
SOLUTION. Since, through our hypothesis, potential ascent at indivi-
dual instants is equal to actual descent, an increment of the former
while a volume element flows out will be equal to an increment of
the latter because it develops in the same short time. Therefore if,
again, the surface of the water, after a given quantity ofit has flowed
out, is set equal toy, the area of the vessel at any place whatever, as
it pleases, is assumed equal to m, the area of the orifice equal to n,
the height of the water above the orifice equal to x, the quantity
N is determined by that law which was indicated in §6, and vis under-
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 41

stood to be the height due to the velocity of the water at the assumed
place where, indeed, the area of the vessel is m, then, through §6, the
.
increment . l ascent w1.11 b e (Nd v - mmvy
o f potentza nn dx + mmv dx)/M,
-y--

and the least actual descent will be -~dx (through the preceding§),

from which there results (x dv - mm:n dx + mrr; dx) / M =

mmvy dx mmv dx . .
-yx dx/M, or N dv - - - - + - - - = -yx dx, which equat10n
nn y
can generally be integrated since the terms N and y are given functions
of x itself, and the term vis of only one dimension.
§9. COROLLARY I. Since the velocities are in inverse proportion to
the areas, it is evident that the height which corresponds to the
velocity of the water flowing out will be mm v; therefore, if this is
nn
. mmMZ~
called z, one will have nnN dz - mmzy dx + - - - - = mmyx dx.
y
§rn. COROLLARY 2. If the orifice is very small in proportion to the
areas of the vessel, then n = o, and the entire equation reduces to this:
-mmzy ~ = -mmxy dx, or z = x; accordingly, therefore, the water
constantly flows out at that velocity by which it can ascend right up
to the height of the uppermost surface, the only case that Geometers
had understood correctly to this time; and this proposition is valid
for all vessels, however formed. But when the orifice is not considered
as infinitely small, by no means is the shape of the vessel to be neg-
lected. Nevertheless, one can observe that, unless the orifice is very
wide, it can be considered as infinitely small without any noticeable
error at all.
§11. COROLLARY 3. When the fluid is not everywhere the same,
the computation is to be undertaken in a similar manner: by in-
quiring, surely, both into the increment of potential ascent of the com-
posite fluid and into the actual descent; and by equating these to each
other. Thus if, moreover, the orifice is very small, it is also intrinsi-
cally evident, as the calculation shows, that the fluid will spring forth
at a velocity due to a certain height such that, if the vessel were
refilled to the same height by the same liquid which flows out, the
walls of the vessel would sustain the same pressure.

GENERAL ScHoLIUM
§12. Before we may deduce rather special Corollaries from our
theory about the motion of fluids from cylindrical vessels, it is fitting
 HYDRODYNAMICS, CHAPTER III

here to examine to what extent the assumed hypotheses agree with

the nature of the matter and what other causes diminishing the fluid
motion, of which we took no account in the computation, could
intervene.
At first, as far as applies to the principle of the conservation of live
forces or of the perpetual equality between potential ascent and actual descent,
I see nothing here which can be a notable impediment to it, if only
we disregard friction, viscosity, resistance of air, and other obstacles
of this sort. But certainly it occurs often that the principle cannot
be applied without limitation, which we shall show in the following;
namely, when the individual particles of water are carried by a dif-
ferent motion, as a result of which it occurs that at every instant
something from the motion, or, if preferred, from the potential ascent,
is lost. But in the present case nothing similar happens, since indeed
almost all the particles are moved altogether similarly, and, especially
when the orifice is very small, the motion of the internal particles is
almost nil, and therefore no detriment can develop from this. More-
over, the other principle, by which it is assumed that the velocity of
any particle whatever is that which corresponds to the inverse ratio of
the area, is indeed affected by a twofold disadvantage: first, namely,
because the motion near the sides of a vessel is a little slower than in
the middle, and therefore all particles corresponding to the same
area of a vessel are not carried at an equal velocity; and second,
because water not greatly distant from the base cannot have that
motion which this principle postulates. However, neither carries a
noticeable error with itself inasmuch as in this simple problem the
internal shape of the vessel is of hardly any consequence to the motion
of the water flowing out. By the same reasoning it is understood that
the motion of water flowing out in some other direction cannot be
very different because, to be sure, the internal motion of the water in
only the lower part of the vessel becomes different, and this dif-
ference can hardly be of any importance. Therefore, it appears that
the hypotheses by which the computation of this Problem of ours is
supported thus agree with the nature of the question; hence no error
perceptible to the senses can arise. But surely the hindrances men-
tioned above, attrition, viscosity of the fluid, and other similar ones,
are of greater importance particularly when an orifice through which
fluids spring forth is rather small, or when the height of the water
above the orifice is very great, or, finally, when a pipe is very slender,
concerning which many experiments are found in the writings of
Mariotte in his Traite du mouvement des eaux. But now I progress to
examining the motion of water flowing out of Cylindrical vessels
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 43

through orifices of any size whatever. We shall also consider vessels

placed vertically by reason of a short cut and a more elegant solution.

CONCERNING THOSE THINGS WHICH PERTAIN

TO THE EFFLUX OF WATER FROM VERTICALLY
POSITIONED CYLINDERS THROUGH ANY OPENING
WHATEVER WHICH EXISTS IN A HORIZONTAL BASE

§13. Geometers who have discussed water flowing from a vessel are
accustomed to consider principally cylinders positioned vertically.
Therefore, it will not be out of place at all to deduce those conclusions
which pertain here from our general theory. Let the area of the
cylinder be to the area of the orifice as m is to n, the height of the
water above the orifice when flow begins be a, the height of the re-
sidual water be x, and the height due to the velocity of the internal
water be v; there will be, in the canonic equation of §8, y = m, and
N = mx ( through §6), which therefore transforms into the following
equation:
m3
mx dv - - v dx + mv dx = - mx dx
nn
or
( I - ::) V dx + X dv = - X dx.

Let this latter equation be multiplied by x-mm/nn, so that there

results

( 1 _ ::) x-mm/nn V dx + xl-mm/nn dv = _ xl-mm/nn dx.

Now this equation can be integrated; but the addition of a constant

in the Integration is to be attended to such that at the beginning of
flow, that is, when x = a, the velocity of the fluid is null, and hence v
itself likewise is null. So indeed there arises:

xl-mm/nn V = nn (a2-mm/nn _ x2-mm/nn)

2nn - mm
or
V = nna [(~) 1- mm/nn _ :a].
2nn - mm x

§14. From this equation, therefore, the height generating the

velocity of the internal water is known; here it deserves to be noted
44 HYDRODYNAMICS, CHAPTER III

that, if the vessel is very wide, it can be directly reckoned that

v = .!!:!!:.... x, certainly after the water descends just a little, that is, as
mm
soon as xis at once a little less than a. This rule fails notably only at
the very beginning of motion, and if that first element of motion is
considered (in which certainly the height a - x can be considered as
infinitely small), the equation indicates that now v = a - x. From
this it follows that in the entire cylinder, whatever the orifice might
be, the internal water is accelerated from the beginning of motion
just like freely falling bodies. But if the motion continues a little
while, then this Rule will err the less, the greater is the orifice and the
higher is the water in the pipe. If, further, that height is desired
which corresponds to the velocity of the water flowing out, which in
§g we set equal to z, there will be
mm
z= nnv'
or
Z = mma [(~)1-mm/nn _ :].
2nn - mm x a
§15. When n = m, that is, when the base is null, it appears from the
very nature of the matter that the water falls and is accelerated freely
in the manner of heavy bodies, which very thing the equation also
indicates; indeed, it occurs in this position that z = a - x. But if
the orifice is considered as infinitely small in proportion to the area of
the vessel, which case we have already considered above, it is to be
assumed that n = o, and then it occurs that z = x, which indicates
that the water flows out constantly at that velocity by which it can
ascend to the total height of the water. Finally, with mm = 2nn,
there develops z = mm (x - x); since nothing can be learned from
0
this value, one must go back to the differential equation of§r3, which
now is this:
x dv - v dx - dx
- v dx + x dv = - x dx, or
xx
--,
X

which integrated, with the addition of the required constant, gives

a
~ = ln ~. or v = x In~, and so Z = 2V = 2X ln -•
X X X X

§16. The velocity of the water flowing out increases at the begin-
ning and afterwards decreases, and is a maximum somewhere,
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 45

namely at that place at which the water descends to the height

I(
mm _ nn) nn /(mm - 2nn>
a nn ; Mariotte, also having learned this through
experience, indicated it in his Traite du mouvement des eaux, Part III,
disc. 3, exp. 5, and the maximum velocity itself is that which is due to
the height

mma [( nn )nn/(mm-2nn) _ ( nn )(mm-nn)/(mm-2nn)]

mm - 2nn mm - nn mm - nn

which quantity reduced becomes

mma ( nn )nn/(mm-2nn>.
mm - nn mm - nn

It is seen from these formulas that the time during which the
velocity is changed from nil to a maximum is clearly imperceptible
when the orifice is moderately small and the pipe is not very long, but
that it becomes noticeable when the situation is otherwise, which we
see in leaping fountains to which water is carried through long con-
duits; but these things which pertain to time intervals will be ex-
plained further in the following section, and at the same time it will
be shown how little water is ejected from very large vessels before it
flows at maximum velocity.

FIGURE 17

The nature of the velocities is better understood from the attached

Fig. 17, in which, if AB represent the total height of the fluid above
the orifice at the beginning of flow, the curves A1CB, A2CB, A3CB
 HYDRODYNAMICS, CHAPTER III

A4CB express the scales of the corresponding heights to which the

fluid flowing out can ascend by its own velocity with different sizes of
orifices. For example, the scale approaches the shape A1CB if the
orifice has a small ratio to the area of the vessel, and [approaches] the
shape A2CB when the base is assumed perforated by a greater open-
ing; and if now the ratio of the orifice is to the area of the vessel as
1 is to V2, that scale will be as A3CB (in which case the maximum
velocity becomes less than in any other, and is expressly that which is
2
due to the height a, by understanding that e is the number the
e
logarithm of which is unity, that is, to a height a little less than !a)
and finally the scale will be as A4CB when there is almost no base
remaining.
§17, But now we will illustrate by a certain example that which
was indicated above in §10: namely, that unless the orifice is very
large, it can be considered in the calculation as infinitely small with-
out very noticeable error, and therefore it can be assumed that z = x,
as was mentioned in§§ IO and I 5. It seems that in the works of many
Authors it prevailed that they reckoned only that no proportion of
size of the orifice is ever to be taken, however great the orifice may be
assumed, which matter is certainly ridiculous; at least up to this time
no one whom I know has considered the size of the orifice correctly
with regard to this matter. Therefore, let us consider a cylinder the
diameter of which is only quadruple the diameter of the orifice, large
orifices of which sort customarily occur rarely in hydraulic equip-
ment; and let us consider that the surface of the water has descended
through only one hundredth part of the entire initial height (indeed
I assume that it has descended some little bit, because at the very
beginning no motion can exist in the water, much less enough that the
water flowing out can ascend by its own motion to the entire height);
these assumptions make m = 16n and mm = 256nn, and x = i 9090 a,
from which there develops

128 [ 99 -
z = 127 100
(_22_)255]
100
a = -2:._ a
IOO '

which certainly differs somewhat from the quantity x or / 090 a, but

nevertheless not altogether greatly, and the difference becomes much
less when the orifice is less and when the surface of the water descends
a little further. Therefore, this Theory differs from the common one
very greatly at the beginning of flow, at which time the motion is less
than was stated; on the contrary, towards the end of flow the water is
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 47
thrust out at a greater velocity than it should be according to the usual
principles.
§18. So far we have considered the motion of water having arisen
from its own gravity; let us now consider: that the water has been
ejected by some outside force apart from the force of gravity; that
such a velocity has been communicated to the water flowing out that
it can ascend to a much greater height than if the gravity of the water
alone had produced the motion; that then that other force suddenly
vanishes; and that the water is left by itself. But if this happens,
experience shows that the velocity of the water decreases very quickly
and soon is such that it is not notably greater than that velocity which
would arise from the gravity alone of the water. Thus we see that it
happens sometimes in leaping fountains (the true cause and the
measurement of which I will discuss elsewhere) that the water leaps
up to a triple or quadruple or even greater height than is customary;
when this happens, that leap stops suddenly, and it does not exceed
the customary height, as far as this can be perceived by observation;
but I am speaking about pipes perforated by not very large orifices, for
when an orifice is somewhat larger, the leap of the water does not
decrease so suddenly. And so, we will examine now to what extent
the theory agrees with these phenomena, and we will include the
accurate measures of them which follow hence. In order that we
may indeed follow the matter generally, we will again consider that
the area of the cylinder is to the area of the orifice as mis to n; that the
water is driven forth at that velocity by which it can surge to the
height a; and that at that same instant the height of the water above
the orifice is a, the gravity of which alone now expels the water; that
then the surface of the water descends in the Cylinder through the
vertical height a - x, so that the residual height is x, and then that
the velocity of the ejected water is that which is due to the height z.
With these having thus been set forth, we will make use of the general
differential equation of §g, which is this:

mmnnzdx
nnN dt - mmzy dx +- -
y
- - mmyxdx

(where again, as was indicated in §13,y is = m and N = mx), which

in our particular case becomes

( 1 - ::) z dx + x dz = - : : x dx,
which, multiplied by x-mm/nn and afterwards integrated accordingly,
 HYDRODYNAMICS, CHAPTER III

so that with x = a one has z = ex, will give the final desired equation

z = ( mm + ~) a<2nn - mm)/nn x<mm - nn)/nn _ mm X

2nn - mm a 2nn - mm
or

If this height is compared with that which was indicated in §14, the
x)<mm - nn) /nn
excess of one over the other is found to be ( a
(X; whence

all these Phenomena are now confirmed which were just indicated;
indeed, when the number m is much greater than n, that excess im-
mediately becomes unnoticeable after the water descends just a little
bit, that is, after a very short time, but nevertheless all of it never
vanishes as long as the flow endures; and, finally, it is continuously
more notable, the more the ratio of the number m to n approaches
unity. For instance, let the diameter of the pipe be ten times
greater than the diameter of the orifice, and let the water be expelled
by such a force that by its own velocity it can spring up to a height
which is quadruple the height a, or of the water above the orifice; it is
sought to what height the water flowing can ascend by its own velo-
city after the aqueous surface has descended in the pipe through a
thousandth part of a itself, if at the same time the water is stimulated
to effiux through its own gravity alone, thence what the similar height
would have been if the water had had no motion at the beginning.
Therefore, m = 100 n, mm = 10,000 nn, x = / 09090 a, ex = 4 a, from
which in the former case one obtains

z= 10,000
[ 9998
(-222._ _ ( 999 ) 9999 ) + 4(-2filL)9999] a
1000 1000 1000
or
z= 99,915 a + 18 a .
100,000 100,000 '

but in the latter case it becomes

1
Z = 99,9 5 a
100,000 '

from which example it is evident how small and clearly unnoticeable

is the excess of the former height above the other, and how suddenly
that aqueous thrust is diminished since, indeed, the entire change
occurs while the surface of the water descends through a thousandth
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 49

part of the height a, which time in customary hydraulic machines

cannot be other than very short. Thus what was stated above in §17
is confirmed as well, that certainly z = x approximately when the
orifice is rather small, since in the present case, wherein the motion
begins from rest, the difference between z and x is only fifteen hun-
dred-thousandths of the height a itself; since in the meanwhile the
height z is a little greater than x, it is evident that the water flowing
out can ascend to a greater height even after the water has flowed out
for some time, than the height of the water above the orifice.
§19. Thus, inasmuch as we have deduced from our general Theory
those things which attend the motion of fluids from cylinders placed
vertically, we will now also consider pipes placed obliquely, which are
customarily very long in leaping fountains. In these, certainly, it is
singular that the acceleration of the motion does not occur as suddenly
as when the Cylinders are vertical, and thus by observation one may
perceive the accord of our Theory with the actual motion of water.
§20. Let us consider a conduit curved any way whatever, but
nevertheless Cylindrical, the area of which again is in proportion to
the area of the orifice as m is to n. Let the motion begin from rest,
and let the vertical height of the water above the orifice at the
beginning of motion be a. Let a certain quantity of water have
flowed out, and let the vertical height of the residual water above the
orifice be taken as x; let the length of the conduit, which at that very
moment is full, be g, and then let the internal water (the individual
particles of which I assume here to be carried in a motion parallel to
the axis of the conduit) have a velocity which corresponds to the
height v. Thus, with these things having been set forth, if we make
use of a reasoning similar to the above for seeking indeed the incre-
ment of potential ascent while the volume element flows out, as we did
in §6, and by assuming the same equal to the actual descent, the follow-
ing equation is obtained:

mm
tdv - -
nn
vdg + vdg = - xdg,
or

( I - ::) V dg + g dv = - X dt;

the integral of this, which is evident after the terms have been multi-
plied by g- mmtnn, is this:

V = gmm/nn-1 J_ X g-mm/nn dg.

50 HYDRODYNAMICS, CHAPTER III

If there exists, for example, a straight conduit so inclined to the hori-

zontal that the sine of the intercepted angle between the two is to the
total sine as I is to g, there will be g = gx, from which

v = nna [(~)(nn-mm)/nn _ :a].

2nn - mm x

Since this equation does not differ from the equation given in §13 for
vertical Cylinders, it follows that in each case the velocities of the
water are the same when the vertical descents of the surface of the
water are the same. Therefore, similar accelerations in homologous
places on either hand are in proportion to the vertical heights, and
only this distinction occurs, that in an inclined conduit it happens
more slowly, and in proportion as I tog; therefore, these accelerations
can be perceived easily by observation in greatly inclined conduits
which cannot be [perceived] in vertical ones on account of the ex-
cessive speed of the changes. On the other hand, it is intrinsically
evident from the fact that the frictions are increased by the length of
the pipe that it cannot be that the velocities are not diminished, to
which those should attend in whom there will be a desire to undertake
experiments on this subject.

CONCERNING THE EFFLUX OF WATER FROM

VERTICALLY POSITIONED CYLINDERS WHICH
TERMINATE IN OTHER NARROWER AND
SIMILARLY VERTICAL PIPES

§21. Experience shows that between two Cylinders wholly equal

and similarly positioned, for one of which a rather narrow pipe corre-
sponds to the orifice of the other, that the one is depleted more
quickly which has the pipe attached, and indeed, the quicker it does
so, the more the pipe increases in size from the place of insertion to-
ward the extremity, which Mr. s'Gravesande showed to many in
Physices Elementa Mathematica, lib. 2, cap. 8. Let us consider the entire
matter in the following Problem.

PROBLEM

§22. Let there be a cylindrical vessel AEHB (Fig. 18) positioned

vertically, perforated at FG, by which opening it connects with the
conic tube FMNG, through the orifice MN of which water finally
flows out. The velocity of the aqueous surface CD is sought after it
descends from rest through AC or ED.
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 5I

SOLUTION. Let the initial height of the water above MN, namely
NG+ HE, be a, the height of the aqueous surface at the position CD
above MN, that is, NG + HD, be x, the length of the annexed pipe,
or NG, be b, the area of the orifice MN be n, the area of the orifice

.A B

o iM .

FIGURE 18

FG beg, the area of the upper Cylinder be m; let the velocity of the
aqueous surface at CD be that which is due to the height v. In the
general equation of§8 there will bey = m and N = m(x - b) + ~~m,
vgn
which substitutions, it will be evident, are conformal to the derived
calculation, since in §6, moreover, the remaining conditions are the
same as before. Therefore, the equation of §8 is resolved to
m 3 v dx
m(x - b) dv + .bmm
1_ dv - -- + mv dx = -mx dx
v gn nn
which, further, divided by m, and with one's having established
mb
X - b + . ;- = Z,
vgn
 HYDRODYNAMICS, CHAPTER III

( 1 - : :) v dz + z dv = - z dz - b dz + ":5;z
which, multiplied by z-mm/nn, yields

( I _ ::)z-mm/nn V dz+ zl-mm/nn dv

mbz-mm/nn dz
-zl-mm/nn dz - bz-mm/nn dz+ v- ;
gn
after the integration of which, with the constant C having been added,
there arises
zenn - mm)/nn V = C _ nn ze2nn - mm)/nn
2nn - mm

-- nnb mnnb
- - zenn- mm)/nn + ---------,,= z (nn - mm)/nn
nn - mm (nn - mm) Vgn
in which the value of the constant quantity C is defined from the fact
that at the beginning of flow, when indeed x = a or z = a - b +
mb b . . . . .
• _ , v = o, ecause mot10n cannot anse in an instantaneous point o
f
1
vgn
time; hence, therefore, it occurs that
mb)
C= [( a- b +-- nn
Vgn 2nn - mm
+ nnb Vgn -
mn!:!] (a _b +mb-) (nn-mm)/nn.
(nn - mm) Vgn v'gn
From these equations, indeed, all things are defined; but because the
calculation is more or less involved unless the area of the upper
vessel, indicated by m, be so great that it can be reckoned as being
infinite in proportion to the areas g and n, we will consider this case
alone, and this the more so because a notable error does not arise
from it, even if the number~ or~ be of moderate size.
n g
§23. Thus, if hence we set m = oo and at the same time make use
of the first differential equation of the last paragraph, and if in this
it is assumed that v = !!!!:._ s, so that thus from the value of the letters
mm
the height may be found to which water flowing out through the
orifice MN can ascend by its own velocity, first there will be
nn bnn
- (x - b) ds + ---= ds - ms dx + -nn s dx = -mx dx,
m Vgn m
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 53

and because m = oo it is easily foreseen that the ratio betweens and x

will be finite, and also between ds and dx, this same equation will be
changed, after one's having rejected the appropriate terms into this
again: -ms dx = -mx dx, or s = x, which was already proven as
well in §rn. But after that I decided to prove it again here because
the present case could be seen as different from the former, about
which there is a discussion in the aforesaid paragraph. These things
having been understood, it is no task to explain to many the Pheno-
mena in §21 concerning this matter indicated.by the Author s'Grave-
sande; for it is evident that the water does not flow out through the
composite vessel AEFMNGHB otherwise than it does through the
simple vessel AOMNPB when, indeed, the orifice MN is very small;
and that hence the velocity of the aqueous surface CD is greater than
if the water were flowing through the vessel AEFGHB, after one has
set the orifice MN = FG, and much more so if MN is greater than
FG, which happens when the pipe increases in area towards the
lower end. But, nevertheless, it must be observed that at the begin-
ning of motion the water descends more slowly than has been thus
defined, and that that rule does not hold until the surface CD has
descended through some little space, which, however, occurs in a
short time. We will examine the changes which occur in this case at
the beginning of motion in the following section.
§24. The computation would be undertaken in the same way if in
the vessel, which now we always consider to be of infinite area, were
implanted a small pipe, not vertical but horizontal, just as in Fig. 19,
or in any other direction whatever; moreover, it is always found that
the water, after the surface of the water in the principal vessel de-
scends some little bit, soon flows out at approximately that velocity
which corresponds to the height of that surface above the orifice.
Therefore, it is clear that, with the height of the water above the pipe
GN as well as the orifice FG itself being maintained, the quantity of
water flowing out in a given time is increased by the increased area
of the orifice MN. Accordingly, therefore, we have given a descrip-
tion here of what was mentioned at the end of §5, Chapter I: that
Frontinus had learned from experience, certainly, that more water than
is due is appropriated through a calix of both legitimate size and position to
which pipes of larger size are direct!)! attached. And certainly quantities
of water, all other things being equal, would be expended, approxi-
mately proportional to the orifices MN themselves unless there should
be many hindrances; the latter may diminish this quantity greatly,
about which I will speak soon. These hindrances can act so that
the flow of water is increased very little on account of the increased
54 HYDRODYNAMICS, CHAPTER III

final orifice; nevertheless it will always be increased some little

bit.

R Q
FIGURE 19

§25. From the previous discussion it is evident that the velocity by

which the surface CD of the water in either case about which we spoke
descends, the remaining things being equal, depends on the area of the
orifices MN. Moreover, these things are supported by the hypo-
thesis that the water adheres everywhere to the walls of the pipes GN,
and it flows out from the full orifice MN, which hypothesis cannot hold
further if that orifice should be increased too much. Hence also it is
evident when water flows out through the vertical pipe in Fig. 18
that its flow is accelerated by an increased length of this pipe; never-
theless the latter can also be increased so that finally the water ceases
to be continuous in the pipe, in fact, so that it is rather divided into
columns, which may occur if the pipe has a length of more than
thirty-two feet or even less if at the same time it increases in area
toward MN. Thus if the orifice MN is double the other orifice FG,
the length cannot be greater than eight feet without danger of separa-
tion of the water following in the uppermost part of the pipe, which
matter I will show elsewhere. But there is an additional cause
besides the excessive length of the pipe which can produce separation
of the water, namely, that the height of the water CEHD be less than
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 55

that which can enter the pipe quickly enough, by which it occurs that
air together with water flows in above at the same time, while the
surface of the water assumes the form of a cataract or of a bottomless
hollow, such that not all of the orificeFG is covered by water. Indeed
this causes the water to flow out in a lesser amount, but not at a lesser
velocity, which a certain Italian Author, Carlo Fontana by name,
considered later; he wrote the following about this matter in his own
vernacular Language: "But if here there were not," he says, "as much
water as would be sufficient to maintain the said pipe full, the water will
attract air within itself in as great a quantity as water will be lacking to it for
intermixing within the water on all sides; but the velocity ef the water will be
lacking as much as will be the height ef all the air collected together that will be
in that pipe." Anyone discerns the reasoning of this, because I stated
that the velocity of the water can hence not be diminished, from the
fact that otherwise the potential ascent could not be equal to the actual
descent, and the matter will be confirmed easily by experiment, with
the extremity MN of the pipe being bent so that the water flows out
horizontally, and from the area of the jet the velocity of the water can
be determined. Moreover, it may occur in any manner it pleases
that, with none of the other conditions changed, the air is mixed with
the water around the top of the pipe; thus remember, indeed, if there
is a tiny opening in the pipe not at all far from the orifice FG (Figs. 18
and 19), and if, further, during the flow of water one has closed that
little opening with a finger, pure water will flow through, and if one
removes his finger, soon air will enter through the same little opening
and will mix itself with the water flowing through. These things
having been understood, it will be easy to present the reasoning of
the Phenomena which are observed in chimneys, or smoke ducts;
indeed smoke seeks height, because it is lighter than air, which is
consistent with experiments performed on smoke in a vacuum, where
it was seen to have descended. Therefore, it is the same for smoke
ascending as for water descending; but in Fig. 18 the latter flows
through the orifice MN more quickly, the larger it is, and the lower it
is positioned; therefore, also, the smoke will travel through the
chimney more quickly, the more the fire is kindled in the furnace,
the higher the chimney is carried, and the more it diverges facing
upward, if only it does not diverge too much; experience confirms
each of these. I myself then learned in addition that if the chimney
be perforated somewhere, it is not at all so that the smoke attempts an
exit through that opening, but rather that air rushes in with a great
impetus, and, mixing itself with the smoke, it rises through the
chimney, and not otherwise than as we indicated that the air rushes
 HYDRODYNAMICS, CHAPTER III

into the pipe FGNM (Figs. 18 and 19). So indeed the smoke ascends
certainly in a lesser amount, or at least with more difficulty, and the
fire slackens.
Still, there are two causes in particular, the one extraneous, the
other intrinsic in the nature of the matter, which can greatly retard
the motion of the water in Figs. 18 and 19. The first is the adhesion
of the water to the walls of the pipe, and the other is that when the
pipe increases in area, the velocity of the water, nowhere constant to
itself, is changing in every location in the pipe; if this change is con-
sidered to arise from infinitely small impulses of water moving more
quickly into water moving less quickly, it appears that at every in-
stant by these impulses of flexible bodies some of the potential ascent
is lost, whence necessarily the effiux of the water is noticeably
diminished.
§26. Finally, now I will say something about curved vessels from
which not all the water flows: for the sake of brevity we will consider
a cylindrical conduit, a certain part of which, that the aqueous surface
does not cross, is straight.

PROBLEM

For example, consider the cylindrical conduit CEDE (Fig. 20), the
sufficient portion CE of which is straight, the remaining EDE being

FIGURE 20

curved in any way whatever; the whole conduit is full of water that
will flow out through the orifice E; after the surface of the water has
fallen from C to F, the height corresponding to the velocity of the
water at Fis sought.
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 57

SOLUTION. Let the vertical BH and the horizontals CH, FG, AB be

drawn, and let the sine of the angle HCE be to the total sine as I is to
g. But now, ifwe consider the matter properly, we will see that the
present problem is contained in the other more general one which
we treated above in §20, where we had this equation:
V = gmm /nn-1 I_ xg-mm /nn ds

where, for our present case, v represents the desired height corre-
sponding to the velocity of the aqueous surface at the position F,
g the length BDEF, x the height BG, and also~ the index of the ratio
n
between the areas of the pipe and the orifice B. But if the length
BDA is set equal to a, one will have x = g - a, from which now
g

V = gmm /nn-1 J_(S; a) g-mm /nn df

Let the length of the entire conduit BDEC be indicated by {3, and
one will have

J- (g - a) g
g-mm /nn ds = nna
g(nn - mm)
(g<nn-mm)/nn - f3<nn-mm) /nn)

nn (s<2nn-mm)/nn _ f3<2nn-mm) /nn)

g(2nn - mm)
and therefore
V = nna (r _ (~)(nn-mm)/nn)
g(nn - mm) 1,

-,--_nn_q-,- ( I _ (~) (2nn - mm)/nn) • Q.E.I.

g(2nn - mm) \,

ScHOLIUM

§27. Since these equations are somewhat involved, we will not

tarry in the general contemplation of them, considering, rather, those
particular cases which shorten the calculation and which cannot be
defined by that last equation.
If we assume that the cover at B is wholly absent, it occurs that
m = n, and (which must be determined separately for this and equally
for the other case soon to be discussed)
b - g + alns - alnf3
V = ----------'-
g
58 HYDRODYNAMICS, CHAPTER III

and then the velocity is a maximum at A, and expressly that which

corresponds to the height
f3 - a + alna - alnf3.
g

Finally, the point E corresponding to the maximum descent 1s

obtained with the aid of this equation:
t - alnt = f3 - alnf3.
The other case 1s to be calculated separately, when mm = 2nn,
where there arises

V = - - - -+-tf3lnf3
at-- -af3--- tf3lnt --
gf3
and also, if it is considered, e having been assumed as the number of
which the logarithm is unity, that t = e<«-fJll fJ {3, then the point of
maximum velocity will be determined, of which the generating height
is e<« - fJl l fJ f3 - a, while the maximum descent, which is proportional
to the total water flowing out, is defined by making
at - af3 - tf3lnt + tf3lnf3 = o.
I do not doubt but that these would correspond to practice exactly,
if only the adhesion of the water to the walls of the pipe would not
retard the motion; nevertheless, I consider that the results of the
experiments can be such that they show the truth of these propositions
well enough to the intelligent person, who has an understanding of
these impediments.
§28. Finally, I will show the correct solution of a certain phenom-
enon which at first glance seems to be very much a paradox.
Indeed, after it appears from all these things freely discussed up to
now that it cannot happen that the water flows out at a much greater
velocity than that which is due to the height of the water above the
orifice (nevertheless they can be somewhat greater, especially if the
orifices are large; refer to what I said in § I 6 concerning maximum velocities),
it will seem to many perhaps a wonder that it occurs sometimes in leaping
fountains that for an instant water makes a far higher thrust than seems
possible according to our rules. It is far from true that these [rules]
therefore lose some of their power; in fact, they are rather exceedingly
strengthened. Moreover, the solution of the paradox consists in this:
so far we have considered the water as continuous and not separated
by any air void; and Mr. de la Hire rightly observed that irregular
spurts of this sort do not occur unless air together with water has
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 59

entered the pipe somewhat like bubbling water, which, as I indicated

in §25, occurs frequently. Indeed, that air is carried together with
the water right up to the orifice of effiux, through which it then erupts.
While this occurs, the aqueous mass acquires an impetus which it
employs exclusively in expelling the water, and in this way it produces
an enormous thrust. Soon I will explain this cause of the phenom-
enon more clearly together with the required measurements, after
I have presented some statements which appear concerning this
matter in the History of the Royal Academy of Science of Paris for the
Year r 702. It states in the place cited: "It is seen sometimes that water
discharging through an orifice springs forth three or four times as high as the
height of the reservoir would permit, and that it also comes back quite quickly to
the height which the laws of hydrostatics prescribe to it. But how could it
deviate from it for an instant? Mr. de la Hire attributes this to the air en-
closed in the conduit, which, being compressed by the continuously descending
water and thus gaining a spring force, is released against the rising part of the
water and imparts to it this instantaneous velocity."
And so Mr. de la Hire noticed correctly that the spurt is due to the
air, and there is no doubt but that he would have extracted the
correct reason by which air can produce this if he had considered
more carefully the phenomenon to which he referred incidentally, and
he would have easily perceived, certainly, that the air within the water
sustains no pressure except that of the water lying above (on the
contrary, not even this much in flowing water, as I will show below in
Chapter XII), and that therefore the compressed air cannot expel
the water preceding it more strongly than if water had been in its
place. I certainly saw in advance (which I found often afterwards
by very simple experiment) that it is not the water located in front of
the air but that which follows the air that rises unusually high, which
I will now show more clearly.
Therefore, in Fig. 20 let the aqueduct CADE be cylindrical, as is
customary, and let the whole of it be filled with water, except the
small part mnB filled with air. Let the horizontal and vertical lines
CH and HE be drawn; let us assume for the sake of brevity that the
gravity of air can be considered as null with respect to the gravity of
water, so that the transition of the air through the orifice B offers no
resistance to the flow of the water, although concerning the remaining
it would be easy to take into account the inertia of the air, unless we
wish to avoid an abundance of calculation in a matter in which we
require no prec1s10n. Let the length of the conduit CADf or CADm
(indeed we assume the differential nif filled with air to be very small)
be fi; mf or ng be o, HE be a, the area of the pipe be m, the area of the
60 HYDRODYNAMICS, CHAPTER III

orifice B be n. Finally, let us state that the water has no motion when
the surface is at mn; that the height due to the velocity which the
surface mn has when it arrives at the positionfg is to be sought; let
that height be v, and the potential ascent of all the water at that very
instant will also be v; moreover, the actual descent is, through §7, equal
to the fourth proportional with respect to the total mass of water, the
volume element of water mngj, and the vertical height HE, that is,
8 8
"ff a; therefore v = ~ a. Indeed, this height is at once diminished
faster than stated, and the water is forced to flow through the orifice
B, which I showed in §18; but, nevertheless, at the first instant the
water will retain the motion which it acquired, and thus the volume
element closest to the orifice will be ejected at a velocity which is due
to the height mmf3S a. However, this height can be not only triple or
nn
quadruple a itself, but howsoever great; indeed, with glass tubes I
created thrusts ten or twenty times as high as a itself at will. For
instance, if 8 = 100 feet, (3 = one inch, but the diameter of the
tube is tenfold the diameter of the orifice, then one will have
nnf3 = 10,000 a, so t hat m
mm8 . t hese circumstances
. t he fi rst vo 1ume e1e-
1200
ment must spring forth, with the resistance of the air removed, to a
height more than eight times as great as the customary height a. In
addition, there are many hindrances, and these are of greatest im-
portance, which restrain huge thrusts; indeed, something from the
motion is lost by the impulse of the aqueous surface mn against the
wall jg, then also by the enormous friction which the water experi-
ences, having been carried so quickly through the little orifice, which
has to be very small; much is also lost from the fact that the water
CADm is not moved with all its velocity on account of the adhesion of
the water to the walls of the pipe, which adhesion is clearly noticeable
in so long a reach.
Meanwhile, there can be no doubt that this is the correct solution
of the phenomenon, and the experiments which I performed satisfy
that solution in every extent. Then, as well, by this theory the other
aspect of the phenomenon is solved correctly, namely, that that thrust
is quasi-instantaneous, and after the shortest little time interval it is no
greater than usual, according to observation. Thus in the present
case that we just considered, if, with the rule of §18 changed a little
(for there the only case discussed is that concerning vessels placed
vertically), we investigate how much water must flow out in order
that the thrust does not exceed the customary thrust by more than a
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 61

thousandth part (which in any case can by no means be observed in

experiments of this sort), ifit was eight times as great as the same at
the start, we find that that quantity is so small that the time in which
the whole ofit is ejected can in no way be perceived.

EXPERIMENTS WHICH PERTAIN TO CHAPTER III

FOREWORD

Indeed there are many things in this Chapter, and these quite
extraordinary, which can hardly be subjected to experiments imme-
diate!;)I. And indeed, since Authors up to now have not considered
any motion in the efflux of fluids other than that which occurs through
very small orifices, and accordingly, since our theory which we gave
for arbitrary areas of orifices is new, this is the very thing the con-
firmation of which should be most gratifying. But I do not see in
what way in vertical Cylinders, which we treated the most, the
velocity of the water flowing out can be observed, especially when
the orifice is very large (indeed, on the contrary, some judgment of the
velocities can be made from the time of depletion). Thus, consider-
ing this, I reasoned at last that §§16 and 20 could be useful to our ob-
jective: in the former the maximum velocity of water flowing out of
cylinders placed vertically had been determined; in the other, more-
over, it was shown that the motion is the same from obliquely placed
and vertical cylinders if both vertical heights are assumed alike.
Therefore, we will make suitable use of cylinders placed obliquely in
order that from the maximum area of the aqueous thrust the maximum
velocity of the water or the height due to the same can be obtained by
experiment; and indeed, by this reasoning that maximum velocity,
whatever it really is, can be investigated, even if the orifices are as
large as one wishes. Accordingly, if this is observed to agree with our
rules, no doubt can remain regarding the entire theory. But before
I attack the matter itself, the mechanics theorem that follows is to be
set forth in advance.

LEMMA

Let the line AB (Fig. 21) be vertical, ED horizontal; in addition

let the line AD have any direction whatever, in which direction a
body at A is understood to be projected, describing the arc AC of a
parabola, the tangent of which at A is certainly the straight line AD;
the height due to the velocity at which the body at A was projected
 HYDRODYNAMICS, CHAPTER III

FIGURE 21

2 2
. (BC) · (AD)
will be AB-BD-CD; and if AD is horizontal, or BAD is a right
4
angle, that same height will be (!if· But now I will show those
things observed by me.

CONCERNING THE MAXIMUM VELOCITIES OF

FLUIDS FLOWING OUT THROUGH VERY LARGE
ORIFICES
PERTAINING TO §§16 and 20

FIRST EXPERIMENT. I placed the Cylindrical Pipe FA (Fig. 22 ), of

a length of four inches, obliquely to the horizon, and I secured it in
that position; moreover, the area of the pipe was to the area of the
opening at A as 2 is to 1, and the diameter of the pipe was equal to
seven lines, more or less. Then, after measurements of the lines FE,
AB and BD had been taken in equal units (the law of which is evident
intrinsically from the figure itself), I found them to be 81, 619, and
74o.
With these things so prepared, I filled the pipe with water, having
closed the orifice A meanwhile with a finger, and with this suddenly
removed, all the water flowed out in the shortest bit oftime; however,
I was able to observe that the first and the last [portions] had fallen
nearer to the vertical AB than the intermediate [ones]; moreover,
that the drops projected the furthest fell at the point C. And I found
after rather frequent repetition of the experiment that BC was 235
of the units which I had used before.
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 63

But now, if through the previous lemma the height EC is desired to

which the drops ejected at the maximum velocity can ascend, it is
found that EC = 56 units; however, by dint of§§r6 and 20 it should
be 62, unless the friction of the water and its adhesion to the sides of
the pipe contribute an impediment to the motion; I did not expect
a greater agreement.

.................. :1:1

FIGURE 22

II. With things arranged as previously, only with the orifice A

diminished to half so that the area of the pipe was quadruple the area
pertaining to the opening, I observed that BC = 252. Hence it is
deduced through experiment that EC = 68, but by theory it should
be = 70; these numbers differ less than the preceding because here
friction was a far lesser impediment on account of the diminished
velocity of the internal water. However, each experiment actually
confirms the theory excellently.

CONCERNING THE VELOCITY OF WATER FLOWING

FORTH FROM A VERY LARGE VESSEL

PERTAINING TO §17
In that paragraph we say that if a vessel is very large, soon after
the internal surface descends some little bit, water flows forth at a
 HYDRODYNAMICS, CHAPTER III

velocity which corresponds constantly to the height of the water

above the orifice. However, one may allow that the water flows in
any direction whatever (for indeed, in very large vessels any direction
of the stream cannot change the velocity), and one may observe at
any arbitrary instant at what distance from the vertical the stream
impinges on the horizontal, and from there one may seek through
the previous rule the height corresponding to the velocity of the water
flowing out at that instant; thus one always finds that height equal to
the height of the water above the center of the orifice, if only one
overlooks the first few drops which, by dint of§16, must flow out and
actually do flow out at a lesser velocity. And the hindrances, which
we have mentioned rather frequently, will cause no noticeable delay
to the flow if only the diameter of the orifice equals at least two or
three lines and the diameter of the vessel itself is not less than a few
inches, and, finally, the height of the water is not excessive, such as
very many feet.
I tested all this often, but the nature of the experiment is too
trivial for it to merit being described fully.

CONCERNING VESSELS WHICH ARE PROVIDED

WITH VERTICAL PIPES

PERTAINING TO §§22 and 23

Concerning these things the honored s'Gravesande, in Physices

Elementa Mathematica, undertook experiments which I repeated; in
fact, those which apply to the present matter are brought out es-
pecially at this point.
Namely, in Figs. 23, 24, 25 and 26, the individual apertures
denoted by the letter A are equal to each other, with B alone being
little greater, in the proportion of 16 to 25, and also the areas as well
as the heights of the cylinders are equal, except for the last, the height
of which is quadruple; however, the pipes annexed to the two inter-
mediate cylinders have triple the length of the cylinders. Therefore,
after these vessels were filled with water, it was observed concerning
its efflux:
I. That the surface of the water from the beginning does not
descend more quickly in Fig. 23 than in Fig. 24; but that, after some
water has flowed out from each, the motion becomes much quicker
in the composite vessel than in the simple; I predicted both at the
end of§23. But the matter is understood better and more accurately
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 65

from the differential equations which we gave in §§22 and 23 if we

make use of them for finding the first increments of motion in the
simple cylinder of Fig. 23 as well as in the composite one of Fig. 24,
and if to this end we assume the areas of the cylinder and the pipe to

LJ

A
A B
FIGURE 23 FIGURE 24 FIGURE 25 FIGURE 26

be as m to n, the increment, which we called dv in the simple vessel,

will be to the increment in the composite vessel as 1 + 3nm is to 4, and
therefore greater by far in the former case than in the latter. There-
fore, ifit should be granted to perceive that first motion correctly, we
would observe at once that that is quicker which occurs in the simple
Cylinder. Since, in fact, in §§15 and 23 it was demonstrated further
that the water surfaces, after they have descended a little in each
vessel, are approximately such that they correspond to the heights
!!:!!_ x, by understanding by x the heights of the water above the orifices
mm
through which it flows; it follows directly that the water descends at
a much greater velocity in Fig. 24 than Fig. 26. Thus, therefore, the
Theory clearly agrees with the observations.
II. That the aqueous surface descends considerably more quickly
in Fig. 26 than 24, so that the velocity in the case of Fig. 24 is some-
what halfway between the cases of Figs. 23 and 26. Here, indeed, it
is evident again that the first accelerations occur much more slowly
in the cylinder of Fig. 24 than that of Fig. 26. Therefore, in this
respect the theory itself indicates what was observed; certainly some-
thing is missing that as great a difference as I found can hence de-
velop, and it should no longer be noticeable after either surface
66 HYDRODYNAMICS, CHAPTER III

descends a little, according to §23. However, that difference must be

attributed to a hindrance which arises from the friction of the water
in Fig. 24; indeed, the water is carried at a great velocity through the
pipe AA, and thus as much on account of the increased velocity as on
account of the diminished area of the vessel a very effective impedi-
ment is offered to the motion of the water.
III. Finally, that the aqueous surface descends very quickly in the
Cylinder of Fig. 25, if one excludes the first instant, and notably more
quickly than in Fig. 26.
Indeed, this conforms to those things which have been shown in
§23; however, soon after the common beginning of motion, namely,
after the heights of water above the orifices of effiux have been set
almost equal, the velocities in Figs. 25 and 26 should be approxi-
mately as the areas of the orifices B and A, that is, as 25 to r 6; and
since a smaller difference of velocities is observed, this is again to be
attributed to the impediment of friction more than to the other cause
indicated at the end of§25.

CONCERNING THE SAME VESSELS, IN WHICH

HORIZONTAL PIPES ARE INSERTED

PERTAINING TO §24

When water flows from a rather large vessel such as CDG (Fig. rg)
through the horizontal pipe GM larger at the extremity NM than at
the origin GF, the former is carried through the orifice GF at a greater
velocity (if again one overlooks the first drops) than if the pipe were
either absent or Cylindrical. Even Frontinus, taught by experience
without doubt, affirmed this, but several modern men have denied it.
Therefore, as something worth the effort, I undertook to investigate
the matter by experiment. Now the height of the vessel which I
used was 5¼ English inches above the axis of the pipe, the length of
the pipe GN was 2 inches 5 lines, the diameter of the orifice GF was
3.36 lines, the diameter of the aperture MN was 5.48 lines; thus the
areas of the orifices were approximately as 3 to 8; the area of the
vessel was large enough that it could be considered infinite with
respect to the area of the pipe. I have wished to dispatch all measure-
ments so that anyone can repeat the experiment. Now, after this
vessel had been filled with water, I observed the area of the jet, and
from this, after I had acquired all the measurements needed, I made
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 67
a calculation of the height which should be due to the velocity of the
water flowing through at CF as well as at NM; I found the latter to
be approximately eleven lines, and thence the other to be I inch
6f lines, which same heights I found as well in another type of ex-
periment. But since the height 6 inches and 6} lines is greater than
5¼ inches, our theory is confirmed about the acceleration of internal
water by amplification of the pipe towards the extremity, although,
as I predicted chiefly because of the two reasons dispatched in §25, it
may be far from actually being accelerated as much as it should
according to §24, after the obstacles have been subtracted which have
not been considered in the calculation.

PERTAINING TO §25

In this paragraph I mentioned in passing that it can occur in many

ways that air is mixed with water flowing through pipes. But from
this it will happen that water flows out in a lesser amount certainly,
but not at a lesser velocity; in order that I might prove one as well as
the other, I first made a very small orifice in both the pipes AA and
AB (Figs. 24 and 25) not far from their Origin; it is a fact that the
water was carried through the pipes with some noise and flowed out
in a turbulent state; moreover, the surface descended much more
slowly than is customary. Then I perforated the pipe of Fig. 19
somewhat similarly, not far from G, and again I observed that the
internal surface descended a little more slowly, of which matter I was
certain, since I counted the oscillations of a certain pendulum while
the surface descended through a given length; but with an under-
standing of the flowing out of the water, I saw that sometimes the
water flows out from a full orifice and then the water is less clear than
usual, but it makes an ordinary jet or one greater by a little bit than
ordinary; however, most often the water and air are carried side by
side, the former in the lower part of the pipe along the wall FM, the
latter in the upper, along GN, and then the water is clear and is
ejected at a velocity not only by no means smaller than the usual but
even much greater; I had foreseen that this could occur by no means
obscurely. Concerning this matter, in the following Chapter I will
discuss another experiment undertaken with greater precision.
However, a place will probably be given elsewhere for showing that
water mixed with a sufficient quantity of air flows out almost at that
amount at which it would flow from a pipe cut off at that place where
it is perforated, to which matter I noticed that my own experience
corresponds also.
68 HYDRODYNAMICS, CHAPTER III

CONCERNING CURVED CONDUITS

PERTAINING TO §27

The horizontal MN (Fig. 27 ) having been drawn on a wall, I

placed the cylindrical pipe CDB, having both legs parallel to each

H .r. .
/';,::~~,
't ! '

---·-' --- N

FIGURE 27

other, wholly filled with water, so that the one extremity B would just
touch the horizontal MN, and also that the legs would be vertical,
while at the same time I blocked the orifice C with my finger, thus
restraining the flow of water.
Then I observed, after the finger had been removed, the maximum
height BP to which the water flowing out ascended, and at other times
I noticed the point E to which the surface of the water descended;
however, I performed the experiment under two different conditions;
indeed, in the first instance I had not placed a cover at B, then I used
a cover perforated by such an opening that it had an area in ratio to
the area of the pipe as I to V2. Meanwhile, the measurements were
such: CA = 345; ADE = 530; BP= 33; and AE = 88 units, 375 of
which were equal to the length of a London Foot. Things were so in
the prior case, but in the other, with the rest of the things being un-
changed, I observed BP = 64 and AE = 54. I will note here in
passing that, desiring to determine the maximum descent AE in
another way, I inclined the pipe after the experiment was ended until
the water seemed just precisely at efflux through B, at which instant
 VELOCITIES OF FLUIDS FLOWING OUT OF A VESSEL 69

I measured the distance of the surface from the point A noted pre-
viously; that distance which I considered to be the same as the
maximum descent AE was far less than expected; whence I learned
that a part of the water which had already flowed out through B in
the experiment had entered the pipe again.
Thus, after those things had been observed, I sought the magni-
tudes of BP and AE by calculation, according to §27, by setting
first m = n, and then mm = 2nn; but I found in the former case that
BP = 79, which in the experiment did not exceed 33, and I dis-
covered the maximum descent AE approximately equal to 250, which
the experiment gave as 88. Next, for the case of mm = 2nn, BP
appears more or less double that which had been observed, and AE =
186, which had been observed as 54 units.
I attribute these enormous differences for the most part to the
adhesion of the water to the walls of the pipe, which adhesion in cases
of this sort can certainly exert an incredible effect. In fact, I used a
pipe of hardly more than two lines in diameter, and certainly I will
experience a greater agreement with a larger pipe. Meanwhile, it is
likely that the curvature of the pipe in the lower region also takes
something away from the motion.

PERTAINING TO §28
I made use of the same curved pipe which I just described; but I
placed a cap at B perforated by a very small orifice. I filled the
whole thing with water except the small region FGB, in which loca-
tion I detained the water with the help of a finger placed on the ori-
fice C. After the finger was removed, the water descended, and
when it had arrived at the position HDB, a number of drops were as
if exploded at so great an impetus through the little orifice at B that
they ascended to a height of more than ten feet, although the height
HA hardly exceeded a height of half a foot. However, on account of
the smallness of the little orifice the water encountered so much re-
sistance while it went through the orifice that, after the impetus had
been weakened, the water not only did not ascend to the height AH
(above which nevertheless, with all hindrances removed, it should
have continuously sprung a little) but hardly a drop or so was pressed
out in a noticeable passage of time, so that I am convinced that if,
apart from impetus, so great a thrust were to be produced from the
natural pressure of the water alone, this would not occur except under
a height of at least one hundred feet.
Further, I observed as well that the thrust of the water is diminished
more, the smaller is the space GB before the experiment; all these
 HYDRODYNAMICS, CHAPTER III

things conform to the theory. It would have been superfluous to

take measurements, because on account of the excessive hindrances
certainly the thrust of the water cannot be as great as it would be
with the [hindrances] removed. But nevertheless, in order that I
might confirm that these things in the experiment also agree with the
formulas, I took a larger pipe CDB in order to eliminate the hin-
drances of adhesion for the most part; the region DFB was very small,
and the region GB, which I left free of water in the experiment, even
smaller; and finally the cap was perforated by an orifice not alto-
gether small. And then I saw that the leap was not very much less
than the height mm:
nn,-,
a, which I gave in §28 for this situation, and I
even remember that I had predicted the height of the leap correctly
to a Friend who was present after I had considered approximately
how much should be given to the hindrances in the calculation.
One will obtain a similar instantaneous explosion of water very
easily, and this arises from a similar cause with fountains which
eject water through a pipe with a full orifice. If, for instance, one
places a finger suddenly over the orifice of the pipe, so that part of the
orifice remains open, one will see directly that the water is expelled
with a great impetus, and soon the thin thread of water is reduced to
within the original limits of the velocity. One will observe also that
the water is projected further and with a greater impetus the less one
leaves the orifice open with the finger, and, for the same orifice
having been left open, that the unusual thrust is drawn forth more
(but always very quickly) and is made more noticeable to the eye, the
longer is the pipe, so that in leaping fountains to which water is car-
ried from a reservoir through very long conduits, if the conduits are
not very large and if the water flows out of the full orifice, I do not
doubt but that through a noticeable period of time a vigorous thrust
of water can thus be produced, returning gradually to the usual
velocity. All this conforms to what has been shown in §§28 and 18.
I remember that I performed this experiment at some time or other,
and for the first time, indeed, in the presence of the most honorable
gentlemen, Messrs. De Maupertuis and Clairaut, with whom I had
previously gotten into a violent discussion on those hydraulic matters.
But although on this occasion there is no air which can be blamed, in
truth, nevertheless, this phenomenon does not differ from that which
Mr. de la Hire has observed, and each develops from the fact that the
motion of the water contained in the conduit, which the enormous
thrust of the water itself constitutes, or at least part of that motion,
cannot be lost without any effect arising thence.
 FOUR TH CHAPTER
Concerning the Various Times which can be
Expected° in the Eifiux of Water

§1. It will seem to many to be a completely Geometrical matter,

which certainly has no concern with any physical consideration, that,
when water flows from a given vessel through a known aperture at
velocities determined in every position, the time be defined in which
a given quantity of water flows out. Nevertheless, experience shows
the contrary; for water flows out through orifices which exist in a thin
place at a much lesser quantity than should follow from the simple
consideration of the velocities, and this for the most part (for the
matter was not self-consistent in different circumstances) in the ratio
of I to V2; this moved Newton to affirm in the first edition of Principia
Mathematica Philosophiae Naturalis that water flows from a vessel at that
velocity which is generated by half the height of the water above the
orifice, which opinion all experiments undertaken on velocities con-
tradict immediately. Exploring the origin of this contradiction a
little later, this great Man himself observed that it was located in the
contraction of the aqueous stream, which contraction customarily
occurs immediately in front of the orifice. Also, another change in
the stream, now similar, now contrary to the former, was observed
by me. Indeed, when water flows out not through a simple orifice
but through a pipe, the stream is again contracted if the pipe converges
toward the exterior, but is dilated if the same diverges. Concerning
the contraction of the aqueous stream flowing out through convergent
pipes, Giovanni Poleni performed very accurate experiments in the
book de castellis [reservoirs] p. 15ff. The contraction of the stream
was observed by this Most Celebrated Man to be greater, the greater
was the internal orifice of the conic pipe, the external orifice and the
length of the pipe being maintained, which is the reason that a similar
quantity of water will flow out more slowly, the remaining things
being equal, the greater the internal orifice will be, although the
 HYDRODYNAMICS, CHAPTER IV

impediments from the adhesion of the water to the sides of the pipe
continually will have a lesser effect; however, those diminutions of the
impediments would cause the water to flow at a greater velocity at the
place where the stream was contracted the most, and [the water]
would be expended no less sparingly; truly, that is understood to
occur from the observed times of efflux and the areas of the streams
where they are contracted the most. Therefore, since the crux of the
matter turns on these changes of the stream, one will be able from this
to examine and explain the phenomena more fully.
§2. Let us assume, for instance, a vertical cylinder which has an
orifice in the middle of its horizontally situated base, but let the
internal water be considered as divided into horizontal strata. With
these things thus assumed, we consider that the motion of every
stratum whatever is the same, and certainly such that a horizontal
position is preserved in them; however, I have warned that this hypo-
thesis cannot be extended to the strata near the orifice, but that, since
thence no noticeable error can arise by reason of the velocity of the
flowing water, it is not worth the effort to take this matter into
account. But now, since other phenomena depend upon the oblique
motion of the internal water, especially such as when it is in the pre-
viously mentioned strata near the orifice, we will illustrate this in a
few words.
§3. Moreover, it seems to me that the motion of the internal water
is to be considered such as it would be if the water were carried
through infinitely small pipes placed next to each other, of which the
intermediate descend nearly directly from the surface towards the
orifice, the remaining being curved gradually near the orifice, as
Fig. 28a shows, from which it appears that the individual particles
descend in this way with a motion very nearly vertical until they
approach the base closely, and they then turn their course gradually
toward the orifice, so that the particles near the base flow with an
almost horizontal motion, the others more vertically, toward the ori-
fice. I was able to observe this sort of motion often by eye when wax
particles, which they call Spanish, were immersed in the water.
Thence it is also known that the individual particles existing at the
orifice cannot preserve their entire direction, and nevertheless they do
not turn it so that they assume a motion clearly parallel to the axis,
but rather the stream of water flowing out will be contracted right to
de, where accordingly [the streamJ will be noticeably more slender
than at the origin near the orifice ac. But this contraction of the ver-
tically flowing stream is not to be confused with the other contraction
which occurs from the acceleration of the water. Next, it is also
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 73

evident that when the direction of the individual particles near the
orifice is different, unavoidably from the impetus which the same par-
ticles create mutually between themselves, the stream is compressed,

A D

C
....-:

d •· •· e
FIGURE 28a

and thus it thins. And from that compression it occurs, which other-
wise would involve a contradiction, that the water already gone out is
accelerated even in front of the orifice, and thus the potential ascent
increases, even if we pay no attention to the other acceleration
common to all falling bodies, as if not pertinent here, and we will not
make mention of it from now on. But unless I am mistaken, this
matter ought to be treated further in the following way.
I. At the outset the stream of water is to be considered while the
velocities of the particles are not being changed further, which, al-
though it never happens in all rigor, nevertheless is to be understood
to occur not far from the orifice, such as at de. But if this were to be
so and the water were assumed to flow out of the vessel ABCD through
the orifice ac, in place of the simple vessel ABCD there is to be under-
stood some other composite one ABadecCD.
Therefore, anything that was set forth in the preceding section for
determining velocities everywhere will apply fully if in place of the
proposed vessel a vessel is considered, as I said, provided with a small
contracted pipe. Nevertheless, by reason of our previously mentioned
method of determining velocities of water flowing out, this correction
cannot produce a noticeable change on account of the shortness of the
small pipe adce, but it can produce an exceedingly noticeable one with
74 HYDRODYNAMICS, CHAPTER IV

respect to the quantity [of water], because the water is to be considered

as flowing out through de rather than through the orifice ac.
II. Thus the velocities in different places in the stream itself will be
reciprocally as the areas of the corresponding sections, and since in
very wide vessels the velocity at de is such that it compares to the total
height of the water, and at the same time it is known from experiments
that the areas ac and de are approximately as V2 is to r, thus Newton
thought his theory could be confirmed in which he stated that the
water truly flows out from an orifice at a velocity that is due to half
the height of the water above the orifice, although, in progressing,
the velocity of the water increases; with regard to this, it seems to me
that he adhered too much to a preconceived opinion: for neither is
the ratio of the orifice ac to de always the same, nor can the motion of
the water from the vessel to which a small pipe is attached thus be
explained; in a word, the attenuation of the stream is by all means
accidental, for the whole ofit can be avoided by applying a very small
cylindrical pipe to the orifice or by merely increasing the thickness of
the plate in which the orifice lies, and then without that correction the
theorems which were shown in the preceding section apply as much
with respect to velocities as to quantities [of water].
III. It is also evident from the very explanation given above on the
contraction of the stream that it cannot remain unchanged by diverse
circumstances; thus, experiments show that the-same is diminished by
an increased thickness of the walls of the orifice; I am not sure
whether the height of the water above the orifice contributes anything.
I should almost believe that the contraction increases some little bit on
account of the increased height of the internal water, although I
readily foresee that it will be slight. Also, it is probable that the lesser
will be the contraction of the stream, especially a vertical one, with all
remaining things equal, the greater a ratio the area of the orifice will
have to the area of the cylinder, because the motion of the internal
water near the base becomes less oblique there, so that if the orifice
occupies the entire area of the cylinder, certainly no attenuation of the
aqueous stream can develop. To this I wish those would pay atten-
tion who will think perhaps that this contraction should be taken into
account in the very determination of velocities. For when the orifice
is not much less than the area of the vessel, no noticeable contraction
can arise, and when the orifice is small, again hardly any difference
arises concerning velocities whether the orifice is increased a little or
is diminished.
§4. The reasoning is almost the same for water flowing out hori-
zontally, so I am saying nothing about other directions: for water
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 75

flows toward the orifice in a similar way from every region; finally, it
ascends even from a lower region right up to the orifice so that it can
flow out, which I myself often observed. Therefore, for a similar
cause the attenuation will become similar in a stream flowing out,
which is rather easy to observe by eye, because here the other attenua-
tion developing from the acceleration of the water already having
gone out does not apply. And on account of this reason, if anyone
begins to make observations on the contraction of a stream, he will
do better in my judgment by using streams flowing out horizontally
rather than in any other direction.
§5. Moreover, it is possible to determine how great the contraction
is, that is, what ratio exists between the area of the orifice and the
minimum section of the stream flowing out horizontally, either by
actually taking measurements of the diameters corresponding to those
areas, or also indirectly from the quantity of water flowing out in
a given time and at given velocities, where, nevertheless, the velocities
are to be deduced not so much from the height of the water above the
orifice as from the area of the jet, since certainly the hindrances,
now greater, now lesser, never permit the full velocity of the water
which it should acquire by dint of the theory, in which no account
is taken of these hindrances.
§6. Now I think it is evident enough from the previous statements
that there will be a perfect agreement between the quantity of water
flowing out and its velocity, if only there is substituted for the orifice
which is in the vessel some other orifice diminished just to that degree
that it would not exceed the section of the maximum contracted
stream; and it will be equally so at whatever place in the stream or
at whatever depth from the surface of the water this orifice is under-
stood to be, whether at ac or at de, since indeed the velocities will
always correspond approximately to the total height of the water
above that place at which the orifice is assumed; henceforth I shall
call the area of this orifice, to be conceived in the mind, the Section ef
the contracted aqueous stream.
§7. Thus, if that Section about which we just spoke would have a
constant ratio to the orifice, the orifice of efflux would have to be
considered diminished in the same way, and afterwards a calculation
of the quantity of water flowing out in a given time would have to be
undertaken. Thus, indeed, with that ratio taken as~ and with the
ex
area of the orifice called n, the Section ef the solid stream would have to
be considered as!!-.
ex
 HYDRODYNAMICS, CHAPTER IV

But since it is variable under different circumstances, rules cannot

be given a priori in this matter; moreover, the section is changed
greatly by an increase or decrease of the thickness of the plate in which
the orifice lies; something, although only a little, can also be contri-
buted by the size of the orifice, the areas of the vessel (and these are
absolute as well as relative), and perhaps by the height of the water
above the orifice. Meanwhile, there having been assumed that the
wall is thin, that the vessel is very wide, and that the orifice develops to
4 or 6 lines in diameter, the ratio between the orifice and the Section of
the contracted stream does not customarily depart much from that which
Newton stated, namely as V2 to 1. But often it has been observed
to be more by some and also less by others.
§8. But whatever it may be, in any case we will indicate it, as
ex
before, by-. And now, for this situation we will develop the com-
I
putation for the times; but for the sake of brevity we will consider only
cylindrical vessels, and in these we will examine two kinds of time
especially: the first which defines the point of maximum velocity, the
other which corresponds to depletion. But in each case we will
assume that the motion begins from rest.
§9. Therefore let a cylindrical vessel be placed vertically, full of
water, and let the height of the water at the beginning of flow be a, the
area of the cylinder be m, the area of the orifice be n, the Section of the
solid stream be:!; let the water already have flowed out for the time t;
ex
and then let the residual height of the water above the orifice be x, and
at the same instant oftime let the surface of the internal water have a
velocity which corresponds to the height v; the velocity itself will be
vv,
-dx divided by the velocity from which dt = -:t
but the element oftime dt is proportional to the element of space
Vv,
Indeed, the value of v itself was determined in Chapter III where we
used the same notation which we now use. But, since for a correct
measure of the expended water it is required that the section of the
contracted stream :! be substituted for the orifice n, it follows that in the
ex
value of v itself the same substitution is made, and thus it is stated
that

V = nna [(~) 1- mmaa /nn

2nn - mmexex x
_ :a].
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 77

dx
But if this value is substituted in the equation dt = - vi/ there

appears:

_ /J
dt - -dx -- nna
--
2nn - mmaa
[(a)l-mmaa/nn
-
x
- -x]
a

with the help of which equation all desired times can be defined
through approximations or series, if only at individual points the value
of a itself be known. But let us assume it to be of constant value,
since indeed in the present case there is nothing by which it could be
changed except the different heights and velocities of the fluid, which
contribute little or not at all to this aspect of the problem, as far as it
can be perceived by observation.
§10. Now, in order that the desired equation can be expressed
through a series, we will consider the quantity

in the following form:

nnx )-112. [r _ (:)mmaa/nn-2]-112

( mmaa - 2nn a

and the latter factor we will resolve through the customary rules into
this series:

r+--
I (X)mmaa/nn-2 I. (x)2mmaa/nn-4
+ -3- -
2 a I · 2·4 a

+-1-3 -
5 -. . (X)3mmaa/nn-6 + etc.
r-2-3-8 a

from which now 1s obtained a very slightly changed form of the

equation:

-
dt -
dxv mmaa
. ;-
- I (X) mmaa /nn - 5/2
+ -2 -a
nva
~ (:)2mmaa/nn-9/2+ 1.3 . 5 (:)3mmaa/nn-13/2+ etc..]
+ 1-2·4 a 1·2·3· 8 a
 HYDRODYNAMICS, CHAPTER IV

This equation is to be integrated so that for x = a there will be t = o;

there arises, accordingly,

t = [2 + 2mmaann- 3nn
+
I 6mmaa
3nn
- 28nn
+ etc.]
v' (mmaa
x--------
- 2nn )a [ (x)l/2+ - -nn- - - (x)mmaa/nn-3/2
2- -
n a 2mmaa - 3nn a
3nn
+ ------- (X) 2mmaa /nn - 7/2 + etc. ] ·V-(-
mmaa - 2nn ) a
16mmaa - 28nn -a -----,
n

where 2v'; expresses the time which a body uses while it falls freely
through the height a. But if in that equation it is assumed that
_I
X-a
(mmaa _ nn) nn/(mm aa -
-----
nn
2nn),
which is the height of the water when the velocity is a maximum
(through §16, Chapter III and §8, Chapter IV), then the time is
obtained which elapses from the beginning of flow right up to the
point of maximum velocity; and when it is assumed that x = o, the
time appears in which the whole vessel is depleted; and, finally, if xis
assumed equal to any quantity c whatever, t will express the time
which the surface takes in descent through the height a - c. More-
over, we will see for these cases what should happen when the vessel is
very wide, and the number m thus contains the other one, n, several
times.
§11. At first let the number ~ be infinite, then the height of the
n
water corresponding to the point of maximum velocity will be

but since mmaa is an infinite number, this might be considered as

nn

(m:arn/mmaa = I + (ln m:;a) ;m:;a,

of which matter the proof is this: let an infinite quantity A be pro-
posed, and, as in our example, let A i tA be considered, and anyone
readily sees that the latter quantity is little greater than unity, and
indeed the excess, which we will call z, is infinitely small; and so one
has A 1 1A = I + z; let the logarithms be taken on both sides, and
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 79

there will be h~/ = ln(1 +z) = (on account of the infinitely small
A
value of z itself) z. Therefore, A 11A = 1 + ln
A; and from there,
similarly, as we said,

(m::a) nn/mmaa I+ ( 1n - - mmaa) ;mmaa

--•
nn nn
Furthermore, because this quantity added to unity is infinitely small,

aI(m::arn,mma«.
or

a [ 1+ ( 1n - -
nn
mmaa) ;mmaa]
-- -
nn
(i n
_a - a mmaa)
- -;mmaa
--,
nn nn
/
therefore the distance through which the surface of the water de-
scends, while the maximum velocity develops from rest, is

a(ln mmaa)/mmaa 2nna ln ma.

nn nn
or
mmaa n

This equation indicates that the descent of water in an infinitely

wide vessel is infinitely small when the water has already attained its
maximum level of velocity. Moreover, this notwithstanding, it
could have been questioned whether or not in the meantime a finite
quantity of water flows out, since indeed a cylinder erected above an
infinite base of infinitely small height might have an infinite volume;
but it follows from our equation that this quantity is also infinitely
. 2nna
small, and nommally equal to - - ln - •
ma
maa n
And this indeed agrees splendidly with the phenomena which we
discover in the efflux of water from reservoirs through a simple orifice
for a whole day. For when we cover the orifice with a finger, as soon
as the finger has been removed we determine that the water flows
horizontally, and we observe that halfway between the longest thrust
and the point which lies on a plumb line from the orifice not a drop has
fallen to the ground.
§12. Just as in the last paragraph we determined quantities how-
soever small, such as the descent of internal water and of water flowing
out while the water reaches the maximum level of velocity, so now we
will show the same thing by reasoning of the time. But I say in the
equation of§rn, expressing the time, that it suffices that only the first
80 HYDRODYNAMICS, CHAPTER IV

term be accepted in each series, which will be evident if one extends

the calculation to two terms; therefore the desired time interval is

x)
y' (mmaa - 2nn)a
t= ( 2 - 2 - - - ~ - - - - ~ ,
J a n
whence, after the value pertaining here, which we defined in the pre-
ceding paragraph, has been entered for x, there occurs

t -_ [ 2 - 2J 1 - (rnmmaa)/mmaa]
--
nn
- - · J(mmaa
nn
- - -- -2nn)
nn
- a

or, after mmaa)/2mmaa

( In -nn- - nn
I - - - has been entered for the correspon-
ding quantity under the radical sign, it yields

but finally, after the quantity 2nn under the radical sign has been
. 2nv'a ma
rejected, there appears t = - - In-·
ma n
But this interval of time is infinitely short, because, as is known, the
logarithm of an infinite quantity is infinitely less than the quantity
itself. But if, indeed, from the beginning of flow, water is expelled at
once at its maximum velocity, it will seem remarkable at first glance
to some, perhaps, that a finite motion is generated in an instant;
nevertheless, no one will consider it absurd that an infinite mass, of
which sort the quantity of water contained in an infinite vessel is,
in an infinitely short time can produce finite motion, and this by the
action of gravity alone.
§13. If, furthermore, in the case of the infinitely wide vessel we
wish to express the time of depletion, which will of course be infinite,
it will have to be assumed, as it was indicated above, that x = o in the
equation of §ro, and at the same time only the first term of the series
is to be applied, and again ma is to be used for v'mmaa - 2nn; and
2ma y'-
thus it occurs that t = - - a.
n
Then at last the time which is spent in the descent of the surface
through the height a - c is expressed in a similar hypothesis by the
following equation:

2ma y'-
t = - ( a - y'-c).
n
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 8I

§14. The previously mentioned equations are satisfied, certainly

not accurately, but closely, when the vessel is not of an infinite but
nevertheless of a very great area; finally, they are not very greatly
deficient when the number m moderately exceeds the number n. It
may be possible here to add some words about the experiment which
I indicated at the end of §11, and let this indulgence be given to our
purpose which is most strongly directed toward illustrating and
examining the phenomena of motions discovered by experience. But
I said in the cited paragraph that, when the water flows out hori-
zontally, the first volume element at once obtains the entire length of
the thrust, and indeed theory certainly indicates that same thing for
very wide vessels; but, truly, in moderately wide vessels a few drops
must flow out at a lesser impetus before the point of maximum
velocity appears, and these drops should strike at some point halfway
between the maximum thrust and the point which corresponds verti-
cally to the orifice ; and I observed that this even occurred from
vessels of area about ten times as large as the orifice. Indeed, when
at one time I undertook an experiment concerning a vessel half a
foot high which had an area more or less one hundred times the area
of the orifice, not even the least particle of water, as much as I was able
to observe, withdrew noticeably from the full thrust of the water.
And so we may see what quantity of water should flow out in this case
before the instant of maximum velocity [is reached]; indeed it will
be as great as that which a cylinder of the same area contains in the
height

(see §10 at the end ) ; nor does this very small height differ greatly from
. 2nna ma
the followmg much shorter one, namely - - ln - (see §11 ), where
mmaa n
1
now by !!:. is understood - - and by a half a foot, while for a bne may
m 100

substitute V2 (for we do not desire the greatest accuracy here) and

through ln is indicated the hyperbolic logarithm; but thus there
occurs

2nna
- - ln -ma = - 1- ( ln 100 + -21 ln 2 ) = 0.0002475 foot
mmaa n 20,000

or 0.00297 inch, and since I found the area of the vessel equal to
6½ square inches, I knew that the desired quantity of water which
 HYDRODYNAMICS, CHAPTER IV

indeed should have flowed out before the maximum thrust developed
was equal to about a fifty-second part of one cubic inch, or, after it has
been assumed that an average drop contains six cubic lines, more
than five drops. But in the experiment I observed none, the reason
of which matter I suspect to be that the first drops, although already
ejected, are nevertheless still propelled by the water that is following;
for the others follow too quickly for it to be possible that the first ones
are removed from them. But it happens here that the interval of
time from the beginning of flow right up to the maximum expulsion
2
(which indeed through §13 is approximately n-V~ In ma, where
ma n
through 2 v~ is understood here the time in which a body falls through
a height ofhalfa foot, which is about ff second), I say, that interval
of time does not extend beyond the one-hundred-fifty-eighth part of
one second.
Perhaps the fact that one's finger cannot be removed from the
orifice quickly enough contributes something. But it pertains here
particularly that the greatest part of that water which flows out before
the maximum velocity is attained so approaches the maximum thrust
that no difference can be observed, and thus hardly a single drop
would have defected by a noticeable interval from the former if it
could have separated itself freely from the water following.
§15. So much for water flowing out through orifices; let us progress
now to the efflux of water from vessels through either converging or
diverging cones. Moreover, if water flows thus through a converging
pipe, the same ratio sought in §3 from the converging motion of
particles explained for simple orifices dictates that the stream of water
will be contracted in front of the orifice and its particles will still be
accelerated, and thus the quantity of water flowing out in a given
time is less than the measures of the orifice of efflux and the velocities
indicate if the contraction of the stream has not been taken into
consideration. But that contraction is customarily small in rather
long pipes. In diverging pipes all things occur in a reverse way: for
the stream is dilated in front of the orifice, the motion of the water is
retarded, and a greater quantity of water flows out in a given time
than would follow from the observed area of the orifice and the
velocities of the water flowing out through the former without that
dilation. Finally, the aqueous stream flowing out from cylindrical
pipes is neither contracted nor dilated.
And thus one must properly attend to these contractions or dila-
tions in estimating the quantities of water flowing out in a given time,
which question we will treat in passing at the end of the section.
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 83

But now it is pleasing to study the changes which occur in the

effiux of water from the beginning of motion. But in these things,
for the sake of brevity, we will not consider the changes of the stream;
for neither is the matter so well established that it can be confirmed
accurately enough by experiments, nor are the previously mentioned
changes of great moment here; but the matter itself is worth being
sought after persistently so that the nature of it can be understood
correctly in one's mind.
Concerning vessels which have pipes attached, we just discussed
them in the above Chapter, §§31, 32 and 33; and indeed in §31 we
gave rather general equations, whatever the ratio might be between
the areas of the vessel and pipe; but they are overly involved, and they
require highly troublesome calculation. In the paragraph which
follows that one, I treated the hypothesis which makes the vessel
everywhere of infinite area in proportion to the pipe, in which hypo-
thesis I said that the water flows out at a velocity by which it can
ascend to the full height of the water above the orifice of effiux;
nevertheless, at the end of the paragraph I expressly warned that at
the beginning of motion the water descends more slowly than was thus defi,ned,
and that that rule does not apply before the suiface has descended through
some little space, which matter is intrinsically evident, since indeed the
maximum velocity cannot be produced in an instant from a state of
rest in a pipe, although it may occur in a vessel perforated by a simple
orifice.
So, holding these things in my mind, I began to investigate the
initial changes and to reduce them to certain measurements. But the
previously mentioned rule, in which those initial changes are not taken
into consideration, does not suffice for this at all, although otherwise
it is exactly true in an infinitely wide vessel; for all changes which
precede the state of maximum velocity occur while the surface de-
scends through an infinitely small space: nevertheless, if only the
vessel is infinite in a Geometric sense, then that descent not only
does not occur in an infinitely small time, as in the case of a simple
orifice, but occurs in an infinitely great time, and meanwhile an
infinite quantity of water flows out as well, while through an orifice
an infinitely small quantity flows out, the other remaining things being
equal. But in order to show these things, I made the effort of bring-
ing forth another equation from the general equation of§23, Chapter
III, which is this very simple one: s = x, withs taken for the height
corresponding to the velocity of the water flowing out and x for the
height of the water above the orifice of effiux; but anyone under-
stands that for our purpose the matter is to be brought about thus so
 HYDRODYNAMICS, CHAPTER IV

that the ratio of the increments of velocity is obtained, which was not
required previously.
§16. Therefore, let there be the cylinder AEHB (Fig. 18) as in §22,
Chapter III, and let this be considered as infinitely wide and full of
water, and let it have the pipe FMNG attached, of finite area and in
the form of a truncated cone, either increasing or decreasing in area
toward the orifice MN through which water flows. Let the initial
height of the water above the orifice MN, namely NG+ HE, be a;
the height of the aqueous surface at the position CD above MN, that
is NG+ HD, be x; the length of the attached pipe, or NG, be b; the
area of the orifice MN be n; the area of the orifice FG be g; the area of
the cylinder, which is infinite, be m; and finally let the velocity of the
aqueous surface at the position CD be such that it conforms to the
height v, which height indeed will be infinitely small. After these
things had been established, we saw in the place cited that the
following equation generally obtains:

m3
m(x - b) dv + .bmm
1
_ dv - - v dx + mv dx = -mx dx
v gn nn

in which it is evident that now the first term m(x - b) dv can be

neglected with respect to the second ~7m dv, just as the fourth mv dx
vgn
3
with respect to the third - ~ v dx, and thus there can be assumed
nn

3
bmm dv - m v dx = -mx dx
Vgn nn

in which equation if again the first term is neglected, which can be

done unless the changes are desired as well which occur during the
first descent, even ifit is infinitely small, the common rule will arise of
the potential ascent of the water flowing out to the full height of the
water; but now for our purpose, in which we desire those first changes,
that [first] term will have to be retained, and thus the last equation
will have to be treated in its entire extension. However, for separat-
mm
ing the unknowns from one another, let - v - x = s, or
nn

v = !!!!.... (s + x), and dv = !!!!.... (ds + dx), and so it will occur that
mm mm
-nnb ds
dx = b
nn -ms gn
V ' which is to be integrated so that, with x = a
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 85

having been established, it produces v = o, and hences= -a; but

thus it occurs that

x _ a = nnb ln nnb - ms Vgn

mVgn nnb + maVgn

and the assumed value mm v - x having been taken for s, there results
nn

x _ a = nnb -ln n4 b - m 3 vVgn + mnnxVgn_

mVgn n4 b + mnnaVgn
Here again in the quantity under the logarithmic sign the term
4
n b, certainly infinitely less than the term mnnxvgn, can be eliminated
from the numerator, and indeed from the denominator the term n 4 b,
likewise infinitely less than the other, mnnaVgn, [can be eliminated].
And so there occurs

x _ a = nnb ln nnx - mma.

mVgn nna

From this there is obtained, after e has been employed for the number
of which the logarithm is unity,

v = nnx - nna em<x - a). -lgn/nnb •

mm mm '

or, with a - x = z having been assumed, so that z denotes the dis-

tance through which the surface of the water has already descended,
this form can be obtained for the equation:

- nn(a - z)
V-------
nna; e(mz /nb)-lg /n
mm mm

from which again it is clear that, when z has a minimum ratio to b,

the denominator of the other term may become infinite, and

nn(a - z) nnx
V=----
mm mm'

but truly the matter is otherwise as long as the descent z is infinitely

small, which case we now consider.
86 HYDRODYNAMICS, CHAPTER IV

§17. With these things having been set forth, it is now easy to define
through what little distance the fluid descends while it acquires the
maximum velocity, namely by making dv = o, or,

-nn
- --dz + _
na .~
Vg/n; e<mz/nb) ,;-
g/n = 0
mm mb
that is,

z = nb . 1- (ma . 1- )
m v n/g -ln nb v g/n .

But this height multiplied by the height m of the cylinder gives the
quantity of water flowing out in the meantime, namely nbvn/g ·
ln (:; Vg /n), which quantity, as I hinted above in §15, is infinite,
although only logarithmically, an infinity of which sort is less than the
root of any sort of dimension given from the same infinity: that is to
say, lnoo is less than oo 11 n, however great a number may be assignable
to n. And it should hence be understood that, if we reason from a
true infinity to very great quantities, this quantity of water becomes
small enough. Finally, the corollaries to the formula are these:
I. If the attached pipe is cylindrical, it occurs that z = nb ln mba.
m n
Therefore, with the remaining things being equal, this quantity
appears as the length of the attached pipe, which is generally true

also: for from a changed value of b itself the quantity ln : ; J!

is to be considered as not changed on account of the infinite value of
the number ~.
n
II. For the same orifice g, and with the remammg things also
equal, the quantity z follows a three-halves-power ratio of the final
orifice: and if the same pipe with first a narrower and then a wider
orifice is applied to a vessel, the quantity of water in the former case
will be to the similar quantity in the latter as the square of the wider
orifice is to the square of the narrower orifice.
III. Finally, it is to be observed that the whole reasoning is valid
for all directions of the pipe, which anyone will perceive who will
properly examine §22, Chapter III. Therefore, the pipe can be
furnished either horizontally or in any other direction whatever, and
however curved, to which one will have to pay attention especially in
undertaking experiments. However, let it always be understood that
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 87

b is the length of the pipe, and a is the vertical height of the water
above the extreme orifice.
§18. Now I come to the time in which those changes from rest to
the maximum velocity occur. But I say that in a calculation of times
. sort one can s1mp
o f t h 1s · 1y set v = -nn a. F or t h e remammg
· · quant1-·
mm
ties in the last equation of §r6 vanish, however small the height z is
assumed, if only it has the minimum assignable ratio to that infinitely
small height which corresponds to the maximum velocity, namely to
nb . 1- (ma .1- )
m v n/g ln nb v g/n . Thence it follows that the time is predicted,
which I will call
bvn
t = --= -ln (ma.
1- )
- v g/n
Vga nb

and that in a like manner it is infinite, although the same time is

extremely small when the area of the vessel is not infinite but very
large, which again is to be deduced from the nature of the infinite
logarithm.
§19. Because the height of the velocity, as we saw in the last para-
graph, can at once be reckoned as!!:!!.... a (that is, equal to the maxi-
mm
mum when the surface descends through the minimum assignable
portion of the infinitely small descent, after which the full maximum
velocity is present), it follows that most changes from rest to the state
of maximum velocity are not noticeable (that is, infinitely small ), by
all means not only the majority, but also all except an infinitely small
portion; to be sure, the matter occurs thus: the velocity is clearly null
at the very beginning, and, after the water descends through an in-
finitely small distance, it is already very nearly maximum; then,
while it descends again through some little distance, infinitely small
but nevertheless infinitely greater than the former [distance], it con-
tinues to be moved at its own velocity, taking on infinitely small incre-
ments, and then at last it truly attains a maximum velocity. But
since these latter or infinitely small changes cannot be perceived by
observation, we will treat differently those theorems which we gave
in §r 7 by considering, in place of the changes from rest right up to the
point of maximum velocity, the same changes right up to a given
rate of speed.
§20. And so we will investigate through how great a distance z the
surface of the water descends from the state of rest, and how much
water flows out, and finally how much time must pass in order that the
88 HYDRODYNAMICS, CHAPTER IV

internal water be moved at a velocity which is generated by free fall

through the given height, which we will call .!!!!:.._ c, so that c itself
mm
denotes the similar height for the velocity of the water flowing out.
For this it is required that in the last equation of§r6 nnc be entered
mm
for v, so that there will be
nnc nn (a - z) _ nna/e<mz/nbJ· ..lgfii
mm mm mm

mz. - a
and hence it is deduced that -b v 1g /n = ln - - - - ; but when
n a-c-z
c is assumed here to be noticeably less than a, the letter z under the
logarithmic sign can be rejected, whence there is obtained
nb .
1 - a
z = -·vn/gln--·
m a-c

But this equation now indicates a space which is infinitely small and
through which the surface of the water descends while the velocity of
the water flowing out from rest is that which is due to the height c;
and this little distance is to that indicated in §r 7, by which indeed the
velocity becomes maximum, as ln _a_ is to ln (mba Vg fn), so that the
a - c n
first is infinitely less than the other, although both are infinitely small.
If, further, the defined quantity z is multiplied by m, one obtains the
quantity of water flowing out while that velocity due to the height c
is produced, which quantity, accordingly, is equal to

nbV n /g ln _a_
a - C

and thus is of finite magnitude, and indeed is greater the longer the
pipe is assumed and the greater a thrust is expected.
And, finally, the time in which the same occurs, if the terms to be
rejected are selected correctly, is discovered to be equal to

2
Jnbb
-n 1 -a-
ag a - c

and thus is finite but very small, and in no case is it to be extended

easily beyond one second.
§21. I wished to examine and pursue all these things accurately,
first because the solution of many phenomena which are customarily
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 89

observed in the effiux of water depend thereon, then as well in order

that we might understand correctly in our mind those changes which
are clearly imperceptible through observation. There have been
many who, not having followed correctly the transition from the in-
finite to the finite or, in turn, from the finite to the infinite in flowing
water, were not able to extricate themselves from the many diffi-
culties which elsewhere easily permit a solution; but ifin place of the
almost infinite vessel, of which there is none, a very wide vessel is
assumed, or even, since it suffices in many cases, a moderately wide
vessel, the formulas will be approximately true, and they will approach
the truth more or less in accordance with the nature of the question;
concerning these matters I will point out certain things in the follow-
ing experiments. Thus, meanwhile, it is already apparent enough
from theory, because I had prepared to explain especially why water
flows from a simple very wide vessel at once at the entire velocity, and
why it is different for water ejected from a vessel through a pipe. But
the precise measures concerning these questions are to be deduced
from the equations themselves.
§22. Finally it is evident, since it pertains to the time of depletion,
that, when the area of the vessel moderately exceeds the area of the
attached pipe, the former can be considered without noticeable error
to be ma 0, by understanding by 0 the time in which a body by falling
n
freely from rest falls through the height which the water had at the
beginning of flow above the final orifice of the pipe, and by assuming
for ma the ratio which exists between the area of the vessel and the
n
section of the stream, whether contracted or dilated. But the hindrances
which accidentally occur in addition in these cases increase that time
somewhat. But if the time is desired in which the surface of the
water descends through a given height, that [time] is to be taken as
ma. ( 0 - T), after there has been assumed for T the time which a
n
body takes in falling freely through the height which the water has
above the orifice at the end of the flow.

EXPERIMENTS WHICH PERTAIN TO CHAPTER IV

Since a large part of this chapter was employed in the contraction
of an aqueous stream flowing through an orifice in a thin plate, I
undertook to begin accurate experiments concerning that contraction,
certainly not by undertaking measurements of the diameters, which
90 HYDRODYNAMICS, CHAPTER IV

method I found cannot be used with sufficient accuracy, but by ob-

serving the actual velocities from the extent of the thrust and the
quantities flowing out in given times. In the experiments I used an
automaton which pulsed 144 times in the period of the first one
minute, and I assumed it to be thus in the following [minutes].

PERTAINING TO THE THEORY OF THE CONTRACTION OF AQUEOUS

STREAMS

EXPERIMENT 1. I furnished a cylindrical pipe, the diameter of

which was 4 inches 3 lines, English measure, made of a thin plate and
which had an orifice in the side, that is, in the cylindrical surface; the
diameter of the orifice was 4-f-l-5 lines. The water flowed out hori-
zontally from the vertically positioned cylinder, and the height of the
water above the center of the orifice was 4 inches 8 lines at the be-
ginning of flow, and the similar height at the end of flow was 3 inches.
However, the whole flow took place in the interval of eleven pulses of
the automaton, which constitutes a time of approximately 4½ seconds.
Further, after the experiment had been repeated often, and after
the height of the orifice above the horizontally placed table and the
extent of the thrust, and this both in the beginning and at the end of
flow, had been observed, I saw from the L emma indicated in the beginning
ef the Experiments ef the preceding Chapter that the velocity of the water
flowing out at the place of the maximum contracted stream had con-
stantly been that, indeed as much as one could judge by observation,
which should be due to the height of the water above the very place
which is at the same height as the orifice.
Therefore, if we assume that the contraction of the aqueous stream
was the same everywhere, we apply to this case the last equation of
2ma • ;- • ;-
§13, namely t = - - (v a - v c), there will have to be established
n
t = 4½ seconds;~n = 133; 2v; (the time which a body takes in
falling freely through the initial height of the water) = o. 1483, and
2V~ (the similar time for the final height of the water) = 0.1246; it
occurs that 4½ = 3.15a, from which a= 1.43. Hence it is a conse-
quence that the area of the orifice was to the section of the contracted
stream as 143 to roo; this ratio is a little greater than that which
exists between V2 and 1, that is between 141 and roo; but if the
velocities could have been observed very accurately, there is no doubt
but that they would have been a little less than those which are due to
the entire height of the water; and when this matter is taken into
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 91

account, it is discovered that the value of a itself is thus to be dimin-

ished a little bit; therefore, from the entire experiment it can be
concluded very safely that the previously mentioned ratio was as
V2 to 1.
EXPERIMENT 2. Next I wished to find out by experiment whether
in all jets flowing in any direction whatever the contraction is the
same, and to this end I reckoned that the matter was to be attacked in
such a way that, except for the change of that direction, all the re-
maining circumstances should be the same thereafter. Indeed, I
obtained it in this manner.
Obviously, I used the same cylinder as previously, but I attached
it to a prismatic box, positioned vertically, so that the axis of the
cylinder would be horizontal, and thus I revolved the attached
[cylinder] so that the center of the orifice selected for the effiux of water
would occupy first the highest place, then the middle, then the lowest;
in the first case the water flowed out vertically upward, in the second
horizontally, in the third vertically downward; but in the individual
[cases] I made the heights of the water in the box above the center of
the orifice perfectly equal; the result was this:
I observed that in equal times the surface of the water in the indivi-
dual cases descends through equal distances in the box. Therefore,
in streams projecting upward, the water above does not offer notice-
able resistance to the water following below, which same thing I
understood in a different way, because certainly if I intercepted the
aqueous stream in any direction whatever at a small distance from
the orifice, such as 3 lines, say by a coin, so that the stream would
strike against the coin perpendicularly, the effiux of the water would
not be retarded. Further, neither does the water below in streams
descending vertically draw the sequent water after itself: and the very
contraction of the stream is everywhere the same, since the retarda-
tion and acceleration of the water ejected upward or downward,
which cause the stream either to swell or to become slender at some
distance from the orifice, were not considered. For here certainly the
discussion is only about that contraction which arises from the
oblique motion of the particles in the region of the orifice.
EXPERIMENT 3. I used the same device, prepared in the previously
mentioned manner, for finding out whether the contraction of the
stream, all remaining things being equal, would be changed by an
increased height of the water above the orifice. To this end I fixed
two needles to the internal walls of the box along a plumb line; the
first projected above the center of the orifice I 3 inches Io lines, the
other I 2 inches I¾ lines, English measure; the area of the box was to
 HYDRODYNAMICS, CHAPTER IV

the area of the orifice as 404 is to I. Moreover, I saw that the surface
of the water had descended from the upper needle to the lower after
an interval of 24 pulses of the automaton, which gives a time of IO
seconds.
But if indeed the same time is sought according to the Hypothesis
that the stream had not contracted and that at the same time the
water had flowed out at the whole velocity which it should have
had by dint of the theory with no alien hindrances being present, that
[time] is ascertained to be 61 seconds.
Thus, therefore, it can be concluded that the area of the orifice
was to the section of the contracted stream as IO to 61; that is, a = 1.45
while in the first experiment for the same orifice, all circumstances
having been considered, one found a = 1.41.
After I had tried this in that way, it remained to discover whether
the water would flow out at the whole velocity, according to observa-
tion, about which matter I doubted all the more, because with the
velocity of the water increasing, the hindrances increase at the same
time, and accordingly they can be noticeable at the greater heights of
water while they are not so at the lesser.
And so with all care applied (because it is required especially for
the precision of the experiment), I made the water flow out in a per-
fectly horizontal direction, and after the measures of both the extent
of the thrust and the height of the orifice above the horizontal table
had been taken, I saw, after performing a calculation, that when the
height of the water was 13 inches and IO lines, or 166 lines, the water
would flow out, or rather would flow across the section of the contracted
stream at a velocity which corresponds to a height of 158½ lines; there-
fore, the velocity is to be diminished in the calculation in proportion
to the square roots of these heights, and the discovered value of the
letter a, which thus is a little less than 1.42, or again 1.41, decreases in
approximately the same ratio, and thus it is allowable to deduce that
the changed height of the water does not alone change the contraction
of the stream according to observation.
EXPERIMENT 4. I used a cylindrical pipe of height 4 inches, the
section of which through the axis is represented (Fig. 28b) by CABD;
the area of the cylinder was to the area of the orifice ac as 1 IO to 1.
This entire cylinder full of water was evacuated in a time of 21½
seconds. However, it must be noted that efflux is not to be granted to
the water until no turbinate motion is observed in it; for otherwise the
water is soon changed into a whorl, somewhat swift during efflux, and
the efflux is greatly retarded, and all the more so the faster the internal
water is driven in a circle; further, because all the water never flows
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 93

out, I considered the time of efflux [to extend] until it began to flow
out drop by drop.
C ..D

A a c B

FIGURE 28b

This experiment indicates that here the contraction of the water was
less than in the ratio 2 to I. I had expected the time of evacuation to
be only about 23 seconds, but the occurrence was a little different, as
I said, of which matter I noticed a little later that the reason was that
the elongated lips of the orifice more or less formed a short pipe, al-
though very short, as the Figure shows, which hindered the contrac-
tion of the stream; however, the length of those lips did not attain two
thirds of a line.
EXPERIMENT 5. I made water flow horizontally through a small
pipe from a very wide vessel; but the pipe was very short, indeed not
exceeding a length of 3 lines, and it was almost 5 lines in diameter.
A given quantity of water flowed out in a time of r r ¼ seconds
which should have flowed out in rof seconds if it is assumed that
neither was the stream contracted nor were any hindrances present.
I did not consider the [determination of the] true velocities of the
water to be a task that I should undertake, not doubting that they
were such as they must be in order that a given quantity of water
might flow in an observed time through an observed orifice, with no
attention having been given to the contraction of the stream.
In addition, I furnished other small pipes of different diameter and
length, and I saw that the quantities of water flowing out in a given
time and at given velocities correspond directly to the orifices of
efflux, but that the velocities were more deficient from the velocity
due to the entire height of the water, the narrower and the longer
was the pipe, and also the higher was the water.
94 HYDRODYNAMICS, CHAPTER IV

PERTAINING TO THE THEORY OF WATER FLOWING OUT THROUGH PIPES

EXPERIMENT 6. The cylindrical vessels, of which Figs. 24 and 25
represent the sections through [each] axis, had a height of 4 English
inches and pipes attached of one-foot length; the areas of the cylin-
ders were to the areas of the orifices A as I 10 to 1. But the orifice B
was to the orifice A approximately as 25 to 16; the time of evacuation
before the cylinders were emptied was, in Fig. 24, 6½ seconds, in the
other, 4½ of these units of time, more or less.
In these cases the vessels were large enough with respect to the
attached pipes that they could be considered as infinite; and accord-
ingly the water should have flowed out, according to the Rules
indicated by us within the text, through the final orifices at velocities
corresponding to the total height of the water, if only one excludes the
first instants of flow, which themselves are so short here that they can-
not be observed. And since in addition, as I advised in passing, the
quantity of water flowing out in a given time through the pipes is to
be estimated simply from the speeds and the size of the orifices, I
found, through the rule shown in §22, the time of evacuation in the
first case as 4½ seconds, in the latter, almost 3 seconds.
The fact that these things were observed to be a little greater in the
experiment in Fig. 24 is to be attributed for the most part to the
adhesion of the water to the sides of the pipe, but in Fig. 25 to a dif-
ferent reason in addition, indicated in §34, Chapter III.
Other Phenomena are to be noted in these vessels: namely, when
the vessels are not quite evacuated, a certain sound is perceived from
the air which then mixes with the water in the upper [part of the]
orifice; in fact, I heard this sound to the last instant of flow; further,
it occurs easily that the effiux of water may be permitted before the
latter has been reduced to perfect rest (for the water is agitated by
filling and moved in a whorl) ; but then the effiux is retarded very
much, and a kind of cataract is formed internally, and air is continu-
ally intermixed with the water flowing out. Thus the effiux can be
retarded at will if the water is agitated into a vortex before it flows out.
EXPERIMENT 7. I used a Prismatic vessel, to which a small pipe
was attached horizontally as in Fig. rg. The orifice GF was precisely
5 lines in Diameter, the other, NM, 6½ lines. Accordingly, the very
areas of the orifices GF and NM were as 100 to 169, but the area of
the vessel contained the area of the orifice NM 201 times. The length
of the small pipe GN was 4 inches.
Then I filled the vessel with water right up to CD, the height of
which above the axis of the pipe was 13 inches IO lines. After the
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 95

orifice NM had been opened, the water flowed out, and the surface
descended right to EH in a time of 8¼ seconds, and the difference of
the heights, CE or DH, was 2 inches 8 lines.
After the calculation has been reduced according to the pattern of
§22, where attention was paid neither to the hindrances nor to the
change of the stream, we see that the predicted time of descent should
have been approximately 5 seconds, almost 5½, Therefore, it is to be
stated in this way: that the mean total velocity was to the entire
velocity which the theory indicates as 5½ is to 8½, or approximately as
2 to 3; and hence it can be concluded that the water flowed out
through the orifice MN at a velocity which compared with (!) 2 , or
four ninths of the height of the water above the orifice MN, but
through the other orifice GF it flowed at a velocity due more or less
to five fourths of that same height.
And so it appears again that the efflux of water is increased by the
increased area of the orifice of the pipe toward the exterior, although
neither the opening at which the pipe is implanted in the vessel nor
the position of the pipe is changed.
Further, on the horizontally placed table PQ, I observed the ampli-
tude of the thrust PQ, for the height oP, which was 4 inches 8 lines.
Moreover, I found PQ, = g inches 6 lines.
It follows from this observation that if consideration of the dilation
of the stream is set aside, the water would be required to have a
velocity at NM which is due to a height of 4 inches IO lines, while,
nevertheless, by dint of the aforementioned experiment it certainly
had a velocity due to a height of almost 6 inches 2 lines. This obser-
vation confirms what I said in §15, namely that in divergent pipes the
aqueous stream is dilated as at m, and the motion of the same is re-
tarded. But in the present case, in order that both observations might
agree, it will have to be said that the stream was so dilated that it had
an area in proportion to the orifice NM reciprocally as the afore-
mentioned velocities or reciprocally as the roots of the heights due to
these velocities, namely as V 74 to V 58, and thus that the diameters
of the dilated stream and of the orifice were as {/ 74 to {/ 58, or as
IOO to 94.I.
EXPERIMENT 8. I performed another experiment which, although
it does not yet pertain to this, nevertheless I will recount: namely, at
the origin near the orifice GF [Fig. rg] I perforated the pipe with an
opening e of almost two lines, and again I observed the descent of the
surface from CD to EH, the water flowing through NM, and at the
same time I examined the amplitude of the thrust.
96 HYDRODYNAMICS, CHAPTER IV

I saw these two things which at first glance seem almost to contra-
dict one another: the descent from CD to EH was made more slowly
than it was in the preceding experiment, and now it lasted IO seconds;
and nevertheless the thrust PQ_ was greater for the same height oP, for
now PQ_ was Io inches Io lines.
I explain both Phenomena thus: on account of the orifice e, which
was made near CF, because it allows the free transit of air, the pressure
is removed which at other times the water exerts within itself in the
pipe, and the water accordingly does not flow differently where the
small opening e is than if the pipe were cut off at that very place; but
the water would flow more slowly, which I showed in passing, if the
pipe GNMF, as if diverging, were made shorter. Further, since the
water can flow, although at a lesser quantity, nevertheless with a
greater impetus, through the unchanged orifice NM without implicit
contradiction, the reason is the mixing together of air with water;
for air perpetually rushes into the pipe through the small opening e,
and together with the water it flows out through NM. And, finally,
it seems to me that the phenomenon according to which the water
flows actually through MN more quickly with the orifice e opened
rather than closed cannot be explained otherwise than that the
hindrances from outside act less on water rarefied by air than on
natural [water].

PERTAINING TO THE THEORY OF WATER WHICH FLOWS OUT OF VERY

LARGE VESSELS FROM THE POINT OF REST RIGHT UP TO A GIVEN
DEGREE OF SPEED

ExPERIMENT g. When water flows from a very large vessel through

an orifice made in a thin plate, the first drop bursts forth immediately
at the entire velocity which is due to the height of the water above the
orifice.
This conforms to the theory indicated in §11 if the vessel is truly
infinite; and even though it may not be infinite in a Geometric
sense, as long as it is very large, in a like manner no drop can be
observed at the beginning of flow which will not flow out at the
maximum velocity; I explained this Phenomenon in §14, since indeed,
by dint of the theory in some particular case reviewed there, scarcely
one or two drops should have defected noticeably from the maximum
thrust. I said that such a little quantity of water could not separate
itself from the subsequent water on account of the mutual attraction
or cohesion of the aqueous particles.
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 97
EXPERIMENT IO. When in fact the water flowed out from a very
large vessel through a pipe inserted horizontally in the vessel, I
observed before the stream flowing out formed the maximum thrust
omQ_ (see Fig. 19) that a noticeable enough quantity of water fell on
the horizontal table situated below, halfway between P and Q_;
that the greater is this quantity, the longer is the pipe GN and the
more it diverges toward N; and finally that that water is distributed
unequally, and that it obviously falls away more abundantly in the
place which is more remote from the point P than in that which is
nearer; however, in the understanding of the time in which all these
changes take place, l saw that that was very short and such that the
measure of it cannot be obtained.
All these phenomena follow as a unit from propositions which we
gave from §11 right up to the end of the section. But the measures
shown in that place cannot be confirmed directly by experiments,
especially those which are indicated in §§15, 16, and 17, where, as
you know, the formulas are communicated which express the quan-
tity of water flowing out while the maximum thrust is attained from
rest; the reason is: first, because the first drops which should have
fallen near the point P on the table did not separate themselves freely
from the water following; second, because the quantity of water
following the stream oQ_ (which certainly constitutes the maximum
portion according to the theory itself) cannot be intercepted; and
finally, because the motion of water through pipes is customarily
retarded very much by outside hindrances, particularly if the pipes
diverge, and thus the real motion is very different from the motion
which the water should have if all the hindrances were removed.
The remaining measures indicated by us are subjected to fewer diffi-
culties, and these of lesser moment; but they are contained in §20,
and they express especially the quantity of water which flows out at
the first instant of motion while the water attains a given degree of
speed.
Although on account of the reasons just mentioned, especially in
the case of divergent pipes, a perfect agreement of theory with ex-
periments cannot be expected at all, nevertheless I found such success
that I would understand that complete agreement would have oc-
curred easily if all hindrances together with the mutual adhesion of the
aqueous particles could have been prevented. Indeed, I performed
experiments on a divergent pipe as well as on a cylindrical one; let
me now explain them individually.
EXPERIMENT I I. In Fig. 19 a pipe in the form of a truncated cone
was inserted horizontally in a vessel; I filled the vessel itself with water
98 HYDRODYNAMICS, CHAPTER IV

right up to CD, so that the height of it above the axis of the pipe was
equal to 433 parts identical to those which I used in the entire ex-
periment. For that height I sought by experiment the point Q
corresponding to the maximum thrust, and PQ was 287 parts while
the height oP was 146 parts. Thus I saw that the motion of the
water, both on account of the adhesion of the water and on account
of the shape of the pipe, had been greatly retarded, which must occur
in these cases, as I warned several times. However, it should have
been, if nothing had opposed the motion, that PQ was 503 parts.
Next I placed a Pan on the horizontal table, the edges of which were
at S and R; but first I moistened the Pan, and I allowed all the
water to rain down from the former [height] again; and after the
measure of PR was taken, I found that to be 206 parts.
And finally the diameter CF was 13 parts, and MN 17 parts, but
the length of the pipe was 125 parts.
After all these things had been thus prepared, during which time
I covered the orifice MN with a finger, after the finger had been
suddenly removed, the water was ejected, and some part of this fell
on the pan; I collected this anxiously in a cylindrical glass pipe of
which the diameter was 8½ parts; that pipe was filled to a height of
2 IO parts, therefore the quantity of water having fallen on the pan
was 11,922 cubic parts.
But now that quantity, through §20, should be nb-V n/g ln _a_,
a - C
where by n the area of the orifice NM is understood, or 227 square
parts, by g the area of the orifice CF, or 133 square parts; further, b
denotes the length of the pipe, which was 125 parts; by a is properly
understood the height of the surface CD above the axis of the pipe,
but here rather the height complying with the velocity of the water
striking at the point Q is to be understood, or 141 parts, and simi-
larly for c is to be assumed the height complying with the velocity of a
particle striking at the point R, namely 73 parts. Finally, the abbre-
viated expression ln signifies the Hyperbolic logarithm. After these
numerical substitutions have been made, there occurs

• 1- a 17 141
nbvn/gln-- = 227-125---ln- = 26,830.
a - C 13 68

Therefore, the quantity of water found in the experiment was to the

quantity which the theory indicates after the consideration of hin-
drances has been set aside as 1 1,922 is to 26,830, which numbers, al-
though they differ by not a little, nevertheless confirm the theory
 TIMES TO BE EXPECTED IN THE EFFLUX OF WATER 99

splendidly, which very thing I now place clearly before one's

eyes.
In the formula nb-vnfg ln _a_, we took for a the height due to the
a - C

maximum velocity of the water flowing out, as in fact it was in the

experiment, not as it would have been if the obstacles had been re-
moved; certainly we set a = r4r, but in theory a = 433. However,
if that latter value is assumed, by retaining the value of the height
C = 73, One has nb'Vn/g ln _a_ equal to approximately 6700, which
a-c
number is now much less than the number found through the experi-
ment, since before it was so much greater. But so it occurs when the
height c is assumed to preserve its value: just as in fact the height a
was increased from r4r to 433, so also the height c is certainly to be
increased, and each height should be increased in the same ratio if the
hindrances resist equally to the first drops and those following; but,
the remaining things being equal, the particles experience less re-
sistance the slower they are moved, and accordingly also the drops
which fall on the near side of the limit R are retarded less than those
crossing that limit. From this it is easy to conclude that the height
c is to be increased in a lesser proportion than the height a, but the
proportion itself we cannot give, unless a posteriori by obviously
making the theory agree with the experiment; thus it is ascertained
that c is to be set equal to 120, which number clearly suffices, after all
circumstances have been well considered.
Thus, therefore, it seems manifest to me that the success of the
experiment was such that it clearly agrees with the theory. But
examples of this sort show wholly that we have transmitted the true
laws of motions into fluids, and these [examples] I selected from among
infinite others, because they have no relationship nor any affinity to
the common rule which states that fluids flow everywhere at a velo-
city due to the entire height of the water above the orifice, and they
cannot be solved by the usual principles. And in remainder, since
the motion of the water in this experiment was retarded, I wished to
undertake another one in which all hindrances would be altogether
diminished, so that thus it would appear that the numbers of the
experiment and of the rules would agree the more with each other,
the less would be the hindrances.
EXPERIMENT r2. And so I now used a cylindrical pipe through
which the flow might pass more easily, and for that very reason it was
wider: in addition, the box to which the pipe was attached was much
wider, and finally, the height of the water contained in the box above
l00 HYDRODYNAMICS, CHAPTER IV

the axis of the pipe was much less, so that the water would flow out at
a lesser velocity, and thus it would encounter obstacles of lesser
moment. The remaining things were as before.
Therefore the height of the water above the axis of the pipe was
130 parts, oP = 553 parts, PQ = 453 parts, PR = 297, the dia-
meter CF or MN= 19 parts, and the length of the pipe was 130
parts.
I saw that the water, having fallen into the pan, had filled up the
cylinder, which had a diameter of 8½ parts, to a height of 281 parts,
the capacity of which was then 15,950 cubic parts. In this case a is
to be set equal to 453 · 453 = 93 parts, c = 40 parts, n = g = 284
4-553
square parts, and b = 130. After these substitutions have been made,
it occurs that

nb-Vn/g In _a_ = 284,130 In 93 = 20,760,

a - C 53
to which the number in the experiment, 15,950, as we saw, corre-
sponds. But the latter number is almost four fifths of the other,
and thus it agrees approximately to the same, since in the preceding
example on account of the applied reasons a similar number defected
from a similar one by more than half.
Therefore, now it is amply evident that it is to be attributed to out-
side obstacles only that the experiments do not correspond to the
formulas correctly; meanwhile, nevertheless, the experiments are such
that they cannot demonstrate the strength of these formulas any
better.
 FIFTH CHAPTER

Concerning the Motion of Water from

Constantly Full Vessels
§1. Vessels are maintained full when as much water is continuously
poured in as flows out: but the pouring in can be in the same direction
as the motion of the aqueous surface and at the same velocity at every
instant, as if certainly a new surface were created continuously which
already possesses the velocity of the adjacent water, or [it can be]
lateral and without impetus, just as if the surface which is assumed
continuously to be created anew were provided with no motion and
finally is to be stimulated into motion by the water below. I will
pass over the remaining methods of supplying new water, which are
infinite.
Meanwhile, the rule about this motion, especially in the latter
phase, is accepted: that the water flows out at a velocity complying
with the height of the surface above the opening; nevertheless, it is
easy to see in advance that this cannot be valid unless for a vessel
infinitely wide everywhere, but in the remaining [vessels] it happens
that the motion beginning from rest will be increased very gradually
through some interval of time, and finally after an infinite time it will
acquire the entire velocity. Nevertheless, if one is to say what the
reason is, those accelerations for the most part occur so quickly that
the entire velocity is not present for only the shortest time. But the
situation is otherwise in very long aqueducts, in which the increases in
velocities do not escape ttotice, and they can be observed with separate
measurements.
But whatever the matter may be, since mathematical accuracy is
never objectionable, I undertook to consider and follow the motion of
water from the beginning to some given limit.
§2. All properties of this motion may be reduced to essentially
three equations: first, between the quantity of water ejected and the
corresponding velocity; second, between time and velocity; and third,
102 HYDRODYNAMICS, CHAPTER V

between the quantity of water and the time. If one ofthi.se equations
is obtained, the remaining follow from it spontaneously.
Therefore, we will scrutinize only the first one rather closely. But
here let us be mindful of those things which were advised in the pre-
ceding section about the contraction of a stream flowing out through
simple orifices or converging pipes, and the dilation of the same when
it is ejected through diverging pipes. However, we indicated in §3,
Art. I, Chapter IV, that the stream is to be considered until that time
when the velocities of the particles (diverting one's mind from the
changes which gravity produces on the particles beyond the vessel)
are not changed any further, and all that portion of the stream is to
be considered as moving inside the vessel, just as if the surface of the
stream became uniformly hardened there. Therefore, from now on
when the discussion will be about a vessel through which water flows,
that ideal vessel is to be considered, the orifice of effiux of which is the
section of the stream subjected to no further change except that which
is due to the descent or ascent of the stream.

PROBLEM

§3. To find the velocity of water flowing out of a constantly full

vessel after a given quantity of water has already flowed out.
SOLUTION. There are two methods of supplying water especially
worthy of considering, either of which postulates another solution to
the problem: for either the water is assumed to rain down vertically
into the vessel, and such indeed that it flows in at precisely the same
velocity which the surface of the water has, or the water flows in later-
ally and thus lacks the impetus by which the water could follow the
surface on its own and finally is to be excited into motion.

CASE I

In order that for the first case we may find the equation between
the quantity of water ejected and the corresponding velocity, this is
to be pursued, with a single circumstance changed, along the same
paths which we followed in the first paragraphs of Chapter III.
Therefore, as in §6, Chapter III, let the vessel aimb (Figs. 15 and 16)
be proposed, which is kept constantly full right up to cd by the inflow
of water; but let water flow out through the orifice pl; and let it be
established that that quantity of water has already flowed out which
can be contained in a cylinder erected above the orifice pl to a height
x, but that the last drop has flowed out at the velocity by which it can
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS 103

ascend to the height qs or v; thus, the equation between x and v will

now have to be shown.
Let the curve CGI be the scale of the areas, such indeed that, for
HL denoting the height above the orifice, HG expresses the area of the
vessel at that place. Next let a third curve tru be drawn, the ordinate
Hr of which is everywhere continuously equal to the third propor-
tional with respect to GH and PL, or, the ordinate Hr of which is
(PL) 2 /GH.
Let the space DCIL = M, the space DtuL = N, and the potential
ascent of the water contained in the vessel, after the previously men-
tioned quantity has already flowed out (through §2, Chapter III),
will be Z v. Further, let the particle plan be understood to flow out,
and the surface cd to descend to ef; now the height of the velocity for
the particle plan will be v + dv; and if now the parallelogram LxyO is
constructed, the side LO of which is lo, and the other, Lx, is PL, the
potential ascent of the same water in the position efmlonpie will be equal
to the fourth proportional with respect to the space EFLONPIE
(which again is M, because PLON expresses the magnitude of the
volume element plan, while CDFE expresses the minimum quantity
cdfe equal to that volume element), the space wuxyOLF (which is
equal to the space N - DtwF + LxyO, from which, if PL or Lx is set
equal to n, CD = m, and LO = lo = dx, there will be Dt = nn,
m
n n3
DF = - dx, hence the small space DtwF = - dx, and the space
m mm
n3
LxyO = ndx and finally the space wuxyOLF = N - - dx + n dx ),
mm
and the height v + dv. Therefore the potential ascent just mentioned
is (N - ;: dx + n dx) · (v + dv) /M, equal to the rejected differen-
. N N n3 n
t1alsofthesecondorder Mv + Mdv - mmMvdx + Mvdx,suchthat
the increment of potential ascent, which was added to the water while
the volume element plan was flowing out, is Z dv - m:~ v dx +
~ v dx, where the spaces N and Mare of constant size on account of the
continuous pouring in of water. In this first case we do not consider
the potential ascent of the volume element cdje, which is filled while the
other equal [volumeJ plan flows out, because that ascent is not gene-
rated by an internal force, nor indeed is the lower water considered to
 HYDRODYNAMICS, CHAPTER V

draw the particle cdje after itself, but rather we consider this to be
poured in continuously by a certain outside force, and this at neither
a greater nor a lesser velocity than that which pertains to the surface
eJ. Therefore the entire increment to be considered here is, as we
said,
N n3 n
- dv - - - v dx + - v dx.
M mmM M
But that increment must be equated to the actual descent of the center
of gravity. And that descent, after DL has been set equal to a, is,
from §7, Chapter III, n~x; therefore, the following equation results:

N n3
- dv - - - v dx
M mmM
-n
+M na dx
v dx = - - ,
M
or
dx = N dv / ( na - nv + ;: v).
But if this is so integrated that v and x vanish together, it gives
mmN mma - mmv + nnv
x = n3 - nmm 1n mma

which equation, after e has been established as the number the loga-
rithm of which is unity, is equivalent to

V = mma [ I _ e<n3 - nmm)x/mmN].

mm - nn
But this solution is suitable for the first case, where the water is
poured in from above with a motion which is common with the
descent of the nearest surface.

CASE II

Thus if now the particle cdje is considered continuously to be poured

in laterally, then on account of its own inertia it stays behind the
motion of the lower water, and accordingly the potential ascent of
the same enters differently into the computation. Moreover, first the
potential ascent of the aqueous mass cdmlpic increased by the volume
element soon to be poured in is to be considered; then the potential
ascent of the same water at the position cdmlonpic is to be investigated,
indeed, after the volume element has already flowed out, and their
na dx
difference is to be equated to the actual descent M . Truly the
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS l05

potential ascent of all the previously mentioned water before the pouring
in of the particle and after the pouring in of the same is thus deter-
mined: certainly the potential ascent of the water cdmlpic is,:;, and the
potential ascent of the particle ready to be poured in is nil, because,
poured in laterally, it does not yet have a common motion with the
lower mass. Therefore, the potential ascent of each [volume of] water
(which one obviously determines by multiplying the respective mass
by its own potential ascent and dividing the sum of the products by the

sum of the masses) is (M-1;; + ndx-o)/(M + ndx) = M :vndx·

But at that very time when the particle n dx was poured in from above,
it acquired a common motion with the water just below, and thus the
potential ascent of the same water in the position cdmlonpic becomes
equal to the fourth proportional with respect to the space
CDLONPIC-(M + n dx), the space DtuxyOLD (N + n dx), and the
height v + dv, that is, (N +;; dx)(vd + dv), the excess of which over
+n X
. . . Ndv + nvdx + ndxdv Ndv + nvdx
the pnor potential ascent 1s M + dx = M
(after the differentials of the second order have been rejected).
Therefore, the following equation is obtained
Ndv + nvdx na dx
M M
which, handled as previously and carried out to the end, gives
N a
x = -ln--
n a- v
or
V = a(I - e-nx!N)

which solution is valid for lateral pouring in.

ScHOLIUM 1

§4. These equations are altogether different from each other; more-
over, the greater the difference, the less the area of the vessel; and if
indeed the uppermost area of the vessel at cd is more or less infinite
with respect to the area of the orifice, n vanishes with respect to m,
and it occurs in the former case, just as in the latter, that
V = a(I - e-nx!N).
I06 HYDRODYNAMICS, CHAPTER V

Therefore, in this hypothesis the motion is the same on either hand,

which everyone could have seen in advance with no difficulty at all.
But the motion in the former pouring in is always swifter than in the
latter, the remaining things being equal.
Here it is convenient to explain the matter physically as well, in
order that we can perceive it more distinctly in all phenomena.
In place of any vessel whatever having any direction whatever, for
the sake of a shorter delineation, let there be a vertical cylinder with an
orifice at the base, namely GHND (Fig. 29), and then let the vessel

FIGURE 29

EFPQ, be perforated at RS; let the orifices RS and GD be assumed

perfectly equal and corresponding to each other perfectly at a mini-
mum distance apart, so that all the water flowing out from the upper
vessel flows into the cylinder placed below.
Let the water begin to flow from each vessel, but let it be assumed
to flow constantly from the upper at that velocity which the surface of
the water in the cylinder below has.
Thus it is evident that this is satisfactory for the first condition of
filling. But now, to see whether they agree with the preceding, we
will investigate the phenomena of this motion.
Therefore, let us consider the upper vessel as if it were infinite, so
that the water flowing through RS at every instant has a velocity
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS 107

which complies with the height PB or FA; thus it will have to be con-
sidered that this height PB is infinitely small at the beginning, because
then the water must flow at an infinitely small velocity, but that it
then increases gradually, and this continuously more and more, until
after an infinite time the motion remains uniform; however, it is
asked whether the height PB of the water will finally become infinite
or if indeed it will not pass beyond a certain limit. This is deter-
mined as follows.
Let the height CH or RH (for it is not to be considered that they
differ from one another) be a, AF= x, the area of the orifice LM be n,
the area of the orifice RS be m; because, indeed, as is manifest, the
two vessels can be understood to cohere and thus become one, the
velocity of the water at LM after an infinite time (from §23, Chapter
III) will be '\/i:i+x, and at RS it will be vx
(which is evident after-
wards if now the vessels are considered separated again, for either can
be assumed without error), but the velocities must be in the inverse
ratio of the areas of the orifices; and so Va + X : Vx : : m : n, from
which (a + x) : x: : mm: nn, or a: x:: (mm - nn) : nn; therefore,
x = nna and a +x = mma ; therefore, we see that the height
mm-nn mm-nn
mma
due to the velocity of the water at LM is, in this way, - - - - ,
mm - nn
certainly after an infinite quantity of water has already flowed out;
but above we had the same height, or
V = mma [ I _ e<n 3 - nmm)x/mmN]
mm-nn '
where if one sets x = oo (for in an infinite time an infinite quantity
flows through), the exponential term vanishes if only mis greater than
mma
n, and thus equally there appears v = - - - - . That agreement is
mm - nn
remarkable, because the paths which we followed are greatly dif-
ferent. On the other hand, if m is not greater than n, the motion
never becomes steady, not even after an infinite time, for the velocity
then increases to infinity, while otherwise the height of the velocity
. h mma
never surpasses t h e h e1g t - - - - . Therefore, concerning the
mm - nn
latter cases there is nothing we may say.

ScHOLIUM 2

§5. Now another question occurs here worthy of being noted:

indeed, what can be the mechanical method of filling in order that the
108 HYDRODYNAMICS, CHAPTER V

vessel above remain full to the required height during the entire flow?
That Problem would be difficult on account of the inconstancy of the
desired height unless a peculiar artifice were applied here, which I
shall now treat.
However, it goes beyond the fact that the water in the minimum
space RSDG undergoes no compression, either positive or negative,
because from the hypothesis it is moving at a common velocity with
the water just below, and thus no particle tends to propel or retain
any other.
Therefore, let each vessel be made as I said, and let a pipe be
attached to the vessel above (for no other purpose than demonstration
did we previously consider it separated), but let the pipe have a
small opening at the highest point a (Fig. 30), to which the short pipe
!J

le

---~-_J J

FIGURE 30

am corresponds; in this short pipe let the curved glass tube abcdg,
sealed at the mouth mn by wax, be inserted; let the horizontal ae be
constructed, and the point e be marked. With these things prepared
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS 109

thus, one will have to make the highest level of the water during the
entire experiment remain constantly at the point e; and for this, one
will see that it is required that at the beginning the surface of the
water be near the base FP, further, that it be elevated continuously,
and finally that after some time, albeit infinite, it never surmounts the
height nna ; but it will be easy to moderate the pouring in of
mm - nn
water so that the surface does not diverge very much from the point e,
if only the circumstances are not so matched that the water is to be
added overly quickly at the beginning.
But thus, if one notices that the surface in the small tube is elevated
above the point e, one must restrain the pouring in a little, which I
will show is to be done elsewhere; ifit should be the opposite, one must
pour in the water more abundantly.
That kind of experiment, of which sort I performed often, involves
no difficulty, but, lest any error creep into the experiment, the
capillary effect of the glass tube is to be examined; one finds this
effect if, after the orifice LM has been blocked, and before the tube has
been moistened, the cylinder is filled with water right up to the
summit, and thus one finds that the surface of the water in the tube
extends right up tof, a point certainly higher thane; however, one
substitutes this point J for that about which we just spoke in order to
disregard the character of capillary tubes.
Therefore, in this way the filling will be properly done according to
the rule of our hypothesis, and finally experiments can be performed
concerning this motion. But thus, after we have explained the
matter freely enough, I think it is unnecessary to warn that the
vessel above does not pertain otherwise to the cylindrical vessel
below, which we consider alone, than to the extent that the cylinder
there is kept full in that manner in which it must be done, and thus
by m is not to be understood the area of the upper vessel, but the area
of the orifice RS, which, especially to us, is the surface of the water,
since the water above RS serves only for the proper supplying to the
cylinder below.

ScHOLIUM 3

§6. Here I must not overlook the fact that thus a case occurs which
pertains to h;ydraulico-statics, concerning which science I advised certain
things in Chapter I, §8: certainly we know now at what velocity the
water must flow by at a in order that its pressure against the sides of
the pipe be precisely null. But while I was writing these things, I
110 HYDRODYNAMICS, CHAPTER V

had already formulated the general laws of hydraulico-statics, and I saw

not without pleasure that that case deduced as a corollary from a
clearly different theory acquires a similar solution from the general
theory. Thus all things conform everywhere to a mutual relation-
ship, and they show a legitimate application of the principles.

SCHOLIUM 4

§7. Now certain things follow concerning another method of

adding water. Let the cylinder RHNS be assumed as any vessel
whatever, and let it be kept constantly full by lateral pouring in; this
can be done by injecting a sufficient quantity of water through the
short pipe ma; but although this does not occur without motion,
nevertheless, because it is horizontal here, soon all is removed, and on
its own it neither advances the flow through the cylinder nor retards
it; but there is, furthermore, another method which we find, after the
calculation has been performed correctly, to reduce to the same:
namely, ifwe consider the vessel EFPQinfinitely large, and we under-
stand the base ofit [to be] continuously covered with water, but such
that the height of the water in the vessel above is to be taken as in-
finitely small, the vessel above will furnish water to the pipe attached
to it, and no other motion will arise thereby than from lateral pouring
in, if only the orifice RS always remains covered over; but it easily
occurs that a certain cataract is formed there if the orifice LM is
large and the pipe RSNH is long. Since here the other method must
put forth the same effect as the former in the motion of water, every-
one sees from this that in each method the inertia of all the water
flowing into the pipe is to be exceeded by [that of] the water below.
But the same could also be shown a priori by inquiring into the motion
which must arise thence according to the equation of§8, Chapter III,
which is this:
mmvy dx mmv dx d
NdV - --'---
nn
+ ---
y
= -yx X.

But it will be accommodated to the present case if form, x, and -dx

one substitutes, respectively, n, a, and n dx (the reason for which act
y
will be evident if one matches these with the others), and at the same
time one setsy to be infinite; for then the third term of the equation
vanishes, and it occurs indeed that for the present treatment above
we find
N dv + nv dx = na dx.
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS I I I

After we showed in these scholia the nature of each motion, as much

as a simple physical consideration of the matter permits, and the dif-
ference between them, and at the same time we treated the mechanical
method of advancing those hypotheses to law, it remains that the rest
of the rather notable phenomena be indicated as well, which I will
now do.
§8. COROLLARY r. If in the vessel RSNH the entire base is absent,
the orifice LM will equal the orifice RS; the latter can even exceed the
former if indeed the sides of the vessel diverge. But in these cases
the height v has no limit in the equation

V = mma [1 _ e<n 3 - nmm)x/mmN]

mm - nn

and it becomes infinite if the quantity of ejected water, indicated by

nx, is infinite.
This indeed is evident intrinsically from the equation when n is
greater than m; but when the areas of the orifices are equal, one is to
return to the differential equation of§3, from which the next equation
was deduced, namely,

N n3 n n
-dv - --vdx
MmmM
+ -M vdx = -adx
M'

which, for n = m, gives N dv = na dx, that is, v = n;, where v 1s

manifestly infinite if xis infinite.
§9. COROLLARY 2. But if the proposed vessel has a base and an
orifice in it, the area indicated by n of which is less than the area
of the orifice RS, expressed by m, v has a value which indeed it never
attains, but nevertheless reaches approximately, and to which it
converges so quickly after a minimum perceptible time of flow that
it is not noticeably different unless by special effort vessels contrived
contrary to this matter are furnished. But that term is as follows:
v = mma ; therefore, in the case of Scholium 2 of§5, the last term
mm - nn
nna
PB is v - a= I will illustrate by example the very quick
mm - nn
accession of the velocity to its ultimate limit, after I have applied
the equation between v and the time corresponding to the height v.
§10. COROLLARY 3. In the case of the pouring which we call
lateral, the ultimate height becomes v = a, whatever ratio may exist
between both orifices of the vessel.
112 HYDRODYNAMICS, CHAPTER V

§11. COROLLARY 4. If the vessel is cylindrical and its length is set

equal to b, there occurs (see §3) N = nnb; but let it be noted that the
m
values of the letters a and b are not to be confused, for the first ex-
presses the height of the uppermost orifice above the lower, the other
the length of the conduit. And so the values thus agree with each
other at least in that case in which the axis of the vessel is a straight
line and vertical; but if the axis is tortuous, or at least not vertical,
they differ from each other. Therefore, I wished to advise this ex-
pressly, lest anyone be misled by the shapes of the vessels, the axes of
which I made straight and vertical everywhere. If, therefore, for
cylindrical vessels one sets N = nn b, for vertical pouring in
m

v
mma
= -- - - (1 _ e<nn - mm)x /mnb)
mm - nn
and for lateral, v = a(1 - e-mx/nb).

PROBLEM

§12. To find the velocity of the water flowing out from a constantly
full vessel after the flow has taken place for a given time.
SOLUTION. With the hypotheses and all the notation retained which
we applied in §3 and, further, with the time elapsed from the beginning
of flow having been taken equal to t, we will have to change the
equations given in that paragraph into others which express the rela-
tion between t and v, after the quantities x and dx have been elimi-
nated. But the element of time difference dt is proportional to the
very small space dx which it passes through divided by the velocity
vu; therefore, we will set dt = ~~, and thus the equation
dx = Ndv/(na - nv + ;:v)
which was given for determining the required velocity for vertical
pouring in, will be changed into the following:

(I) dt = N-y·dv I( na-Vv - nv-Vv + ;: v-vv);

but the other, serving for lateral pouring in, namely, dx = N dv/
(na - nv), is changed into the following after the same substitution:
(II) dt = N·y·dv /(na-vv - nv-vv).
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS I I3

But these equations, integrated in the required manner, give for the
first,

mNy mva + vmmv - nnv

(a) t = --======= ·1n -,=--,=====
nvmma - nna mva + vmmv - nnv
and for the other, which is deduced from the former, after one has
set m = oo,

(f3) t = Ny -ln Va + vv. Q.E.I.

nva va - vv
ScHOLIUM

§13. If the vessel which is being discussed is cylindrical and twisted

and inclined in any way whatever, and the length of it is set equal to
b, the height of the aqueous surface above the orifice remaining equal
to a, there will be again, as in §n, N = nn b.
m
But since, as is well known, 2yVA expresses the time which a body
takes in falling freely and from rest through the height A, it is evident

that the quantity :::~ ( = 2y Jb:) expresses the time in which a

body beginning to be moved from rest descends freely through the

height bb; we will accept that time as a common measure and we will
a
set the same equal to 0, and the equation (a) for vessels or cylindrical
conduits will be changed into this:

t= n0 -ln mva + vmmv - nnv.

2Vmm - nn mVa - Vmmv - nnv'
the other suitable one, designated by ({3 ), becomes the following:

t = n0 ln va + vu,
2m Va - Vv
from each of which it appears that the water cannot but acquire
almost the full velocity in a very short time, and this all the more
quickly the larger is the pipe, the shorter, and the more nearly ver-
tical; and that the accelerations are not perceptible in any way,
unless the aqueducts are made very long; and then also almost all
grades of accelerations are passed through in a short time, each of
which I shall now illustrate by an example.
I 14 HYDRODYNAMICS, CHAPTER V

I. The time is sought in which fluid from a constantly full vertical

cylinder, 16 English feet long, the diameter of which is five times the
diameter of the orifice, acquires a velocity which is due to the height
.filL a according to the hypothesis to which the second equation per-
100
tains; thus!!._ = _:_, v = .filL a, and b = a, whence the time which a
m 25 100
body takes in falling freely through the space bb or 0 equals one
a
second; hence t becomes _:_ ln 399, that is, approximately the ninth
50
part of one second, which short time is indeed imperceptible. But
when the time is considered notable, the changes of the heights v
become unnoticeable. If the similar time (in which certainly the
velocity due equally to .filL of the height which results after an infinite
100
time is generated) is sought in the first hypothesis, namely, the time in
which there is obtained v = .filL · mma , then that is found to be
100 mm - nn
slightly greater than the preceding, but of unnoticeable excess; whence
it is evident in vessels of this sort that the water cannot be poured quite
quickly enough into the upper vessel to satisfy the hypothesis, and
therefore, by reason of the same, that other experiments cannot be
performed with regard to the hypothesis in order to show whether the
height BP in Fig. 30 is truly as great as it should be by virtue of§5, in
order that the point e or f retains the position during flow which it had
before flow when the orifice LM was blocked off and no water existed
in the upper vessel.
II. Now the same time is sought again for the second hypothesis, if
the pipe is of the same area and has the same orifice attached to it and
has an oblique position and a length b of 184 poles or 1104 Paris feet,
while the height of the aqueous surface above the orifice of effiux is
r6 Paris feet. Thus it will occur that b = 104, bb = 76,176, and
I
a
0 = 72 seconds, more or less, from which the mean desired time is
between eight and nine seconds, which certainly is noticeable enough.
But if the time is desired in which the height vis equal to only a fourth
part of the height a, that will be found equal to H ln 3, or approxi-
mately one second and a half.
I do not know whether these things agree with those observed by
Mariotte to which he refers in his Traite du mouvement des eaux, Part V,
Disc. 1, where he mentions a certain leaping fountain, which is at
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS I I5

Chantilly, to which the water descends through a conduit 184 poles

long, if only I concluded correctly from the antecedent [discussion],
and the greatest height of the aqueous surface above the orifice of
efflux had been indicated by a as 16 feet; the diameter of the aqueduct
was 5 inches, but the orifice had a diameter of one inch. Mariotte
seems to me to speak thus, as if the accelerations had been much
slower than is indicated by our formula, although I do not know
whether this is to be attributed to the fact that the water perhaps may
have had some other exit in addition to the orifice which is being dis-
cussed here, or that the aqueduct had not been full of water when
flow was beginning, which many things occurring afterwards caused
me to believe; ifit has been neither, I believe that the phenomena, as
they had been observed by Mariotte and can be observed again daily,
clearly agreed with our calculation. Finally, these words are
Mariotte's: "Furthermore, this," he affirms, "happens to the same individual
thrust: when an orifice has been blocked by hand for a space of time of ten or
twelve scruples of a second, and afterwards the same has been opened, water does
not burst forth at once, but, surging slowry, the thrust ascends to 3 inches,
afterwards to a height of a foot, and.finalry to two feet, at successivery noticeable
intervals. . . . But, nevertheless, .finalry the water springs forth with its full
impetus."

PROBLEM

§14. To find the quantity of water flowing through a given vessel,

constantly full, in a given time.
SOLUTION. With the positions and notation of§§3 and 12 applied
again, now the equation between x and tis to be found; but because,
as we saw in §12, dt is rJ~, y1; will be yd~X, and this value has to be
substituted in the integrated equations which we gave in §3; the former
of these equations was this:

V = mma [ I _ e<n3 - nmm)x/mmN],

mm - nn

which, according to the previously followed custom is changed into

this:

(I) yy(dx)2 = mma (1 - e<n3-nmm)x/mmN).

r;Ji)2 mm-nn
II6 HYDRODYNAMICS, CHAPTER V

The other of the mentioned equations from §3 was the following:

v = a(1 - e-nxfN)

which therefore supplies the following in the present case:

(II) yy(dx)2 = a(I - e-nxfN)

(dt) 2 •

Now the equations (I) and (II) have to be integrated, which is

certainly easy, and because the former contains the latter (for each is
the same if m = oo), we will treat that one alone, and we will now
consider it in the following form:

dt = yVmm - nn·dx/Yr - e<n3-nmm)xt mmN

mVa

But let it be established, so that the method of integration is ex-

plained all the more, that e<n 3 -nmm>xt mmN = z, and, accordingly,
dx = ( 3mmN dz) , from which, for the sake of brevity, the constant
n -nmmz
. yv'mm - nn mmN -ymN . . d" db
quantity . ;- , or -,=====, 1s 1n 1cate y
3
mv a n - nmm nv' (mm - nn)a
a, an d t h ere WI·11 resu1t dt = . a dz
1
, . w h"IC h , I"f In
In . a dd"Itlon
. .
It
ZV I - Z
occurs that r - z = qq, or z = r - qq, dz = - 2q dq, there develops

-2a dq -adq a dq
dt = - - -
1 - qq I + q I - q
the integral of which is
I - q
t = a ln ( I + q) + a ln ( I - q) = a ln - - ·
I + q

And there is no need of a constant, since indeed from the nature of

the matter t and x must vanish simultaneously; but for x = o it occurs
that z = r and q = o; therefore, equally t and q must begin simul-
taneously from zero, which condition the derived equation
I - q
t = a ln - - satisfies. It remains that we reassume the original
I + q
values in reverse order; in fact, it thus occurs that

1
t = a ln - - ~ - - -
v1-z
1 + VI - Z
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS I I 7

or
t = y
mN r+v~
ln-----
nY (mm - nn)a r - Vr - z

or, finally,

(I) t = ymN ln [ r + Yr - e<n3 - nmm)x/mmN]

nY(mm - nn)a
_ ln [ r _ yr _ e<n 3 - nmm)x/mmN].

And that equation, after one has set m = oo, gives the other
desired equation,

(II) t = ~~In [r + Yr _-nx /N] - ln [r - Yr - e-nxlN]

nva
Q.E.I.
§15. COROLLARY r. If one sets x = oo, in order that the nature of
the matter might appear when an infinite quantity of water has
already flowed through, and m is assumed greater than n, just as is
mostly customary, the exponential quantity is to be considered to
vanish in both places if the logarithm has been assumed positive, and
ln 2 will occur on either hand. But indeed, if the logarithm has been
assumed negative, it is to be stated that
yr _ e<n 3 - nmm)x/mmN = r _ ½e<n 3 - nmm)x/mmN

and accordingly
ln [ r _ yr _ e<n 3 - nmm)x/mmN] = ln ½e<n3 - nmm)x/mmN
n3 - nmm
= mm
N x - In 2.

If these substitutions are properly made for the first method of

pouring in which we devised,
3
(I ) ymN ( mmn - n )
t=-===== 2~2+ N X
nY (mm - nn)a mm
which, for m = oo, gives for the other case

(II) t = :~ ( 2 ln 2 +ix)·
It follows from the above formulas: that certainly water flows at a
lesser quantity than if it would flow from the beginning at the entire
118 HYDRODYNAMICS, CHAPTER V

velocity which it acquires in either case after an infinite time; that the
difference, however, never surpasses a certain limit; and that after an
infinite time it is described in finite terms.
§16. COROLLARY 2. When we convert the derived equations, we
obtain

(I) x = -2mmN
- - -3 [ In (1 + e-tlx) - ln 2 + -t]
mmn - n 2a

(II) X n
= 2N [ In (1 + e-t//3) - ln 2 + _!___]
2(3
-ymN -yN
where, as above, a = ----=======, and f3 = . r ·
nV (mm - nn)a nv a
If in addition, as in the last corollary, one sets t = oo, unity vanishes
with respect to the exponential quantities, which are infinite beyond
all degree, and there results
t
a'
and ln (1 + e- 1113 ) = - ~)

from this then, after the values of the letters a and f3 have been
considered again,

(I) X = -,====
mtva 2mmN
1n2
yVmm - nn mmn - n3

(II) X
tva
= -- -
2N
-ln2.
y n
Therefore, if suddenly from the beginning of flow the water would
flow constantly on either hand at the entire velocity which it can
acquire, its quantity should not exceed after an infinite time the
quantity corresponding to the theory for the same time except by the
2
very small quantity which in the first case is expressed by mmN ln 2
mm - nn
and in the second by aN In 2. But if in place of an infinite time one
n
takes a time of only a few scruples of a second, the same theorem will
hold approximately, so that, for instance, if after the first ten seconds
2
the quantity Q has flowed out, Q + mmN ln 2 will thereafter
mmn - n3
flow out in approximately the same number of seconds, or in the other
case, Q + -2N
n
In 2.
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS I 19

ScHOLIUM

§17. The motion of water through siphons pertains as well to the

theory set forth thus far. However, the theory indicates that the axis
of a siphon can be inflected in any way whatever and the motion of
the water thence will not be altered, if only the height of the aqueous
surface above the orifice of efflux remains the same; since in addition
aqueducts, siphons or diabetes, and other vessels of this sort are usually
cylindrical, it is to be established, as I advised in §13, whenever this
applies, that N = nn b, by understanding that b is the length of the
m
conduit or siphon; also in the formulas of§§r4, 15, and 16, the quan-
tities are to be so interpreted, where the question concerns time, that
2yVA represents the time which a body having begun from rest uses
in a descent through a vertical height A.
For the rest, as I said in passing, the theory of this chapter indicates
nothing unique which falls under observation except in very long
aqueducts, greatly oblique to the horizontal, and having not very
small orifices; for these three things combine to the end of retarding
and thus effecting noticeable accelerations, the measures of which
very strongly support the theory.
Nevertheless, even in these circumstances some average is to be
observed lest the hindrances developing from the adhesion of the
water are excessive.
As far as the pouring in of water is concerned, I seemed to notice
for myself that, if it occurs vertically and with impetus, the motion
hence is so far from being accelerated that, rather, it is retarded,
unless the pouring in of water occurs equally over the whole surface,
in the matter which I showed in §4; for, if it is poured in otherwise,
the motion of the water in the vessel is disturbed, and this disordered
motion retards the efflux.
§18. Finally, to this point there pertain to some extent the experi-
ments undertaken by the celebrated Giovanni Poleni, as he reports in
the first book of De Motu Aquae Mixto, p. 12ff., which, therefore, I
thought should be brought forth here, since they show splendidly that
everywhere the ultimate speed in constantly full vessels is that which
agrees with the entire height of the water, if the vessels are not sub-
merged, or with the difference of the heights of internal and external
water in submerged vessels, although as for the rest there is nothing
in them thus far which is new, because there no accelerations are
considered.
Take a cylinder oflength as if infinite, the axis of which has a vertical
120 HYDRODYNAMICS, CHAPTER V

pos1t10n; let the base be whole; but in a wall let there be a fissure
parallel to the axis, rendering an orifice [in the form] of a rectangular
parallelogram which extends from the base right up to the summit of
the cylinder. Further, consider water to be poured into the cylinder
steadily, so that in equal times equal quantities are injected; the water
flows from the cylinder through the fissure; nevertheless, from the
beginning it does not flow at the same quantity at which it is poured in
from above, but at a lesser one; therefore, the surface of the water in
the cylinder surges right up to a certain height asymptotically; but if
this limit is now known to be reached, the height of the water will
remain unchanged, and the water flows out constantly at the same
quantity at which it is poured in. It appears also that the greater
will be the height of the water in the cylinder, the more [water] is
poured in. And so there is sought, for increased quantities of water
to be poured in in a given time, in what ratio the heights to which the
water surges in the cylinder must increase.
The solution is this: Let the height of the water, when it is in a
permanent state, be a, and let the part which is x be terminated at the
surface, together with the differential dx; let the width of the crack
be n, and we will have [something] just like an orifice of area n dx,
through which water flows at a velocity -Vx, therefore, the quantity of
water flowing out here in a given time is, for instance, n dx -vx, the
integral of which is jnx-Vx; this expresses the quantity of water flow-
ing out in a given time through the partial length x of the crack; and
thus the quantity of water pimring out in the same time through the
entire crack will be expressed by Jnav;; but only as much flows out
as is poured in; hence if the quantity of water poured in in that given
time is called q, there will be ~na-V; = q. This indicates that the
quantities of water to be poured in in that given time follow a three-
halves-power ratio to the heights to which the water ascends from the
base of the cylinder, or on the other hand that the heights follow a
cube-root ratio to the squares of the quantities at which the water is
poured in in a given time.
§19. With this problem having been solved, I come to the other
considered by the celebrated Poleni.
Let the cylinder be the same, but submerged in water standing in a
trench, as if in an infinite vessel, and let the depth of submersion be
designated by a; now, with the same things having been assumed as
before, again the equation is sought between the height a of the internal
aqueous surface above the external, and the quantity q to be poured
in in a given time.
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS 121

With respect to that portion of the crack ex which ejects water and
rises above the external water, we already saw that it expends the
quantity -inexv'; in a given time; but the remaining submerged
portion of the crack transmits water everywhere at a common
velocity, as will be evident from things to be discussed below, and
certainly at the velocity v';, so that, after this velocity has been
multiplied by the magnitude of the submerged crack na, the quantity
naVa is obtained which it ejects in a given time. If both quantities
are gathered in a sum, it will result that (!ex + a)nv'; = q.
With the help of this equation, q is known from the given heights
a and ex; or, on the other hand, the height ex [is known] from the
known quantities a and q.
Moreover, the very celebrated author himself, whose solution does
not differ from this of ours, shows that this equation agrees very
accurately with the experiments. It follows from the equation that
the elevations ex are the greater for the same pouring in of water, the
less is the depth of submersion a.

EXPERIMENTS WHICH PERTAIN TO CHAPTER V

PERTAINING TO §5

I used the vessel described in §5 with a glass tube (Fig. 30). But
first I covered the orifice LM and filled the pipe RN with water until
its surface touched the little opening at a; then I observed that the
water having gone into the tube reached the pointf at the extremity;
after the orifice LM had been opened and the water was flowing,
I poured new [water] into the vessel EFPQ, above, care having been
employed so that the extremity of the water atf meanwhile neither
ascended nor descended. While these things were occurring, the
surface AB was elevated, but it never exceeded a certain limit;
indeed, the maximum height PB or FA, as far as I could see, was
nn a, with!: denoting the ratio between the lower orifice LM
mm - nn m
and the upper RS, and a the vertical height of the latter orifice above
the other.
But this is the only experiment which I myself undertook; although
there are many propositions contained in this section which would
merit attention, and these sufficiently unexpected, nevertheless I was
not able to perform experiments concerning them; for things in the
shorter vessels are so composed that whatever unusual [property] they
122 HYDRODYNAMICS, CHAPTER V

possess escapes observation; moreover, I was not able to prove the

matter aptly in long aqueducts; when the opportunity will arise for
others to examine this theory, they should turn their attention to the
following:
I. In leaping fountains the whole height of the thrust may be
observed; after the orifice has first been closed, and the same soon
opened, the quantity of water which flows out may be seen, while the
water reaches half the height of the entire thrust, or any part what-
ever, which certainly will happen in the shortest time; let the measure
of that quantity be the length of the cylinder erected above the orifice
through which the water springs forth, which length we have called
x, but we have called the entire height of the thrust a, and the ob-
served height of the thrust which has not yet reached the total height
we have designated by v. Then at last, after the calculation has been
performed, let it be examined whether these quantities correspond
correctly to the equations for either method of pouring in shown in §3.
II. Let all things be done as before, with this difference only, that in
place of the quantity flowing out let the time of effiux be noted, so that
thus the formulas of§13 can be examined, and finally let the quantity
be compared with the time of flow, in order that this may show
whether it compares properly to the formula of§14.
III. Then expecially let that kind of experiment be performed
which I indicated in §16, by observing certainly the quantities of
water corresponding to the halftimes; but I said that, however great
a time is taken, the difference of these quantities never equals
2
mmN In 2 in the former method of pouring in which we estab-
mmn - n3
2
lished, or N In 2 in the latter. But these differences, although they
n
will never develop perfectly, will nevertheless be reached closely in a
very short time.
The remaining things in this section are Corollaries and Scholia;
anyone will see easily how they can be subjected to experiments.
But let me wish, before judgment is made, that attention be paid to
all circumstances with regard to the hindrances, the contraction of the
stream, and other things which I do not care to repeat everywhere.
To§§17 and 18: For the confirmation of the problem of§17 pertaining
to vessels not submerged, see the experiments on p. 26, lib. cit. of
G. Poleni.
But since in the submerged vessel the height a should be 55 Paris
lines (which height is called dead by him), he undertakes five experi-
ments in which the height which he calls live, or ex, was successively
 MOTION OF WATER FROM CONSTANTLY FULL VESSELS 123

8¾, 25, 42, 58, and 73½ lines; after these values have been substituted
in the equation shown in §r8, it follows that the quantities of water
poured in in a given time were as r oo, 199, 299, 396, and 495; actually
they had been poured in in proportion as roo, 200, 300, 400, and 500;
the difference is so little that it cannot be doubted but that the
agreement would have been perfect if all measurements had been
taken most correctly.
Also, the remaining experiments undertaken by that celebrated
man agree perfectly with the theory; the calculation of them is seen
among the works of that same Author. But I undertook in the
interest of the matter to include them here because they pertain to the
argument of this very section, although as for the rest, I may say
freely that I long more for those experiments which by calculation
depend on a change of instantaneous [conditionsJ, considered to this
time by no one that I know, rather than those which assume the
permanent state.
 SIXTH CHAPTER
Concerning Fluids not Flowing out but Moving
within the Walls of Vessels

§1. Up to this point we have considered water flowing out; but

now we will contemplate the motions of water which does not flow
beyond the bounds of vessels. Let me reduce all these motions to two
kinds, each to be treated separately:
First: When the fluid in an infinitely long pipe is moved continu-
ously in the same direction.
Second: When it is driven in reciprocal or oscillatory motion.

CONCERNING THE MOTION OF WATER THROUGH

INDEFINITELY LONG CONDUITS

CASE I
§2. First let there be a conduit placed horizontally but varying with
respect to [cross-sectional] areas according to some given law; let a
fluid be so placed in it, as it customarily occurs in rather narrow pipes,
that both end surfaces obtain an alignment perpendicular to the axis
of the conduit and thus begin to be moved at a certain given velocity.
If these things are so, and clearly no impediments to the motion are
assumed to be present, it is obvious that there will be no end to the
motion, in the same way that a sphere progressing freely on a hori-
zontal table continues its motion without end. Nevertheless, a sig-
nificant difference arises between the two motions; namely, all
portions of the sphere progress continuously at a uniform velocity; in
water they perpetually change motion. And it will not be difficult to
define that motion, if we will consider that the motion must be such
that the same potential ascent of the entire water which existed at the
beginning of motion is conserved. But we have determined the
potential ascent of water moved at a certain velocity in any conduit
 FLUIDS MOVING WITHIN THE WALLS OF VESSELS 125

whatever in §2 of Chapter III: therefore, nothing further remains

with respect to the solution of the problem. Nevertheless, it will not
be useless to have introduced one or another example of this matter.
EXAMPLE I. Let there be, for example, the conduit BgfC (Fig. 31),

FIGURE 31

which has the shape of a truncated cone; let the portion BGFC of it be
considered filled by fluid moved toward gf; and let the particles of
fluid at CF have a velocity due to the height v; and, finally, let the
fluid have arrived at the position bgfc. With these things established,
the velocity of the fluid at gf is sought. Moreover, I will assume that
V equals the height due to the velocity of the water at gJ. Let the
vertex of the cone be at H, the diameter at BC = n, the diameter at
CF= m, the length BG = a; if GB = b, the diameter gf will be
ma - mb + nb. Hence, because the solid BGFC is equal to the
a
solid bgfc,
(BC) 2 -BH - (GF) 2 -GH = (bc) 2 -bH - (gf) 2 -gH,
from which
(bc) 2 -bH = (BC) 2 -BH - (GF) 2 -GH + (gf) 2 -gH;
BH
but bH = BC· be, therefore,
2
(bc) 3 = (BC) 3 _ (GF) ~~H-BC + (gf)~r:-BC

= (BC) 3 - (GF) 3 + (g/) 3

or
3
(ma - mb - nb)
bc =
3J
n3 - m 3 + a
•

But, from §3, Chapter III, the potential ascent of the water in the
3
position BGFC is ( 3m v
) ; and similarly the potential ascent
n mm+ mn nn +
of the same water in the position bgfc is found equal to
3asv
fJ(aa + a{J + fJfJ)'
126 HYDRODYNAMICS, CHAPTER VI

where, for the sake of brevity, a and f1 have been inserted for the de-
termined values of the diameters gJ and be. Therefore,

V = m 3 • (aa + af1 + f1f1 )f1v.

a3 • (mm+ mn + nn)n
From this formula it is easily understood that the anterior particles
are moved at a continuously greater velocity, the posterior at a lesser
and such that, if the little orifice gJ is considered infinitely small, the
velocity of the water at g.fbecomes infinite and at be infinitely small.
EXAMPLE 2. Let there be a conduit composed of two cylindrical
pipes BN and OP (Fig. 32 ) of unequal area; in the wider branch let

FIGURE 32

the fluid BGFC be assumed to be moved toward Pat a velocity which

corresponds to the height v. Thus it is obvious that no change will
occur in the motion before the surface GF will have reached MN; but
from this point of time the motion will be varied continually until all
the fluid will have entered the narrower pipe. Therefore, one seeks
the velocity of the surface Jg when the fluid occupies the position
bgfc; moreover, we will designate the height required for this velocity
by V.
Let the diameters GF and g.fbe as n and m; let the length BG = a;
[if] bM = b, Og = !!!!.... (a -
b) ; the potential ascent of the water at
mm
BGFC is v; the potential ascent of the water at bgfc is
n4 a - n4 b + m4 b
4
na
V;
therefore,

From these [relationships] it is understood that the velocity of the

first drop bursting forth into the narrower pipe corresponds to the
4
height n 4 v, but that this velocity decreases very quickly, so that, after
m
 FLUIDS MOVING WITHIN THE WALLS OF VESSELS 127

a very small portion of fluid has flowed through, it can then be con-
sidered that V = __!!__b v and after all the fluid has flowed through,
a-
it assumes the former velocity. For example, let the diameter of the
wide pipe be ten times the other, and the first drop will flow from the
wider pipe into the narrower at a velocity due to the height 10,000 v;
but if one assumes that -h of the fluid has already flowed through, one
will find that the height which conforms to the velocity of the fluid
moving in the narrower pipe is approximately equal to .ll v.
If one seeks the time in which the transflux of the fluid Of occurs,
one finds that this is equal to

2(n 4a - n4 b + m4b) 312 - 2m 6 av'a

3mm(n 4 - m4 )v'av
Therefore, all the fluid flows through in the time

2n6 av'a - 2m 6 av'a 2(n 4

+ mmnn + m4 )a
3mm(n 4 - m4 )v'av 3mm(nn + mm)v'v '

where by ;v the time is defined in which the fluid, moving freely in

the wider pipe, traverses the distance a. But, as I said, these things
will behave this way if there are no impediments to the motion, and
at the same time the velocities in the full extent of the composite con-
duit are assumed reciprocally proportional to the areas. Meanwhile
I have already showed elsewhere that the water near the boundary
MN cannot satisfy this law. And so, since the situation is such, the
more nearly the actual motion will agree with theory, the longer the
portion bm will be and the fewer obstacles will be present.
§3. Thus, if now the conduit is placed not horizontally but obliquely
to the horizon, it is evident that all things occur similarly except that
the potential ascent of the water in every position is to be equated to the
initial potential ascent augmented by the actual descent, that is, by the
vertical descent of the center uf gravity. And so if the water begins
to move on its own without any impulse, the actual descent will simply
be equal to the potential ascent.
Therefore the water keeps on progressing continuously as long as
the center of gravity is located at a lower point than it was at the
beginning of motion. But when the pipe has been so formed and
curved and filled with such a quantity of fluid that the center of
gravity can assume its previous height again, then the fluid will
develop a retrograde motion and will oscillate without end. We will
 HYDRODYNAMICS, CHAPTER VI

soon discuss that motion comprising the principal portion of this

section. Meanwhile it may be observed that it can occur that all
water flows on its own accord from a lower place through a higher
without previous suction if only all things occur in the required
manner.

CONCERNING THE OSCILLATIONS OF FLUIDS IN

CURVED TUBES

CASE II

§4. In the Commentaries ef the Imperial Academy ef Science ef St. Peters-

burg, Book II, my Father presented certain theorems which manifest
the significant use which the theory oflive forces possesses in mechani-
cal matters. Indeed that which was given third is as follows:
Let there be a cylindrical pipe ABCH (Fig. 33) open at either end, bent into
H

FIGURE 33

two legs BA and CH [attached] to the horizontal portion BC; let the sine ef the
angle ABC be p, and the sine ef the angle HCB be q, with the total sine,
certainly, being 1; further, let that pipe be filled with water right up to the
horizontal MN, and let the length ef the portion ef the pipe MBCN filled with
water be called L. All the oscillations ef the fluid agitated within this tube,
greater as well as lesser, will be tautochronous with and ef the same duration as
the very small oscillations ef some simple pendulum the length ef which is ____!:__ ·
p+q
The following is a corollary to this theorem, by the same author.
if ABC and HCB are right angles, which is the only case solved by Newton,
the length ef the simple pendulum which is isochronous to the oscillating water
will be -~ L,just as Newton found.
§5. These are the things which have been communicated to the
public up to this time on the oscillations of fluids, and certainly first
 FLUIDS MOVING WITHIN THE WALLS OF VESSELS I 29

by Newton, in order to show the nature of waves, and by my Father,

in order to show the fruitfulness of the principle of live forces. But
since it is our intention to give a more complete theory concerning the
motions of water, it will be to the point to follow that type of argument
to its full extent. Therefore, let me inquire by which ways the un-
equal oscillations of a fluid may become isochronous, and by which
they do not. Then for the former I will give the length of the simple
tautochronous pendulum, and for the other I will indicate the time of
duration; moreover, I will consider pipes that are bent in any way
whatever and unequally large.

LEMMA

§6. Let cAd (Fig. 34) be a leather bag or a conduit of any given
shape whatever, full of water, terminating at either end in two
B

FIGURE 34

cylindrical conduits ac and fd inclined to the horizon in any way

whatever and of any area whatever, one of which let me assume
full of water up to a, the other up tof; let it be necessary to determine
the height of the center of gravity of all the water from the given
height of the center of gravity of the water contained in the bag
cAd, with as many of the remaining things known in advance as is
sufficient.
130 HYDRODYNAMICS, CHAPTER VI

SOLUTION. Let the center of gravity of the water contained in the

vessel cAd be at C; and let it be understood that the vertical AB is
drawn through that point C, and then let the horizontals am, cg, fn,
and dh be drawn, together with the verticals cb and de. Let it be
established that ac = a, fd = o:, be = b, ed = (3, the area of the pipe
ac = g, and the area of the pipefd = y; further, let the aqueous mass
or the volume of the conduit cAd be M, the line Ag = j, Ah = cp, and
AC = m. Let the lines mg and ng be divided in two at the points D
and E, and thus the centers of gravity of the water contained in the
cylindrical pipes will be at the heights of the points D and E.
After these things have been set forth, it occurs that AD = f = ½b;
AE = cp + ½/3; the mass of water at ac is ga, and atfd it is yo:. There-
fore, if the sought center of gravity for all the water acAdjis understood
to be located at the height F, AF will be obtained, as is well known in
mechanics, by multiplying the mass of water in ac by DA, the mass of
water [in] fd by EA, and the mass of water in cAd by CA, and by
dividing the sum of these products by the sum of their masses. From
this it is found that
AF= ga(f + ½b) + yo:(cp + ½/3) + Mm .
ga +yo:+ M

PROBLEM

§7. To determine everywhere the velocities of the oscillating water,

after it has been established that the oscillations do not extend
beyond the limits of the cylindrical pipes.
SOLUTION. Let the water beginning the oscillation be in the position
acAdj, and after a while it will have arrived at the position ocAdp;
and with the designations made in the preceding paragraph retained,
let ao be set equal to x; thenjp will beg\ whence (if indeed the center
y
of gravity of all the water is understood to have descended from
F to 0), by virtue of the preceding paragraph,

g(a - x)(f + ½b - bx) + y(o: + gx)(1> + ½/3 + (3gx) + Mm

AO = 2a y 20:y .
ga +yo:+ M
Hence the descent of the center of gravity or the actual descent is
deduced
(b - (3 + J - cp) gx- ( -bg + -bgg) xx
FO = 2a 20:y .
ga +yo:+ M
 FLUIDS MOVING WITHIN THE WALLS OF VESSELS 131

Now let the velocity of the water in the pipe ac (namely, when the
surface is at 0 ) be such that it corresponds to the height v, and then
the potential ascent of the water in the other pipe will be gg v; similarly,
yy
the potential ascent of the water cAd will be proportional to the height
v, and therefore we will set it equal to Nv (where N depends upon the
shape of the bag cAd and can be determined through §2, Chapter III).
But now if, after the potential ascents everywhere have been multiplied
by their proper masses, the products are divided by the sum of the
masses, the potential ascent of all the water ocAdp will be obtained as
3

( ga - gx + a.gg
y
+ g x + MN) v
yy
ga +ya+ M ·
And because this potential ascent is equal to the actual descent FO
found a little earlier,

(b - f3 +J -
cp)gx - (bg + _bgg)xx
2a 2ay
v=--g-a___g_x_+_a._g_g_+_g~3 x_+_M_N__ , Q.E.I.
y yy

§8. COROLLARY 1. Because the line mn = mg - nh + gh = b -

f3+ J - m, we will set mn = c, and at the same time we will multiply
the denominator and the numerator by 2yya.a.. Thus indeed we will
obtain
2gyyaa.cx - (gyya.b + ggyaf3 )xx
V=-----~---~~--~~~----c--=-•
3
2gyyaaa. - 2gyyaa.x + qggyaa.a. + 2g aa.x + 2yyaa.MN
§9. COROLLARY 2. If v = o, it is evident that the value x then
denotes the total displacement of the surface of the fluid in the pipe
ac, which is thus found equal to
2yaa.c
ya.b + gaf3'
but in the other pipe it becomes
2gaa.c
ya.b + gaf3·

Therefore, the water in the narrower pipe can be elevated to any

height whatever if only the ratio of the areas g and y is taken large
enough.
 HYDRODYNAMICS, CHAPTER VI

§10. COROLLARY 3. That portion of the vessel cAd which we assume

is never reached by either of the surfaces contributes nothing to either
increasing or diminishing those paths of the fluid; nevertheless, as
is shown below, it can serve for accelerating and retarding the
oscillations.
§11. COROLLARY 4. Let each pipe be assumed of a common size;
there will be, namely, for g = y,
2gaacx - (gab + gaf3 )xx
V = ---''-------'-"------''---'---'---,~. •
2gaaa + 2gaaa + 2aaMN
In this case the maximum velocity of either surface occurs when
they are located at the middle of the total displacement, but it occurs
differently when the pipes are of unequal area.
It is to be noted also that the retardations and accelerations are
similar to each other at similar distances of the surfaces from the
points at the middles of the paths, that is, from the points of maximum
velocities.

THEOREM

§12. When the areas of the cylindrical pipes are equal in the pre-
viously mentioned manner, greater as well as lesser oscillations will
be Isochronous to each other if only the surfaces never descend below
the orifices of these pipes.
PROOF. It is known from mechanics that if an oscillating mobile
[object] has passed through a distance x, and if at individual locations
it has the element of time dt = m dx , with m and n understood
Vnx - xx
to be constant quantities, this [object] makes its respective oscilla-
tions, whether greater or lesser, in the same time.
But because in our case
2gaacx - (gab + gaf3)xx
V = -----------,
2gaaa + 2gaaa + 2aaMN

and because the velocity itself is equal to vv, there will be

2gaaa + 2gaaa + 2aaMN/J 2aacx
dt = dX· - ---------- ----- - XX,
J gab+ ga(3 gab+ ga(3

where in a like manner all letters have constant values except x,

which denotes the distance traveled through; it is evident that these
oscillations of the fluid will also be isochronous. Q.E.D.
 FLUIDS MOVING WITHIN THE WALLS OF VESSELS 133

PROBLEM

§13. To find the length of the simple pendulum which is tauto-

chronous with the previously mentioned oscillations of a fluid.
SOLUTION. It is shown in mechanics that, with dt = V m dx ,
nx - xx
the length of a simple tautochronous pendulum is ½mm. Therefore,
in our case which is under discussion the length of the pendulum
. b gaaa + gaaa + aaMN Q EI
soug h t w1 11 e b f3 . . . .
ga + ga
§14. COROLLARY I. If the conduit cAd is assumed to be of the same
area as the attached pipes, and if its length is called l, the mass of the
water contained in it, which we have called M, will be gl; and the
potential ascent of the water contained therein, which we have set
equal to Nv, will be v, so that N = 1. Moreover, after those values
have been substituted for the letters M and N, the length of the
tautochronous pendulum becomes, for that particular case,
aaa + aaa + aal aa a+ a+ l
ab + af3 ab + af3
(a + a + l) = b f3
- + -a
a
But since a + a + l is the length of the entire system filled with
water, and~ signifies the ratio of the sine of the angle bac to the total
a
sine, and, equally,~ denotes the ratio of the sine of the angle efd to the
a
total sine, we see that our solution does not differ from that which my
Father gave for that case, and which I recounted above in §4.
§15. COROLLARY 2. If the conduit cAd is assumed to be of infinite
area everywhere, MN= o (through §2, Chapter III ) and the length
of the tautochronous pendulum will be : + ;, certainly just as if the
-a + -a
entire intermediate conduit cAd were absent and the cylindrical pipes
were connected immediately to each other.
Nevertheless, something special is to be considered here, which I
will show below.

ScHOLIUM

§16. This theorem includes all cases which cause tautochronous

oscillations where the pipes ac and pd are straight; but when these
 1 34 HYDRODYNAMICS, CHAPTER VI

pipes in which the surfaces of the fluid are traveling are curved, other
. cases of tautochronism are given in addition which would be easy to
solve if we should wish to delay here for a longer time. Finally,
when these pipes are of unequal area, the times corresponding to
oscillations of different magnitudes become unequal also, and how
such a time must be defined is apparent to everyone from §8, where we
gave the velocity of the fluid at any arbitrary point.
But this concerns finite oscillations. If now we consider that the
oscillations are very small, we will see that they all become mutually
tautochronous for the same quantity of fluid and the same conduit
being retained, whatever might be the shape of the conduit and the
areas. Let me show this in the following paragraph.

THEOREM

§17. Very small displacements of a fluid oscillating in any conduit

whatever, although they are unequal to one another, are all Iso-
chronous.
PROOF. When the oscillations are very small, those small portions
of the conduit in which the surfaces of the fluid are agitated can be
taken as cylinders; therefore, with all the designations kept the same,
the value which we assigned to the letter v in §8 will remain, and
from the same reasoning it follows that the letters a, b, a, f3, and x can
M
be neglected as being of infinitely small value with respect to - ,
g
so that in the present case it must be considered that
2gyaacx - (gyab + ggab )xx
V = ----~~~=-c-~~~.
2yaaMN
Therefore, by virtue of§ r 2, all oscillations, as far as they are very
small, are Isochronous to one another. Q.E.D.

PROBLEM

§18. To determine the length of a simple pendulum tautochronous

with the very small oscillations of a fluid agitated in any conduit
whatever.
SOLUTION. Because in the entire motion the element of time is

dt = :v' now there will be

d d
t = x·
J 2yaaMN
gyab + ggab
/J 2yaacx
yab + gaf3 - xx.
 FLUIDS MOVING WITHIN THE WALLS OF VESSELS 135

Therefore, by virtue of§r3, the desired length of the pendulum tau-

tochronous with the previously mentioned oscillations will be
yaaMN
gyab + gga{3° Q.E.I.

ScHOLIUM

§19. Although I may have already advised in passing what is to be

understood by the quantities Mand N, nevertheless let me set forth
the entire construction here, so that the nature of the matter is all the
more evident to everyone.
Let there be a conduit ABCDE of any kind whatever (Fig. 35a

FIGURE 35a

and b), filled with water right up to B and D; let the total sine be
assumed as r, the sine of the angle DEC as ~ = m, the sine of the angle
a
BDC as @=
(X
n; the length of the tautochronous pendulum will be

yMN , where g denotes the area of the conduit at B, and y the

mgy + ngg
area ofit at D.
Now let the length BCD of the conduit filled with fluid be considered
as extended in the straight line bed, above which, as ifit were an axis,
is formed the curve FGH, which let be the scale of areas in homo-
logous places, so that, after be has been set equal to BC, eG is to bF as
the area at C is to the area at B. Therefore, if bF represents the area
at B, then the area bdHF will represent the magnitude M. Then,
from the same axis bd let another curve LMN be constructed, the
 HYDRODYNAMICS, CHAPTER VI

ordinate cM of which is everywhere (b-:;r, and N(from §2, Chapter

III) will be the area bdNL divided by the area bdHF, so that M-N is
the area bdNL, which, multiplied by will give the length
mgy y+ ngg,
of the tautochronous pendulum.
H

b1--------..;,+--------l,b-l

.N

FIGURE 35b

§20. COROLLARY r. If the pipe BCD is of the same area everywhere,

and its length is called l, FH will be a straight line parallel to bd itself
and LN equally; hence the distance bdNL = gl and the length of the
tautochronous pendulum is _l_.
m+n
§21. COROLLARY 2. Let BCD be a conical conduit of length l;
cG (after be has been set equal to x) will be [1 (Vy - Vg) + Vgr;
gg / [7(Vy - Vg) + Vg]
2
from this cM = ; therefore, the area

bcML = v~g~ g - Vy g:l v g/ [7(Vy - vg) + vg], and thence

. ggl ggl ggl
the entire area bdNL = V- + V- . _ . Therefore,
gy - g gy - y V 1gy
 FLUIDS MOVING WITHIN THE WALLS OF VESSELS 137

the length of the pendulum tautochronous with the oscillating water

IS
tvgy
my+ ng
Hence it is understood, the remaining things being equal, that the
water is oscillated very slowly when the areas at Band Dare in recip-
rocal proportion to the sines of the corresponding angles DEC and
BDC; hence, the longer the portion full of water and the less the angles
just mentioned, the slower the oscillations become as well.
Further, after the cylindrical and conical pipes have been compared
to one another, and the angles BDC and DEC have been set equal, it is
clear that the water, the remaining things being equal, oscillates more
quickly in the conical than in the cylindrical [pipe], namely, because

lVgy is always less than ½l, whatever unequal ratio exists between
y+g
g and y. If, further, the previously mentioned angles are made
unequal, it can happen that the water oscillation is slower as much
as faster in one kind of pipe with respect to the other; so, in order that
I may show this by an example, let me assume DEC to be a right
angle, that is, m = 1, and the sine of the other angle BDC, or n is
equal to ¼, then the length of the pendulum for cylindrical pipes will
be il. But if under the same remaining circumstances one substi-
tutes for the cylindrical pipe a conical one which has an area at B four
times as large as the area at D, one will have, after y = }g has been
established, the length of the pendulum equal to l; therefore, the
remaining things being equal, the tautochronous pendulum for the
conical pipe is longer than for the cylindrical, and the oscillations in
the former occur more slowly than in the latter; but if now, once
again the remaining things constant, we assume the conical pipe
narrower at B than at D, the situation will be the opposite; for ex-
ample, let y = 4g, and the length of the pendulum will be -/'-7 l, and
accordingly less than if the pipe were cylindrical; and again it will be
less if one assumes the area at B altogether greater than it is at D;
thus, if y = -hg, the length of the pendulum will be -bl, as before.
It is notable that we saw also in the preceding example that, with the
area at B, the position of the conduit BCD and the length of the same
being maintained, two distinct areas can always be defined at D for
the same length of a tautochronous pendulum, except when the angles
DEC and BDC are equal. A particular example of this matter is
that, if either the area at Dis equal to the area at B, or it has to the
same a squared ratio of the sine of the angle BDC and the sine of the
angle DEC, the oscillations of the fluid are completed in either pipe
within the same time.
 HYDRODYNAMICS, CHAPTER VI

GENERAL ScHOLIUM

§22. I performed experiments on the oscillations of fluids in such a

way that by trial I often found the length of the simple Isochronous
pendulum, and I was able to observe in different cases that this
length was more or less such as the theory in this section indicates;
nevertheless, once I found that length to be a little greater than re-
quired; I saw with no difficulty that the reason for this situation is that
the frictions of the fluid not only diminish the paths, but also retard
them, and this because the pipes are customarily narrower in that
place where they are bent. Subsequently, if this is avoided with all
care, and if the very inflections are not made at an angle, but slowly,
and if finally the purest mercury is used as the oscillating liquid, I
have no doubt that the experiments will confirm the previously
mentioned theory to the letter, so that I would not think it is worth the
effort to inquire anxiously about them.
Nevertheless, I will add this as an explanation of the experiments
undertaken by me, that before the experiment I accurately deter-
mined the areas of the pipes in their different places with the help of
a column of mercury: while it slowly ran through the entire length of
the pipe, it disclosed the variations of the areas everywhere by its
different lengths, of which I continually took measurements. And
certainly these areas are thus to be determined in the pipe after it has
already been bent, for the areas are decreased somewhat by the bend-
ing. This was the reason that in the first experiment undertaken by
me in this regard success failed my expectation: Indeed, I bent a
glass tube of the sort that is customarily used for making barometers,
wide enough and almost perfectly cylindrical, more or less as Fig. 27
shows, and then, after a very great portion had been filled with
mercury, I saw that its oscillations occurred much more slowly than
I had expected, because I did not pay attention to the fact that the
tube was constricted significantly by the bending at D, especially
where the angles are formed. Therefore, in order to take this matter
into consideration, I made use from then on of gradually curved
tubes, such as Fig. 35a shows, and in these I carefully determined the
areas after the bending.
 SEVENTH CHAPTER
Concerning the Motion of Water through Submerged
Vessels, where it is Shown by Examples how
Significantly Useful is the Principle of the
Conservation of Live Forces, even in those
Cases in which Continually some Part of
Them is to be Considered Lost

FIRST PART: CONCERNING THE DESCENT OF

WATER

§1. Assume there is a cylinder full of water, the base of which is

perforated, submerged to a certain depth in standing water of in-
finite extent, and one easily realizes that the surface of the water con-
tained in the cylinder will descend, and indeed below the surface of
the exterior water, then it will ascend again, and so on. But these
oscillations differ completely from the oscillations considered in the
preceding chapter, in which certainly the motions are always recipro-
cal in inverse order to those motions which have preceded. But some-
one may presume here that the reflux of water, or the ascent, will be
the same as the descent was. If one would state such things, he
would certainly be in error, even if, for instance, the motion is dimin-
ished not at all by the adhesion of water to the sides of the vessel and
other hindrances of this sort; but the rules of motions for percussion
for elastic bodies are otherwise not very different from these which are
valid for pliable bodies, in whatever manner in either case bodies are
considered to be moved completely freely. I use the following similar
[procedure] which illustrates our argument splendidly: For just as the
rules of motions are determined correctly in pliable bodies, if after
collision that part of the live force is considered lost which was ex-
pended in the compression of the bodies (for this is not restored to the
 HYDRODYNAMICS, CHAPTER VII

progressive motion as in elastic bodies), so the ascent of the fluid will

be defined no less correctly if one examines accurately how much of
the live force in individual instants is communicated to the internal
motion of the aqueous particles, never to return to the progressive
motion, which is the subject of this discussion.
§2. Since, therefore, the matter is reduced to the fact that it should
be investigated how much of the live force is lost continuously in those
reciprocal motions, we will begin the investigation from this point.
But first it is evident that all the live force which the particles flowing
out possess is transferred to the external water, and in no way does it
promote the subsequent ascent or influx of external water into the
pipe. This hypothesis is all too clear to warrant a greater explana-
tion; but it pertains to the efflux of water and in this case it is the only
one to be considered. Then we come to the other, which pertains to
the influx of water.
Secondly, therefore, it is no less clear to me that with the water
rushing in through the orifice at a greater velodty than that which is
present in the internal rising water, that excess again produces a
certain internal motion in the same internal water, adding little or
nothing to the ascent. If this is so, and if the area of the orifice is set
equal to r and the area of the cylinder equal to n, the potential ascent of
a volume element flowing is equal to nnv, and its velocity is nvv, this
particle will in its own motion, which it has in common with the
remaining internal water, retain the velocity vv, and accordingly it
will conserve the potential ascent v; but it must be considered that the
remainder of the potential ascent, namely, nnv - v, was transferred to
the internal motion of the particles. This hypothesis, although it is
physical and only approximately true, nevertheless has great useful-
ness for determining the motions of fluids without noticeable error
whenever in a vessel uniform continuity, which was assumed so far, is
broken off, as when the water is forced to go through many orifices.
Finally, I should believe that this [hypothesis] is unique, with the help
of which amazing phenomena of this kind of motion can be explained
correctly. On this account let me implore that it be pondered
properly before the reader is diverted to other things.
§3. Therefore, we will now examine this very question by beginning
with the descent of the water. Let the cylinder AIME (Fig. 36) be
considered full of water up to XY and submerged in the infinite
water RTVS, so that it is in a vertical position and its base has the
opening PL through which water from the vessel can flow into the
water surrounding. The velocity of the internal water is sought
after its surface descends through the given space XC or YD, after one
 MOTION OF WATER THROUGH SUBMERGED VESSELS 141

has set MY or IX = a, MV = b, MD = x, the area of the orifice

equal to 1, and finally the area of the cylinder equal to n .

X ...... y

FIGURE 36

The solution will be the same as that which we gave for a similar
but very general question in Chapter III; only let it be noted that,
with the infinitely small particle of water CDFE assumed equal to the
volume element PLON ejected in that same time, the actual descent
must now be estimated from the height DVor CT, while in the other
case it was to be defined from the total height DM.
Indeed, let the velocity of the aqueous surface CD be that which is
due to the height v, and in the infinitely close position EF the same
velocity will correspond to the height v - dv. And since the potential
ascent of the water CDMLPIC is v, the potential ascent of the same water
in the next position EFMLONPIE will be obtained if the mass
EFMLPIE (nx - n dx ) is multiplied by its proper potential ascent
(v - dv) , as also the volume element LONP (n dx) by its proper
potential ascent nnv in the same way, and the aggregate of the products
is divided by the sum of the masses, (nx ) ; and so that potential ascent is
(nx - n dx )(v - dv ) + n dx-nnv
nx
or
xv - v dx - x dv + nnv dx
X
 HYDRODYNAMICS, CHAPTER VII

. h . f .l . - vdx - x dv + nnv dx
A ccord mg1y t e mcrement o potentia ascent 1s - - - - - - - - -
x
(see §6, Chapter III). But that increment must be considered equal
to an infinitely small actual descent, which (accordingly to §7, Chapter
· JUSt
III an d t h e notat10n · ·
given ) 1s (x - b) dx
· ~--'--. A n d so t h e c: 11
10 ·
owmg
X
equation is obtained
- v dx - x dv + nnv dx = (x - b) dx

which, integrated in the proper way, is changed into this:

v = nn -
I (x -
2
xnn-1)
ann - 2 -
b
nn - I
(
I -
xnn-1)
ann -l .

But from that equation certain corollaries follow.

§4. Let the area of the cylinder be as if infinite in proportion to the
orifice, and it must be considered that v = x - b; and the very height
nn
corresponding to the velocity of the water, while it flows out, is x - b.
Hence it is a consequence that the water flows out at the velocity
which a heavy body acquires by falling from the height of the internal
surface above the external, and it will flow out until both surfaces are
placed at a level, and then all motion will cease; and therefore the
water flows out by the same law as if the base changed the position
IMwith TV.
But when the orifice cannot be considered as infinitely small, the
surface of the internal water descends below the external; and in
order that it may become known to what depth xy the surface CD is to
descend, v must become null, or

but the internal surface will never descend so far below the external
surface as it had been elevated above the same; that defect arises from
the potential ascent of the water ejected during descent, to which it
must be proportional.
§5. It is noticeable that although the water descends more deeply
in the cylinder the greater it has been elevated at the beginning of the
descent and the greater the opening is by which the base is perforated,
nevertheless all the water can never flow out of the cylinder, no
matter how much it has been elevated before the descent and however
small is the submerged portion of the cylinder, and at the same time
the orifice itself or the entire base is assumed to be discharging.
 MOTION OF WATER THROUGH SUBMERGED VESSELS 143

§6. The velocity of the surface of the internal water is a maximum

when it is assumed that

X _ ( ann-1 )l /(nn-2)
0

- nna - nnb - a+ 2b

Accordingly, if n = 1, the orifice of the cylinder of course being

fully opened, it occurs that x = b, and the velocity is a maximum
when both surfaces are positioned at the same height.
But because there are many things which cannot be learned from
these equations in the two cases, namely, when nn = I and nn = 2,
and these have many particulars, let me now attack them separately.
§7. First, let nn = r, whereupon -x dv = (x - b) dx (by §3), or
- dv = dx - b dx, which, integrated so that simultaneously one has
X

v = o and x = a, gives -v = x - a + b ln '.:,

X
or v = a - x - b ln '.:.
X
From these the following can be deduced.
I. In order that the maximum descent be obtained, it must be
established that a - x - b ln '.: = o; moreover, it is evident from that
X

equation that the letter x never obtains a negative value; on the con-
trary, it certainly does not vanish completely without [causing] a con-
tradiction, unless Eis set equal to oo, which indicates that it cannot
occur that all the water flows out during descent in that case, and
much less in the others, which §5 confirms.
II. The maximum velocity is that which is due to the height
a - b - b ln E' and if the difference between a and b, which I set
equal to c, is very small, the expansions of the fluid being indeed
insignificant in proportion to the depth to which the cylinder is sub-
merged, ln i - 7b' and therefore the height
Ecould be considered as
2

itself due to the maximum velocity as b - b ln i = ~~' which

a -
indicates that the motion will be very slow.
Moreover, I shall demonstrate in what follows that the entire
motion remains the same, the rest of the things being equal, when the
cylinders are considered infinitely submerged, by whatever orifice the
base may be perforated, so that the motion of the internal water is not
retarded by the diminished orifice; although this may seem at first
1 44 HYDRODYNAMICS, CHAPTER VII

glance to be altogether a paradox, nevertheless, the true physical

reasoning ofit cannot escape the mind that is attentive to these things.
It is certainly concerned with the fact that the liueforce which is gener-
ated in the pipe is as if infinite with regard to the live force of the water
flowing through the orifice, and therefore the consideration of this
orifice does not make the computation different.
We will also show that the reciprocal motions are similar, and that
both the greater and the lesser oscillations are Isochronous with each
other, and for these we will determine the length of a simple tauto-
chronous pendulum.
§8. Now let nn = 2 be assumed; but thus it occurs, by virtue of§3,
x du - u dx (b - x) dx .
that u dx - x du = (x - b) dx, or - - - - = - - - - , which,
xx xx
correctly integrated, changes into this:
bx a
u = - - b + x In -·
a X

If it occurs that bx - b + x In~ = o, x will give the position of

a X
maximum descent; but the position of maximum velocity will be
obtained by setting x - ae<b-a)Ja, where by e is understood the
number of which the logarithm is unity.
After we have so glanced over various cases for different magnitudes
of orifices, it remains that we also consider what can happen in
different cases of the heights a and b.
§9. And first, certainly, if b is considered null with respect to a,
which occurs when the base of the cylinder only touches the surface of
the exterior water, then there results

U = nn -
I ( xnn-1)
2 X - ann - 2

which equation, certainly, differs only in form from that which was
given in §14, Chapter III, for that case in which the water is con-
sidered to be ejected from the cylinder into the air. And I also often
found that the cylinder is evacuated in the same time whether the
water is ejected into the air or the base is submerged a very little bit in
standing water. This experience shows that the external air offers
little or no hindrance to the effiux, since a resistance more than eight
hundred times greater does not have a more noticeable effect.
Therefore, because that case contains nothing in particular which has
not been mentioned in the place cited, we will not dwell on it any
further. Rather, we will inquire into what must occur when the
 MOTION OF WATER THROUGH SUBMERGED VESSELS 145

elevation of the internal water above the external, such as it is at the

beginning of the descent, is assumed very small and should be neg-
lected with respect to the immersion of the cylinder; this hypothesis
is satisfied when the excess of the height a over the height b (which
excess again let us call c, as in §7) is very small.
§10. Therefore, if one sets a - b = c, one must also set a - x = z,
and then both quantities, namely, c and z, are to be neglected with
respect to the quantities a and b; but if a - x = z, then x = a - z and
xnn-1 = (a _ z)nn-1 = ann-1 _ (nn _ 1) ann-2z

(nn - 1)(nn -
+ [- ------ 2)] ann-s zz
2

[
(nn - i)(nn - 2)(nn - 3)] ann- z
4 3
+ etc.
2·3
This series is to be continued as much as it suffices for our purpose;
however, up to three terms will be sufficient. Therefore, in the
integrated equation which we gave in §3, we will assume x = a - z,
and

xnn-1 = ann-1 - (nn - I)ann-2z + [(nn - I)2(nn - 2)]ann-3zz,

and thus there will be

v=
nn-2
I {a - z - a+ (nn - i)z - [(nn - i)(nn -
2
2)] zz}
a
_ _ b_ { 1 _ 1 + (nn _ I)~_ [(nn - i)(nn - 2)]zz}·
nn-1 a 2 aa
If in this equation the terms cancelling themselves are deleted, and
a - c is put in place of b, and the term which is affected by the
. -czz 1s. reJecte
quantity . d , t h ere resu1ts s1mp
. 1y
aa
2CZ - ZZ
V =- ---
2a
from which formula, since the letter n has vanished, we have evidence
that the size of the orifice pertains not at all to the motion of the
internal water, the origin of which matter I already indicated above
(§7).
In the following [paragraphs] we will show further that this
motion does not differ from the subsequent reflux motion, and hence
that the oscillations become tautochronous. But before I proceed to
 HYDRODYNAMICS, CHAPTER VII

other things, I consider that one must be advised that in this calcu-
lation the quantities : and i were taken as infinitely small, not only
a a
with respect to unity, but also with respect to_:_, of which one has to
nn
take proper notice in undertaking experiments; certainly it is allow-
able to apply the theory of the infinitely small to experiments without
noticeable error by greatly diminishing the quantities which have
been considered as infinitely small in the theory, but it must be done
so that all things in the experiment are subject to this law. Thus, for
example, if in the cylinder the entire base is absent, n = I having
been established, and it is considered submerged to a height of 35
inches, the experiment may be regarded as sufficiently accurate if the
water has been elevated before the oscillations only to a height of one
inch above the surface of the water surrounding; the error will not yet
be noticeable even if the orifice below is half obstructed, with :
a
then being to _:_ as I is to 9, which ratio so far can be safely neglected
nn
in our experiment; but if one now assumes the diameter of the pipe as
two times the diameter of the orifice, with three-quarters of the entire
aperture closed, it occurs that n = 4, and: will be to _:_ as 4 is to 9,
a nn
which ratio will now be small enough that the experiment can
be affirmed to satisfy the conditions of the theory with sufficient
precision.
Therefore, it will now be appropriate to inquire further what
should be stated here concerning these cases in which: and_:_ indeed
a nn
have an appreciable ratio to each other, but each quantity is very
small, which certainly occurs when the cylinder is submerged very
deeply and at the same time the base is perforated by a very small
orifice.
§n. But that case which we just treated is better deduced from the
differential equation of§3, rather than the integral [equation], as was
done previously; however, under these circumstances the term
-v dx can be rejected with respect to nnv dx, and thus it can be
assumed that -x dv + nnv dx = (x - b) dx, in which, if it is again
established that a - b = c and a - x = z, there results adv +
z dv + nnv dz= (c - z) dz, the second term z dv of which can again
be neglected with respect to the first one, and thus one obtains
adv + nnv dz = (c - z) dz,
 MOTION OF WATER THROUGH SUBMERGED VESSELS 147

Let it be assumed here (after e has been taken as the number of

which the hyperbolic logarithm is unity) that v = __:_ e-nnzlaq; in this
nn
way the last equation will be changed into this:
e-nnz/a a dq = nn(c - z) dz,
or
a dq = nnennz/a (c - z) dz.
But this must be integrated so that z and v, or as well z and q,
vanish at the same time; thus there will result

q = (c + !:.nn _ z) ennz/a _ c_ !:.,

nn
or, finally,

v = __:_
nn
(c + !:.nn - z) ___:nn_ (c + ..::.)
nn
e-nnzla.

But from that equation it is deduced:

I. That it again develops, as was found by another method in §10,
2cz- -
t h at v = - zz, 1'f.1n d eed t h e numb er -nnz 1s
- - . establ'1sh e d as
. agam
2a a
very small. But in order that this be evident, the exponential
quantity e-nnz/a must be resolved into the series which is equivalent
4
. nnz n6 z 3
to 1t, 1 - - + n-- zz
- - -3 + etc., from which the first three
a 2aa 2-3a
terms are sufficient for our purpose; therefore, with this value sub-
stituted and the term to be rejected having been rejected, there is
obtained, as I said,
2CZ - ZZ
V = ----
2a

II. But if, on the other hand, nn is assumed infinitely greater than
I

a or ~ because then e-nnz/a - o and also !:. = o, it occurs that

z c' - nn '
v = c - z, or v = x - b, as in §4.
III. But it is evident that neither of the aforementioned formulas
stands without noticeable error when the number nnc is moderate,
a
that is, neither infinite nor infinitely small, and nevertheless both
.. -nn an da- are m
quantities ' f imte.
'
I C
 HYDRODYNAMICS, CHAPTER VII

For example, let an elevation of one inch be indicated by c, let the

immersion of the cylinder b be 80 inches, and let a itself be 81 inches;
then let the diameter of the pipe be established as triple the diameter
of the orifice, that is, nn = 81, and there will result v
z -
nn
= 2
-
2
e-\
and if further it is established that z = c = 1, in order that the
height for the velocity be obtained when both surfaces are positioned
2
at [the same] level, one will have v e - , that is, v = ah- inch,
=
nne
approximately, although according to §w it should have developed
that V = f i inch, and according to §4, V = O. Jn the same example
it happens that the entire space through which the surface passes is
not fully eight-fifths of one inch, and the point of maximum velocity is
more or less sixty-nine hundredths of the same measure below the
initial height.
§12. It should not be any more difficult to extend those things which
were said so far to all shapes of vessels, finally even to finite spaces by
which the external water may be terminated; however, the formulas
become mostly so obliging that I would consider it rather well
advised to pass in silence over them and to show only by an example
the particular manner in which the theory should be applied for
eliciting any number of other cases.
Deserving more particular attention are those things which I indi-
cated concerning the motion of water in pipes opened considerably at
the bottom and submerged very deeply, because in these the oscilla-
tory motion, as in pendulums, is of constant period, and the flow of
waves in the sea is illustrated by them. But I thought that the back-
flow of water in submerged cylinders in general is to be treated first,
and it is to be shown that according to this hypothesis the backflow
does not differ from the preceding flow, before the entire oscillatory
motion is examined. Therefore, we will now comment on the back-
flow, and then we will combine both motions in different cases, lest
anything could be wanting in the proof.

SECOND PART: CONCERNING THE ASCENT OF

WATER

§13. After the water has descended in the submerged vessel as

much as the nature of the situation permits it, two things especially
offer themselves for consideration: first, the excess of the height of the
 MOTION OF WATER THROUGH SUBMERGED VESSELS 149

external surface above the internal, and second, the live force, or the
product of the potential ascent and the mass of that water which was
ejected from the cylinder into the surrounding water during the
descent; indeed, this live force, which cannot return to the water in the
cylinder, principally causes the water to fail considerably in approach-
ing the original height from which it had fallen; nevertheless, this is
not the only reason, even if the hindrances of tenacity, adhesion, and
others of this sort interfere not at all: the other reason was indicated
in §2. Indeed, the extent of that reason is to be deduced from the
ascent itself, since the former pertains to the descent, and this is the
only reason, disregarding the external hindrances, that the water is
not elevated in ascent above the external surface as much as it had
been depressed below the same. For it must be noted that, even with
the water flowing in through a very small orifice, it would ascend at
the same velocity as if the entire base were missing, and it would rush
in through the entire orifice if only after the inflow it would exert
the entire impetus which occurs in the internal water for promoting
its ascent. Truly anyone who considers this matter properly sees
easily that the major part of that entire impetus is expended wholly
in some internal motion which provides nothing to the ascent; but
I say clearly the major part (I wish it to be noted well), because, when
the orifice is very large, it is not difficult to see in advance that the
impetus of the entering water is produced so suitably that the internal
motion is thence increased by no means slightly; but when the orifice
is smaller, it is clear that the situation is otherwise. Therefore, our
hypothesis is applied correctly when either the whole base is absent
or is almost completely perforated (for thus the excess of the velocity
of the water flowing in over the velocity of the internal water is nil or
very small, and that does nothing to this impetus) or even when the
orifice is very small, because thus all the impetus is overcome. But if
the orifice should have a ratio to the area of the pipe such as V2 to 1,
or as 2 to 1, or thereabout, the motion will be a little greater than that
which follows from that hypothesis, for then the water rushing in
produces a noticeable impetus, and not all of it is lost because of the
nature of the matter.
Therefore it is easy to see in advance, without performing any
calculation, the following relationships for backflow in water after it
has fallen from a certain height.
I. Certainly no noticeable backflow will occur if the orifice is very
small.
II. When the submerged portion of the cylinder remains un-
changed, the water in backflow will never pass a certain limit, even
 HYDRODYNAMICS, CHAPTER VII

if it had been elevated to infinity in the previous descent; indeed,

from whatever height the descent begins, not all the water ever flows
out of the cylinder, as we saw in §5 and §7.
III. If the descent is understood to begin at the height XY [Fig. 36]
and the subsequent ascent to be produced right up to CD, the product
of the actual descent of the mass of water X YDC right up to TV by the
mass will be a measure of the combined effect of both of those causes
which, as was said in §2, make the ascent differ from the preceding
descent, and when the cause reviewed in the second place vanishes,
if the entire base IM is removed, that product will then be equal to
the live force of all the water ejected during the descent, so that,
without any other calculation except those things already considered
so far, the ascent of the water in the entire open cylinder can be
defined.
IV. The ascent will be equal to the descent if the cylinder is under-
stood to be infinitely submerged, the previously mentioned causes of
diminution then vanishing.
V. Therefore, these oscillations will be endless, because the last
oscillations are always just as if infinitely small in proportion to the
heights of submergence; however, the alien hindrances, of which we
have taken no account up to now, soon cause all the motion to cease
altogether.
§14. With these things having been generally shown in advance,
let us submit the problem to more accurate calculation; however, I
will give a double solution, one accommodated to the principles just
explained, the other different in kind to some degree.
Therefore, with the figure [36] and the notation of§3 retained, we
r
will consider that the water has descended from the height X all the
way to xy, and from this terminal point it begins its ascent; let My or
Ix be called a, and after it has already ascended to cd or ef, let Md = (
and dj = dg. After these things have been so prepared for the cal-
culation, and with the height due to the velocity of the water at cd
again designated by v, and the similar height in the adjacent position
ef by v + du, we will inquire in to the increment of potential ascent of the
water entering while the volume element LONP goes into the cylinder
and the surface ascends from cd to ef. Moreover, it is clear that when
the potential ascent of the internal water multiplied by its proper mass
is expressed by n(v (indeed, no attention is to be paid to the internal
motion), the increment of the same product will be n( du + nu dg.
But if in addition one were to consider the potential ascent nnv - v
(see §2) which the volume element n d( flowing in loses, and which
 MOTION OF WATER THROUGH SUBMERGED VESSELS 151

equally is due to the actual descent of the aqueous particle n d( through

the height b - x, it is evident that one must write
n(dv + nvd( + (nnv - v)nd( = (b - flnd(
or
g dv + nnv d( = (b - () d(.
But the same thing is found differently as follows.
Indeed, let it be considered that the volume element LONP has
almost no velocity before it begins to flow in, but that, once it begins
to flow in, the same acquires a potential ascent nnv, although soon after
its influx (according to the notation following §2) it is to be considered
as continuing its motion at the common velocity v;. From this fact,
the reasoning is this: before the influx of the volume element, the
potential ascent of the water cdMLPlc (the mass of which is n() is v, and
the potential ascent of the volume element LONP (the mass of which
is n d() is o; therefore, the potential ascent of all the water cdMLONPle is

n (v (v
n g + n d( = g + d(
But, indeed, after the volume element LONP has flowed in and
taken the position LonP, its potential ascent is nnv; moreover, the
potential ascent of the remaining water efMLonPle ( the mass of which
indeed again is nfl is v + dv; therefore, the potential ascent of all the
water considered here after the influx of the volume element is

n d( nnv + n((v + dv) (v + g dv + nnv d(

n( + n d( g + d(

while before the same influx it was g !v d(; consequently, it took on

. g dv + nnv d( . g dv + nnv d(
the mcrement g + d( , or, more simply, g . But
that increment is to be equated to the actual descent which the water
makes in changing position from cdMLONPlc to efMLONPle, which
descent is equal to the fourth proportional of the mass n( of the in-
ternal water, the volume element ndg, and the height VJ orb - (, so
that the previously mentioned descent 1s ( b - /) d(; from which
again the following equation results:
g dv + nnv d( = (b - fl df

- -----------~ -- - -
 HYDRODYNAMICS, CHAPTER VII

But the integral of this, after the addition of the required constant,
becomes the following:

nn + 1

which we will now consider under varied circumstances.

§15. And, indeed, when the area of the pipe is an infinite number
of times greater than the area of the orifice, it is evident that
v = b-
nn
t, and that accordingly the water rushes in at a velocity
which is due to the height of the external surface above the internal,
and then the ascent will not occur beyond the surface of the external
water.
But when the area of the orifice has a finite ratio to the area of the
pipe, the ascent occurs beyond the surface RS, for example, right up
to st; but Vt will always be less than ~' except when the entire base
is missing, for then Vt = ~- Just as we warned in §5 that in descent
the difference between VY and Vy is proportional to and has its origin
in the potential ascent of the water ejected during descent, so it can now
be observed in ascent that the difference between ~ and Vt has its
origin in the collision of the volume elements LonP with the mass of
water lying over it, which collision indeed does not promote the
ascent but is lost in useless internal motion, just as was indicated in
§2. Therefore, when the entire base IM is absent, the water flows into
the pipe at the same velocity at which the water having previously
entered the pipe moves, and no collision occurs, which is the reason
why in that case the water ascends as much above the surface RS as it
had been depressed below it, which the equation indicates, as we shall
soon see.
§16. The maximum ascent st will be found by making v = o.
Therefore, in order to define the entire motion correctly, the formulas
brought forth in §3 and §14 will have to be applied alternately, which
I will now show by the single example in which nn = r.
Accordingly,ifnn = r,itoccursthatv = b(r - ~) - ; (t - ex;);
and v = o will result when it is assumed that ( = 2b - x, that is,
when it is assumed that Vt = ~- Therefore, if, for example, the
pipe ABM/is full of water, destitute of any base and immersed up to
its midpoint in the exterior water, and if the entire length of the same
is called a, the water will be set in motion so that at first it descends
through a distance 0.297 a below TV, then it is elevated through a
similar distance above the same TV, and again it is depressed through
 MOTION OF WATER THROUGH SUBMERGED VESSELS 153

a distance 0.240a below it, and it transcends that line again in the
same way, and so on.
§17. It is evident as well that when a = o, with the pipe of course
empty of all water, generally there will result v = !!_ - __g_; and
nn nn+r
1
consequently the entire ascent will be nn +
nn
b, or the ascent above

the exterior surface of the water will be !!.....

nn
§18. I come now to infinitely submerged pipes, the descent within
which we have determined with the appropriate relationships of §10.
Moreover, let us use clearly the same method which we used there for
defining this case; therefore, for us the initial depression will be
Vy ( = b - a) = c, and the ascent thence produced will be yd(= g -
a) = z. Thus f = a + z, and b = a + c, where the quantities z
and c must be considered as infinitely small in proportion to the
quantity a. Hence there results (""t
cx)nn = (a +a Z )nn ( 1 + ;z) nn,
which, by applying a known series and taking the first three terms of
. nnz nn (nn + 1)zz
1t, one makes equal to 1 - -
a
+- -- - - . After these values
2aa
have been substituted for b, q, and (i) nn' the last equation of §14 is
changed into this:
v = a + c (nnz + nn(nn +
nn a 2aa
1)zz)
--- I
- ( a+z-a+nnz- M0n+
nn + 1 2a

= (a + c) [~ _ (nn ::)zz] _ ( z _ n::z)

cz _ zz _ (nn + 1)czz.
a 2a 2aa

But that last term can be neglected, and thus one has simply
2CZ - ZZ
V=
2a

in which equation n no longer appears. This does not differ from the
.
equat10n r d
1or .
escent given . §10, name ly, v = -
1n 2CZ - ZZ . . d d
- - - , since 1n ee
2a
the quantities a and a differ only by the very small quantity 2c.
For the rest, here as well are to be understood all those things which
1 54 HYDRODYNAMICS, CHAPTER VII

were said in the same §10 on the fact that a pipe must not be overly
obstructed.
§19. Consequently, the descent and ascent are equal to each other;
for from our equations it is evident that the liquid is balanced equally
beyond the surface of the external water. But then it follows es-
pecially from those formulas that even the unequal oscillations are
isochronous to each other, if only all of them can be considered in-
finitely small with respect to the submersion, and that moreover the
simple tautochronous pendulum is of the same length as the sub-
merged portion of the pipe.
That theorem differs from the one which was cited in §4, Chapter
VI, concerning the oscillations in the cylindrical pipe composed of
two vertical legs, as follows: there all oscillations are tautochronous,
the oscillations of finite magnitude not having been excluded, while
in the present case finite oscillations are of unequal duration; further,
there the length of the pendulum is equal to half the length of the
pipe, while here it is equal to the whole; however, if the matter is
pondered properly, this should be considered as consistent rather
than inconsistent, on account of the duplication of the pipe which
occurs in the former case.
§20. In either type of oscillations the nature of waves agitated by
the wind is illustrated; for they are not being moved otherwise than
that the water in them continually ascends and descends again.
Thus, what Newton says is evident, that the times of the oscillations
are in proportion to the square roots of the lengths of the waves, be-
cause he assumes that the form of the waves is constantly similar to
itself and accordingly that their length is proportional to the depth to
which the water is agitated. Moreover, it is probable that the depth
is that of a simple pendulum tautochronous with the waves, that is,
for example, 60¼ Paris feet if the ascent or descent of the waves occurs
every two seconds.
§21. Although I should not want, for the sake of avoiding the
abundance of calculation, to pursue this argument to its full extent,
and in view of these things I should only treat cylindrical vessels,
nevertheless, since in the case of infinite submersion the propositions
and theorems lose little of their elegance, let me extend the general
theorem for the case of oscillations of water in an arbitrarily shaped
pipe. However, the proof has been omitted, since it will be obvious
to everyone from things said elsewhere, but especially from those
which were presented in Chapter VI, §6, §7, etc., up to §20. How-
ever, that upper part of the vessel in which the oscillations occur must
have a cylindrical form.
 MOTION OF WATER THROUGH SUBMERGED VESSELS 155

§22. Therefore, let bd be the length of a submerged vessel (Fig.

35b). Let bF represent its area at the location of the surface, and let
the vessel be so shaped that the curve FGH is the scale of the areas.
Let the line be be drawn and the curve LMN be formed, the ordinate
cM of which is everywhere equal to (b::J2, and the length of the
pendulum isochronous with the oscillations of the aqueous surface
will be equal to the area bdNL divided by bL.
§23. COROLLARY. It follows from the preceding paragraph that if
the submerged pipe were conic, and if it had an area in the region of
the surface of the water which is to the submerged orifice as m is to n,
the length of the pendulum Isochronous with the vibrating water will
vn,
be to the length of the submerged pipe as -y1;;:; is to that is, as the
roots of the previously mentioned areas, but if the same pipe is sub-
merged not quite fully, once in the correct and then in the inverted
position, the lengths of the isochronous pendulums will be in recip-
rocal proportion to the submerged orifices.

GENERAL ScHOLIUM

§24. Attempting the things which are contained in this chapter by

experiments will be worth the effort all the more, since the majority
of them arise from new hypotheses. I indeed performed several
[experiments], but there was no time to execute some individual ones
which I had planned; those which I did I will recount below. Mean-
while, in order that judgment can be passed more safely on the
agreement of experiments with theory, first there is to be understood
for the circumstances of the matter, generally, whether and how much
the contraction of the stream flowing out (the nature of which I ex-
plained in Chapter IV) can disturb the calculation; this incon-
venience can be removed for the most part if the walls of the final
orifice form some small cylinder of barely half a line in height;
concerning this let the fourth experiment pertaining to Chapter IV
be brought to mind. Thence as well one has to pay attention to the
resistances arising from the adhesion of the water, which indeed retard
the motion little if one considers the times of the oscillations, but they
detract much from the displacements, especially if rather narrow
and rather long pipes are used. Therefore, more faith is to be put in
experiments which are performed on the times of oscillations because
these times are not altered much at all by the diminution of the dis-
placements. With respect to the first kind of experiments, in which
the displacements of the fluids in the pipes come from seeking and
 HYDRODYNAMICS, CHAPTER VII

observing the descent as much as the ascent, I used this foresight:

I wrapped a thread around the pipe at that place to which I expected
the water to ascend or descend, and thus I finally located this thread,
after frequent repetition of the experiment, so that the surface of the
oscillating fluid would run neither above it nor below it. And I also
marked the rest of the places which were to be observed in the pipe
equally by wrapping a thread around it. Then, what pertains to the
times of oscillations, because these decrease very quickly and become
imperceptible and clearly null, I was not able to investigate otherwise
than by examining the length of the simple isochronous pendulum,
after the experiment had been repeated very often; while this [pen-
dulum] was oscillating, I placed a finger over the orifice of the pipe
and removed it therefrom precisely at such an instant of time that both
the pendulum and the fluid would begin oscillation together.

EXPERIMENTS WHICH PERTAIN TO CHAPTER VII

EXPERIMENT I. I acquired a cylindrical glass tube of almost four

lines in diameter, entirely open below. I submerged it in water
standing in a very large clear vessel to a height of 44 lines, and I put
a finger over the opening above, so that in extracting a portion of the
tube the water would not descend in it; then I extracted the tube to
a height of 22 lines, so that the submerged portion of the tube as well
as the height of the internal water above the external was 22 lines,
and, as soon as the finger was removed, I observed the descent of the
surface in the tube below the surface of the standing water, and I saw
that it was 9½ lines.
But according to §§7 and I 7 it should have descended thirteen lines.
It seems that the defect of three and a half lines is to be attributed
almost solely to the adhesion of the water to the walls of the tube.
After the descent had been observed, I repeated the entire experi-
ment in order that I might discover the next ascent also. However,
it seemed to me to be 8 lines, which, according to§ I 6, with the previous
descent having been considered, should have been 9½ lines, that is,
just as much as was the preceding descent. But here the experiment
failed by only one and a half lines, while in the first part of the ex-
periment a defect of up to three lines and a half was present, because,
indeed, there the displacement was greater, and this at a greater
velocity, so that one finds altogether greater hindrances, which in-
crease together with the velocities.
EXPERIMENT 2. I used the same tube, but [the end of it was]
covered by a thin plate which was perforated by an orifice with an
 MOTION OF WATER THROUGH SUBMERGED VESSELS 157

area having a ratio of V½ to the area of the tube; when the surface of
the tube was elevated 18 lines above the standing water and the base
submerged the same number of lines, I saw that the surface of the
tube in descent fell almost five lines beneath the standing water.
However, §8 argues a descent of 7½ lines; the defect, which was more
than 2½ lines, I again ascribe to the adhesion of the water to the walls
of the tube.
Then I immersed this tube, completely empty of water, furnished
with the same plate, with a finger placed over the top, to a depth of
18 lines into the water; after the finger had been removed, the surface
in the tube emerged above the standing water a full eight lines, while
§17 indicates nine for this case.
I attributed the fact that here the defect was altogether less than in
descent to the reason which I indicated freely in §13, where I said that
a slightly greater motion would develop when the orifice would have
a considerable area with respect to that of the tube, such as in the
ratio V½ to I or thereabout, than that which follows from the hypo-
thesis; and in order that I might be made clearly certain of this
matter, I applied a shorter and wider tube, so that the effect of almost
all outside hindrances would be forestalled, and I performed the
experiment which follows.
EXPERIMENT 3. I provided a tube the diameter of which was more
than seven lines, which I took pains to have made of iron, because
sufficiently cylindrical glass was not at hand; the length of this was
four inches and six and a half lines; its area was in a ratio of 1 .860 to
that of the orifice, indicated by n, and nn was 3.458.
With that tube I performed the experiment thus:
With the upper orifice closed off, of course, I tried many times [to
determine] to what depth it should be submerged in the water stand-
ing in a very large tank so that, directly after the finger had been
removed which covered the orifice, the water would ascend precisely
to the edge of the same orifice, and nothing would flow past. Indeed,
I found that depth to be 3 inches and three lines; therefore, the ascent
above the external water was one inch and three and a half lines,
whereas, even with all the hindrances removed, the ascent should have
been just a little beyond eleven lines, according to §17. Accordingly,
one was correctly advised in §13 that the ascents cannot be a little
greater in cases of this sort than the hypothesis postulates. I then
applied another base to the same tube; now n = 3.68 and nn = 13.54;
it was difficult to determine correctly the success of the experiment,
because the surface ascending in the tube was always bubbling;
nevertheless, it seemed that now the tube had to be immersed to a
 HYDRODYNAMICS, CHAPTER VII

height of 4 inches and two or three lines, with four lines more or less
thus remaining above the water, precisely as the theory indicates.
EXPERIMENT 4. I immersed a cylindrical glass tube, which had a
diameter of three lines more or less, to a depth of 20 inches, and I
caused the water in it to be held in equilibrium after it was first
elevated to a height of almost one inch. It did not produce beyond
four or five clearly noticeable departures and returns, and therefore
I was not able to determine the length of the simple isochronous
pendulum with all rigor; nevertheless, it seemed to me that it was
22 or 23 inches, from which I inferred that the adhesion of the water
to the walls of the tube not only diminishes the displacements, but it
also delays the times of the oscillations slightly; for according to § r g
the previously mentioned length should have been only 20 inches.
I found the same thing in the oscillations which we considered in the
section above.
Finally, with the lower orifice approximately half blocked off, I
was not able to observe that the displacements had been diminished
or the oscillations retarded, which agrees with those things which are
found in §§7 and 18.
EXPERIMENT 5. I immersed a conical tube with a length of 21
inches with the wider orifice in the water, so that one inch extended
beyond the water; moreover, one orifice was a little more than twice
the other. I found that the length of the pendulum isochronous with
the vibrations of the water balanced in the tube was r 5 inches, but
according to §23 the same length should have been a little less than
14 inches. Finally, using the same tube similarly, but in an inverted
position, I discovered that the length of the isochronous pendulum was
a little more than double that which it had been before, just as 1s
indicated in the cited paragraph.
 EIGHTH CHAPTER

Concerning the Motion of Homogeneous as well as

Heterogeneous Fluids through Vessels of Irregular
and Abrupt Shape, where from the Theory of Live
Forces, a Part of which is Continually
Absorbed, are explained Excellently Singular
Phenomena of Fluids driven through
Several Orifices, after General Rules have been
Set Forth for Defining the Motions of Fluids
Anywhere

§1. So far, except in the chapter immediately preceding, we have

not made use of any principles other than those two: that the velocities
offluids are everywhere reciprocally proportional to the areas ef the vessels, with
the aid of which is found the potential ascent of all the water from the
given potential ascent of any particle whatsoever; and, further, that the
potential ascent of all the water always remains equal to the actual
descent. So often do these two principles apply that it is by no means
to be questioned whether the motion of fluids is correctly defined by
the method furnished by us. Nevertheless, I will not deny that
vessels in which fluids are moved can be made of such a shape that
neither one of these principles develops correctly. The former
[principle] indeed rarely or never varies noticeably from the truth,
because wherever it does not apply the water usually has almost no
motion and can without noticeable error be considered as standing;
but the other principle is regarded otherwise by far, which will be
apparent from the following example, and those things that we
offered in the preceding chapter on the backflow of water can serve
as a splendid testimony to this matter; indeed, it is far from possible
that water in a submerged vessel, having fallen from a given height,

- ----=------------=-
160 HYDRODYNAMICS, CHAPTER VIII

rises again to this height, as it should by virtue of this principle after

external hindrances have been subtracted; rather, its ascent is often
barely noticeable compared with the descent which it made before;
on the contrary, indeed, the water surface certainly cannot rise as
much above the water in which the pipe is immersed as it had been
depressed below the same unless the entire pipe is open; but that sur-
face is depressed much less than it had been elevated before. The
reasoning on these things we gave in the preceding chapter. Because
they are so, I will now give two rules for defining the motion of water
everywhere, and then I will illustrate them with certain examples
which could not be explained by any theory up to this point, but which
conform most excellently with ours.

RULE I

§2. One must discover what the velocity will be in the other parts
of the fluid after the velocity anywhere in the vessel under considera-
tion has been assumed as known. For thus the potential ascent of the
entire fluid and its differential may be found. So far we have con-
sidered fluids to be divided into infinite parallel layers, or rather
layers everywhere perpendicular to the walls of the vessel, and we
have stated that the velocities are reciprocally proportional to these
layers; certainly it is easy to fashion a vessel wherein the fluid is
moved differently; I would believe, though, that the fluid in these
places never has a conspicuous motion, so that a noticeable deviation
from this hypothesis can hardly ever arise; nevertheless, for the sake
of greater accuracy the above-mentioned rule could be applied.
Indeed, this pertains especially to the contraction of jets whenever
fluid is forced to go through orifices made in very thin plates, in which
matter great care has to be taken. I believe that the effects of con-
tractions of this sort will be understood in advance quite properly if
what I said about them in Chapter IV is considered correctly.

RULE 2

§3. One must discover at any instant how much of the live force
or what product of potential ascent and mass may develop without con-
tributing anything to the main flow, the nature of which is sought.
But this in turn is to be left to anyone's careful estimation. Then the
product is to be added to the product of the mass and the potential
ascent that the main motion contains, and the sum of the products
finally is to be considered equal to the total mass of the water times
its actual descent.
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS I6I

This rule is of great progress-making importance and I believe

almost unique for obtaining measures of motions which occur in
irregular vessels divided into several cavities connected to one
another, which I will now illustrate by several examples.

PROBLEM

4. Let there be proposed a vessel ACRE (Fig. 37) of area every-

where infinite, so to speak, in proportion to [the areas of] the orifices

FIGURE 37

to be defined presently, divided into two intercommunicating cavities

by some diaphragm EF, with the orifice G in the center; furthermore,
let that vessel have another orifice D in its lowermost part; then let it
be supposed that the vessel is full of water up to PQ,, so that the lower
cavity CEFR will be filled completely with liquid, and that further-
more the other part PQ,FE lies above the diaphragm. With these
things having been set forth and the fluid already starting to be moved,
the velocity of the water flowing out through the orifice D into the
air, or the height creating this velocity is sought.
SOLUTION. Let the height of the surface PQ, above the orifice D be
x, the area of the orifice D be n, and that of the other one, G, be m.
It is also very clear that the potential ascent of any drop flowing through
G promotes nothing toward the effiux through D, and that all is used
 HYDRODYNAMICS, CHAPTER VIII

for exciting some internal motion which is soon absorbed without any
other effect; therefore, it is necessary that in every instant a new
motion be generated in the particles that will pass the orifice G, and
no less in the particles flowing out through D. But if the potential
ascent of the volume element flowing out through D is called v, that is,
if the water is assumed to spring forth through D with a velocity of
which the generating height is v, then the similar height, in reference
to a volume element (equal to the former in bulk) flowing through G
·
at t h e same time, nnv. Ar.1ter t h ese potentza
w1·11 b e - · l ascents h ave b een
mm
multiplied by the mass which they have in common and which I shall
Mnnv
call M, the sum of the products will be M v + --. And since,
mm
because of the infinite area of the vessel no other motion is generated,
the aforementioned sum (by Rule 2) is to be considered equal to the
product of the total mass of water and its actual descent. But if now
the total mass of water is called µ, the actual descent, which occurs as
long as the volume element M flows out, will (per §7, Chapter III ) be
Mx, so that the product of the two is Mx. Hence one obtains
µ
Mnnv mmx
Mv +- -
mm
= Mx, or v = ---
nn+mm
Q.E.F.

§5. ScHOLIUM 1. It is evident from this example that the motion

can be determined without differential calculus, since the shape of a
vessel that is very wide everywhere cannot affect this motion. Mean-
while, it would not have been difficult to define the flow, with con-
sideration having been given also to the areas of the vessel, and only
for the sake of brevity did we avoid it and will we similarly omit it in
the future, unless perhaps the motion be noticeably changed by the
varying shape of the vessel, which can happen in vessels in which
fluid is moved that are wide enough but very long, particularly if the
motions to be determined are oscillatory. Finally, we have seen in
the preceding Chapter that, if there are very small oscillations in
extremely deeply submerged pipes, then it is so far from necessary
that one should pay attention to the orifice in the base alone, the areas,
although large enough, having been neglected, that rather one should
take those [areas] alone into consideration.
§6. ScHOLIUM 2. Because in the calculation which we have per-
formed the live force of any volume element flowing through G must be
absorbed by the water in the lower cavity, it is evident that the
proposition must not be extended to those cases which oppose the
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS 163

hypothesis, as, for example, when the diaphragm EF is close to the

base CR and consequently the orifices lie directly opposed to one
another; thus indeed it is not hard to predict that the motion will be
different by far from that which the present theory indicates. How-
ever, if the distance DG is large, and if as well the position of the
orifices is oblique and the walls of the orifices do not allow the aqueous
stream to contract, then there is no doubt that the theory corresponds
accurately to all phenomena.
§7. COROLLARY. If the orifice G is fairly large compared to the
other, v becomes almost equal to x; but this height v, to which cer-
tainly the velocity of the water flowing out through D corresponds,
decreases considerably with the orifice D increasing, so that if, for
example, there would be twice the orifice G, v would be ½x, but the
total [height] almost vanishes when the orifice G is extremely small
with respect to the orifice D.
Thus, with these things having been effected, anyone will now per-
ceive the true understanding of those motions which Mariotte first
observed and concerning which very admirable [findings] he states that
he was overly pleased; and simultaneously one will understand how
far this Author, most clear-sighted in other matters, diverged from the
[right] path in these treatises. I believe that it will not be irrelevant
to insert the observations of Mariotte here.
§8. He used a vessel such as Fig. 38 represents, which differs from
the former only in this, that in the lowermost part there is inserted in

A ,.
I!
I

FIGURE 38
 HYDRODYNAMICS, CHAPTER VIII

the cylinder ABC a horizontal pipe MD, perforated by an opening D

through which water springs forth vertically: but the Diaphragm EF
is, as before, perforated in the middle by the opening G; below this
there was a very small orifice K, [placed there] in order that the
lower cavity could be filled with water more easily; after this had been
done, [the small orifice] was closed and the rest of the vessel was filled.
Having prepared things in this manner, with water flowing out
through D, Mariotte observed that it soon ascended up to I, then,
gradually, with the velocity diminished, up to H, and finally, with
total depletion of the upper cavity ABFE impending, up to 0, and
then, new forces suddenly having been added, jumps almost up to F.
He also noticed, if I remember correctly, that the height of the initial
thrust is the smaller, the smaller the orifice G with regard to the other,
D. This may be seen in his Traite du mouvement des eaux, Part IV,
disc. I. Moreover, he believes that the changes in these motions can
be explained by imagining that to the very wide vessel ABFE a rather
narrow pipe GLMD is connected through which the water flows.
But we have certainly demonstrated and experience teaches daily
that the motion of water out of the vessel ABGLMD is very different
from that which has just been indicated. One would be no less
wrong if he believed that the water springs forth with the same velocity
through the orifice D as if the latter would have been located in the
diaphragm EF, for it can happen that the height of the initial thrust is
larger or smaller than the height FE. And, finally, the water does
not flow out in that quantity, as one might easily suspect, in which it
would, at the same time, flow out of the upper vessel alone through
the opened part EFDC, although this is approximately the case when
the orifice G is so much smaller than the orifice D.
§9. In truth our equation, namely v = mmx , corresponds alto-
nn + mm
gether correctly to the phenomena: it indicates indeed that the water
ascends soon after the beginning of flow to a certain height, which is
less, the smaller the orifice in the diaphragm is with respect to the other
orifice; that this ascent is then gradually diminished until all the water
has flowed out of the upper cavity; that at this very instant it imme-
diately experiences an increase and [the thrust] reaches not quite the
total height of the water lying above, because then the water is to be
regarded as flowing out of a simple vessel which is infinitely wide;
nevertheless, even now the water is retarded a little by the transition
of the air through the orifice G, and understandably it is retarded
noticeably if the upper orifice is very small, about which subject we
will soon say something when the discussion will concern hetero-
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS 165

geneous fluids. If Mariotte's figure stands in proper proportion in the

argument brought forth, then it is necessary that one make the orifice
G a little more than half of the other.
§10. Our formula indicates further that it could perhaps have
seemed to be a paradox to many, when this theory had not yet been
understood, that either a higher or a lower position of the diaphragm
EF in no way changes the impact or the velocity of the water flowing
out; however, I believe that the understanding of this phenomenon is
now manifest to everybody.
§11. Now, however, we will examine further the motion of water
when there are many diaphragms perforated by orifices through
which the water is forced to flow in order that efflux through the
orifice D may occur. This can be solved by the same method which
we have used in the problem of§4. Moreover, after the calculus has
been correctly applied, and with the notations applied in the same
manner having been retained, there appears

V =XI ( I + :: + ;; + ; + .... ),
where by a, f3, y, etc., are understood the areas of the orifices which
are in the diaphragms, while n expresses, as before, the area of the
orifice D through which the water flows out.
§12. If then in place of one diaphragm there are in the same
vessel, which Fig. 39 represents, many diaphragms, let us say B, C,
R, etc., through which the water flows as long as it flows out through
the lowermost orifice D, then the velocity of the outflowing water will
be changed and increased immediately every time some cavity is de-
pleted: further, the proportion between the heights AB, BC, CR, RE,
etc., and the areas of the orifices D, G, F, H, etc., can be such that
every time when a new chamber starts to be depleted, the outflowing
stream always rises to the same height 0, or it flows out at the same
velocity. But this is obtained (the areas of the orifices D, G, F, H
etc., having been designated by n, a, f3, y, etc.) by setting
nn
BC= nn AB; CR = f3f3 nn
AB; RE= -AB; etc.,
aa yy
so that after the orifices. have been set equal to one another, the lines
AB, BC, CR, RE, etc. are similarly to be made equal to one another.
It will also be easy in a,cylindrical vessel to ascribe such a magnitude
to the orifices that the surface of the fluid descends in the same time
from one diaphragm to any subsequent one, and since these dia-
phragms are spaced equally from each other and from the base, a
uniform construction of clepsydras can be invented.

- -------

1
166 HYDRODYNAMICS, CHAPTER VIII

FIGURE 39

§13. Indeed, if all diaphragms are placed very high, it will be a

pleasant hydraulic game to observe the discharging stream DO, which
increases repeatedly by equal increments and in equal intervals of
time, which can be done.
§14. Let it be proposed now to investigate the motion of the dis-
charging fluid when different fluids flow through all the individual
orifices. But evidently the successively lighter fluids have to be
placed so that they are located higher in order that the motion does
not become disordered, which happens when a lower fluid ascends at
the same time that an upper one descends through a common orifice.
In this manner one may determine what the motion is in the water
flowing out of a vessel closed everywhere except for some small
orifice located at the top which allows air to enter. But we will
retain the hypothesis of the infinite area of the cylindrical vessel in
relation to the orifices, and further we will designate the specific
gravity of the fluid discharging through D by A and that of the one
which flows through G we will denote by the letter B, and similarly
we will indicate the specific gravities of the fluids flowing through
F, H, etc., by the letters C, D, etc., respectively. Finally, since here
also one must consider the heights of the different fluids, of which
only the lowermost discharging one changes height, of course, be-
cause of the cylindrical shape of the vessel, we will let x be the height
of the lowermost fluid above the orifice D; the heights of the remain-
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS 167

ing fluids we will designate, in that order in which they lie on top of
each other, by b, c, d, etc., respectively, and we will retain the other
designations of§11; with these things having been thus prepared, the
computation may be performed as has been done in §4, for there is
nothing else further to be observed than that the masses of the volume
elements passing through the different orifices in the same small time
intervals are estimated not simply from the bulk, but also from the
specific gravity; the actual descent for the individual fluids will have to
be taken separately, though. By following this path an equation is
found at first in this form:
nn nn nn
Av + -aa Bv + f3f3 Cv + -yy Dv + · · · · = Ax + Bb + Cc + De + · · · ·
which, reduced, gives
v = (Ax + Bb +Cc+ Dd + · · · ·)/
nn nn nn
( A+-B+-C+-D+
aa f3f3 yy
... · )·
§15. If there are two liquids, two terms will have to be taken in the
numerator as well as in the denominator, and three terms if there are
three liquids, and so forth; if then the liquid flowing out were mer-
cury, for example, and if water were lying on top of it, and if the
specific gravities of these liquids were established as 14 to 1, it would
occur that
l4X + b
v = 14 + nn/aa'
and if the ratio of the orifices D and G should be, for example, as
3 is to 1, it would occur that

V
I4X + b
= --"---•
23
§16. It is also evident that that reasoning does not exclude those
cases in which the upper fluids are specifically heavier than the lower
ones, only that the lower fluids do not ascend through the same ori-
fices through which the upper ones descend; but I presume (however,
I do not confirm) that this will not happen when instead of a simple
orifice there is some little pipe of small height through which the
upper liquid may descend into the lower cavity, just as in Fig. 40,
where indeed only two liquids are considered.
Here, though, the height CR is variable, the height AC is constant;
meanwhile, nevertheless, for the sake of uniformity of nomenclature,
168 HYDRODYNAMICS, CHAPTER VIII

we will set the height AC = x, the other one, CR equal to b; the speci-
fic gravity of the fluid going out through D we will set again equal to

FIGURE 40

A, and that of the other fluid, passing through G equal to B; the height
DO will result, or
Ax+ Bb
v=-----·
A+ Bnn/aa
Therefore, if water and mercury flow through the orifices D and G,
respectively, now there will be
X + 14b
v = -I-+--14_n_n_/_a_a·
§17. In order to understand further the motion of a simple fluid out
of a vessel that admits air through a very small orifice on top, it is to
be observed that here the height b is null; because the air can be con-
sidered to lie above each orifice up to the same height, hence there
will be
Ax
V=------,
A+ Bnn/aa
and if ~ were 850, which is more or less the usual proportion be-
tween the specific gravities of water and air, there will be
85ox
V = ~--=--~·
850 + nn/aa
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS 169

§18. All these principles which we so far have applied are, as I have
said already, easily extended to vessels which have a finite area in
proportion to the orifices; but the truth of these things can be proved
also in another, very different manner, as I shall show when I come
to hydraulico-statics, because, by that other method of proof, the pres-
sures of the fluids on the individual parts of the vessel become more
clear; however, the statical rules of those fluids differ strongly from
the laws which are due to standing fluids.
Otherwise, these things have their use in correctly understanding
hydraulic machines; indeed the professional men seem not to have
attended enough to this; occasion will also be given to elaborate on
these things more copiously in the following chapter, where we will
perform a calculation of how much force applied in propelling water
may be lost from the passing of water through many orifices, and we
will simultaneously show the remedies to be applied in order that that
loss of forces be diminished as much as possible. But we will consider
certain other composite vessels in this chapter before we turn to those.
§19. It happens sometimes that vessels put next to each other
receive water from one another that finally will flow out of the last
one. We will now illustrate those motions by an example.
Let there be proposed a vessel ACME (Fig. 41) of any shape what-

FIGURE 41

ever, which is kept constantly full up to AB by a new supply of water.

Meanwhile, let the fluid be understood to go from this very vessel
through an orifice Minto another adjoining vessel BMNC, and from
that again into another one, CNRD, through the orifice N, and so on,
170 HYDRODYNAMICS, CHAPTER VIII

until finally the water is emitted into the air, and let the locations of
the surfaces HL, PQ, etc., be sought after they have been reduced to
a state of permanence. The question is solved as follows:
It is certainly clear from the fact that the surfaces AB, HL, PQ,
etc., remain in the same position that the water goes through the ori-
fices M, N, R at the velocities which are due to the heights BH, LP,
QR, if only the transit of water through the one orifice does not
accelerate its flow through the next orifice, which certainly does not
occur unless an effort is made expressly in order that this happen
somehow. But furthermore it is to be considered that the velocities
of the water flowing through the orifices are reciprocally proportional
to the orifices, because in the state of permanence the same amounts of
water are released in the same time through the individual orifices.
From this it is recognized, once the areas of the orifices M, N, R have
been designated by m, n, p, that LP will be mm BH; QR = mm BH.
nn pp
But BH + LP + QR is equal to the height of the surface AB above
the last orifice R, or [equal to] DR; therefore,

mm mm
BH + -BH
nn
+ -BH
pp
= DR
and thence

BH = DR /( I + -mm
nn
+ -mm)
pp
;
and similarly

and

QR= mm DR/(1 - +mm)

+mm
nn
-,
pp
pp '
or

BH=DR/(1 - +mm)
+mm
nn
-
pp

LP= DR/(1 +-+-

nn
mm
nn)
pp

QR= DR/(1 +pp+ PP)

nn mm

and thus are determined the invariable locations of the surfaces HL,
PQ, etc. But on the other hand we will examine below, together
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS I 7I
with other questions pertaining to this, how much time it takes for
this to occur if those surfaces are located differently and in the in-
terim some certain quantity of water flows through the individual
orifices. However, we will deduce [some] outstanding relations
arising from the presented values of the heights EH, LP, QR, etc.
§20. I. If the individual orifices are equally large, there will be
EH= LP = QR, etc., and something of these heights will be con-
tained in the height DR as long as the vessels are open.
II. But if some orifice is infinitely small in relation to the others, all
surfaces which are located upstream from the orifice will be at the
same height as the first surface AB; but the remaining will be close to
the base GR.
III. If a continuous conduit passing through the individual orifices
M, N, R, etc., is assumed, then it is recognized that the water must
flow out through the orifice of the conduit at a velocity which is due
to the total height DR. But in our case that velocity corresponds
only to the height QR, of which matter the reason and origin is this,
that the potential ascent of the individual volume elements flowing
through the orifices-except only for the orifice of efflux-is ab-
sorbed. Therefore, the live force which is lost at any individual
instant is to the live force which is generated at any individual in-
stant as DQ is to DR. But the heights EH, LP, etc., represent
respective!)! the live force which is continually withdrawn separately
from the volume elements flowing through the orifices M, N. Never-
theless, if the orifices are almost equal and if their centers are located
in a straight line, and if, finally, the walls BM, CN, DR are placed not
very far from each other, [then] I believe it can happen that the
water springs forth at some higher velocity than this theory indicates.
In the remaining cases I do not doubt their accuracy, neglecting the
often indicated hindrances.
IV. Finally, it is evident that every time the water surfaces HL,
PQ, etc., change their position, whether many or one alone, soon all
surfaces will change their locations until they have been brought back
to equilibrium in the manner that has been mentioned. But to
define these changes generally is [a matter] of equally intricate as
well as laborious calculation, unless the vessels are taken as prismatic
and as practically infinite with respect to the area of the orifices,
namely in order that the increments of the potential ascents of the water
ML, NQ, etc., which change their locations, can be neglected in
comparison to the potential ascents which are perpetually generated in
the volume elements flowing through M, N, R. And this restriction
at the outset must not affect us, since we already saw in passing that in
 HYDRODYNAMICS, CHAPTER VIII

moderately wide vessels the increments of the motion of the internal

masses can be left out of the calculation without noticeable error.
Therefore, let me omit the general solution, which is mine, on account
of its overly great involvement, and for greater elegance let me, as I
have done so far in this chapter, assume the vessels as infinitely wide
and certainly prismatic. But let me start with a sectioned vessel.
§21. A sectioned vessel of this kind (Fig. 42) is represented, the part
AM of which is assumed full of water; the other, BN, is assumed to be

FIGURE 42

filled only up to HL when flow starts through each orifice Mand N;

water is also poured in at AB in order to keep the vessel constantly
full, and thus it will occur that the water in BN rises (or even de-
scends, according to the conditions). When this is so, we will seek
the velocity of the surface of the water when it arrives at the position
hl.
To this end we will express the area of the orifice Mby m, that of the
orifice N by n, and the area hl (which certainly is taken everywhere
the same) by g. Then let us set BM= a, HM= b, Bh = x, and
hence hM = a - x. But thus it is evident from the assumption of a
practically infinite area of the vessels AM and BN that when the
variable surface of the water is at hl, then the height due to the velocity
of the water flowing through M will be Bh = x, and the velocity
itself will be Vx, and with respect to the orifice N the similar height
will be hM = a - x, and the velocity of the water flowing through
N will be Va - x; therefore the quantity flowing into the vessel BN
through Min a given time element is to the quantity flowing out of
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS I 73

the vessel in the same time element as mVx is to nVa - x, and the
difference of these quantities divided by the area g gives the velocity
of the surface hl; hence, this velocity, which we will call v, will be
expressed by the equation:

mV°x - nVa - X
V=-------
g

§22. Now, in order to find the time in which the surface of the fluid
-dx
rises from HL to hl, we will call this time t. But because dt = --,
V
there will be, after the value just found has been substituted for v,

-gdx
&=--=-~--c,==·
mVx - nVa - x

Certainly this formula can be made rational at once by putting

x = ( 4aqq ) 2 and then arranging in the required manner. But this
I + qq
method is slightly more favorable than that other in which the quan-
tity to be reduced is divided into two parts that are consequently to be
integrated; certainly the equation set forth does not differ from the
following:

mg dxVx ng d x ~
dt = ---"-,------
nna - (mm + nn)x
+ --"-~--~-·
nna - (mm + nn)x

And also

mg dxVx 2mg v;
Jnna - (mm + nn)x = mm + nn
mng-va nva + vmm + nn -vx
+ -:-------"-==== 1n -t=---====----;=.
(mm+ nn)Vmm + nn nva - vmm + nn vx'

.
the mtegral of the other part, namely
J ngdxVa
(
- x ) , becomes
nna- mm+nnx

-2ng . 1- - mngVa
Va - X + ----- ====
mm+ nn (mm+ nn) Vmm + nn
mVa + Vmm + nn Va - X
X ln - , = - - , = = = = - ~ , = = 0

mVa - Vmm + nn Va - x
1 74 HYDRODYNAMICS, CHAPTER VIII

It is evident that, after the required constant has been added, this
will yield

2mg~ - 2mgVx + 2ng-Vb - 2ng-V a -x mng-Va

!=----------------+----=--;::===
mm+nn ~-- (mm+nn)Vmm+nn
mna +(mm+ nn) Vax-xx+ m-V mm+ nn-Vax+ n-V mm+ nn-Vaa-ax
xn-----'-----===---====---=--====-===
I
mna +(mm+ nn) V ax-xx-m-V mm+ nn-Vax-n-V mm+ nn-Vaa- ax
mngVa
(mm+nn)Vmm+nn

mna+ (mm +nn)Vab-bb + m-V mm+nn-Vaa-ab +n-Vmm+nn-Vab

xn--~--~~==~-~===~==~-~===~~-
I
mna + (mm+ nn) V ab - bb - m-Vmm+ nn-Vaa - ab - n-Vmm+ nn-Vab
§23. From §19 it is clear that the surface hl remains in its position
since Bh( = x) = nna . But if in the integrated equation of the
mm+ nn
prece d mg - - , t h e d enom1nator
. paragrap h one puts x = - -nna · · the
1n
mm+ nn
logarithmic quantity becomes = o, and hence the quantity itself is
infinite. The time of the total motion, therefore, is infinitely greater
than that of any part.
But in order that we may determine another case beyond this, we
will see in how much time the surface of the water would ascend from
its lowermost position MN (namely by setting b = o) by the quantity
½a, but setting m: n = 4: 3, one has

Bg-Va - 14gV ½a
t -- --=-----==------''-- l 2g-Va I
+- -- n (49 + 35 V2)
-==-------"---''--~ -
l
---
2gVa In ( -4 ) ,
25 125 49 - 35v2 125
or,

t = Bg-Va - 7gV2a + 12gVa In ( 49 +_ 35V2 )

25 125 14ov2 - 196
15
that is, approximately t = g 2-Va, which indicates that this time is
100
to the time during which a heavy weight falls through the height Blvf
approximately as 15g is to 100: equally, the time of descent is found,
if in the beginning the surface hl would have been located above the
position of equilibrium. Let, for example, either one of the vessels
be completely void of water, and let the orifices Mand N now have a
ratio 3 to 4, and let the time be determined in which the surface
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS I 75

descends from B through half of BM: these hypotheses make m = 3;

n = 4; b = a; and x = ½a, so indeed one has

Bgva
t = - - -- -7gv2a
-- 12gva 1n
+ --- (49 + 35v2)
-=-=------=-='---~ -
12gva
- - - 1n
( -4.
)
25 125 49 - 35v2 125
From this it is apparent that the time is the same in either example.
§24. Before we get onto manifold vessels, it is convenient to have
investigated what quantity of water flows through each orifice Mand
N while the surface of the water goes from the position HL into hl.
And first of all, certainly, it is evident, as far as the orifice Mis con-
cerned, that the quantity of water flowing through it in a given time
interval (dt) is proportional to the velocity (Vx) multiplied by the
magnitude of the orifice (m) and the same small time interval dt, so
that this quantity is (on account of dt = V- -g ~ - - per §22)
m x-n a-x

v-mg d~Vx ' and hence the entire quantity which has flowed
m x-n a-x
out from the beginning is

But

-J mgdx-Vx = mnga ln(ma-mb-nb)+ mg (a-b-x)

m-Vx-n-Va-x (m+n) 2 mx+ nx-na m+n ·

In the same manner one evaluates the quantity of water flowing out
meanwhile through the orifice N, which, of course, is

ngdx-V~
- J
m-Vx - n-Va - x'
is equal to

mnga 1 (ma - mb - nb) _ ___!!L (a _ b _ x).

(m + n) 2 n mx + nx - na m + n
And from here the quantity of water which is poured into AB
becomes known also, and it certainly does not differ from that which
flows through M; finally, the water collected in the vessel EN is
represented by g(a - b - x), and when the difference of the water
flowing through M and through N is determined, then that same
quantity g(a - b - x) appears.

·- - ----------------
 HYDRODYNAMICS, CHAPTER VIII

§25. Just as in §2 1 we have determined for a two-part vessel the

velocity of a surface continuously changing its position, so let us now
define the velocities of the individual surfaces for manifold vessels.
Certainly one may set the height of the uppermost surface above the
next equal to x, the height of that one above the following equal toy,
then equal to z, and again the next height equal to s, and so forth.
But let the areas of the orifices be designated by m, n, p, q, etc., let the
areas of the second, third, fourth, etc., vessel be M, N, P, etc. Thus
· ev1"d ent t h at t h eve1oCity
· 1s
1t · o f t h e secon d sunace
r-. w1·11 b e mV~ M
- n-Vy ,

the velocity of the third surface will be nVy 7.r PVz, the velocity of the
.
fourth surface will be
PVz p- q-VS, etc.
Further, since the small spaces passed through by the surfaces in
the same small time intervals are in proportion to the velocities, it is
thus apparent that at any instant the position of these surfaces is
determined, although the equations are almost intractable. This is
evident by itself, or, if a single surface were put off its position of
equilibrium defined in §19 above, then all the remaining are agitated
by reciprocal motions, until after an infinite time they will simul-
taneously go back to their original position.
§26. Further, let a vessel be formed such as Fig. 43 shows, divided,
of course, into two parts ABEG and LQNE, communicating with each
other through the middle orifice M; and let there be, furthermore, the
orifices Hand N through which water springs forth as long as the same

---~-~ -- ...,~.... --
~ ":' ... ~~
·~--~~-::.
-~~-- :---~
-~ -~-~ _;_..~
~~ ~~-~
·-,~

_:--r· - -~

FIGURE 43
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS 177

amount is poured into AB. Also, let the areas in either vessel be in-
finitely large in proportion to the orifices M, H, and N; after these
things have been established, let it be proposed to find the velocities
with which the water is ejected through Has well as through N, or the
heights due to these velocities. But the velocities will be invariable,
because the vessel is kept full of water and at the same time the areas
of the vessel are taken as infinite with respect to the orifices.
The solution of this problem may be easily understood from the
preceding, if only the orifice M is conceived to be divided into two
parts o and p, of which the one, o, sends water to the orifice H, the
other, p, to the orifice N: the parts o and p, however (because through
either one the water flows at the same velocity) will have the ratio
which the quantities of water flowing out through H and N at the
same time have, that is, a ratio composed of the ratio of the area H
to the area N and of the velocity at H to the velocity at N. After
these things have been admonished, then, if the areas of the orifices
M, H, and N are indicated by a, {3, and y and the heights due to the
velocities at H and N are designated by x and y, and hence the velo-
cities themselves by vx
and -vy,
it is clear that one will obtain the area
f3Vx rvy
o = v
f3 x+y y
v a and the areap = v
f3 x+y y
v_a.
Now let the height of the surface AB above the orifice H be given
equal to a, and x will result, as it was proven in §4, if the square of the
orifice o is divided by the sum of the squares of the orifices o and H,
and if the quotient is multiplied by a; and so it will occur that
x = aaax
.r r:
. 1 , irom w h'1ch t h'1s equation
. resu lts:
aax + (f3v x + yv y) 2
(A) aax + (f3Vx + yVy) 2 = aaa.
In the same manner, from the ratio of the orifices p and N, after the
height AB above N has been set equal to a + b, this other equation is
obtained:
(B) aay + (f3vx + yvj) 2 = aa(a + b).
After equation (B) has been subtracted from equation (A), there
results y = x + b, from which it follows that, if both streams are
directed vertically upwards, each one springs up to the same position.
Hence, if in equation (A) the value of x + b is substituted for y, then
(C) aax + (f3Vx + yVx + b) 2 = aaa,
from which the value of x itself is deduced from the quadratic equation.

. --------------------- - -----
 HYDRODYNAMICS, CHAPTER VIII

§27. From the equation of the preceding paragraph the following

conclusions result.
I. Because the velocity of the water flowing through M is
fvx+yv_;;
a
h h . h . h. i .
, t e e1g t generatmg t 1s ve oc1ty w1
·11 b e (fivx+rv_;;)2
a

but if the equations (A) and (B) are added, there results:

fiVx + yVy) 2-_ -------'::...

2a + b - X - y
-_ (smcey
· -_ x + b) a -
( IX 2
x.

II. If the orifice His very small in proportion to the orifices Mand
N, that is, if fi can be assumed as zero in proportion to a and y, then
equation (C) changes into this:
aaa - yyb
aax + yyx + yyb = aaa, or x = --~---'----
aa + yy

But this agrees splendidly with§ 19, since it is manifest that the water
springs forth through a very small orifice to the same height which the
water would have ifit pressed the section LQas much downward as it
is pressed upwards by the internal water; but this mentioned height
is, by virtue of§19, aaa - yyb_ Further, in this hypothesis one finds
<XIX + yy
the height of the velocity of the water at N, or
aaa + aab
x+ b = - - - - ,
<XIX + yy

and finally the height of the velocity of the water at M, or

yya + yyb
a - X = -'-'--~-'-,
<XIX+ yy
which latter equations could have been immediately understood or
predicted in this particular case from §19 as well.
III. But if now another orifice N, sufficiently small, is placed in
front of the remaining two, there will be, after one has set y = o,
aaa aaa + aab + fifib fifia
x = thenx +b= and a - x
<XIX + r:u:i;
/J/J <XIX + RR /J/J
, =
<XIX + RR. /J/J

IV. If yyb = aaa, x becomes null. Therefore, in this case the

various portions of the section LQ sustain no pressure: in fact it is
pressed downward if y is larger than at and the section is not per-
forated anywhere.
But, similarly, all these things are understood easily from §19.
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS I 79
V. Thus, also, by means of the same paragraph it could have been
predicted without new calculations what should happen if, the ori-
fices H and N having been located at the same height, the sum of
these orifices or a unique one of area (3 + y can be considered.
Certainly §rg as well as §26 indicates that

a.a.a
X = ------,-.,,----,--=,•
Ct.Ct. + ((3 + y)2
VI. It can also be noted that, when the value of x itself becomes
imaginary, it happens not only that does the water not flow out
through H in certain cases, but also that the surface LQ descends;
whence it can happen that it descends below the orifice M, in which
case the continuity of the water ceases, contrary to the hypothesis of
the proposition. Moreover, if the value x is real, then it is doubly
expressed, but the other value is to be considered useless; accordingly,
therefore, care has to be taken lest the absurd root be taken as useful.
VII. Finally, in order that we may treat a very special case, let us
set all orifices equal to one another, and there will result

5xx + (2b - 6a)x = -aa + 2ab - bb

or
3a - b - 2-V (aa + ab
x-- ~ - - - - -5~ - - - - - ·'
- bb)

and if, furthermore, a = 3b, then x will be (approximately) 1kb,

hence the height of the velocity at the orifice Nor x + b = Hb and
the height due to the velocity at M or a - x = f lb. And so the
velocities or even, since the orifices are equal, the quantities of water
flowing in the same time through the orifices M, Hand N, are approxi-
mately as V4r: 2: Vrg.
§28. From all this the method is evident for determining the motion
of fluids, even when the amount of live forces is not conserved; and the
computation is always performed in a similar manner as long as it
can be presumed from the nature of the subject of investigation (as
could be done accurately in the investigations in this chapter), how
much of the live force vanishes that is useless for determining the motion
at any instant. Certainly the cases are not singular which we have
examined so far; it is plea~ing, therefore, to add another one which
treats oscillations of fluids, in order that one may know for this how
much the displacements of the fluid decrease.
Let there be two pipes, equal in size and cylindrical, AL and BH
180 HYDRODYNAMICS, CHAPTER VIII

(Fig. 44), inserted vertically into a very large horizontal vessel

ABOP. Let that vessel be completely full of water, but let the pipes
contain water up to C and F; then, with equilibrium having been
disturbed, let one surface stay at G and the other at E; and let the
water, left to itself, soon begin to move. These things having been

FIGURE 44

set forth, the surface G should descend as far below the position C and
the other surface E should ascend as far above F as the height CG or
EF is, if the entire live force were conserved (we disregard the hindrance
of frictions and other similar things); in truth it is evident that the
live force of all the water flowing through A in the horizontal vessel is
absorbed without any other effect from the water standing there, and
hence it follows that the descent of the surface G and the ascent of the
other will be less than was just mentioned; therefore, we will now
explore this decrement.
To this end let it be assumed that the surface from G has reached
M, and let GM = x, GC = b, CA = a; it will occur that BE =
a - b, EN = x, MC = FN = b - x. Further, let the height due to
the velocity of the surface at M be v, and at the next position let it be
m = v + dv; and the increment of the live force of the water (while the
surfaces run through the elements Mm, Nn, or dx) will be 2a dv, to
which is to be added the live force of the volume element which is
absorbed by the water in the horizontal vessel, namely v dx, and the
sum 2a dv + v dx will be equal to the actual descent of the water multi-
plied by the mass of water, which product is equal to the actual descent
of the volnme element dx, multiplied by 2b - 2x. Therefore,
 MOTION OF FLUIDS THROUGH IRREGULAR VESSELS 181

2a dv + v dx = 2b dx - 2x dx. But this equation, integrated cor-

rectly, transforms into the following:
v = 4a + 2b - 2X - e-xt2 a(2b + 4a),
whence, if one sets 4a + 2b - 2x - e-xt2 a(2b+ 4a) = o, the value
of x itself will give the total displacement; if b is subtracted therefrom,
the residual will indicate the descent below the point C of equilibrium.
§29. But in order that by a certain example one may show how
much the oscillations are diminished by this reasoning, let us set
a = b, having made, of course, CA = GC and BE = o.
Thus arises

or x = 3a
2a ln - - - ,
3a - X

in which equation the value x = -¼a is almost fully satisfactory.

Therefore, the decrement of the displacement, or a - b, is equal to
the fourth part of the elevation of the fluid above the middle point; if
it is observed to be greater in the experiment, the balance will have to
be attributed to the adhesion of the water to the walls of the pipes.
§30. This reasoning of the diminished displacements clearly should
not be withdrawn, as I suspect, if the horizontal pipe becomes equal
in area to the vertical ones, on account of the changed direction of the
fluid at the points A and B.
Furthermore, infinitely many other cases could be invented to be
solved by these principles, just as the nature of the oscillations is to be
investigated in the vessel of Fig. 44 when in the diaphragmatic hori-
zontal section it is split into two parts communicating with one
another through the single opening which the diaphragm has, and
other cases of that sort. But I believe that this suffices already, so
that anybody can easily form for himself the general rules for solving
questions of this type.

EXPERIMENTS WHICH PERTAIN TO CHAPTER VIII

EXPERIMENT 1. The fourth paragraph, in which it is said that the

height for the velocity of the water flowing out through the orifice D
(Fig. 37) is mmx , I confirmed in this manner, that either of the
nn + mm
orifices G and D has an edge like a little belt, very slightly elevated, so
that there be no place for the contraction of the stream, and a safe
judgment could be made of the velocities from the quantity of water
 HYDRODYNAMICS, CHAPTER VIII

flowing out in a given time. Then, having taken the measurements

accurately and having observed the time in which the surface de-
scended through the given space AP, I saw that this time corresponds
correctly to the velocities defined in said paragraph; I also observed
that the motion is not at all changed by an elevation or depression of
the diaphragm. The remaining matters pertaining to the experi-
ment have slipped my mind, and I did not keep a record of them;
however, it seemed superfluous to me to repeat the experiment, since
it will be easy for anyone to imitate it; but it is the basis for the re-
maining matters, which therefore hardly need any further experi-
mental investigation; nevertheless, I wanted to try the following
things in addition.
EXPERIMENT 2. I used a vessel exactly of the kind which Mariotte
applied (see Fig. 38) and I confirmed our equation again in this
manner: I made the water flow out of the orifice D horizontally, and
then I took measurements of the height of the orifice D above the
floor and the distance of the spot where the stream struck the floor
from the point on the same floor vertically above which the orifice D
was located; from this I found the height due to the velocity of the
water flowing out at D; moreover, this very same height I found by a
related experiment, which [height] the theory of this chapter indicates
in §4. Similar experiments I may add at the end of the experiments
pertaining to Chapter XII, which at the same time will confirm our
hydraulico-static theory.
Finally, since there are many things in §§26 and 27 which would
have to be evaluated by individual calculation, it will be worthwhile
also to perform experiments concerning them, particularly since in the
same effort the other experiments which will be enumerated in Chap-
ter XII could be performed also, if, to this end, one would care to
make a vessel such as Fig. 43 shows.
Furthermore, this theory is also confirmed by the experiments
listed in Chapter VII which I performed concerning the oscillations
of fluids flowing into pipes through openings.
 NINTH CHAPTER
Concerning the Motion of Fluids that are Pushed
forth not by their own Weight but by an Outside
Force) and particularly concerning Hydraulic
Machines and their Ultimate Grade of Per-
fection that can be Attained) and how
this could be Perfected further
through the Mechanics of Solids
as well as of Fluids

§1. In this chapter, in which I have chosen especially to examine

hydraulic Machines and to perfect the use of them as much as this can
be done, let us disregard the variations of the motion which take their
origin from the force or inertia of the internal fluid because, as we
have seen, the motion of the internal water is as much not uniform
from practically the first instant of flow, if the orifice is small in pro-
portion to the internal areas, as is the case in most hydraulic Machines.
It would be ridiculous to be concerned in practical cases about the
changes which occur in the first instants of flow, and which we have
already determined in Chapter IV, since there it had been worth the
effort in order that the whole force of the theory might hence be
brought into the light. Therefore, for the sake of brevity, let us
assume that during the entire motion the water is constantly expelled
with a velocity that is proportional to the root of the internal pressing
force, after that force will have been reduced to the weight of an
aqueous cylinder lying over the opening; because, whatever that force
should be, there will have to be considered the weight of a vertical
aqueous cylinder lying over the internal water surface, and the
height of that cylinder will give the height due to the velocity of the
water springing forth, if only there are no extrinsic obstacles and
 HYDRODYNAMICS, CHAPTER IX

the water is emitted from a very wide vessel. This is to be understood

in such manner that, if the lid AB loaded with the weight P (Fig. 45)

!I .c -
.,ii_Y ')l
•1 t
' 'i
I' ·I·,

FIGURE 45

expels the water through the orifice F, and, moreover, if the weight P
is equal to the weight of the aqueous cylinder HABI, then the aqueous
jet FG ought to attain the height HI.

DEFINITIONS

§2. By moving potential then let me understand that acting principle

which consists of a weight, an activated pressure, or other so-called
dead forces of this kind.
Moreover, the product which arises from the multiplication of this
moving potential by its velocity and also by the time during which it
exerts its pressure I shall designate by absolute potential. Or, because
the product of velocity and time is simply proportional to the distance
covered, it will be permitted also to understand the absolute potential as
the moving potential multiplied by the distance which the same moves
through. But this very product I call absolute potential, because from
that finally is to be estimated work endured by day laborers elevating
water, which I shall soon show, proven in rules which were observed
by me in this matter. Meanwhile hydraulic Machines seemed to me
apt to be conveniently reduced to two types, of which the one emits
water with impetus, and the other transports it, so to speak, smoothly
from one place to another. I will treat an example of each in the
proper order, and, finally, before the end I may add something about
the diverse moving potentials.
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 85

FIRST PART: CONCERNING MACHINES EXPELLING

WATER UPWARD WITH IMPETUS

RULE I

§3. The work of day laborers which is applied to hydraulic Ma-

chines for elevating water is to be estimated from the absolute potential,
that is, from the moving potential or pressure which they exert, from the
time, and from the velocity of the point to which the moving potential
is applied.
PROOF. (ex) Concerning the moving potential the matter is clear:
namely, everything else being equal, the work is in any case propor-
tional to the number of laborers or to the moving potential. ({3) With
respect to time the matter is no less manifest from the reproduction
of all circumstances which arises from a duplication of the time.
(y) Finally, the matter that pertains to the velocity is to be deduced
from the fact that, whether one doubles the moving force or its
velocity, the effect is no different from twice [the effect] of either
part. Imagine that the weight P [Fig. 45] by its descent ejects water
through the orifice F to the height FG; then, the rest remaining the
same, imagine the orifice F to be doubled, and one sees that twice the
quantity of water will be ejected to the same height FG in the same
time from the same moving potential P, but with the latter descending
twice as fast. Equally, the quantity of water will be doubled, the rest
remaining the same, if one doubles both the orifice F and the area AB
and the weight or the moving potential P, but then the velocity of this
doubled potential remains unchanged. Therefore, in either way the
effect is doubled. Q.E.D.

ScHOLIUM

§4. The preceding proposition is not to be interpreted in a physio-

logical, but in a moral sense: morally I estimate neither more nor less
the work of a man who exerts at some velocity a double effort than that
of one who in the same effort doubles the velocity, because certainly
either one achieves the same effect, although it may happen that the
work of the one, despite being no less strong than the other, is very
much greater in a physiological sense. If someone advances in an
effort of 20 pounds a distance of 200 feet in the first minute, he will
easily _be able to double the effort, but with great difficulty double the
velocity. From this it follows that for every kind of machine it is to
186 HYDRODYNAMICS, CHAPTER IX

be considered particularly how it should be constituted in order that

for the minimum fatiguing of the men at the same time the product of
their effort by the velocity of all be a maximum: and hence it will be
evident what length should be prescribed to the levers in windlasses,
how large the radius should be made in wheels or rollers for tread-
mills, how great a length should be considered for oars, and so forth
regarding other machines.
Moreover, by the reasoning of the use of treadmill rollers, whi<:h
are very frequently applied in order that the moment become clearer
to us, let this experiment be considered.
Let us suppose in Fig. 46 a vertical height of many thousands of
feet, to which a man ought to ascend in a given time; further, let us

take a time of ten hours, because such is usually the limit of a day for
workers; finally, let us consider several paths, AC, AD, etc., inclined
differently to the horizontal BD; having supposed all this, we under-
stand that a walker must progress the faster, the less inclined a path he
will have chosen, so that he reaches the top of the mountain A in the
same time, and it is evident that there will be some path, as, for ex-
ample, AC, along which he travels the way with the least fatigue,
insofar as nobody can either proceed up a vertical plane or travel in a
given time an infinite distance; let us state that this path of least
fatigue makes an angle ACB or 30 degrees with the horizontal.
If this is so, the treadmill roller will have to be fabricated such that
the weight is raised with the desired velocity when the man in the
treadmill is constantly 30 degrees away from the lowermost point of
the roller.
According to the same principle a selection is to be made between
machines of a different type: thus, for example, if on a windlass the
operator exerts a potential or a horizontal pressure which has the
effect of the fourth part of his own weight, and by this effort travels a
distance of 200 feet in the first minute, he will, as I believe, hardly be
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 87

fatigued in the same manner as ifhe were treading on a rotating roller

with the same velocity at an angle of 30 degrees; meanwhile the man
in the treadmill will nevertheless in this manner carry double the
weight in the same time to the same height, because he exerts double
the pressure, other things being equal.

RULE 2

§5. With the same absolute potential existing, I say that all machines
which suffer no friction and generate no motions useless to the pro-
posed end maintain the same effect, and that one is therefore not to
be preferred to the other.
PROOF. From mechanics it is certain that any composite machine
can be reduced to a simple lever: therefore, it will be pleasing to
represent all hydraulic machinery by the simple pump supplied with a
lever (Fig. 4 7), where, for example, by aid of the lever MN movable
/~
_,11[ <,,__
....---------~'·);::::;
--::..:..-.=::.::-===~"'~

FIGURE 47

around the point M, a piston is pushed down, and thus water is ex-
pelled through the orifice F. But if the moving potential P applied to
the lever is understood [to beJ at N, we may see from the preceding
proposition that no benefit comes to the absolute potential from an in-
creased or diminished length of the lever MN; and certainly, whatever
this length may be, it can occur that the same moving potential, moved
at unchanged velocity, expels the same quantity of water with the
same impetus as long as the area AB of the pump has a constant ratio
to the length MN of the pole. From this it is very clear that all
188 HYDRODYNAMICS, CHAPTER IX

machines maintain the same effect under the same absolute potential as
long as one disregards friction and motions which are useless to the
destined end.

ScHOLIUM

§6. There are some who believe that a machine can be invented
by aid of which a maximum quantity of water can be elevated to any
height with a minimum of work, and they torture their minds with
investigating wheels, levers, and weights to be applied: but they
waste their effort, and proponents of this kind ought not to be heard,
since what do these great [men] seem to have found for themselves?
The best machine is, if we consider its effect alone, that which suffers
the least friction and creates no useless motions, the precepts concern-
ing the avoiding of either one of which we shall treat below.

RULE 3

§7. In pumps such as are represented in Figs. 45 and 47, in which

the internal surface AB of the water is at approximately the same
height as the orifice F, the absolute potentials for the same instants are
in a threefold ratio to the velocities of the water springing forth.
PROOF. The moving potentials are certainly in a twofold ratio to the
velocities at which the waters flow out through the orifice F, and the
velocities of the moving potentials follow the same ratio as the velocities
of the water springing forth; but for the same instants the absolute
potentials are as the moving forces multiplied by their velocities, hence
the proposition is evident.

ScHOLIUM

§8. It follows from this rule that, if it be our will to elevate water
through the orifice F to the height FG, a large part of the absolute
potential is wasted fruitlessly, since the water springs forth with a
greater impetus than corresponds to the height FG; for example,
arrange for water to be expelled at twice the velocity, and an eightfold
absolute potential is required, and, nevertheless, according to reason the
limit of the proposed effect is not to be considered [ to be] more than
double, because certainly at the same time twice the quantity of
water is elevated: and this effect could have been obtained with a
quarter of the absolute potential by expelling the water at the simple
velocity through double the orifice; therefore, on this account three
quarters of that potential must be said to have been wasted uselessly.
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 89

I have indicated the origin of this loss in §5, and it consists of the
motion which is generated that is useless to the proposed end: namely,
the entire motion which remains in the water after it has attained the
height G is to be called superfluous in our case.

RuLE4

§9. When water is expelled through the conduit DF (Fig. 48) and
has at the orifice Fa velocity which is due to the vertical height CF,

G-

FIGURE 48

the absolute potential applied at the same time is proportional to the

velocity of the water atFmultiplied by the height G above AB.
PROOF. The moving potential P is certainly proportional to the
indicated height, and the velocity of that potential is proportional to
the velocity of the water at F.

ScHOLIUM

§10. The absolute potentials increase at a greater rate than the

velocities of the outflowing water, that is, than the quantities ejected in
 HYDRODYNAMICS, CHAPTER IX

the same time; but, nevertheless, the difference in ratios is almost un-
noticeable, since the height FG is very small in proportion to the
height FD of the conduit. For example, let FG be equal to ¼FD
(neglecting the height ED ) ; then let the water be emitted at twice
the velocity, such that now FD = FG; thus the absolute potentials
will be as I x ¾to 2 x 2 or as 5 to 16, so that a more-than-threefold
absolute potential is required for emitting twice the quantity of water.
But if the former FG is set equal to 1 hFD, and then again the water is
assumed to be expelled at twice the velocity, [then] the absolute poten-
tials will be now as I x 101 to 2 x 104 or as 101 to 208, which ratio
is just less than one half. It follows thence that the less the speed at
which the water is discharged, the greater the success with which I
have applied the absolute potential; and then finally I have applied
approximately all of it usefully when the water flows out through the
orifice F at almost unnoticeable velocity; furthermore, the size of the
orifice could compensate for the scantiness of velocity, so that in a
given time a noticeable quantity of water can be discharged. Let the
loss of absolute potential be so defined.

RULE 5

§11. Let the pump ABDF [Fig. 48], furnished with a little valve at
the base and put into water, transfer water from a lower region AD
to a higher region F, and let the median velocity of the water flowing
out at F be due to the height FG; then the loss of absolute potential will
be to that entire potential as FG is to the height G above AB.
PROOF. Let us imagine that the orifice Fis enlarged very much,
with the velocity of the water flowing out through F decreased in the
same ratio; thus the quantity of water flowing out in a given time will
not be changed if the velocity of the moving potential is the same, and
thence the effect will be the same. But if the velocity is so diminished
that the height due to it is unnoticeable, the moving potential may be
expressed by the height F above AB, since previously the moving
potential was equal to the height of G above AB; and since in either
case the velocity of the moving potentials is the same, the absolute potentials
for the same times will be as the height G is to the height F above the
common [base] AB. Therefore, the difference of the heights G and F
will express the loss, since the entire height G above AB represents the
total absolute potential.
§12. The same reasoning is valid for every kind of machinery:
Indeed, whenever water, having been conveyed to the location to
which it is to be elevated, has a noticeable velocity, the loss of absolute
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 9I

potential becomes great; for if one sets the height of the elevation equal
to A, the height due to the velocity of the water at the place at which
it is emitted equal to B, and the entire absolute potential equal to
P, the quantity A ! B · P will be lost.
It can also be noted that when water has to be conveyed over some
height the culmination of which is k _:ated at F by means of a pump
attached to a pipe, the pipe DF is to be continued downward as much
as it may please and is not to be discontinued at F, just as it appears in
Fig. 49. Because if, let us say, the point Fis located twice as high as

FIGURE 49

the extremity G of the pipe, twice as large an absolute potential is re-

quired for transferring water through the conduit discontinued at F
than through that continued to G, even if in either case it flows out
at very low velocity, [and] its generating height is indeed small in
proportion to the heights FD or GD.

RULE 6

§13. When in pumps which we have considered so far the covers

AB, or rather the pistons, do not correspond well to the sides of the
 HYDRODYNAMICS, CHAPTER IX

machines, an opening is left, and from this arises another kind of

loss in absolute potentials, which is determined thus in pumps in
which the height of the orifice above the piston can be neglected. As
the sum of the orifice of effiux and the aforementioned opening is to
this very opening, so the absolute pote.,tial which is exerted is to that
part of it which is useless, or to its loss.
PROOF. The water is indeed pressed equally through the orifice and
the opening, and it flows at an equal velocity; but the entire absolute
potential that forces the water through the opening is lost, and this is to
the complete absolute potential as the opening is to the sum of the
orifice and the opening.

ScHOLIUM

§14. It certainly suits the piston to be well formed and smooth; it is

also necessary that the cavity of the pump be exactly cylindrical and
its sides be very smooth as well. But I should hardly believe, unless
it is done for another purpose, that it is of importance that the pistons
fill the cavities with ultimate accuracy, because thus perhaps a greater
loss of forces arises from friction than if a more or less very small
opening would have been left. For if that opening amounts to, let
us say, the hundredth part of the orifice of effiux, there will hardly be
any place for friction, and thence only approximately the hundredth
part of the absolute potential is lost, and perhaps a larger loss arises
from the friction of a piston occupying the cavity of the pump exactly.
Therefore, it is not in this respect that we only too carefully avoid the
transit of water through an opening left by the piston. But this
consideration does not refer to those machines in which the water is
to be drawn into the pump by the retraction of the piston. For here
the correct and full size of the piston is entirely necessary.

RULE 7
§15. In machines which have several orifices transm1ttmg water
from one cavity to another, something of the absolute potential is lost,
the reason for which we said in the preceding chapter is that the
potential ascent of the individual volume elements flowing from one
cavity into another through a common orifice vanishes.
· The more and the smaller the orifices of this type are, the greater a
loss of absolute potential arises, which usually is of great importance, and
this perhaps apart from the common opinion, in the machines which
Vitruvius names after their inventor, Ctesibius. Indeed, I speak of
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 93

orifices located such that all the water that will flow out must go
through them. That type of loss may now be determined by the
following calculation.
Let the area of the last orifice emitting water into the air be n, but
the areas of the remaining orifices through which the water is driven
inside the machine be designated by the .ctters a, fi, y, etc., and when
the same moving potential has been assumed in either case, the height
due to the velocity of the outflowing water will be to the similar height
· h no mterna
wit · 1 on·fices o b structmg . to r + -nn + RR
. as r 1s nn + -nn +
aa ,-,,-, yy
etc. (by §r 1, Chapter VIII), and hence it follows that with these
heights having been made equal to one another, the moving potential
w1·11 b e as 1
nn + RR
+ -aa nn + -nn + etc. 1s
. to r, an d b ecause m
. e1t
. h er case
,-,,-, yy
the velocities of the moving potentials are the same, the absolute
potentials will also have an equal ratio for these instants. Therefore
.
t h e portion -nn + RR
nn + -nn + etc. 1s
. superfl uous, w h ence t he l oss o f
aa ,-,,-, yy

absolute potential will be to that entire potential as :: + ;; + ; +

etc. is to r + -nn + -nn + -nn + etc.
aa fifi yy

ScHOLIUM

§16. Whenever the idea of a machine requires orifices through

which water flows from one small container into another (which
happens in every kind of pump, such as aspirating ones, aspirantes in
French, or pressing ones, foulantes, etc.) those orifices are to be made
very large, as much as the remaining circumstances permit, so that the
area of the orifice of effiux is very small with respect to those internal
orifices. But, in order that the use of the rule be more clearly evi-
dent, we will consider examples of other, no less useful machines.
ExAMPLE r. Let a machine be proposed (which Fig. 50 represents)
in which the pistons C and F are alternately depressed, and by which
water is introduced into the small container BEH through the passage
AB and DE, in order that a continuous jet may thus discharge through
the orifice H. Since here the pistons act alternately, we will consider
one or the other alone, so to speak, but acting continuously; and so
one must consider the orifice of effiux H, of area n, and one or the
other of the orifices o and p, each one of which let have the area a;
1 94 HYDRODYNAMICS, CHAPTER IX

so the loss of absolute potential will be equal to nn, the whole poten-
cxcx

tial having been set equal to 1 +

nn, which quantities are as nn is to
aa
nn + aa. Certainly that loss is considerable, if one may trust the
representations of those machines in which often the orifices o and p

FIGURE 50

are smaller than the orifice of effiux H, because if this were so, more
than half of the absolute potential would be lost. The conduits AB and
DE will also have to be enlarged throughout their entire extents, as
much as this is permissible, in order that the machine may lose little
of its excellence.
As for the rest, this machine was thought up in order that a con-
tinuous jet emerge through H. Nevertheless, because it cannot
happen but that some interval of time occurs between the last point
of the elevation of the piston and the beginning of the instant of its
depression, it will not be possible for the jet to be completely continu-
ous and steady. However, the inventor of that machine presents an
optimal remedy for this inconvenience, which Mr. Perrault mentions
in Commentarii ad Vitruvium, p. 318, edition 2, Paris, which he says is
kept in the Royal Library in Paris; this machine will serve us as
another example: also, let me take the figure together with its descrip-
tion from Perrault himself.
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 95

EXAMPLE 2. " There is a machine," according to the aforemen-

tioned Perrault, "in which water is expelled from the small container
A (Fig. 51 ) by means of the piston B into the jar FG, out of which the

A I

I
i
Ir··

FIGURE 51

air cannot discharge as long as there is already some water present,

because the pipe EF descends almost down to the bottom: indeed, it
happens thus that the water, propelled from the small container A
through the passage D and occupying the lowermost portion of the
jar, closes the orifice of the pipe at F and prevents transit of the air.
Therefore, when the piston brings new water into the small container,
filled partly with air, partly with water, this newly supplied water
exerts a force on either fluid, and since the water cannot spring forth
through the pipe FE at the same velocity at which it flows in from the
pump through the passage D, because naturally (these are Perrault's
words) the pipe FE is perforated at its extremity Eby an orifice much
smaller than the orifice of the pipe D , the water accumulated in the
vessel compresses the air and, pressed reciprocally by the latter,
springs forth through the pipe FE even while the piston is raised."
In this machine a large part of the absolute potential is lost by the
transition of the water through the passage D, and that loss will be
greater, the narrower that little pipe is; therefore, it should be made
wide, or even several pipes transmitting water may be constructed;
this annotation is of greater importance in the present case, since a
196 HYDRODYNAMICS, CHAPTER IX

much greater loss arises from a narrow passage D than in other

machines; indeed, make the area of this passage the same as the ori-
fice E, and assume furthermore that the piston is depressed and
retracted in equal time intervals: now no only is one-half of the
absolute potential lost as previously, but clearly four-fifths will become
useless. But since there are many [things] in this machine which
postulate individual calculation, it is suitable to illustrate that one
separately.

A DIGRESSION CONTAINING SOME COMMENTS ON THE

HYDRAULIC MACHINE WHICH FIG. 5r REPRESENTS

(a) The aqueous jet through E cannot be completely steady during

the entire agitation of the piston. Indeed, while the piston is ele-
vated, no new water flows in, and thus the quantity of water contained
in the vessel GE is diminished: hence the water discharges also at a
continuously smaller velocity until it is accelerated again by the
intruded piston.
But if the space which the air occupies in the vessel is set [to be]
much larger than that space occupied by the water which is ejected
during a single elevation of the piston, this entire inequality almost
ceases, it having been assumed that the piston is agitated uniformly
and has been agitated for a long time previously, which latter hypo-
thesis is necessary insofar as the first ones differ very much in agitation
from the following. Therefore, for the sake of brevity let us satisfy
all these hypotheses, that is, let us assume everywhere what is called
the state ef permanence.
(fJ) Therefore, since the velocity of the water flowing out through
E is increased noticeably by the first agitations of the piston, it hap-
pens soon that the aqueous jet attains almost the entire velocity; with
this state of the matter having been assumed, it is evident that during
the depression of the piston as much water is pushed into the vessel as
is ejected out of the same during the total agitation of the piston.
During the first agitations, however, more is pushed in than is
ejected, and this not for the reason, as Mr. Perrault believed, that the
orifice at Eis less than the other one at G (for the same would happen
if it were larger), but that the generating cause does not immediately
exert its total effect in ejecting water.
(y) It would seem perhaps that it will not be sufficient for investi-
gators that, with everything assumed in the permanent state already
and no outside obstacles being present, the water springs forth from
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 97

the orifice E at a velocity which can ascend to the height of an

aqueous column assumed in equilibrium with the pressure of the
piston; and it would be reasonably so if the pressure of the piston were
present without interruption, and if no potential ascent were lost in the
w~er; but because in both [cases] the situation is different, it is not
possible that in the aqueous jet no other estimation of the velocity
arises; hence everyone sees not obscurely that attention has to be paid
to the consideration of the time in which the piston is depressed and
retracted, [and] then also to the consideration of the areas in the
small conduit D and of the orifice E.
( S) Let us therefore set the time in which the piston is depressed
equal to 0; the time of one entire agitation equal to t, the area of the
orifice E equal to fl,; and [that of] the passage D equal to m; then,
after the force pushing down the piston has been compared with the
aqueous column lying over it, let us make the height of this column
equal to a, but the height due to the velocity of the water springing
forth equal to x. After these [things] have been prepared thus for the
calculation, it will be permitted to investigate in two ways the ratio
which will prevail between the velocities of the water at the orifice E
and at the passage D, and from here to elicit the value of the unknown
x. First, namely, it is evident that in the time 0 (in which certainly
the piston is pushed down) as much water flows through the passage
Das flows out through E in the time t (in which the piston is depressed
and retracted). The velocity at D is therefore to the velocity at
E as __!_ to _:_; and since this latter velocity is equal to Vx, the other
m0 f.d

one will be equal to ~~ -v'x. Second, because the velocity of the out-
flowing water is due to the pressure of the air in the jar, it follows that
this pressure is equivalent to the weight of an aqueous column of
height x; but if one subtracts the pressure of the air from the pressure
of the piston, one will have the pressure which generates the velocity
of the water at D; hence, because the difference of pressures is ex-
pressed by a - x, the velocity of the water at D will be represented by
-v'a - x; therefore, the velocity of the water at D is now to the velocity
of the water at the orifice E as -v'a - X to -v'x. After combining the
ratios found by either method,
. ;- - .r I I mm00
va-x:vx=-:- or X - -----·a
m0 fl,t - mm00 + f1,f1,U •
It is evident from this equation that the height of the jet is deficient
for a double reason from the height a of the pressing column; indeed, it
198 HYDRODYNAMICS, CHAPTER IX

is deficient by more when the piston is depressed faster or elevated

more slowly, and then also when the orifice E increases in proportion
to the area of the small conduit D. For example, let the area of this
orifice be equal to the area of the small pipe D, and let the piston be
depre" ed and elevated at an equal velocity, and there will result
x = ½a, such that the outflowing stream rises only to the fifth part of
the height a.
(E) The loss of absolute potential is now evaluated in the following
manner after it has been assumed beforehand that no work is done in
elevating the piston. Let the velocity at which the piston is de-
pressed be v, and the absolute potential expended in the time of one en-
tire agitation will be av0 (by §3), but because the effect consists in the
fact that effiux occurs through E during the time t and the water
00
itself is elevated to the height ;m a, the simple pump of
mm + µµtt
Fig. 45 could have managed this, if in the latter as the pressing potential
00
an aqueous cylinder of height ;m a had been taken, and
mm + µµtt
this potential had acted during the time t at the velocity ~ v;
whence
t
the required absolute potential in this simple machine in which nothing
of the former is lost would have been

mm00 0 mm00
-----·a·-v·t = ------av0.
mm00 + µµtt t mm00 + µµtt

The total absolute potential is therefore to the uselessly wasted part of

· as av 0 1s
1t · to av 0 - mm 00 · av 0, or as mm 00 + µµtt 1s
· to µµtt.
mm 00 - µµtt
Therefore, if the entire absolute potential is designated by P, the loss of
. Wl·11
lt be µµtt P.
mm 00 + µµtt
Therefore, it is necessary in this rather than in other pumps that the
passage at least exceed the orifice E in area, or that it be many times
as great. Indeed, if there is a single one, and this is equal in area to
the orifice E, and at the same time the piston is assumed to be agi-
tated upward and downward at uniform velocity, a loss of four-fifths
of the total will arise; and if it were made twice as large, then still
balf of the absolute potential would be lost.
(l) Finally it is clear that the sides of the jar GE sustain a lesser
pressure than [those of] the small container AA; indeed, these pres-
sures are as x is to a, that is, as mm00 + µµtt is to mm00, from which
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE I 99

ratio engineers will judge the strength of the sides which is required
for either one.

- - - - - - - - - [End of7Jigression] - - - - - - - - - -

RULE 8

§17. When the piston in pumps is extracted and the water flows
into the small container, not only excited by its own weight but for
the most part drawn by the piston, then all the absolute potential ex-
pended in this attraction comes into the problem in addition, because
a pump placed under water, as it happens, would be filled on its own
if sufficient time for filling were allowed; thus that attraction does not
especially pertain to the ejecting of water with a certain velocity, so
the entire [attraction] could be avoided, and on this account the work
expended in that [attraction] is called useless by me.
But as the inflow occurs partly by the water's own weight, partly also
by the lifting of the piston, the loss of absolute potential cannot be esti-
mated from the effect; indeed, the calculation is to be set forth rather
so that, after the force elevating the piston to a certain position has
been set equal to TT, the velocity of the piston equal to v, and the small
time interval corresponding to the quantities TT and v equal to dt, the
entire absolute potential expended in elevating the piston is called
f f
TTV dt or TT dx, if by dx is understood an element of space traversed in

the small time interval dt. It follows hence that if the effort by which
the piston is raised is of constant magnitude, as it is almost, the
absolute potential will be equal to the moving potential multiplied by the
traversed space; but since a similar consideration is valid also for
the depression of the piston, and also the piston is as much raised as it is
depressed, it is apparent that the absolute potentials which are exerted in
alternately attracting and expelling the water are approximately in
proportion to the moving potentials in either case; whence a loss arises
which is equal to _TT_ P, after one has set, of course, the elevating
TT+ p
potential equal to TT, the depressing potential equal to p, and the
absolute potential exerted in the elevation and depression of the piston
equal to P.
The loss of absolute potential can be estimated approximately in a
different way from the fact that the whole potential ascent of the water
flowing into the pump must be thought of as generated uselessly.
But if the piston is moved upward and downward in the same time
intervals, or at the same velocity, the velocity at which the water is
200 HYDRODYNAMICS, CHAPTER IX

taken in will be to the velocity at which it is expelled reciprocally as

the corresponding orifices, and the potential ascents themselves in either
case will be in the inverse-square ratio of the corresponding orifices;
if, further, the elevation and depression of the piston occur in different
time intervals, the velocities are reciprocally as the time intervals, and
the potential ascents reciprocally as the squares of the time intervals.
Therefore, the potential ascent generated by the inflow of water is to the
potential ascent which arises from the efflux, and which alone is inten-
ded, in an inverse-square ratio composed of the ratio of the orifice of
inflow to the orifice of efflux and of the time in which the water is
drawn in to the time in which it is expelled.

ScHOLIUM

§18. From either means of estimating, it follows that the piston is

to be raised slowly; for thus the moving potential becomes small, by
reasoning of the first method, or the time of elevation becomes large,
by reasoning of the second, and thus the laborers may recover from
the exhausting effort of the preceding depression during the indivi-
dual intervals of the elevation of the piston. The latter method
indicates further that the orifices through which the water is drawn
are to be made larger and more numerous; but this is also in accord-
ance with the former method because thus an almost sufficient quantity
of water flows in on its own, and so less moving potential is needed.

RULE 9

§19. Finally, it is to be observed that the aqueous stream, nsmg

vertically, never attains that height which would be due to the
initial velocity of the water; that is, if the stream of fluid would start to
rise vertically from its origin with as great a velocity as a weight
falling freely from the height a would acquire, the fluid could not
ascend to the total height a, even if one were to remove the resistance
of the air or whatever one may think might retard the motion in this
case. Indeed, the very nature of the matter inevitably demands
some defect, the physical reason of which is this: certainly any
volume element whatsoever, even though beginning a vertical ascent,
can nevertheless not help but be deflected noticeably to the sides, and
finally, when it reaches the summit, it is carried by a horizontal
motion, which must be noticeable, because through the uppermost
limb or section of the aqueous stream all the water passes, which has
fl.owed out through the orifice; assume, therefore, that the velocity of
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 201

any volume element a the instant of time at which it is moved

horizontally is that which a weight acquires by free fall through the
height b; thus one sees that the stream cannot ascend beyond the
height a - b. And for this reason a loss arises in proportion to
the total absolute potential as b is to a.

ScHOLIUM

§20. It has been observed that among quantities of water ejected

at a common velocity from differently formed small pipes, some rise
higher than others; therefore, attention is to be paid here to the most
apt configuration of the final pipes emitting water (des ajutages).
Mr. Mariotte set up experiments on this matter in his Traite du
mouvement des eaux.

GENERAL ScHOLIUM

§21. So far we have examined the hindrances which appear in the

case of hydraulic machines ejecting water with impetus; I consider
those which I exposed to be the outstanding ones; nevertheless, still
others could be considered, but, as I believe, only of less importance.
Almost everywhere we gave completely geometric measures and have
indicated simultaneously the extent to which these hindrances could
be counteracted for the most part. He who reaches for greater ones,
believing that the effect expected in elevating water by the least work,
or (which I have shown in §3 to come back to the same thing) by the
smallest absolute force, can be exceeded, is tricked by his opinion and
wastes [lamp] oil and effort. For if one disregards the indicated
hindrances and other similar ones that might perhaps be considered,
in the nature of things the most perfect machine will be the simple
pump of Fig. 45, and if water projected upwards by means of it is
collected at G, [then] I say that it could not have happened that the
same amount of water was elevated to the same height FG with less
work.
There is, furthermore, another kind of machine which differs from
the machinery treated so far in that the latter ejects the water with
impetus, while the former transfers it quietly without noticeable
motion. But also in the latter the ultimate degree of perfection which
can be reached comes back to the same thing. But most [of them]
are subject to many hindrances of very great importance. There-
fore, these will have to be treated by us directly.
202 HYDRODYNAMICS, CHAPTER IX

SECOND PART: Ccr.i"CERNING HYDRAULIC

MACHINES TRANSPORTING WATER WITHOUT
NOTICEABLE IMPETUS FROM A LOWER POSITION
TO A HIGHER

RULE IO

§22. If a given weight is elevated through a given vertical height a

by a moving potential [that is] arbitrarily variable but applied directly,
and if the body retains no motion at the summit of the proposed
height, the absolute potential expended in the lifting of the weight will
always be equal to the product of the weight of the elevated body and
the height a of elevation.
PROOF. Indeed if a weight, which I shall call A, ascends through
the height y and is assumed to be moved with the velocity v, and
animated by a variable moving potential P directly applied in this
position, the small time interval in which the weight is elevated
through the element dy will be dy' which multiplied by the moving
V
potential P and its velocity v, gives the element of the absolute potential
(by the definition of §2) equal to P dy, therefore fP dy will give the
total absolute potential, if after the integration one setsy = a; but during
the entire motion the increment of velocity dv is equal to the exciting
or moving potential, which here is p ~ A multiplied by the small

time interval which is now d:; therefore, we have dv = ( p ~ A) d:

or Av dv = P dy - Ady, that is, ½Avv = f P dy - Ay, or JP dy =
½Avv + Ay, where one is to set y = a and v = o (by hypothesis) so
f
that P dy = Aa.
f
Furthermore, because P dy expresses, as we have seen, the entire
absolute potential expended in elevating the weight, this very potential
will constantly be the same and expressly equal to the product of the
weight A and the height a, just as the proposition states. Q.E.D.
§23. COROLLARY. From the proof it appears to us also that the
absolute potential is the same whenever the velocity at the summit is the
same, that is, whenever the height to which a body can ascend at its
residual velocity, namely ½vv, is constant; and if this height is called
b, the absolute potential will be equal to A (a + b). Therefore, it is now
evident how large a portion of the absolute potential is lost when one
intends to elevate the weight A to the height a, and when the same has
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 203

at the summit a residual velocity due to the height b; certainly the

loss of force will be to the entire force as b is to b + a.

ScHOLIUM r

§24. And so precautionshave to be taken lest the machines are con-

structed such that the water is transported to the determined location
with a violent motion. Usually, however, this kind of loss is small in
most machines.

ScHOLIUM 2

§25. Everything behaves similarly if the body is not elevated verti-

cally but along a plane however inclined or even curved in any
manner whatever; indeed, the total absolute potential will always be
equal to A (a + b), that is, to the product of the weight by the height
of elevation augmented by the height due to the residual velocity of
the body at the summit, the proof of which matter I omit since it
differs little from the preceding proof.

GENERAL ScHOLIUM

§26. Because the effects of all machines, however composite, can

be reduced to the nature of the inclined plane, it is evident that if we
disregard frictions and these losses of absolute potentials which we have
dealt with so far, all machines come back to the same, because the
absolute potential simply depends on the height to which a body is to
be elevated and its weight. The absolute potential has this in common
with the live force or with the actual ascent or descent. And this is the
ultimate level of perfection of machines, which cannot be exceeded,
on the contrary, not even be reached, for a larger weight could always
be elevated to the same height by the same absolute potential when the
frictions and losses have been removed. In order that some com-
parison can be made of the loss in those machines which, so to speak,
project water to a desired height as well as in those which transport it,
we will now indicate also the most greatly noticeable losses in the
latter ones.
I. In most machines of this kind friction is of so great a hindrance
that it alone absorbs the largest part of the potential, particularly
when square blades or oval bowls connected to a chain moving
around in a circle elevate the water passing through the conduit to
which they are fixed.
 HYDRODYNAMICS, CHAPTER IX

II. Most machines, but particularly those which we have just de-
scribed, usually designated by the name of water wheels, are so joined
together that continually, while the water is elevated, part of it
trickles d~wn, or plainly runs back to the place from which it was
drawn or at least from a higher position to a lower, just as in water
wheels; ifin these the bowls or blades are well adapted to the conduit,
the friction becomes almost unsurmountable, but if they are less
[well adapted], a very great amount of water drips through the open-
ings that are left, from the higher portions into the lower, so that a
very small part of that quantity of water which they received in their
entire traverse is left when they reach the culmination point. There-
fore, or for this reason only, it seems that these machines are to be
strongly condemned, and particularly if clear water is to be elevated
which could be drawn by pumps.
III. Machines are also customarily of such a nature that they lift
up the water beyond the proposed height; but the potential which
corresponds to the excess is wasted, and if the water is to be raised
through labor, [then] that which I indicated in §12 is obtained with
difficulty.
IV. There are also machines which do not allow direct application
of a moving potential, from which drawback again some loss arises.
§27. These are the obstacles, more or less, which seemed to me of
notable importance; I do not know, however, whether those can be
counteracted so much as we have shown regarding the first kind of
machines; the mechanics know certain tricks of diminishing frictions;
I would prefer, to water wheels, machines which draw and lift the
water in buckets; but the buckets are to be constructed so that, if this
can only be done, they are filled immediately in the lowermost
position and emit nothing before they have reached the uppermost
pos1t10n. Since the water is to be transferred through the higher
location to another one, less high, an effort has to be made that the
impetus of the falling water promotes the motion of the roller or wheel
acting in a circular course, although thus the entire absolute potential
is far from being expended usefully, as we have shown to happen in
the pump of Fig. 49 (§r 2). The principle of action exists, ifl judge
correctly, most aptly in the treadmill: for these men are best accus-
tomed to work. That which I advised in §4 on occasion of the first
rule about the angle of inclination according to which a walker can
attain a certain vertical height in a given time with the least fatigue
pertains here. I would believe that a man of ordinary stature,
healthy and robust, marching on a path inclined at 30 degrees will
accomplish 3600 feet in a single hour without difficulty, and therefore
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 205

he will elevate to a vertical height of 1800 feet the weight of his body,
which I may assume [ to be] 1 4 pounds or two cubic feet of water.
Such a man, therefore, could by means of a treadmill machine,
acting in a circle and being most perfect (in which of course nothing
of the absolute potential is wasted) elevate in a single hour two cubic
feet of water to a vertical height of 1,800 feet, or, which is the same, in
a single second one cubic foot to the height of one foot; machines
which are of much lesser effect, doing a favor to the workers, I believe
to have little to recommend them; meanwhile, having set up an experi-
ment with a pump in the house of the illustrious General de Coulon,
which I shall account for at the end of the chapter, I experienced an
effect by no means less, by which I am confirmed in my statement that
workers usually accomplish more with a treadmill: I easily foresee,
moreover, that in very composite machines a far lesser effect is
achieved, because in these the greatest part of the absolute potential is
expended uselessly. Notably, I shall now contribute to this matter
the example of the very well-known machine de Marl;Y showing what an
almost incredible loss of absolute potential arises from all the collected
hindrances.
Weidler published a treatise about hydraulic machines in which he
gives a full description of the machine de Marl;Y, and reports that all the
water is elevated by the motion of 14 wheels, the blades of which are
propelled by the impetus of the Seine; this makes the impetus for all
wheels equal to a weight of 1,000,594 pounds, and this is what we
have designated by the name of moving potential. Furthermore, I
could understand from some circumstances that the blades are
carried by a motion by which they travel 3¾ feet in a single second,
and this velocity is to be taken for the velocity of the moving potential.
Then he adds that in a single day 11, 700,000 pounds of water are
elevated by means of this machine to a height of 500 feet. These
things having been so assumed, let us see now how great a potential
P, similarly moved at a velocity of 3¾, would be required for this
effect in the very simple machine of Fig. 45, in which it is assumed
that none of the absolute potential is lost. Indeed, the height FG will
be 500 feet, and since now in the time of 24 hours 11, 700,000 pounds
must be ejected through the opening F, that is, 162,500 cubic feet, the
size of this opening will have to be taken as 0.0108 part of one square
foot. The velocity of the water is so great that it travels 173 feet in
a single second. Therefore it contains the velocity 3¾, which the
weight P is assumed to have, 46 times, and the area AB of the pump
has to exceed the area of the opening F just as many times. Conse-
quently, the area AB will have to be taken as 0.4968 part of a square
206 HYDRODYNAMICS, CHAPTER IX

foot, from which follows that the wei~ht P will be equal to the weight
of an aqueous cylinder constructed over the base AB to a height of
500 feet, or to the weight of 248.4 cubic feet of water, that is, to a
weight of 17,885 pounds, which brings about only the fifty-sixth part
of the moving potential which Weidler shows to be applied to the
[water] moved at the same velocity. Thus, therefore, a loss occurs
in the entire machine which equals U of the entire absolute potential.
After we have so examined the nature of hydraulic machines, as
much as it can be done in general, by no means will it be irrelevant to
treat some special example more accurately, and because the water-
screw of Archimedes possesses many outstanding properties which, as
far as I know, no one has exposed sufficiently, I want to take the ex-
ample from it, and this all the more willingly since there are many who
believe contrary to our rules that this waterscrew has a singular virtue
for elevating a large quantity of water in a short time and by a small
force; but those who think so are deceived; for if no account of acci-
dental hindrances is taken, this vouches for the same absolute potential
as all other machines.

SPECIAL COMMENTS ON THE WATERSCREW OF ARCHIMEDES

I. There are various authors who taught a method of constructing

this waterscrew: the summary comes back to the fact that some con-
duit or several cylindrical surfaces are bent around, and certainly
so that the conduit has everywhere the same inclination in relation
to the axis of the cylinder, which Vitruvius, beyond necessity, orders
to be made at a half right angle [45°] in all waterscrews. Therefore,
it is required first of all that on the surface of the cylinder a spiral
line be drawn, to the normal of which the conduit is to be put, which
can be done most easily, in my judgment, on a very smooth surface
(particularly since the helices have to be no little distance from one
another) by winding a string around the same several times. For
here the tension will produce the desired line on its own, and the
spiral indeed cannot be everywhere similar to itself or have a constant
inclination to the axis of the cylinder unless the arc spanned between
two points is the smallest of all arcs having the terminal points, which
is shown to be the case with a stretched string; but if friction is a
hindrance, the string need only be extended to smaller intervals.
But this is not why we are hesitant in a matter that is intrinsically
very simple in many respects.
A primary law of the spiral is that it is everywhere equally inclined
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 207

to the axi of the cylinder, following which law the construction is

undertaken, which I shall add for the benefit of the things to be said
below.
Imagine a right cylinder MaJN (Fig. 52-1), on the surface of which
is to be drawn the desired spiral arb2c3d, etc., and consider the same
surface to be laid out in the plane given by the shape of the rectang-
ular parallelogram AaJF (Fig. 52-2), [and] let here be taken from the
one part AB, BC, CD, DE, and EF, from the other ab, be, cd, de, and ef,
each one to be equal to the corresponding one; the points B, C, D, E,
and Fare joined by straight lines with the points a, b, c, d, and e: if,
after these things have been done in this way, the plane surface is
again rolled up into a cylindrical one, with the lines AF and af
joined, and the points A and a, B and b, etc., coinciding, it will
happen that the lines aB, bC, cD, etc., form a continuous line on the
cylindrical surface which will be the desired spiral itself. For easier
understanding I marked homologous points in either figure with
common letters.
II. The cylinder MaJN (Fig. 52-1 ) was already proposed, having
as a duct the curved conduit of the spiral just described, the diameter
of which we shall assume as infinitely small in proportion to the
diameter pertaining to the cylinder; and thus the waterscrew of
Archimedes will be obtained; if we want to use this for elevating water
from M to N, the cylinder will have to be inclined with respect to the
horizontal, and certainly so that the angle aMH (the intercept be-
tween diameter Ma of the base, which is in the vertical plane, and the
horizontal MH) is greater than the angle sao, which the tangents
to the circle and the spiral form at the common point a. Then, after
the cylinder has been turned around its axis in the direction aghMs,
the water will flow in through the lower orifice of the bent conduit
and flow out through the upper one.
III. So that we may understand the nature of this elevation cor-
rectly, three points in any arbitrary helix of the spiral are to be
examined by us, namely the points o, p, and q, the first of which, o,
is the farthest away from the horizontal, the other, p, is closest to it,
and q is located at the same height as the point o taken in the next
lower helix; through the individual points o is drawn the straight line
gn, through the points p the straight line hm, and through the points q
the straight line st. But the locations of these lines will be determined
below.
IV. Let the radius that pertains to the base of the cylinder be r and
let it be taken as the total sine; the sine of the angle sao equals m; its
cosine equals M; the sine of the angle aMH equals n, its cosine equals
 l'J
0
, N m n .f co

11 - .e

:i::
~
~
0
ti
~
z
---rt----J-~ I >
s::
( ....
I 0
~c:n

C 0
:i::
>
...,"d
-?I t'1
~
....
~
pl ' }
l
.1
~~
-d v· .H I I
~ J .M
?[::::hJa
(I) FIGURE 52 (2)
 MOTION OF FLUIDS PUSHED BY A_~ OUTSIDE FORCE 209

N; the arc ag equals X, the cosine of that arc equals x, [and] the per-
pendicular drawn from o down to the horizontal will certainly be
mNX
or= -x:T + n(I + x). But because or is a maximum, it happens

that mx:X + n dx = o, and since from the nature of a circle there is

dX- -dx -mNdx

there will be - , = = = + n dx = o, therefore
- VI - xx' mVI - xx
mN
VI - xx=-· Mn
Therefore, the sine of the required arc ag is ,:.,

.
or cosme x = ±
V nnMn
- mm h . . h h
; t e upper sign gives t e arc ag, t e
lower the arc ah determining the lowest points p.
And so we have determined both the uppermost points o and the
lowermost p, and it is evident that the arcs Mh and ag are equal to one
another, but simultaneously it is understood from the irrational
quantity V nn - mm determining the value of the letter x that it
cannot occur that m is larger than n; and, indeed, in this case the
lowermost point is not given, since the entire spiral ascends continu-
ously everywhere. Indeed, the waterscrew will not serve thus in
elevating water, hence the reasoning is now evident which I pointed
out in the second article of this digression concerning the required
excess of the angle aMH over the angle sao.
V. Let us suppose now that a sphere is located somewhere in the
cavity of the conduit and that the waterscrew is fixed in its position;
thus the sphere is certainly not at rest unless it is located at some
point p. But if the waterscrew is assumed not to be held back, the
sphere will descend, and by its descent it will drive the waterscrew
around, ancl if, furthermore, it is imagined that the waterscrew is of
no weight and that the motion of the sphere occurs very freely with
no hindering friction, [then] the sphere descends on the straight line
mh by no other law than a sphere descending freely on an inclined
plane. And so it is evident that a potential is required for holding
back the descent of the sphere and fixing the waterscrew. Let us
assume that that potential is applied at the pointf in the plane of the
circle and perpendicularly to the radius in question in the ratio which
it has to the weight of the sphere resting at some point p.
Let the weight of the sphere be p; but, because the action of the
sphere is vertical, it will have to be resolved into two others remaining
perpendicular to each other, the one of which let have a common
direction with the axis of the waterscrew, the other let be perpendic-
ular to it; the former will have to be rejected, since it contributes
210 HYDRODYNAMICS, CHAPTER IX

nothing to turning the waterscrew, and the latter will have to be con-
sidered alone; but that residual action equals np, and it acts on a lever
which is equal to the sine of the arc Mh or the arc ag, and this sine
(by Art. IV) is~- Therefore, the moment of the action is ~-np =

m,:; if one divides this by the radius of the base, which is the lever
pertaining to the potential applied at f set in equilibrium with the
action of the sphere, one will have the required potential equal to
mNp
M . Therefore, others customarily derive from a foreign principle
that which can thus be deduced directly from the nature of the lever.
With these things having been set forth which were to be set forth,
let us now begin to consider the use of the machine for elevating
water.

PROBLEM

VI. It is asked what the maximum quantity of water is that a

given waterscrew can discharge in a revolution.
SOLUTION. Let us consider an entire helix aib, and let the quan-
tity of water which it contains [when it is] full be q: it is also to be
noted that the helix cannot be entirely filled with water; for if the
entire conduit were full, water would flow out through the lower
orifice: therefore some branch, which is a1b, is occupied partly by air,
partly by water; also one extremity of the water will be at o, or the
uppermost point, the other at q, or a point situated at a level with the
former; therefore, the part full of water is opq, and if this part is
assumed to be in proportion to the length of the entire helix aib as g
is to h, the maximum quantity of water to be discharged in one
revolution will be equal to gi" Q.E.I.

ScHOLIUM 1

VII. Because, as we have said, it cannot happen that the water is

continuous through the full extent of the conduit, care has to be taken
that no separation is imparted to the water, which can easily be
accomplished if the entire base of the cylinder is immersed in the
water, because thus air cannot enter through the lower orifice of the
conduit; neither must it happen that too large a part of the base
projects from the water, because then the waterscrew does not draw
 MOTIC,., OF FLUIDS PUSHED BY AN OUTSIDE FORCE 2I I

all the water that it could otherwise [draw] in one revolution; on the
contrary, it draws nothing if the immersion does not reach the point
h; but due immersion should occur up to the point g, because thus the
arc opq of the helix, which is capable of retaining water, becomes
largest. Although indeed I never conducted an experiment on the
matter, and most authors seem to speak differently about it, I would
rather trust in reasoning than in the authority of those who did not
pay attention to this immersion.
Therefore, this rule for the ratio of immersion will be observed, namely
that the base is submerged until the chord of the arc projecting from
the water is v;:,;:, where the letters m, N, M, and n signify the same
[variables] as in the fourth article.

ScHOLIUM 2

VIII. It is apparent, indeed, after light contemplation of the

matter, that the ratio between the arc opq of the helix and the entire
helix arb, that is, between g and h, is greater, and hence a larger
quantity of water is discharged in a single revolution, the rest being
the same, the smaller the angle sao and the larger the angle aMH, or,
the smaller the distance between two adjoining helices and the
more the waterscrew is inclined towards the horizontal; but it is not
possible to express that true ratio algebraically; nevertheless, in every
particular case this is obtained by an easy approach.
Let me select an example of the preceding rule from a waterscrew which
Vitruvius shows how to construct and apply. He makes sao a semi-
right angle, and thus m = M = V½ = o. 707 IO; then he sets a ratio
between NG and MG, which is as 3 is to 4; whence one deduces the
angle GNM or aMH is equal to 53° 8', and the sine n of it equals
0.80000 and the cosine N = 0.60000; therefore (by Art. III ), the

sine of the arc ag defining the highest point o equals~ = ¾, and the
arc ag itself equals 48° 35'. And thus, by virtue of the rule of Art. VII,
the arc projecting from the water at the base must be 97° ro', and an
arc of 262 ° 50' is immersed.
Furthermore, in order that we may now define the ratio between
the arc opq of the helix and the entire helix arb, it is to be noted that
the ratio is the same as that which exists between the circular arc
ghMs and the circumference of the circle, which is manifest from the
accompanying figure. But the arc ghMs may be determined in the
following manner : for example, the arc ghMs = arc aghMs - arc ag.
212 l-:--'DRODYNAMICS, CHAPTER IX

But we have seen in the third article that, if from any point of the
spiral, such as o and q, perpendiculars are drawn down to the hori-
zontal going through M, such as or and qx, that perpendicular will be
mi.JX + n(1 + x), or in our case o.6ooooX + 0.80000(1 + x), with
X denoting the circular arc corresponding to the assumed point on the
spiral, namely the arc ag or the arc aghMs, and x denoting the cosine
of that arc. But the arc ag = 48° 35' = 0.84 794 (because the radius
is expressed by unity), and the cosine of it equals o.66153; therefore,
in our case or becomes 0.50878 + r.32922 = r.83800. Further,
because the points o and q are located at the same height, and the lines
or and qx are equal to each other, it is apparent that the question is
now reduced to this: that the other arc aghMs corresponding to the
point q be found so that, if it is called X and its cosine x, then
o.6ooooX + 0.80000(1 + x) =or= r.83800; for this condition the
arc aghMs is found [ to be] approximately 175½ degrees, intersecting
the cut agM at the points. And since the arc ag will be 48° 35', the
arc ghMs will finally be 126° 55', which thence will be to the circum-
ference of the circle approximately as IO to 29: the same ratio prevails
between the arc opq of the helix and the entire helix.
From this follows that in a single revolution there is discharged by
the waterscrew described by Vitruvius approximately ~ ~ of that
quantity which the full helix contains, or very little more than one
third.

SCHOLIUM 3

IX. It is nevertheless to be noted that whatever be the quantity

of water which enters the conduit at the bottom at any revolution of
the waterscrew and flows out of the same at the top, it imparts neither
a loss nor a gain to the absolute potential if no consideration is given to
friction, because the moving potential, the rest being equal, is propor-
tional to that quantity. But because friction always hinders and is
almost the same on account of the very weight of the machine whether
a larger or smaller quantity of water is pumped, and certainly an
effort is to be made that that quantity becomes great, the rest being
equal, this matter I shall now treat a little more expressly.

ScHOLIUM 4

X. I already hinted above that the ratio of the arc ghMs to the
circumference of the circle increases with decreasing angles sao and
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 213

N MG; either one should therefore be built very small unless other
inconveniences interfere, particularly in the consideration of the
angle N MG. As far as the angle sao is concerned, it can be dimin-
ished almost arbitrarily, and thence no other inconvenience results
apart from the fact that the sides of the conduit to be curved cannot
come too close to one another: on the contrary, from the diminishing
of that angle another benefit is obtained, namely, that then the ma-
chine can be erected more nearly vertical, and the water itself is
elevated higher, for truly the angle aMH must always be larger than
the angle sao; from the more nearly vertical position of the water-
screw, moreover, it occurs simultaneously that the very weight of the
machine is of less inconvenience, and that the latter is more easily
supported.
Considering these things accordingly, I should believe that it usually
suffices for the conduit to make an angle of 5 degrees with the base of
the center. Cardano also made that angle smaller than Vitruvius,
and since the fewer conduits can be wound around the same center,
the more obliquely they are attached, Vitruvius stated that eight are
to be placed, Cardano only three; but the conduits are longer in the
waterscrew of Cardano, so that it contributes in the lengths what it
lacks in the number of conduits. In the consideration of the other
angle N MG it merits being observed that the water can be elevated
higher, the larger the angle becomes, but, on the contrary, the
quantity of water discharged in a single revolution is less. Probably
those who make that angle 60 degrees will reach a just median.
XI. Now we will also perform the calculation of our waterscrew
constructed to the norm of the preceding article as we have done for
the waterscrew constructed according to the concept of Vitruvius in
Art. VIII. But because by hypothesis the angle sao is 5° and the
angle N MG = 60°, the arc ag will, by Art. IV, be found to be 8° 43',
and the vertical line or = 1.00574, to which the other vertical qx will
be equal if 284° 57' is assigned to the arc aghMs; hence, if the arc ag
is subtracted, the arc ghMs remains as 276° 14', which corresponds to
the arc of the helix capable of retaining water; therefore, this part is to
the entire helix as r 6,574 to 2 r ,600 or as 8287 to ro,800, such that in a
single revolution more than four-fifths of the capacity of the entire
helix can be discharged, and two and one-third times as much is
accomplished by this machine as is obtained from similar machinery
constructed according to the understanding of Vitruvius; also, the
water is elevated higher from the same center in the ratio as VS is to
V2. I come now to the moving as well as the absolute potential that is
expended in elevating water.
214 HYDRODYNAMICS, CHAPTER IX

PROBLEM

XII. Given the weight of the water resting in the helix, find the
tangential potential located atf in equilibrium with that weight.
SOLUTION. We have seen how this problem may be solved geome-
trically by reasoning of a sphere resting at the lowermost point p.
But in the present case the situation is slightly different, since the
weight of the water is distributed through a large arc of the helix and
not concentrated at some given point. It is certainly easy to foresee
that in either case the potentials will be the same from the indirect
rules of mechanics; it pleases, nevertheless, to present the desired proof
of this matter from the nature of the lever, because the mechanics
love to reduce everything to that.
We shall consider the helix aib taken separately from Fig. 52 in
order to avoid confusion of the lines, the notations applied in Art. IV
having been preserved. Thus, therefore, in Fig. 53 the angle N MG

I,

,e, ·f
fl .
zi·

FIGURE 53

will again be the angle which the center makes with the horizontal,
its sine equals N, and the sine of the angle aMH equals n; arb is one
revolution of the spiral. The circle acMpa is the base of the center;
the sine of the angle pal, as before, equals m, and its cosine equals M;
but the points l and o are the extremities of the water resting in the
spiral and located at the same height from the horizontal; from these
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 215

points are drawn to the periphery of the base the straight lines le and
op, perpendicular to the base. In the part of the helix which the
water occupies, two infinitely close points m and n are assumed, and
through these the straight lines nj and mg are drawn, again perpen-
dicular to the base. Finally, from the points c, J, g, and p the per-
pendiculars cd, jh, gi, and pq are drawn to the diameter aM; and the
center of the base is assumed at e, and the radius ea = r. Now let
the arc lro of the spiral, full of water, be equal to c, and consequently
the circular arc cMp corresponding to the same be equal to Mc : al = e;
ac = Me; ad (or the sine with respect to the arc ac) equals]; aq = g;
the weight of water in lso equals p; the arc aln = x; nm = dx; aif =
Mx;fg = M dx; ah = y; hi= dy; hf= -V?J> - yy, [and] the weight
of the volume element at nm equals p dx; but if the line lifis multiplied
C

by the sine of the angle aMH and divided by the entire sine, there
results the lever arm by which the particle nm attempts to turn the
waterscrew; therefore, this lever arm is equal to n-v' 2y - yy, which,
multiplied by the weight of the volume element given above, p dx,
C

gives its moment np dx -v' ?JI - yy. But from the nature of a circle
C

M dx = 'V dy ; therefore, after this value has been substituted for

zy -yy
dx, the moment of that same volume element nm becomes n~:,
the integral of which, after subtraction of the proper constant, is
np(~~ f) and denotes the moment of the water in the arc ln; whence,

therefore, the moment of all water in lro is np(~~ f). This divided
by the lever arm of the potential applied at J, or by r, yields in a like
manner the desired potential np(~~ j). Q.E.I.

ScHOLIUM r

XIII. In order that it be apparent that the value of this potential

does not differ from that which we found in Art. V for a sphere of the
same weight p, namely m-:;, the equality between np(~~ j) and

m-:; or between n(g - j) and mNc is to be shown; but this equality

216 HYDRODYNAMICS, CHAPTER IX

is to be deduced from the fact that the extremities land o of the water
are located at the same height above the horizontal; for hence it
follows, as we have shown in Art. IV, that the sum of the arc ac
multiplied by m_: and the line Md multiplied by n equals the sum of
the arc acMp multiplied in the same way by m_:: and the line Mq
multiplied by n. And so, with the notations of the preceding article
having been applied, there results
mN mN
Me· M + (2 - f) ·n = (Me+ Mc)· M + (2 - g) ·n,

or n(g - f) = mNc, which equality has to be shown for demon-

strating the equality of the potentials to be applied for the sphere as
well as for the water atf
ScttOLIUM 2

XIV. Because the potential np(~~ f) does not differ from m-::

an dh .mN
t e quantity · t h e same, w h atever quantity
M remams · o f water 1s
·

drawn in or discharged in one revolution, that potential will be pro-

portional to that very quantity of water discharged in a single revo-
lution, or to the weight p. Also, it is easy to prove that, if the same
quantity of water is elevated by the same moving potential and at the
same velocity to an equal vertical height above the base plane, which
to this end must be appropriately inclined towards the horizontal, it
will happen that the time of elevation is also the same.
Therefore, the same absolute potential is required in the waterscrews
of Archimedes as [is required] on an inclined plane, to which all
machines can be reduced, and this waterscrew does not have any
prerogative over other machines viewed in the theory. Perhaps in
practice it is less exposed to the inconveniences indicated in §26; by
no means do I reject its use, but neither do I prefer it to the pumps
of Ctesibius.
- - - - - - - - - - [ End of Digression] - - - - - - - - - -

§28. From what has been said so far, one understands under what
conditions one machine ought to be preferred to another: namely,
what degree of perfection of the machine these [conditionsJ permit;
to what one should pay most particular attention in their construction
and use; how large a part of the absolute potential is lost; and other
similar things. Of course, we have considered only machines driven
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 2I 7
by animated potentials, as they are called; but it is readily apparent that
those machines that are to be driven by the impetus of water, by
wind, or by the gravitation of water and other principles of this kind
are subjected to the same laws; always, indeed, the moving potential
multiplied by the time and the velocity of the point to which the
potential is applied will give the product of the quantity of water and
the height to which that quantity can be elevated in a given time by
means of the proposed machine, other hindrances having been set
aside. However, I am speaking about machines which lose none of
the absolute potential; it can happen, indeed, that the greatest part is
lost, which we have shown often enough above.
§29. Hence it is apparent that water elevated to a certain height
can by its descent produce the same effect again; but the effect will
have to be estimated from the quantity of water to be elevated and
the height of elevation; for example, by the descent of 8 cubic feet
from a height of one foot, it is wholly possible for 8 cubic feet to be
elevated again to the same height, or 4 cubic feet to a height of two
feet, or one cubic foot to a height of 8 feet, and thus however one
would please. A specimen of a machine which can elevate water to
any height whatever by a minimum descent of water is found among
[the works of] Mr. Perrault in the Commentarii ad Vitruvium, Book ro,
Chapter 12, which machine he introduces as an almost incredible
paradox, and he makes the Italian Mr. Francini the inventor ofit, by
whose industry and planning it was constructed successfully in the
garden of the Royal Library. The basis of the machine consists in
the fact that buckets chained together and moving around in a circle
take up water and transport it to the lowermost point where they
discharge it, while another series of buckets take up water, although
less in quantity, and carry it to a much higher location and discharge
it. It is very clear that if all descending buckets are heavier than all
ascending buckets, the former series will activate the other perpetually
in a circle. There exist also machines which produce the same
through simple pipes by means of flaps that are to be reversed at
regular time intervals, in which conversion certainly no potential is
expended. Carlo Fontana describes machinery of this kind.
But if anyone believes that the same can be obtained from the im-
petus of water falling from a certain height and impinging on the
blades of the machine, he will be far off. Machinery of this kind
would pertain to that class in which the largest part of the absolute
potential vanishes without benefit.
It will not be beside the point to follow this argument more ac-
curately and to show how great an effect can be obtained from the
218 HYDRODYNAMICS, CHAPTER IX

impetus of water or wind, and under which circumstances this effect

may be considered the greatest of all.

THIRD PART: CONCERNING MACHINES WHICH

ARE DRIVEN BY THE IMPETUS OF A FLUID, SUCH
AS BY THE FORCE OF THE WIND

§30. After water elevated to a certain height falls down again

from the same and impinges continuously on the blades of a wheel
to be turned, it cannot happen differently than that the absolute
potential required for so turning the wheel is much less than that
which was expended in the elevation of the water, the foremost
reason for which matter is that the water falling down after the
impulse on the sides still preserves a velocity which contributes
nothing to the rotation of the wheel. Therefore, a large part of the
absolute potential of elevated water would become useless if a machine
were driven by the impetus of this water and finally, by this in turn
other water were elevated to a certain height; and indeed a larger or
smaller part is lost because of different circumstances, but never, as
I shall show, is lost less than -H· of the total if a computation of the
ordinary impulse of water is made according to the norm.
§31. Furthermore, it is commonly stated that if water flows out of
a very wide cylinder through a simple orifice at its total velocity, that
is, that which would be due to the total height of the water above the
orifice, and the stream immediately in front of the orifice impinges
directly on a plane, [then] it will occur that the impetus of the fluid
against the plane is in equilibrium with the weight of the aqueous
cylinder erected above the orifice to the height of the water. Authors
certainly misled by a false experiment have supported this completely
false theory. I nevertheless did not want to withdraw here from the
latter, because I have not yet shown the true theory, and then, after
our theory has been explained, it will be easy to correct the calcula-
tion. May one therefore be allowed to adhere to the common,
although erroneous, statement until we consider the matter more
correctly in its proper place. The greater the impetus of a fluid, the
greater the ratio by which the absolute potential, which we shall give,
will have to be increased.
§32. Consider now (Fig. 54) a vessel ABC or a pump which expels
water through the orifice Cina not quite vertical direction; but the
water is taken up by another vessel EFD when it has reached the
summit. At the base of this other vessel picture an orifice D, equal
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 2 I9

to the former C and located at the same elevation, so that as large an

amount of water flows out through D as is poured in above, and the
vessel EDF is kept constantly full. Assume further that the water
flowing out through D impinges continuously on the blades of some

Jl I

FIGURE 54

wheel, which, turned in this manner, elevates other water; in place

of this machine is described in the simple figure a lever arm rotating
around H, by assuming that continually some one or another dif-
ferent lever arm of this kind is present in front of the orifice D, which
receives the water and draws water at its other extremity and elevates
it to the given height.
After these [things] have so been assumed, I shall first inquire
about the absolute potential that elevates the water flowing through the
orifice C to the height CE; then also about the absolute potential that is
required at G for moving the lever arm at the same velocity at which
it is moved by the impulse of water DG.
§33. Let the area of the orifice C or D be n, the area AB be m, the
velocity of water at C or D be v, the weight of the cylinder erected
above the orifice C or D to the height CE be p, the time of flow be t;
then the weight P will be ~ p; the velocity at which the weight
n
descends while water is expelled equals !!:. v; therefore (by §3) the
m
absolute potential expended in ejecting water through C is
m n
- p ·- V • t = pvt.
n m
220 HYDRODYNAMICS, CHAPTER IX

§34. Now, in order that the absolute potential expended in the gyra-
tion of the lever arm GL around the point H may be determined, it is
to be noted that the former is the least consistent with itself; for it is
changed by the changed velocity at which the lever arm is rotated.
Therefore, let us make the velocity at which its extremity at G is
moved equal to V. But in this manner the water is to be considered
as impinging at G with the velocity v - V, and thus it exerts a pres-
sure which is (v ~ V) p (for the pressures are in a square ratio to the
2

velocities of the impinging fluid, and a pressure equal to p is substi-

tuted for the velocity v). But this pressure exists in place of the
moving potential; in place of the pressure of the fluid we can certainly
substitute the weight lying above the lever arm at G, which is

(
V -V V\2p.
-J But that weight will be moved at the same velocity as
the point G, namely, at the velocity V, and it acts during the time t.
Therefore, the absolute potential required for the rotation of the lever
arm during the time t and at the velocity Vis C~ vrP · v. t.
§35. Thus, if the lever arm LG is not rotated immediately, but the
fluid is elevated to the height CE with the intention that the stream of
fluid, by its impulse at G for rotating the lever, elevates water from the
other part, the entire absolute potential will be to the useful absolute
potential as pvt is to C~ ~ p Vt,
2
or as v3 is to (v - V) 2 V, and it will
be to its useless part as v3 is to v3 - vvV + 2vVV - V3 •
§36. In almost all machines in which the principle of motion con-
sists of the impulse of fluid, it usually happens that the velocity V of
the lever where it sustains the impetus of the fluid is very small in
proportion to the velocity v of the fluid; but in these [machines] the
largest part of the effect that could be obtained from the same
quantity of fluid moved at equal velocity is lost.
§37. The greatest effect from the impulse of fluid develops, or,
which is the same thing, the absolute potential defined in §34 becomes
greatest, if V = ½v, and then this absolute potential is 2~-pvt; and even
then it falls short by twenty-three twenty-sevenths of the similar
potential that is expended in elevating water from C to EF.
If a natural descent of water exists and is to be used for elevating
water or for accomplishing anything else, it must be arranged that the
machine is moved at that place where the impulse occurs at a velocity
of one third of the velocity of the impinging fluid. But this condition
can always be satisfied, which is evident from the cited example of the
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 22 I

lever. For if the point G is moved at a greater velocity, one must

diminish the part HG, the rest remaining the same, or one must in-
crease it if the point G is moved at a lower velocity. Or even with
the length HG retained, one must arrange that the water is drawn in
at the extremity Lin a larger or smaller quantity.
§38. Truly this is the reasonini about fluids impinging perpen-
dicularly on blades; the computation is different for fluids attacking
obliquely the arms of mills agitated by the force of the wind, and other
similar machines. About these let me now add some few things, and
with these I will bring this chapter to an end.
Since the fluid impinges on the surface of the entire blade, arbi-
trarily located and to be rotated in the direction perpendicular to the
motion of the fluid, writers show that the fluid exerts the greatest
pressure on the blade for promoting the rotation when the blade
makes an angle with the direction of the wind the sine of which is to
the total sine as V2 is to V3; but if the same entire stream of fluid is
received by the blade, whether [it is] thus or inclined differently to the
direction of the fluid, then that blade which makes a half right angle
with the direction of the fluid will sustain the greatest pressure in the
direction of rotation.
The first Rule pertains to machines which are driven around by a
wind surrounding everything; the other to those which are moved by
a solitary stream and by a certain determined quantity of fluid. But
either hypothesis depends on the fact that the motion of the blades is
very small with respect to the motion of the fluid; for if one refers to
the motion of the blades, both rules are false; and in the outset this
motion is not to be neglected; indeed, I have often observed on mills
that the tips of the arms are carried at a velocity which almost equals
the velocity of the wind itself.
Since these [things] are so, let us perform a calculation so that we
obtain an understanding of either motion.
§39. Therefore, let there exist the fluid DEBA (Fig. 55) which
impinges on the entire plane AB in the direction EB; moreover, the
plane is assumed to be moved in parallel motion in the direction Bb
perpendicular to EB. Further, let the velocities be of the kind that,
while a particle of fluid moves through the line EB, the point B of the
plane travels the line Bb. Under these assumptions one may con-
sider that the entire system, namely, the fluid and the plane, is moved
from b towards B, and certainly with the velocity bB. But so it will
occur that the plane AB is at rest, yet the particle of fluid striking at
the point B is to be considered as having come from the point e,
Ee = Bb having been assumed, and so accordingly for all volume
222 HYDRODYNAMICS, CHAPTER IX

elements. Therefore, instead of the fluid DEBA striking the moving

plane AB with the velocity EB, one will have to consider a fluid
deBA striking the same but immobile plane AB with the velocity eB.

FIGURE 55
Let AB now be extended up to h, and let DEdeh be moved perpen-
dicular to EB; the motion represented by eB of the particle of fluid
will have to be resolved into eg and gB, remaining perpendicular to
each other, the latter of which does not act upon the plane AB; but
the other, eg, is again composed of two motions, ef and jg, the latter,
Jg, of which tries uselessly to propel the plane AB in the direction EB,
while the former, eJ, alone propels this plane in the direction Eb. It
is therefore shown that any arbitrary particle causes an impulse
proportional to the line ej; then it is also evident that if the line AB
represents the entire plane, the number of particles impinging in a
given time on the plane is to be represented by the line EN, perpen-
dicular to Ad or Be. Whence finally the pressure of water for moving
the plane in the direction Eb is proportional to the line ef multiplied
byBN.
In order that now the inclination of the plane to the fluid be deter-
mined that is most favorable under these circumstances for promoting
the movement of the plane in the direction Eb, let us set AB = 1,
DE or AC= x, ED = V r - xx, the line EB, which represents the
motion of the fluid, equal to v, and Eb or the measure of the motion of
the plane equal to V; and thus, after the calculation has been per-
formed, one finds
ef = xv-V1 - xx - (1 - xx)V,
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 223

and
BN = [xv - VVr - xx] /Vvv + VV;
whence

ef-BN = [xv - Vv r - xx] 2 V

. ; - vr - xx
,
vv + vv
which quantity will be largest when this occurs:
(gv 4 + r8vvVV + gV4 )x 6 - (r2v 4 + 3ovvVV + r8V4)x 4
+ (4v 4 + r6vvVV + gV4 )xx - 4vvVV = o.
§40. The calculation in consideration of the inclination of the arms
in mills is different, because the velocities are different in different
locations on the arms; they are, indeed, proportional to the distances
from the center. But now it will be easy for anyone to perform a
computation for mills. I do not wish to pursue this case any further,
so let it suffice to have noted that it is stated by authors not accurately
enough [that] xx = i, and that the true value of x itself is always less
than V!- For example, if V were equal to v, and all points of the
arm were thought of as being moved at similar velocity, x would
become V½, which indicates that the arm is to be inclined to the
direction of the wind at a half right angle. The best construction of
arms would be if they were curved so that the wind impinges on them
higher up at a smaller angle than lower down, or if it were made that
the arms receive the wind everywhere at a mean angle of approxi-
mately fifty degrees.
§41. I pass on to the other case in which all fluid is assumed to be
received by the plane, whichever way it be inclined. Here, how-
ever, it is evident that because the number of particles impinging in
a given time is always the same, no attention must be paid to the line
BN, and that thus the pressure that the water exerts for moving the
plane AB in the direction Bb is represented simply by ef or
xvV r - xx - ( r - xx ) V. Therefore, this pressure will be made the
greatest by taking xx = !. + V
V , and then the pressure itself
2vv + vv2
will be ½V vv + VV - ½V, if by v one understands the direct pressure
which the stream exerts on a plane which it strikes perpendicularly.
§42. Let us consider now the stream DEBA as if immediately dis-
charged from the orifice D in Fig. 54, and let us call p again the
direct pressure of the stream thus considered, just as in §33; and the
pressure of this water, by which it tries to propel in the direction per-
pendicular to the stream the plane inclined in such a manner that the

. ------------ --
224 HYDRODYNAMICS, CHAPTER IX

pressure becomes greatest, will be f!.... · ( v' vv

2V
+ VV - V ) ; and if,
furthermore, this pressure is multiplied by the velocity V of the
plane and by the time, the absolute potential is obtained by which the
plane can be moved at the same velocity through the same time
interval; so, therefore, the aforementioned absolute potential will be
pVt_ (v'vv + VV - V.)
2V
§43. The absolute potential which we have just defined is so consti-
tuted that it increases continuously with increasing V, and if the
velocity Vis assumed infinite, that same potential becomes ¼·pvt. If
we therefore want to use the stream DC in Fig. 54 for rotating a
machine by an oblique impulse, there can never be obtained more
than the fourth part of that absolute potential which is expended in the
elevation of the water from C to EF. But we have seen in §3 7 that by
direct impulse more than -!-7 is never obtained. Therefore, by an
oblique impulse or by horizontal motion of the wheel an effect can
be obtained almost two times as great as by vertical motion of the
wheel.
But if the impulse of fluids is estimated differently than was indi-
cated in §3 r, one will have to change the value of the letter p every-
where in the same ratio in which the estimation of the impulse was
changed.

The experiment of which I made mention in §27, Chapter IX, is this:

namely, by means of a pump one worker lifted sixteen and a half
cubic feet to a height of fourteen feet within seven and a half minutes.
But this effect, equally distributed, is equivalent to that action by
which approximately half a cubic foot is elevated in a single second
to a height of one foot. Here, therefore, the effect is only half of that
which I deduced from other principles in §17, that a healthy and ro-
bust man can produce on a treadmill. I would not believe that the
entire defect is to be sought in the losses which can occur to the
absolute potential from the different causes exposed in this chapter, but
rather in the fact that the men become more tired from the agitation
of the piston in the pump than from the tread in the treadmill.
Some months ago at Geneva I finally performed an obviously
similar experiment, but with a far more excellent machine con-
structed by a singular craftsman, and with these Most Famous
Gentlemen present: Messrs. De La Rive, Calandrini, Cramer, and
 MOTION OF FLUIDS PUSHED BY AN OUTSIDE FORCE 225

Jallabert, Professors of the Academy of Geneva. The success of the

experiment was such that I found out that one worker elevated four-
fifths of one cubic foot to a height of one foot in a single second, or,
rather, that he achieved an equivalent effect. The experiment is
noteworthy, and I do not believe that an effect greater than this can
be obtained by any other machine. Also curious is the fact that it
thus appears that machines of all kinds, animated by any potential
whatever, achieve, if you remove hindrances, an effect not greatly
dissimilar. Having thought over the matter well, I state that by a
most excellent machine a man can elevate a cubic foot of water in a
single second to the height of one foot, or produce a similar effect.
Here as well, particularly in consideration of§31, would pertain the
experiments that I most accurately performed for estimating the im-
petus of a fluid stream impinging on a plane, by which was confirmed
the new theory which I had established about this matter, and simul-
taneously I learned that in Mariotte's time a common error was
committed. But since at the end of that chapter there was no elo-
quent discussion on this subject, and since the intention is to treat it
expressly in Chapter XIII, let us therefore delay until then these
discussions brought forth from mechanical principles not yet observed.

- ----------
1
 TENTH CHAPTER
Concerning Properties and Motions of Elastic
Fluids, but especially of Air

1. Now being about to consider elastic fluids, we may ascribe to

them a constitution that coincides with all properties known so far,
in order that thus also a path be provided to the remaining, still in-
sufficiently explored properties. But the outstanding properties of
elastic fluids are stated as follows: (I) they are heavy, (2) they extend
in all directions, unless they are confined, and (3 ) they allow them-
selves to be compressed continuously more and more as the com-
pressing forces increase. Air, to which our present considerations
pertain mostly, is composed in this way.
§2. And so consider a cylindrical vessel ACDB, placed vertically
(Fig. 56), and in it a movable lid EF, on top of which lies the weight
P. Let the cavity ECDF contain extremely small bodies agitated in a

FIGURE 56
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 227

very rapid motion; thus the small bodies, while they impinge on the
lid EF and also support the same by their continually repeated
impacts, compose an elastic fluid which expands if the weight P is
removed or diminished; this is compressed if the same is increased,
and it gravitates on the horizontal base CD not at all differently than
if it existed with no elastic property. Indeed, whether the small
bodies are at rest or agitated, they do not change gravity, so that the
base sustains either the weight or the elasticity of the fluid. There-
fore, let us substitute for air a fluid that is consistent with the primary
properties of elastic fluids, and thus we will explain some properties
which have been already detected in air, and we will illustrate
further some others [that are] not yet sufficiently investigated.
§3. We shall consider the small bodies enclosed by the cavity of the
cylinder as infinite in number, and since they occupy the space
ECDF, let us say that the latter forms the natural air, to the measures
of which all [other measures] are to be referred; and thus the weight
P holding the lid in the position EF does not differ from the pressure
of the Atmosphere lying above it, which we therefore shall designate
henceforth by P.
But let it be noted that this pressure is not at all equal to the
absolute weight of the vertical cylinder of air lying above the lid EF
in the atmosphere, which authors so far have affirmed inconsiderately;
but that pressure is equal to the fourth proportional to the surface
of the earth, the size of the lid EF, and the weight of the entire atmo-
sphere on the surface of the earth.
§4. Now the weight 7T is sought which can compress the air ECDF
into the space eCDJ, the velocities of the particles in either air (the
natural and the compressed), of course, having been assumed the
same; moreover, let EC= 1, and eC = s; but since the lid EF is
transferred to eJ, it suffers a greater pressure from the fluid in two
ways: firstly, because the number of particles is now greater in pro-
portion to the space in which they are contained, and secondly, because
any particle repeats the impetus more often. In order to perform
correctly the calculation of the increment which depends on the first
cause, we shall consider the particles as resting, and we shall make n
the number of those which are adjacent to the lid in the position EF,
and the equivalent number for the location of the lid at ef will be
eC)21s
n/( EC ' or n/s21s.
But let it be noted that the fluid is considered by us not more com-
pressed in the lower part than in the upper part, which is so because
the weight Pis just as infinitely much larger than the very weight of

------
 HYDRODYNAMICS, CHAPTER X

the fluid. Hence it is clear that with this designation the force of the
fluid varies in proportion to the numbers n and n/s2 13 , that is, as s2 13
is to 1. But what pertains to the other increment arising from the
second cause is found by observing the motion of the particles; and thus
it is apparent that the impulse occurs the more often, the closer the
particles are located to each other; of course, the number of impulses
will be reciprocal to the median distance between the surfaces of the
particles, and these median distances will be determined as follows:
We assume the particles to be spherical, and we shall call D the
median distance between the centers of the small spheres for the posi-
tion EF of the lid, and the diameter of a small sphere we shall desig-
nate by d; so the median distance between the surfaces of the small
spheres will be D - d; but it is evident that at the position ef of the
lid the median distance between the centers of the small spheres will
be Dtfi, and therefore the median distance between the surfaces of
the small spheres is Dtfi - d. Therefore, with respect to the second
cause the force of the natural air ECDF will be to the force of the
I
compressed air eCDJ as - Ddis to fl~ , or as Dtfi - dis to
- D s- d
D - d; for both causes together, however, the aforementioned forces
will be as s2 13 • (Dtfi - d) is to D - d.
For the ratio D to d we can substitute another, more intelligible one:
namely, if we consider that the lid EF, [when] depressed by an in-
finite weight, descends to the position mn at which all particles touch
each other, and ifwe designate the line mC by m, D will be to d as I is
to fl;i;_, which ratio being substituted, finally the forces of the natural
air ECDF and of the compressed, eCDj, will be as s2 13 • ( tis - fi;i;_) is
.a;- .a;- .a;-
to 1 - v m, or as s - ·v mss to 1 - v m. Therefore,
.a; -
1 - vm
7T = a -P.
s - f!mss

§5. From all phenomena we can judge that natural air can be
compressed into an almost infinitely small space; therefore, with
m = o having been assumed, becomes ~, so that the compressing
TT
s
weights are almost in inverse proportion to the space which the air
occupies when compressed differently; manifold experience has con-
firmed this. This rule can also certainly be accepted safely for air
rarer than natural; but I have not explored sufficiently whether it can
be also [accepted] for very much denser air; and indeed experiments
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 229

have not yet been performed with that accuracy which is required
here; for defining the value of the letter m, there is need of only one,
but of one to be formed most accurately and certainly with violently
compressed air; however, let the degree of heat in' the air, while it is
being compressed, be carefully kept unchanged.
§6. Meanwhile, the elasticity of air is increased not only by com-
pression but also by increased heat, and since it is established that heat
is spread out everywhere by increasing internal motion of the par-
ticles, it follows that an increased elasticity of air not changing volume
discloses a more intensive motion in the air particles, which agrees
correctly with our hypothesis; it is indeed evident that the greater the
weight P required for keeping the air in the position ECDF, the greater
the velocity at which the air particles are agitated. By all means it is
not difficult to see that the weight P will follow the ratio of the square
of that velocity, for the reason that the number of impacts as well as
their intensity increases equally with an increased velocity; but [each
one] separately is proportional to the weight P.
If, therefore, the velocity of the air particles is called v, the weight
which it can sustain at the position EF of the lid will be vvP and in the

_-{½;;
. wi·11 b e
. . eif it 1 - P . 1 vvP b ecause,
position ; - • vv , or approximate y - ,
s - -v3 mss s
as we have seen the number m is extremely small with respect to
unity and the number s.
§7. That theorem which I presented in the preceding paragraph,
by which, namely, it is indicated that in all air of any density whatever
but of the same prescribed degree of heat, the elasticities are as the densities,
and therefore also the increments of the elasticities, which are created by equally
increased heat, are proportional to the densities, that theorem, I say, Mr.
Amontons was taught by experience, and he recorded it in the memoirs
of the Royal Academy of Science of Paris for the year 1702. The sense of this
theorem is that if, for example, natural air of moderate heat sustains
a weight of 100 pounds imposed on a given surface, and then its heat
is increased until it can carry 120 pounds on the same surface and at
the same volume, then it will occur that the same air, compressed to
half the space and possessing the same degrees of heat, can carry,
respectively, 200 pounds and 240 pounds, so that increments of 20
pounds and 40 pounds, proportional to the densities, are generated in
either case by the increased heat. He further affirms that the ex-
pansion of air, which he calls tempered, is to the expansion of the
same air with the heat of boiling water approximately as 3 to 4 or,
more accurately as 55 to 73. However, I have learned from per-
formed experiments that very hot air, such as it is in the hottest
230 HYDRODYNAMICS, CHAPTER X

summer in this country, is not yet of such an expansion as Mr.

Amon tons attributes to tempered air; I believe that not at the equator
itself is the air ever of this heat. But I believe that my experiments
are to be trusted more than the Amontonian ones for the reason that
in the latter the air will not preserve its volume, and to this variation
no consideration was given by the Author in the calculation. I
learned that the expansion of air which was very cold here in Peters-
burg on 25 December 1731 (old statutes) was to the expansion of
similar air attributed with the same heat as boiling water as 523 is
to IOOO.
But in the year 1733, en the 21st day of January the cold was much
more intense, and for this I observed that the elasticity of the air
corresponds to within half of that which the same air has when
heated to boiling water. But when the heat of air in a shadowed
place was greatest in the year 1731, it had an elasticity of approxi-
mately ½and more accurately 1..fj- of that which the coldest air had
and i of that which air of the same heat as boiling water has; there-
fore, the greatest variations in the air in these places are contained
within the limits 3 and 4; I have read that in England they do not
exceed the limits 7 and 8. But I believe that the heat of air, the
elasticity of which equals three quarters of the elasticity of air as hot as
boiling water, is almost intolerable for an animal body.
§8. From the known relation between the different elasticities of
the same air contained in the same space, it is easy to deduce a measure
of the heat which pertains to the air if only we agree in defining twice,
three times, etc., the heat which definition is arbitrary and not fixed
in the nature of things; to me it seems indeed not incongruous that the
heat of the air, if it is of common density, is proportional to the state
of its elasticity. But let the first degree of heat, from which the others
obtain their measure, be taken from boiling rain water, because this
has, without doubt, approximately the same degree of heat everywhere
on earth.
If these [thingsJ are so accepted, the temperatures of boiling water,
of air in the hottest time of summer, and of air in the coldest time of
winter in this country will be approximately as 6, 4, and 3. Let me
tell now how I found those numbers, so that judgment can be passed
on the accuracy of the experiments, the success of which is so greatly
different from the Amontonian.
§9. Indeed, I made use of an ordinary barometer ACEE (Fig. 5 7),
and I took care that it was sealed hermetically at m; in this manner
I changed the instrument into an air thermometer not subjected to
barometric changes, for with increasing heat an expansion of the air
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 23 I

AmF is achieved, and the column of mercury BD which the captured

air sustains becomes higher, and if the space AmF could be assumed as
practically infinite, the heat would be in proportion to the height BD
(by§§7 and 8), and by means of this thermometer the measure of heat
E

.~ ~
! 1l:1

I tl ·e
j
1,D
I
;I
I

Ii ······J;

FIGURE 57

could be defined specifically everywhere. For if the instrument is

immersed in boiling rainwater in a vertical position and the point G
is observed to which the surface of the mercury ascends, then any other
degree of heat whatsoever which has been observed to have sustained
the mercury up to the point D will have to be defined, and in any case
that heat will be to the heat of boiling water as BD is to BG. And
since the ratio BD to BG is constant, whatever the height BG should
be, the same degree of heat that we are discussing can be easily
imitated in any location. Furthermore, BG could be divided into a
hundred or a thousand parts, and the height BD could be defined by
partitions of this kind .
I say nothing about the methods of obtaining more sensitive ther-
mometers of this sort; whoever wants to will easily think up many of
them. Care should be taken, however, that the height BE be not
below 4 feet; on the contrary, it should be larger if one intends to deter-
mine the degree of heat of other boiling fluids, which is often greater
than in water. If smaller thermometers of this kind are desired,
 HYDRODYNAMICS, CHAPTER X

these could be made so that at the time of sealing at m a glass ampule

AF is put into the fire of the torch for rarefying the air contained in it,
and then let the sealing be made immediately, and lest any delay be
introduced in the sealing, the glass ampule could be drawn previously
into a capillary tube which melts together easily when it is brought
close to the flame. In this manner I have obtained thermometers not
more than four or six inches long, but oflittle value. Furthermore, it
is of great importance that the space ED be vacant of all air, as much
as this can be done, and we shall not be sure enough of this vacuum
when we shall have seen that at a horizontal position of the instrument
the mercury reaches the extremity E, because it can happen that the
air which was previously in the space ED retracts into the pores of the
mercury and again occupies the original space when the mercury
descends; the test will be safer by moving the part DE towards a
flame; for if the surface D does not change its position from the heat
of the flame, this will be a certain proof that the space ED is vacant
of air.
§10. In the preceding paragraph we have considered the space
AmF occupied by the air as practically infinite in proportion to the
space DG or DE; but since it will be only eight or ten times larger, it
will not yet be permissible to consider it as infinite without notice-
able error; and from here I suppose that some error has arisen in de-
fining the expansion of moderately warm air in the Amontonian
experiments.
Therefore, in order that the experiment be performed most accu-
rately, one will have to proceed as follows: Let the lower surface of
the mercury be at AF, and let a horizontal [lineJ be drawn at AL;
then, for defining any degree of heat whatsoever, let the instrument
be inclined until the surface of the mercury is at the point g (which is
the same one at which the mercury remained from the degree of heat
of boiling water at the vertical position of the thermometer), and then
let the measurement of the vertical height gh be taken, which will be
in fact to the height GB as the expansion of the air, the heat of which
is to be defined, is to the expansion of air as hot as boiling water.
Thus, therefore, the heats will be accurately in the ratio of the height
gh. Before I discontinue this argument, it will be convenient to have
noted (in case perhaps to some it will seem that the.first degree of heat,
which was defined by us, taken from boiling water is not always and
everywhere completely consistent with itself) that instead of the heat
of boiling water also a thermometer of certain and fixed measures can
be made if in the experiment the density of the air is measured or its
specific gravity is noted together with the barometric height. For if
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 233

the thermometer is inclined until the surface of the mercury is at g,

and at that time the height of the barometer is 28 Paris inches, and a
cubic foot of the air in which the thermometer is located has a weight
of 600 Nuremberg grains, the vertical height gh could be considered
as the first degree of heat. But if at a different location and time the
height of the barometer is 29 Paris inches and the weight of a cubic
foot of the air surrounding the other thermometer (in which it is the
intention to define the first degree of heat) is 500 Nuremberg grains,
and finally the surface of the mercury in the thermometer is again at
g, the vertical height corresponding to the first degree of heat will be

:~~::: ~ · gh. In using the thermometer, let the instrument always be

inclined until the surface of the mercury is at g; I wanted to add this
method so that it be apparent how easy it is in theory to give a fixed
measure of heat; but in practice I shall prefer another much easier and
sufficiently accurate one to this.
§11. Let us come now to considering the atmosphere of air that is
influenced not by an overlying foreign weight, but by its own mass.
But first we shall examine the pressures of vertical air columns and
their equilibria between one another as well as with the mercury
columns in barometers. Second, we shall investigate the elasticities
of air at different heights of the atmosphere above the sea and the
corresponding barometric heights; and these [things] having been
set forth, we shall satisfy many other phenomena pertaining to the
changes of the atmosphere.
12. Let there be two vertical pipes AC and BD (Fig. 58) of equal
area, and each one of indefinite height. Then imagine narrower
horizontal pipes ab, cd, if, gh, lm, etc., practically infinite in number,
open on both sides and connecting to the vertical pipes. Assume
furthermore that the air particles occupying these pipes are every-
where agitated at the same velocity and so have the same degree of
heat; thus there is no doubt that the bases A and B are equally
pressed and simultaneously the same weight (which, of course, is the
very weight of the indefinite air column AC or BD) lies on top of them.
One also understands that if one assumes diaphragms at equal
heights such as g and h and imagines the lower air gA and hB to be
absent, even now these diaphragms are pressed equally on both sides,
and that the weights of the air columns gC and hD lying above the
diaph~agms are equal. Therefore, if the weight of the entire air
column AC or BD is called A and the weight of the air column gC
or hD is taken as B, the weight of the air contained between A and
g or Band h will be A - B, the weight lying above the base A or B
2 34 HYDRODYNAMICS, CHAPTER X

equals A, and the weight lying above the diaphragm at g or h equals

B.

§13. But if the particles in the pipes AC and BD are agitated at

unequal velocity, the situation will be different; nevertheless, what-
ever difference of velocities and temperatures be assumed in the
individual locations, it is evident that nonetheless the parts of a pipe
located at the same height will be pressed equally on both sides, such
as at g and h, and that hence the diaphragms, if they are assumed to
be located at the same height on both sides, will sustain equal pressure.
For if one says that the pressure at g is less than at h, there will be
nothing which may hinder the flow of air from BD into AC through
the small transverse pipe hg, and thus this statement will contradict
the state of permanence which we have supposed.
Therefore, since places located at the same height are equally
pressed by the air lying above, the densities in arbitrary homologous
places, such as at g and h, will be (by §6) approximately in the in-
verse-square ratio of the velocities at which the particles are agitated
in these locations.
§14, It follows from the preceding paragraph that in every place
the pressure of air is the same at equal heights above the surface of the
sea if the atmosphere is assumed to be in a permanent state of equili-
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 235

brium and agitated by no winds, whichever be the difference in heat

between the different parts of the atmosphere. Therefore, it is
obvious that everywhere on earth, at the equator and at the pole, the
height of mercury is the same in barometers which are located at the
surface of the sea or at equal heights above the latter if the atmosphere
is not exposed to changes. I state further that the water terminated
by the surface of the sea is located according to a common equili-
brium, not because this is completely necessary, but because so far no
difference has been observed; in reality, the currents (Les courans ) of
water in many places of the ocean, which are perpetually directed
toward the same region, show that this hypothesis is not to be accepted
with all rigor.
§15. I have already noted that the density of air at any position of
the vertical pipes depends upon the corresponding heat; and since the
degrees of heat can be different, equilibrium being retained, the
densities can also be different. Therefore, let the densities be taken
equal to D at g and equal to 8 at h; and let there be assumed on both
sides two strata of equal and infinitely small height dx, with the height
Ag or Eh taken equal to x. So the weight of the air column Ag will be
f f
D dx and [that] of the column Eh will be odx, and in this manner
one can define the weight of both the entire column and any part
[of it]. Meanwhile it is apparent that the nature of the matter
requires least of all that the weights of the columns AC and ED, or
Ag and Eh, or finally gC and hD are equal to one another, although
(by§r3 ) the pressures at the bases A and E as well as at the diaphragms
g and h are equal to one another. It will perhaps be amazing to
some at first consideration that it can happen that the base A sustains
another pressure than that which is the weight of the indefinite air
column AC lying above it, since indeed, with everything remaining in
its state, as we have just seen, the individual orifices a, c, e, g, etc., can
be conceived as closed, in which case there is certainly no doubt that
the pressure at the base A is the very weight of the overlying air
column. But anyone may examine this point for himself in the
following manner: let us assume either column to be of finite height
(for although they rise without end as long as the particles maintain
some motion, they will nevertheless be terminated if those particles in
the uppermost part of the columns have no motion and thus form a
simple heavy fluid without any elasticity). With this assumption it is
apparent, first, that either column rises to a common height after the
transverse tubes which are everywhere present have been opened;
second, that the uppermost layers are equally dense on both sides
because they are positioned at equilibrium and have a common
 HYDRODYNAMICS, CHAPTER X

height. For this is now obvious why it would not be possible to con-
sider the transverse tubes as closed, which I undertook to show. Also,
it is evident by itself that the pressures are everywhere proportional to
the weight of the uppermost layer, from which it follows, as was
already shown in §13, that the pressures in either part are equal to one
another at equal heights. If now the columns are not terminated
anywhere, it will be allowable to conclude that the last layers are
loaded on either side by an equal weight, or to assume [that] dia-
phragms [exist] at equal heights, so that hence nothing of the power
of the demonstration is lost.
§16. If, therefore, the mercury descends in a barometer transported
from a lower point such as A to a higher one, g, it does not follow that
the weight of the mercury column which descends in the barometer is
equal to the weight of an air column of the same diameter and the
height Ag, which is so asserted by others. And actually, the rest
being equal, the descending column of mercury will be the same in
wintertime as in summertime, since from that statement it should be
less in a warm season than in a cold season. It will also be the same
in southern and northern regions.
Hence it is evident what ought to be thought of that method which
Mr. Duhamel, in the History of the Royal Academy ~f Science of Paris,
reports was used at some time or other in England for investigating
the ratio between the specific gravities of air and mercury. The
height of the mercury having been observed, of course, at a lower
point, then also at a higher one, they announced the specific gravities
in air and mercury to be as the difference of heights of the mercury in
the barometer was to the height contained between the points of
observation. Even if the air is assumed as of the same density from
the lowermost point of observation up to the other one, it is hence not
permissible to pass judgment on its specific gravity in proportion to
mercury. This is all that one may conclude from the experiment.
Let us consider indeed the entire air shell surrounding the earth and
contained between the two points of observation, and the weight of
this shell will be to the surface of the earth as the weight of the mer-
cury column which descends in the barometer is to its base; this is
manifest from the fact that the sum of the bases A and B sustain
certainly the sum of the weights which the air columns AC and BD
have, and that nevertheless neither base is pressed separately by the
weight of its column; and this must also be understood for the columns
gC and hD lying above the diaphragms located at g and h after the
columns Ag and Bh have been cut off. Therefore, the experiment
does not indicate that specific gravity of the air in which it is per-
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 237

formed, but rather it determines the mean specific gravity of all air
close to the earth; the former is greatly variable, the other without
doubt remains almost constantly the same.
Let us compute the mean specific gravity of all the air that surrounds
the earth. Indeed, from many experiments which have been per-
formed in different places elevated slightly above the sea, it is shown
that a descent of one line in the barometer corresponds to an eleva-
tion of approximately 66 feet. It follows hence that the average
specific gravity of the air is in proportion to the mercury as the height
of one line is to the height of 66 feet, that is, as r is to 9504; therefore,
with the specific gravity of mercury taken as 1, the mean specific
gravity of air will be 0.000105. It is indeed noteworthy that this
average gravity of air is so large; for I am sure that even in the most
raging coldness in this country the specific gravity of the air is hardly
yet as large as we have just shown for the mean state of all air sur-
rounding the earth; and at the equator it will be much less, and,
everything having been thought over correctly, I should not believe
that the average gravity of the air which is contained between the two
latitudes of 60 degrees extends beyond 0.000090; this having been
assumed, the average gravity of the air encompassing the earth from
either pole to the 30th degree (which space makes up a little more than
an eighth part of the total surface of the earth) will be 0.0002 ro,
which is twice that of the most dense air in this country; but at the
pole itself, particularly the Antarctic, the air will be very much
heavier and perhaps almost ro times lighter than water, since it is
very cold and very dense.
§17. Let us come now to changes of both the atmosphere and the
barometer. We shall therefore consider two barometers, both
located at the lowest point of air, the one at A, the other at B, and let
us assume that in either one the mercury is suspended at the same
height. Next let us imagine the air at A to be greatly heated; thus
we see that this very air will be rarefied; nevertheless, no change of the
barometer would be produced, if the air had no inertia against the
motion, even if all air were driven from AC over to ED; but, this
inertia having been assumed, a certain pressure develops in all
regions, and this is most noticeable in the region A. Therefore, the
height of the mercury in either barometer increases with time, and it
increases more at A than at B. The contrary will exist, if at once
some great mass of air close to the barometer A or B were compressed
by cold.
§18. This seems to be the unique cause which can effect some change
in the barometers located at A or B, because, if it is removed, the
 HYDRODYNAMICS, CHAPTER X

bases A and E are always equally pressed, certainly either one by a

weight which is one half of the air columns AC and ED added to-
gether, which sum of weights is indeed constant. If we want to
apply this to the atmosphere, it is to be noted that the bases A and E
represent the lowest points of the atmosphere, which certainly would
be located at the surface of the earth if the air could not penetrate the
inner parts of the earth; but because the situation is different, the
locations analogous to the bases A and E will have to be considered
below the surface of the earth.
§19. Let it now be assumed that the barometers are located at g
and h, and let the mercury in both be suspended at the same height;
with these things established, let a cause be imagined to develop by
which the column Ag either alone or together with the other, Eh, is
heated up and expands. From this it is evident that if the inertia of
the air is practically null, the pressures of the air at g and h will
increase, because a larger quantity of air than before now lies over
these places; of course, the weight of all the air which was pushed
upward from Ag and Eh by the heat was involved. And in order to
indicate this by symbols we will set A equal to the weight of the
column Ag before a new degree of heat will have developed, a to
that of the other, Eh, E to the weight of the column gC, f3 to that of
the column hD, C to the weight of the rarefied column Ag, y to the
weight of the likewise rarefied column Eh, l to the height of the mer-
cury at g before the expansion of the air Ag and Eh, and x to the same
height after that expansion; we will then have this analogy:
E + f3: l = E +A - C + f3 + a - y: x;

from this
_E+A-C+fJ+a-y_l
X- E+fJ .

Therefore, the mercury ascends, by having rarefied the lower air,

A-C+a-y A-C
through the height x - l = E + f3 -l = - E - · l (setting
everything equal in either pipe).
But with the air again cooling in Ag and Eh, the mercury descends
again in either barometer.
Here is to be noted that in this manner from a very small change of
heat at Ag and Eh a noticeable variation in the barometer can de-
velop on account of the tremendous density of air in the lower parts,
by which it can happen that much more air is contained in the part
Ag (finally, an infinite number of times as much if air pressed by an
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 239

infinite force is assumed to be compressed into an infinitely small

space) than in the remaining gC, although it is infinite in length.
Whence, if the weight A is very much larger than the weight B, and
simultaneously the cause rarefying the air is maintained, the weight
C follows a given ratio to A; since this usually happens, it is apparent
that the ascent of the mercury on account of the least degree of heat
being added at Ag can be arbitrarily large.
Equally, if it is assumed that the parts Ag and Bh are very much
narrower than the areas at gC and hD, it is recognized that the
variations of the barometer due to the increased or decreased degree
of heat at Ag and Bh thus become less noticeable, because the weights
A and a and C and y, proportional to the former, decrease in this
manner; nevertheless, the barometric variations which originate from
this cause can still be understood to be arbitrarily large.
§20. When this is thus considered, it happens indeed that the baro-
metric variations are for the most part to be sought from quick
changes of temperature in underground caves. It has been known
for a long time that there are many caves of this kind, and that they
are immense; even in solid ground, pores can make something like a
cave. If one collects all cavities (both those which are formed by
caverns and those formed by air-containing pores) up to a height of
20,000 or 30,000 feet below the surface of the earth and compares
their capacity with the solid part of the earth's crust of the same
height, and if one assumes the latter to be a thousand or a hundred
thousand times larger than the former, then this will now indeed be
reason enough to explain the very large changes of the barometer.
From the preceding paragraph I believe that these things will be
clear to anyone.
By the way, places which are closer to caves will be the more
exposed to winds and changes of the barometer because of the inertia
of the air to motion, which is perhaps the reason why toward the
equator, where almost everything is deep sea, smaller variations are
observed on a barometer than in these northern places.
§21. From the same source it is deduced that aqueous exhalations
from the pores of the earth can also contribute something to baro-
metric changes, but this will certainly be little. For if the vapors
would give as much water as can fall in the most severe rain, then the
mercury will hardly ascend a single line in the barometer, apart from
the fact that this matter does not occur so fast but that the effect of it
is distributed almost evenly over the entire atmosphere, and thus
vanishes completely for a certain definite location. For if we con-
sider the entire atmosphere which surrounds the earth, I have noticed
 HYDRODYNAMICS, CHAPTER X

that it certainly cannot be loaded unevenly with vapor. Undoubtedly,

I would prefer the reasoning given in §20 to all others; indeed,
motions of the earth, which are often noticed up to a hundred miles
at the same time, and other phenomena of this kind, indicate that
large and quick changes can occur in the interior of the earth.
In order to explain barometric changes, there is first of all some
sudden cause required; indeed, I have already mentioned that the
slow ones, which I have distributed over the entire mass of air, are of
no effect, and I have demonstrated this in §14. And for this reason
the changes which occur immediately in the atmosphere above the
surface of the earth are to be considered as of little importance.
§22. And this also seems to be the cause of why the moon, which
has such a great effectiveness in agitating the waters of the ocean,
exerts no effect on a barometer that anyone has .b een able to notice
during very careful observations. And if the remaining causes which
may produce some change somewhere in the atmosphere would also
act gradually, one would notice constantly, without doubt, the same
height of mercury at all points equally distant from the surface of the
sea. This height can be called the mean and will be determined
approximately in the manner which Johann Jacob Scheuchzer used,
by observing daily the barometric height through a long period of
time and taking the mean of all of them.
And, having used this procedure, that most famous Author stated
the mean height from the many observations which had been trans-
mitted to him from many places.

At Padua 27 inches 11½ Paris lines

Paris 27 9½
Turin 27 " 1¼ "
Basel 26 IO¼ "
Zurich 26 " 6½ "
On the "
mountain 21 27½
(St. Gotthard) " "

§23. It is known that the diversity of these mean heights stems from
the unequal elevation of the places above the sea. Indeed, in Pascal's
time experiments had already been performed on the descent of
mercury in a barometer which is carried from a lower position to a
higher one. Thence the philosophers inquired into the mutual pro-
portion between cause and effect. Several rules were produced on
this matter from various authors. The foremost of them, to which
many people still cling, is that the heights of the locations are propor-
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 241

tional to the logarithms which correspond to the heights of the

barometer. This rule is established principally on the fact that the
density of air is everywhere proportional to the weight of the air lying
above it; but this principle is wrongly applied here, since it is valid
only for air of constant temperature and is not a definite matter at
every height of air, although it exists in the same vertical column; but
if it were so that the heat would be equal, then it must be acknowl-
edged that thus the rule behaves correctly enough.
But experiments are clearly contradictory to the rule; therefore,
the same degree of heat does not exist everywhere in the entire height
of the vertical air column. In order to make this plain, let me add
now some experiments, performed accurately, as I believe, but none-
theless at different times and locations, which I regret; experiments
performed at the same time on the same mountain, only at different
elevations, would aid our undertaking more; such experiments,
however, exist, as far as I know, only for moderate heights of places,
with all the circumstances which one has to know.
I. At a height of 1070 Paris feet above the surface of the sea the
barometer descends r 6½ lines, when at the surface of the sea it would
hold 28 inches 4i lines (others set it simply equal to 28 inches, but in
the papers which Mr. Delisle exchanged with me it was obtained as
28 inches 4i lines). Therefore, the elasticity of the air at the surface
of the sea having been set equal to unity, as I will always do from now
on, the elasticity at the higher point which I shall designate by E was
found equal to 0.9520.
II. At a height of 1542 Paris feet above sea level the mercury, which
at sea level was clinging to a height of 28 inches 2 lines, descended
2 r ½lines in the barometer; here, therefore, E = 0.9364.
III. On top of the Peak on the Island ofTenerife, 13,158 Paris feet
above sea level, the mercury reached a height of r 7 inches 5 lines,
while at the surface of the sea it maintained a height of 27 inches
IO lines, whence at that place E = 0.6257.
IV. If the descents of mercury are observed accurately at smaller
heights, it is found that a descent of one line corresponds to a height
of 65 or 66 feet. Therefore, at a height of 65 feet E = 0.9970.
These observations are very widespread; indeed, I have a third from
Mr. Delisle, and it was performed by Mr. Feuillee and presented
before the Royal Society of Science in Paris; and that one is the stumbling
block against which all theories collide that have been brought forth
so far.
§24. In order that it now be evident to what extent these theories
agree with the position of the logarithmic, or of the scale of heights
 HYDRODYNAMICS, CHAPTER X

corresponding to the elasticities, let us set the height of a place above

sea level, to be defined by a certain number of Paris feet, equal to x;
the expansion of the air at the surface of the sea we shall designate by
r, and the expansion of the air at the height x we set equal to E. Let
it be noted further that the atmosphere is now considered unchanged
by us, or at least always similar to itself, so that the expansions of the
air at the surface of the sea and at any height x stand in a constant
proportion to each other. For if only the expansions, not constant
with time, would change unequally at different heights of the atmo-
sphere, no rule could reasonably be invented. This having been set
forth, let us now take the equation ex ln E = x, where the coefficient
will be determined from a single observation; let us use the first
observation and there will be ex ln 0.9520 = ro70, and hence (due to
Vlacqian logarithms) ex = - 50,194. Therefore, for this treatment, if
the logarithmic [method] is to be satisfied, one must set - 50,194 ln E
= x, or ln E_: x_; but, according to the norm of this equation
= __
5o,r94
it is found that E = 0.93 r 7 if x = I 542 for the second observation;
the observation itself, however, indicates E = 0.9364; the difference
between hypothesis and observation is more than a line and a half,
which is clearly noticeable in respect, as usual, to the small difference
in the vertical heights.
Furthermore, if now, for the third observation one sets x = 13,158,
there results from the hypothesis E = 0.5469, while the experiment
indicated E = 0.6257; this difference is too large to be supported by
any logarithmic method, for it amounts to more than two inches and
two lines.
§25. The logarithmic law having been rejected, it follows that the
elasticities at different heights of the atmosphere are not at all pro-
portional to the densities, or, which amounts to the same thing, that
the mean degree of heat is different at different heights. Therefore,
different rules were thought up by others, to whom this defect was
well known; nevertheless, none of those rules can be said to be suited
satisfactorily to Experiment 3 (§23). I think one should hardly hope
to find the true law which nature follows; this, indeed, would provide
a ratio of the mean velocities of the air particles different from [that
provided] by simple interferences. Nevertheless, I have agreed
strongly with a certain hypothesis that corresponds not badly to the
phenomena; but before I proceed to this special hypothesis, I shall
give the curve for any law of velocities.
§26. Let the line AD (Fig. 59) be vertical; the horizontal Q,F
touches the surface of the sea; BF denotes the mean velocity of the air
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 243

particles at the surface of the sea, BM the mean density, and BQ. the
elasticity, which is the same over the entire location of equal elevation.
Then let there be conceived the curves EFH, LMO, PQ.S, drawn
through the points F, M, Q., or the scales which at all elevations such

.µ

FIGURE 59

as BC denote by the ordinates CG, CN, and CR the mean velocities of

the air particles, the mean densities, and the mean elasticities. Now
with two curves given, one can determine the third from the fact that
the elasticities (as experience has also shown and was explained in
§§3, 4, 5, and 6) are approximately in proportion to the product of the
square of the velocities just mentioned and the first power of the
densities.
I myself have indeed advised in the above-mentioned place that this
244 HYDRODYNAMICS, CHAPTER X

proportion cannot be exactly true, because air can certainly have an

infinite expansion or can be compressed by an infinite force, but it can
plainly not be compressed into an infinitely small space; nevertheless,
since this property, namely that the elasticities are in proportion to the product
of the square of the velocities of the particles and the first power of the densities,
has been seen to correspond completely to the observations in experi-
ments even in air which is four times denser than natural air, we may
use it without any noticeable error for the natural air of the atmosphere
overlying the sea, and certainly it will be the more true the rarer the
air is.
With these things having been prepared for the calculation, let us
set BF = a, BM = b, BQ = c, BC = x, Cc = dx, CG = v, CN = z,
. cvvz
CR= y, and one will have y: c = vvz: aab, or y = -b-. Because,
aa
furthermore, the weight of the overlying air is a measure of the elas-
ticity, one will have qR( -dy) equal to the weight of the air layer con-
tained between C and c, which is proportional to the density z of the
air and the height of the layer dx; therefore, - dy = Z dx or z = -dn dy;
n X

with this value substituted in the equation (y = c:;n, one obtains

cvv -n dy -dy aab dx
y = aab·~ ory = ncvv
§27. If the velocity of the air particles is taken the same at any
· d e, fcor examp1e, a, t h ere resu lts -
a 1tltu - -
dy = --,
b dx or, aiter
c. t h e require
· d
y nc
. . l C bx
integration, n - = -. But we have seen in §24 that that hypothesis
y nc
is not sufficiently confirmed by experiments. Therefore, having tried
others, I have set v = v1 aa + mx, or vv = aa + mx, which is the law
for motion of freely falling bodies, and this not without success; so
-dy aab dx c aab aa + mx
thus it occurs that - - = - - - - - or ln - = - ln - - - ·
y naac + mncx y mnc aa
In this slightly more general equation in which m and n are still
arbitrary, I further made an attempt to see whether one could not set
aab = 1, and I saw also that this can be done aptly; truly, thus I
mnc
. d c aa + mx c aa + mx y aa Th"
o b tame 1n - = 1n - - - , or - = - - - , or - = - - - . is
y aa y aa c aa+mx
hypothesis indicates that the elasticities of air are everywhere in in-
verse proportion to the square of the velocities at which the air par-
ticles are agitated, or that CR is to BQas (BF) 2 is to (CG) 2 , and since
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 245

by hypothesis EFH is a parabola above the axis AD having a vertex

at a distance aa below the point B, it follows that the curve PQS is a
m

hyperbola. But I have noticed that said distance aa has to be taken

m
equal to 22,000 feet in order to satisfy approximately the observations
of §23. Hence this specific equation now results:
~ 22,000
C 22,000 + X

z aay
But for the curve LMO one finds (by §26) b= cvv, or, because
aa 22,000 = y-, one ob ta1ns,
. a fiter t h'1s sub stitution,
. .
vv 22,000 + X C

2
Z ( 22,000 )
b= 22,000 + X •

§28. In order to show to what extent our hypothesis coincides with

the experiments of §23, in the equation for the elasticities let us set
successively x = 1070, 1542, 13,158, and 65; thus one finds, respec-
tively, -2:'.
C
= 0.9536, ~C = 0.9345, -2:'.
C
= 0.6257, and~C = 0.99705; the

observations, however, indicate~ = 0.9520, ~ = 0.9364, ~ = 0.6257,

C C C

and ~ = 0.9970. The third observation, so very unfavorable for

C

the other hypotheses, agrees clearly with ours, and the others deviate
not more than 6.0019 division, which constitutes three-fifths of a line
in the height of the barometer. But nobody who has experienced
how vague and how little consistent with one another the barometric
observations are will even care about such a small difference. I my-
self, meanwhile, consider this no different from a precarious hypoth-
esis, and I have presented the calculation of §§26 and 27 for no other
purpose than to give the reason by which it can happen that the ver-
tical heights do not correspond to the logarithms of the barometric
heights, as it should occur if the temperature were uniform throughout
the entire atmosphere; indeed, after the calculation has been per-
formed and a comparison ofit has been made with the experiments, it
seemed to me that this matter cannot be sufficiently explained by the
different gravitation of the air particles at different distances from the
center of the earth, such as Newton has attempted by stating that
the gravitations of these particles decrease with the square ratio of
 HYDRODYNAMICS, CHAPTER X

the distances from the center of the earth, which hypothesis affects no
noticeable difference from the hypothesis of uniform gravitation at
heights that do not exceed 13,000 Paris feet. Similarly, I once came
across the opinion that the increased centrifugal force of the air par-
ticles in higher altitudes can contribute something here; but, simi-
larly, after performing a calculation I did not adhere to this opinion
any further. Meanwhile, I do not believe that it is absurd if we say
that the mean temperature of the air is greater, the further it is from
the surface of the sea. But let me wish that it be properly noted
that here we are discussing the mean temperature in the free atmos-
phere; for thus it can happen that the real temperature in the moun-
tains certainly does not rise for other reasons; nevertheless, the
hypothesis is not overthrown thereby, since indeed it has already been
shown in§ I 5 and§ I 6 that the weight of the column of mercury in the
barometer is not to be understood as being precisely equal to the
weight of the air column taken in that region, but equal to the mean
weight of all columns surrounding the earth; therefore, I think
accordingly about the different densities.
§29. If the temperature were everywhere the same, the densities
would be proportional to the elasticities, as far as can be noticed,
and the vertical heights would correspond to the logarithms of the
barometric heights. But I state this to oppose the experiments;
nevertheless, I would not believe that at two places spaced only a
little apart from each other a noticeable difference of temperature
can occur, because heat is quickly distributed uniformly in a body of
rather small density, such as air, unless a perpetual cause exists which
cools the air in the vicinity.
But the situation is different in more remote locations, and indeed
I think it is not absurd to assume the air at the pole as ten times
denser than at the equator, if only the air is accepted in either case
as being close to the surface of the earth; but at great heights the
difference will certainly be less between the density of the air which
corresponds to the pole and that which corresponds to the equator,
other things being equal, and therefore the densities of air decrease
altogether differently away from the surface of the earth, and they
decrease much more at the pole than at the equator; therefore, in
this way it could happen that the real densities of the air at the pole
at small altitudes, let us say, decrease in proportion as (22,000 + x) 4
is to (22,000) 4 on account of the increased temperature, and at the
equator they decrease hardly noticeably, because of the decreased
temperature, which decrease in temperature close to the equator is
confirmed by the fact that the top of the mountain Pico is covered
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 247

with snow through a period of almost ten months, while on the island
of Tenerife it never snows, as they say. Therefore, the mean densi-
ties can be thought of, not absurdly, as being diminished in the ratio of
(22,000 + x) 2 to (22,000) 2 as it has been assumed in §27, while the
elasticities decrease everywhere in the ratio of (22,000 + x) to 22,000;
and certainly these cannot differ at the same height above the surface
of the earth, unless due to causes brought forth by chance and lasting
only shortly.
§30. In countries which lie between the 40th and 60th degree of
latitude, it is probably that the densities decrease in approximately
the same ratio as the elasticities; and for this reason I wanted to per-
form an experiment [to find out] what theory of refractions would
hence arise, about which subject I shall now add something.

DIGRESSION CONCERNING THE REFRACTION OF RAYS p ASSING

THROUGH THE ATMOSPHERE

(o:) It is a well-known property of rays passing from one medium

into another, confirmed by innumerable experiments, that the angle
of incidence maintains a constant ratio to the angle of refraction.
Furthermore, it is also obvious that if the refraction becomes infinitely
small, that is, if the difference of the two sines has an infinitely small
ratio to either sine, the sine of the angle which is contained between
the prolonged ray of incidence and the refracted ray will have the
same ratio to the total sine as the difference of the sines of the angles
of incidence and refraction has to the cosine of the angle of incidence.
But from now on I shall call that angle which I just defined, contained
between the prolonged ray of incidence and the refracted ray, the
differential angle ef refraction. Thence it follows that, everything else
being equal, the sine of the differential angle ef refraction is proportional
to the sine of the angle of incidence divided by the cosine of the same
angle.
(/3) Experiments show further that if a ray passes from air into air
of a density different from the former, the differential angle ef refraction
is, other things being equal, proportional to the difference in densities.
But experiments pertaining to this matter, as much as it is possible,
have been performed most accurately by Mr. Hawkbee on greatly
compressed air as much as on very rare air also, which finally could be
assumed for null. The manner in which they had been performed is
described in the Philosophical Transactions; but the success of all ex-
periments reduces here to the fact that they argue that the sine of the
 HYDRODYNAMICS, CHAPTER X

differential angle of refraction was to the total sine as 5¼ inches is to

2588 feet when a ray passes out of the natural air into a space empty
of air at an angle of 32 degrees, that is, as I is to 6060, and under the
same conditions, with the angle changed from 32 degrees into a half
right angle, as I to 3787 (by §ex). Hence it is deduced that, if a ray
hits a vacuum from natural air at an angle, the sine of the angle of inci-
dence is to the sine of the angle ofrefraction as 3787 is to 3786.
In his Treatise on Optics Newton assumes, instead of this, the ratio of
3201 to 3200, and he deduces it from magnitudes of the refractions
observed by Astronomers; he states, moreover, that the amount of
refraction is the same if the layers refracting the ray are parallel,
no matter in what ratio the densities of the medium decrease, if only
the difference of densities in the first and in the last layer remains the
same (see Newton's Treatise on Optics, page 321, French edition).
Concerning the remainder, the refraction cannot be but greatly
variable under diverse circumstances, since the air that we call
natural is exposed to many changes, as much from heat and cold as
from the pressure of the atmosphere, which both act together in
forming the density of the air, to which density, other things being
equal, the refractions of rays entering a vacuum are proportional.
Mr. Hawkbee has mentioned the same thing in the report of the
experiments which we just discussed, and for this reason he has
defined properly the state of the air that existed when he did the
experiments.
(y) Now let AC (Fig. 60) be the arc of a terrestrial circle drawn
around the center B, in the plane of which lies the ray oflight AG; but
this curved ray AG will be of such a nature that it converges to an
asymptote; let AH be assumed parallel to this asymptote; let the
horizontal AE be drawn, and the straight line AF which touches the
curve AG at A. Thus we may see that the angle HAE will be a
measure of the true height of the star, and the angle FAE will be a
measure of the apparent height, and the angle FAH will be the angle
of refraction; moreover, the angle FAH is the same as the sum of all
differential angles of refraction or of all angles of contact such as the
angle cbo.
Let two elements ab and bo of the curve be considered, and let it be
understood that the arcs exex, f3f3, and yy with the common center B
are drawn through the points a, b, and o; let the density of the air
exexf3f3 be D, and the density of the air f3f3yy be D - dD; then (through
§§ex, f3) the sine of the angle of contact at b divided by the total sine, or
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 249

the angle of contact itself, will be proportional to the difference in

densities dD multiplied by the ratio of the sines of the angles of inci-
dence and refraction, that is, multiplied by!!!.. But if BD is drawn
eo
. 1ar to F'A extend ed , It
perpen d ICU BD b are Iy
. evI"d ent t h at -be an d -D
. IS
eo o
differ, because the ray is almost straight and thus the triangle BDo
can be taken for rectilinear and similar to the triangle beo.

FIGURE 60

Therefore, the required angle FAH will be proportional to J~~ ·dD.

( o) Following these paths, and assuming that the density D is
22,000
everyw h ere - - . na In
- - - G, w h ere x expresses t h e 1Ine . p ans
. 1eet
r
22,000 +X
and G denotes the density of the air at the point of observation, I
found the following: Let the sine of the apparent height of the star be
j, the cosine be F, the radius of the earth be r, to be expressed in
Paris feet; let the number 22,000 be indicated by a; assume further
that the total sine is r, and that the differential angle of refraction for a
ray entering a vacuum from natural air at a half right angle is g;
finally, for the sake of brevity, let 2r - 2a = a, and - FFrr +
2ar - aa = f3; and f3 will be either a positive or a negative number;
it will be positive if the apparent height of the star is small, and
indeed below 2 ° 44 1 , otherwise it will be negative. In the former
case the required angleFAHwill be obtained in this manner: namely,
250 HYDRODYNAMICS, CHAPTER X

let the semicircle MLF (Fig. 6r) be drawn, the radius AM of which

A B
FIGURE 61

a 2f3 - aa
1s 1; let AC= -Ji,
2 r
AB = fi , and let the lines CD and BT
2a r
be drawn perpendicular to MC; and the angle FAH will be

- JFrr Jar Jara· D T

-f3- g
2
+ -f3 g + , I g.
2f3v f3

In the case that f3 is negative, the same angle F AH will be

-Jar JFrr Jara l (a - 2-V~) · (Fr - a + -V~)

- - g + - g + - - _ g ·n - _.
f3 f3 2f3-Vf3 (a+2-Vf3)·(Fr - a - -Vf3)

(E) By means of these hypotheses, by assuming 19,600,000 for the

radius of the earth, for any apparent height of a star one can deter-
mine its astronomic refraction if the value of the angle g has been found
correctly by experiment; but, because it is very difficult to define this
value with sufficient accuracy, it will be wiser to define the refraction
in some particular astronomical case and from this to determine the
other items by calculation. Let us assume, for example, that at a
height of ten degrees the refraction is 5 minutes 28 seconds, which
hypothesis most of the Astronomers of Paris follow. We will find the
following table of refraction [page 2 5 r].
But since the refractions follow the proportion of the letter g, that is,
of the differential angle ef refraction of a ray entering at a half right angle
from natural air into a vacuum, and because this angle is proportional
to the density of natural air, or the air which an observer breathes, it
is evident that even if the air were constantly loaded uniformly with
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 251

Apparent Apparent
height of the Refraction height of the Refraction
star star

o degrees 34 minutes 53 seconds 50 degrees o minutes 53 seconds

5 9 59 55 - 44½
IO 5 28 60 - 36½
15 3 44 65 - 29½
20 2 52 70 - 23
25 2 12 75 - 17
30 I 47 80 - l lt
35 l 29 85 - 5½
40 I 15 go - 0
45 l 3

vapors (which we have neglected so far), nevertheless, it cannot

happen that astronomical refractions are so variable. They will
indeed be larger at the surface of the sea than in the mountains, and
the differences will be noticeable even at moderate heights of moun-
tains; furthermore, they will be greater in a cold season than in a
warm one, and this cause alone can increase refractions at least by a
quarter in this country; finally, the refractions will also be larger
when the barometer is high than when it is low. But if vapors were
ofno hindrance, the refractions could be defined correctly at any time
if the instrument which has been described in §9 and which Fig. 57
represents were applied simultaneously with a barometer; for if one
divides the height of the mercury in the other instrument, one will
have the density of the air, to which, other things being equal, the
refraction is to be made proportional. And I do not doubt but that
the refraction of the sun is less than the refractions of the other stars,
since the heat of the sun expands the air considerably and diminishes
the density of the air.

[End of Digression] - - - - - - - - -

§31. It appears from what has been mentioned about the agitation
of air particles, on which in turn the heat of the air depends, but par-
ticularly from that which has been mentioned in §ro, that the air
possesses the same degree of heat whenever the same ratio prevails
between the elasticity and the density of the former; the barometer
indicates the elasticity; the density we conclude from the specific
gravity of the air; and hence, as we have seen in§ Io, a fixed degree of
heat can be obtained, if the heat of boiling water seems uncertain,
 HYDRODYNAMICS, CHAPTER X

just as Mr. Fahrenheit observed that it depends upon the weight of

the atmosphere lying over it. Instruments which at any single
instant indicate the density of the air can be invented easily and have
been described by many.
It is to be noted here that that ratio just mentioned between the
elasticity of air and its density simultaneously shows the height of
homogeneous air, and since we discuss that height from now on, it
is in order to define the latter correctly before we proceed to other
things.
§32. If we consider a vertical air column of uniform density and
brought to equilibrium with the mercury of the barometer, then the
height of that column will be what I call the height qf homogeneous air
for the given density.
And since the specific gravity of moderately dense air is to the
specific gravity of mercury as I is to 1 1 ,ooo, and since the mean
height itself of the mercury in the barometer for locations slightly
elevated above the surface of the sea is 2½ Paris feet, the height of
homogeneous moderately dense air will be 25,666 feet.
It is evident from this definition that those heights about which we
are talking now are smaller, the denser the air to which the height
must correspond and the smaller the height of the mercury in the
barometer. Therefore, if the degree of heat is the same in the moun-
tains and at the surface of the sea, the height of the homogeneous air
will also be the same in either case, because for the same degree of
heat of the air the density is in proportion to the elasticity of the air or
to the height of the mercury in the barometer. Further, it is apparent
that the height of the homogeneous air at the surface of the sea de-
creases considerably from the equator towards the poles, because the
cold is intensified and the density of the air is increased, with the
elasticity remaining the same, and in the same regions the height is
less in wintertime than in summer.
§33. There are many things which pertain to defining the motion
of air, the solution of which depends upon the height of homogeneous
air; among these also is the propagation of sound and its celerity, for
although the celerity of sound is defined differently by different
methods, we can understand those concerning its propagation in this
way: that at first the celerity seems to be that which is due to the
height of the homogeneous air, then that which corresponds to half
the height, or even to half the height multiplied by the ratio of the
square circumscribed on a circle to the area of the circle; neverthe-
less, all opinions agree on the fact that the celerity of sound is propor-
tional to the root of the height of the air homogeneous with that in
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 253

which it is propagated. If the situation is thus, then sound is pro-

pagated faster in warm air than in cold, with a high barometer rather
than low (to say nothing about favorable or contrary winds); many
experiments have been performed on this matter, partly in Italy,
partly in England, and the latter have shown us that the mean
celerity of sound corresponds to I 140 English feet to be traveled in one
second. But since in one and the same place the height of the homo-
geneous atmosphere is variable, and since especially in this area it
rises, from barometric changes combined with changes of heat,
[relatively] from 3 up to 4, the celerity of sound will be everywhere
variable, even if the winds do not change at all, and in these regions
that celerity will be contained [relatively] within the limits V3 and
V4, or I 73 and 200.
§34. I come now to solving various questions that can be proposed
concerning the motion of air similar to those which we presented
previously on the motion of nonelastic fluids.

PROBLEM

The motion of air discharging from a vessel through a small orifice

into infinite space empty of air, is to be defined.
SOLUTION. From the nature of the question it is apparent that the
local motion of the internal air, which expands itself while a certain
quantity of it flows out through the orifice, is not noticeable; there-
fore, only the potential ascent which an air particle acquires while it is
expelled is to be considered here, and it is to be compared with the
actual descent or rather with the decrease in elasticity which the internal
air has. But in order to reduce the entire matter to our method
applied to nonelastic fluids, we shall consider a vertical cylinder of
common area with the proposed vessel and of the same height as the
height of the air homogeneous with the int~rnal air; but this cylinder,
ifit is considered full of similar but nonelastic air, expels the lowermost
air through the orifice by its own weight at the same velocity at which
the air in the proposed vessel expels itself by its own elasticity. But in
the former case it is ejected at a velocity which is due to the height of
that very cylinder, and hence also in the latter case. It is to be noted
further that the height which we assumed for the cylinder is always
the same, because the elasticity and density of the air are diminished
in the same proportion, but we assume that the temperature is not
changed. Therefore, if the height of the homogeneous air (which
depends on the temperature of the internal air) is called A, then the air
will flow out constantly with the velocity VA. And nevertheless the
254 HYDRODYNAMICS, CHAPTER X

vessel is never emptied, which the calculation shows, because the air
flowing out becomes continually thinner. In order to include this in
the equation, we will set the density or the quantity of air at the
beginning of flow equal to unity; the residual density or quantity of
air after a definite time is x, and the time itself is t; then, since the
velocity is constant, - dx = ax dt, where by a one understands a con-
stant quantity to be defined from the size of the vessel, the area of the
orifice and the height A; hence -dx = a dt, and ln !. = at; moreover,
X X
the value of the coefficient a is found in this manner. Because we
have set - dx = ax dt, at the beginning of the efflux - dx will be
equal to a dt. The first element ( -dx) is now changed into a cylinder
sitting above the orifice as a base; moreover, the height of that little
cylinder will be -L dx, if Lis the height of the cylinder constructed
above the same orifice and having the same capacity as the proposed
vessel; further, this length -L dx is that which is traveled in the small
time interval dt, and because this small time interval is usually set
equal to the distance traveled divided by the velocity, in this case one
will have dt = - : ; \ let this value be substituted in the equation

-aLdx VA
- dx = a dt, and there will result - dx = VA , or a = y·
Thence the final equation is this:

ln.: = tvA_
X L
If one chooses to express the time by a certain number of seconds,
which we shall call n, and if by s is understood the distance which a
movable object travels by falling freely from rest within one second,
then one will have to set t = 2n'Vs, and thus it will occur that
ln _:X = 2nVAs
_L__ _

PROBLEM

§35. The motion of denser air flowing out from a vessel through a
very small orifice into an infinite, thinner, external air is sought,
assuming the same degree of heat in either air.
SOLUTION. Let the initial density of the internal air be D, the
density of the external air be o, the density of the residual internal
air after a given time t be x, the height of the homogeneous air (in
relation to either the internal or to the external air, for it cannot be
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 255

different, if each air is furnished at the same temperature, and thus

the densities and elasticities decrease in the same ratio) be A. At
every point the height of the homogeneous air is sought that has the
same pressure or expansion as the external air and the density of
which is the same as that of the internal air; at the beginning this
height will be 8j;, and after the time tit will be 0:. But it is evident
that the velocity of the discharging air will be such everywhere that it
8A
corresponds to the difference of the defined heights A and
X
therefore, the velocity of the discharging air after the time t is

jA - o:.
Further, the increments of densities ( -dx) are proportional to the
quantities of discharging air, which have a ratio composed of the

j 8
velocity ( A - :), the density (x) and the small time interval
0
(dt); thus, therefore, -dx = aj A - : x dt, where a is a constant
number which, by the method of the preceding paragraph, becomes
I: the significance of this letter as applied there having been retained;
and after this value has been substituted, there develops
dt -dx dtvA
- dx = L V Axx - 8Ax or
vxx - OX= -r-·
After the required integration has been performed this becomes

1 [ vx - vx"="8J. [ VD + v ~ J tvA
n [ Vx + vx - 8] · [ VD - v D - 8] = -Y-'

or, again setting t = 2nvs, as in the preceding paragraph, one will

have:

1
[vx - vx"="8J-LvD + v ~ l _ 2n-vAs_
n [ vx+ V x - 8] · [ VD - VD - 8] - L
§36. COROLLARY I. The entire efflux occurs in a finite time, in
which matter this problem differs from the preceding one; for the air
ceases to flow out when x = 8, and then

n = _L_•ln VD + v ~ .
2vAs VD - VD - 8
 HYDRODYNAMICS, CHAPTER X

For instance, let A be 26,000 Paris feet, let the proposed vessel
contain one cubic foot, and in addition let the orifice have an area of
one square line; L will be 20,736; let it be assumed also that at the
beginning the internal air was twice as dense as the external; moreover,
as is established, s = 15½ Paris feet. Therefore, it will occur that

which indicates that either air will be brought to equilibrium in a

time slightly longer than twenty-nine seconds, and that after this all
effiux will cease. It can also happen, on account of the contraction
which the fluid suffers in front of the orifice (see Chapter IV) and to
which we paid no attention in the computation, that this time is
increased almost in the ratio of 1 to V2.
§37. COROLLARY 2. If one assumes that the air flows out not
immediately through the orifice but through a long pipe, the velocity
will therefore not be changed, if only the capacity of the entire pipe is
as if infinitely small in proportion to the capacity which the vessel
itself has; it seems, moreover, that the density of the air, as long as it is
in the pipe, is the same as the density of the air enclosed in the vessel;
nevertheless-as I shall demonstrate below-the elasticity of the air
in the pipe is not greater than the elasticity of the external air which
surrounds the pipe. It follows, hence, that moved air is denser than
air at rest, but not more elastic; nevertheless, the difference in densi-
ties will also be small; for a wind that moves at about 30 feet per
second will exceed in density the neighboring air, of the same tem-
perature and at rest, hardly by a one-thousand-seven-hundredth part.

PROBLEM

§38. To define the inflow of air through a very small orifice into a
vessel full of rarer air, the same degree of heat again having been
assumed everywhere.
SOLUTION. Let the vessel be completely empty at the beginning,
and after a time t let the density of the internal air be set equal to
x; thus, by remaining on almost the same track which we used m
§35, and with the same notation retained, one finds

dx dtvAl5 t = 2nvs = VA 2 2
Lv~.
or ~ -
L VAD
The number of seconds, therefore, m which the entire vessel 1s
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 257

filled until equilibrium exists between both airs is expressed by v'L ;

· As
and the time of replenishing is twice that in which it would be filled
if the air would flow in constantly at the initial velocity. In the case
when the capacity of the vessel is one cubic foot and the orifice equals
one square line, the repletion occurs in a time of about thirty-three
seconds, unless the repletion is retarded by the contraction of the
inflowing air jet.
§39. We have shown various properties of elastic fluids, whether
moving or at rest; one remains that is not to be omitted by which
elastic fluids differ from nonelastic fluids: namely, that an elastic
fluid at rest possesses a live force, which cannot, like other moved
bodies, elevate itself to a certain height, and indeed we do not con-
sider here the local motion in the former, but that which by its expan-
sion can generate a certain ascent in other heavy bodies. It will be
allowed, I hope, to use in the future the phrase live force contained in a
compressed elastic body, wherein nothing else is understood by this than
the potential ascent which an elastic body can communicate to other
bodies before it will have lost all its elastic force.
It is of merit to note here in advance that just as the descent of a
given body through a given height, whichever way it happens, pro-
duces constantly the same live force in the body, so also an elastic
body or an elastic fluid, after it has been reduced in any manner from
a given degree of tension or compression to a given degree, always
retains in itself the same live force and can again communicate it to
another body by an opposite change.
Let me mention now a little about live forces of this sort contained
in a compressed elastic fluid and a few of their measures; this is an
argument worthy of attention, since to this are reduced the measures
of the forces for driving machines by air or fire or other motive forces
of this kind, of which perhaps several new ones could be developed,
but not without considerable practical mechanical improvement and
perfection.
§40. As we began with air in a vacuum, we shall consider the cylin-
der ABCD (Fig. 62 ), oriented vertically, with the piston EF, which,
devoid of all weight, can be moved upward and downward very freely.
Let air be enclosed in the space EBCF, but assume the entire cylinder
to be located in a vacuum. Let the pressure of the air EBCF be such
that it can sustain the weight p, which will be equal to the pressure of
the atmospheric column if that air is natural. Now let another weight
P appear; thus it will happen that the lid descends to GH and is
agitated by reciprocal motions to the points H and F . In order to
 HYDRODYNAMICS, CHAPTER X

define the motion, let us make use of the common hypothesis that the
pressures of air, other things being equal, are proportional to the
densities.
A .D

FIGURE 62

And thus FC = a, FH = x, the velocity of the piston at the position

GH is v, and the pressure by which the piston GH is forced to further
descent will be P +p - _a_ p, and the force which animates the
a - X

weight lying above the piston must be considered equal to this pres-
sure; therefore, if one divides this force by the mass, one will obtain
the accelerative force, which, multiplied by a time increment, or by
dx, will give the increment dv of the velocity, and thus
V

dv = (p + p - -3:L)
a -
dxj(P + p),
X V

or
a
½(P + p )vv = (P + p )x - ap I n - - ·
a - X

But from the descent of the weight (P + p) through the height x,

the potential live force (P + p )x is generated, and when the piston is in
the position GH, the body (P + p ) possesses the actual live force
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 259

½(P + p)vv, that is, (P + p)x - ap ln a ~ x' which is less than the

former by the quantity ap ln _a_, and this has gone into the com-
a - x
pression of the air.
I say, therefore, that air occupying the space a cannot be condensed into
the space a - x unless a live force is applied which is generated b_y the descent
of the weight p through the height a ln _a_, however that compression may
a - X

have been achieved; but it can be done in an in.finite number of ways. Indeed,
I shall now illustrate this rule by one or two examples.
Let the base of the cylinder be of one square foot, the initial height
FC two feet; in the space BF let air be contained of the type that is
usually the mean on the surface of the earth, which can carry 2240
pounds on the surface EF; let x = l be assumed, in order to obtain
the live force by which two cubic feet of natural air can be driven into
the space of one cubic foot in a vacuum; and that live force will be
2 · 2240 -ln 2 = 3105, that is, as large as is generated by the free fall
of a body of 3105 pounds through the height of one foot. There-
fore, in turn, if one had a cubic foot of air twice as dense as natural
air, then a weight of 3105 pounds could be lifted by means of the
former to a height of one foot in a vacuum while it assumes the
density of natural air.
Further, under the same circumstances let the air be expanded
into twice as much space as it was in before, now occupying a height
of four feet in the cylinder, and let this again be compressed into the
space of one cubic foot, and for this compression a live force will be
required which is expressed by 4 · 1 120 -ln 4, which is twice as large as
the former. Therefore, if one had in a vacuum a cubic foot of air
twice as dense as natural air, then by means of that a weight of fo ro
pounds could be elevated to a height of one foot, while it assumes half
the density of natural air, or a weight of 9315 pounds while it becomes
four times as rare as natural air.
Hence it is a consequence that if air can expand itself into infinite
space and it preserves everywhere an elasticity proportional to the
density, then a finite quantity of air possesses an infinite live force.
§41. However, these things pertain to the estimation of the live
force which is contained in air placed in a vacuum; the computation
becomes slightly different for the denser air which is located in the
atmosphere; here, namely, the maximum degree of expansion cannot
be extended beyond equilibrium with the air of the atmosphere; hence
it is easy to predict in advance, if one has, for instance, a cubic foot of
 HYDRODYNAMICS, CHAPTER X

air twice as dense as natural air, that the live force which can be pro-
duced in the atmosphere from this compressed air is far from infinite.
Other live forces of this kind could also be determined in this manner.
§42. Let the air EBCF be natural and in equilibrium with the
external air; also let the pressure of the atmosphere on the piston EF,
which is certainly in equilibrium with the pressure of the internal air
not yet condensed, be indicated by p. Let the weight P be imposed
onto this piston; let the air now be condensed into the space CECH,
and let the piston, loaded with the weight P, have the velocity vat the
position CH; the remaining notations having been retained, then

dv = ( P +p - - ap- ) -dx; P,
a - X V
or
Pv dv = (p - _!P_)
a -
dx,
X
which, integrated, yields
a
½Pvv = Px + px - apln--·
a-x
But now the live force Px was generated by the descent of the weight
P through the height x, of which force the portion
a
·tPvv or Px + px - ap In - -
a - x
pertains to the same weight moving at the velocity v; therefore, the part
of the live force which transferred to the air is -px + ap-ln _a_,
a-x
which is less than the other one defined in §40.
For instance, let there be a cubic foot of air twice as dense as
natural air; one will find that the live force which this air yields while
it assumes the density of the surrounding natural air is that which is
generated by the free fall of a body of 865 pounds through a height of
one foot.
In the same sense, a cubic foot of air three times as dense as natural
air is found to have a live force such as corresponds to the free fall of a
body of 2898 pounds through a height of one foot, which number
certainly results if, as in §40, one sets p = 2240, a = 3, and x = 2.
§43. It is evident from this correspondence between the conserva-
tion of live forces contained in compressed air and in a body having
fallen from a given height that no advantage is to be hoped for from
the principle of compressing air for improving the use of machines, and
that everywhere the rules shown in the preceding section are valid.
But since it happens in many ways that air is compressed not by force,
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 261

but by nature, or acquires a greater expansion than is natural, there

is certainly hope that from natural occurrences of this kind great
advances can be devised for driving machines, just as Mr. Amontons
once showed a method of driving machines by means of fire. I am
convinced that if all the live force which is latent in a cubic foot of
coal and is brought out of the latter by combustion were usefully
applied for driving a machine, more could thence be gained than from
a day's labor of eight or ten men. For the coal not only significantly
increases the elasticity of the air while it burns, but it also generates
an enormous quantity of new air.
Thus Hales discovered in his vegetable staticks that from a half
[cubic] inch of coal 180 [cubic] inches of air of the same elasticity as
natural air had been generated; therefore, a cubic foot of coal will give
air to 360 cubic feet. But if in §41 the live force is sought which can
be generated from a cubic foot of air 360 times as dense as natural air,
it will be found that the former corresponds to the falling of a weight
of 3,938,000 pounds from a height of one foot; and if, furthermore, it
is assumed that the elasticity of that air is made four times larger by
the heat of the burned coals, then that live force corresponds to the
falling of a weight of 15,752,000 pounds from the same height.
However, it is difficult to devise a machine suitable for this purpose.
There are, furthermore, many natural occurrences which not only
heat up compressed air but can also make the surrounding air more
elastic by warming it; of this sort are quicklime mixed with fresh
water and all fermenting matter; water reduced to vapor by means of
fire possesses incredible force; the most ingenious machine so far, which
delivers water to a whole town by this principle of motion, is in Lon-
don, and the illustrious Weidler has described it. But, above all, the
astounding effect which can be expected from gunpowder deserves
to be considered; indeed, having performed the calculations on some
completed experiments, which I shall add below, I learned that the
elasticity of ignited gunpowder exceeds more than ten thousand
times the elasticity of natural air, and even after everything has been
considered carefully, it is likely that the elasticity of the former is
incredibly larger; let us assume, moreover, that the elasticity of an
expanded blast of ignited gunpowder decreases in the same proportion
as the density; under this -assumption the live force existing in a
cubic foot of gunpowder will be found, if in §42 one sets a = 10,000,
x = 9999, and p = 2240, and if one considers [the live force equal to]

-px + ap l na --a-X,
which quantity thus becomes equal to 183,913,864. Therefore, m
 HYDRODYNAMICS, CHAPTER X

theory a machine is given by means of which one cubic foot of gun-

powder can elevate 183,913,864 pounds to a height of one foot, which
work, I would believe, not even roo very strong men can perform
within one day's span, whatever machine they may use. It is also
probable, as I have said, that the effect of the gunpowder is far
greater; but certainly it is not less, since the calculation is based on
the height to which an iron shot ejected from a cannon can ascend in
a vacuum, in which type of experiment the greatest part of the gun-
powder is wasted.
But these things may be better understood if it is noted that the
same calculation (which we performed before for demonstrating the
effect which arises from condensed air re-establishing itself) appears
also for air which, surrounded by natural air, becomes certainly not
more dense, but nevertheless more elastic from the increased tempera-
ture; so, for instance, every time a cubic foot of ordinary air has
acquired twice its expansion by increased temperature, a weight of
865 pounds can be lifted by means of this to a height of one foot, if
only a perfect machine is applied.
But the effects of all things shown here depend upon both the in-
creased density and increased temperature of the air.
§44. Meanwhile, a live force to be applied for driving machines can
be obtained not only from compressed or heated air, but also from
rarer or colder air. Indeed, wherever the equilibrium is disturbed, a
live force exists which can be applied, if a fitting machine can be in-
vented, for lifting loads and driving machinery. But the method of
determining the live force which can be produced from air of given
density and given temperature occupying a given space, the appro-
priate changes having been made, is the same as that which we have
furnished in §42.
§45. Again, indeed, let the vertical cylinder ABCD (Fig. 63) with
the movable diaphragm EF be considered; assume the air EBCF, as in
§42, as natural and in equilibrium with the external air; also, let the
pressure of air of any sort on EF be called p; next, consider the weight
P, which by means of a rope passed through the two pulleys Mand N
is connected to the diaphragm and pulls the same towards AD; and
thus let the diaphragm arrive at GH from the position EF; finally, one
assumes again FC = a, FH = x, and the velocity of the diaphragm in
the position GH or of the weight in the position Pis equal to v; if with
these assumptions §§40 and 42 are consulted, it will now be evident
that
dv = ( P +- ap- - p) -dx; P
a + X V '
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 263

or
Pvdv = (p - __E_)
a+ X
dx
'
which integrated gives
a+ X
½Pvv = Px - px + apln--·
a
But again, by the descent of the weight P through the height x a
live force Px was produced, during which time, meanwhile, the weight
itself, moved with the velocity v, possesses a live force of only ½Pvv, or
Px - px + ap I na+
a
X
--. Therefore, the live force which remains,

A D

4·

FIGURE 63
 HYDRODYNAMICS, CHAPTER X

namely px - ap In a + x, was transferred into the air, and, with

a
equilibrium restored again between the internal and external air, the
live force could be transposed arbitrarily to other bodies. There-
fore, if one has the space CECH full of air, the density of which is to
the density of the external air as CF is to CH, then the live force
px - ap In a + x will be in effect.
a
But whether that live force is contained properly in the external or
internal air is just a play on words; it suffices that from the disturbed
equilibrium between either air such a live force can be obtained while
the restoring process is acting.
For instance, let a cubic foot of air twice as rare as natural air exist,
to which hypothesis there apply the conditions p = 2240 pounds,
a = ½ foot, and x = ½foot, and the live force which is being dis-
cussed will be 1120 - 1 r 20 ln 2 = 344, that is, the same as is gener-
ated by the free fall of 344 pounds from a height of one foot.
If a cubic foot is filled with air four times as rare as natural, the
required live force will now be (having set, namely, p = 2240,
a = ¼, and x = ¾) 1680 - 560 ln 4 = 904, or as great as that which
arises from the free fall of a weight of 904 pounds through a height of
one foot.
Finally, if a cubic foot completely void of air is considered, one must
set p = 2240, a = o, and x = I; and thus the required live force will
be 2240( r - o In i-) ; but it is well established that o ln i- is infinitely
small as compared to unity; therefore, that number is 2240, which
indicates that by this live force 2240 pounds can be elevated to a
height of one foot.
§46. To the present argument the astounding force of greatly com-
pressed air pertains, but particularly of a blast of ignited gunpowder
in the employment of pneumatic rifles and cannons. Let me add to
this chapter what I have commented separately about these matters.

CONCERNING THE FORCE OF COMPRESSED AIR

AND A BLAST OF IGNITED GUNPOWDER FOR
PROJECTING SHOTS IN THE EMPLOYMENT OF
PNEUMATIC RIFLES AND CANNONS

I. Let AG (Fig. 64) be the length of the barrel in a cannon or

rifle located horizontally, and let it be called a; let AC denote the
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 265

length of the space which the compressed air or the cloud of ignited
gunpowder occupies at the beginning of the explosion, and let
AC= b, and the weight of the shot E to be ejected equal to unity; let
us further assume that the shot fills out the cavity of the barrel

exactly and is moved very freely in the latter; the density of the com-
pressed air in the space AD is to the density of natural air just as n is
to r; finally, let the weight of the column of mercury (the base of
which is CD and the height of which is the same as in the barometer)
be P. Moreover, let us use the hypothesis, whether the shot is pro-
pelled by compressed air or by a blast of gunpowder, that the force of
that propelling fluid is proportional to the density.
These things having been prepared for the calculation, we shall
consider the shot in the position e, by setting Ac = x and the velocity
of the shot at that location = v; thus the force propelling the shot at
the position ewill be (n: - I )P, which, divided by the mass r and
multiplied by the element of space dx, gives half the increment of the
square of the velocity; hence it occurs that v dv = (n: - I )p dx, or

½vv = (b - x + nb In i)P. If one sets x = a, the height is obtained

which is due to the velocity at which the shot is exploded; let that
height be called a, and therefore

a= (b - a + nb In i)P.
II. For instance, let the length of the barrel in a pneumatic rifle,
or a, be 3 Paris feet, the length AC be 4 inches, the air contained in
AD be r o times denser than natural, or n = Io, the diameter of the
barrel or the shot to be ejected be 3 lines, and its specific gravity be
in proportion to mercury as IO to r 7. Therefore, P will be 286, more
or less, and hence it is found that a = 2788, an indication that the
sphere is being ejected at a velocity by which it can ascend in a vacuum
 HYDRODYNAMICS, CHAPTER X

to a height of 2788 feet. From the preceding formula is understood

that the most vehement thrust of the shot occurs from the same
quantity of elastic blast if the length of the barrel is made equal to
nb. But if attention is paid to the other hindrances which the sphere
encounters in its transit through the barrel of the Rifle apart from its
inertia and the resistance of the external air, it appears that a much
smaller length of the barrel is required for producing the most violent
thrust. If the length nb is much larger than the length a, which is
so in stronger thrusts, then one will have, without noticeable error,
a = nbP In~-
b
If the cannon is erected vertically, the calculation becomes some-
what different, but for more violent thrusts the difference cannot be
noticed. Therefore, because from now on we shall consider only very
vehement thrusts, for the sake of brevity we will set a = nbP In i·
III. Just as we have determined the height due to the velocity at
which the shot is exploded in the preceding from the given elastic
force of the blast ejecting the sphere, so it is evident in turn that from
that observed height the elastic force of the blast can be deduced; thus,

As a consequence, the elastic force of gunpowder can be, if not

accurately defined, at least reduced to limits which it will certainly
exceed. But one may ask how the height a can be determined in an
experiment; to this I answer that it can be accurately enough under-
stood from the time which a shot ejected vertically upwards takes
from the instant of explosion until it has fallen to the earth, with the
air resistance taken into account in the calculation. Let me tran-
scribe here the experiments reported in the Commentaries of the Imperial
Academy of Science of St. Petersburg, Book II, pp. 338-39, the calculation
of which I performed after the hypotheses had been formed, consider-
ing air resistance, that the specific gravities of iron and air are as
7650 is to I, and that the air in which the shot ascends is of uniform
density; the ratio of the specific gravities seems to have been assumed
a little higher than was appropriate, but in very high thrusts the
error will be compensated by the decrease of the densities of the air
towards the higher altitudes.

The location of the cannon was adjusted with all accuracy to the
perpendicular, and in every case it was reset to this position and
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 267

fixed; the individual experiments were repeated; the length of the

barrel was, moreover, 7. 7 English feet, the diameter of the shot was
0 .2375 foot; the diameter of the barrel was not measured, nor was
the size of the touch-hole; instead, every quantity of gunpowder
used was weighed, and with a pendulum the time was defined
from the instant of explosion to the instant at which the sphere fell
to the earth; the following table shows both what has been ob-
served and what has thence been deduced by calculation:

Quantity Time of Height of Time of Time of Height of Time of

of gl.lll,,- ascent thrust in ascent in descent in thrust in a ascent and
powder, and resistant resistant resistant vacuum descent in
descent, air from air from air from from cal- a vacuum
calcula- calculation, calculation, culation, from cal-
tion, culation,

Expressed Observed In In In In In
by the in English seconds seconds English seconds
number of seconds feet feet
Holland
ounces

I II III IV V VI VII
½ II 486 5.42 5.58 541 I 1.6
2 34 4550 14.37 19.63 13,694 58
4 45 7819 16.84 28.16 58,750 12[

For the same cannon and the same shot, but with the former
being shortened by one foot and seven-tenths, such that the residual
length of the barrel is precisely 6 English feet, the following table
serves, constructed by the same rule.

I II III IV V VI VII
t 8 257 3.95 4.o5 274 8.2
2 20.5 1665 9.74 I0.76 2404 24.5
4 28 3187 12.5 15·5 6604 4o.5
6 32 .5 43°4 13·9 18.6 11,810 54·3
8 38 5643 15·54 22.46 22 ,394 74

There are many things which render the success of those experi-
ments doubtful insofar as there is nothing which proves the same
elasticity of the blast. I myself would believe the greatest discrep-
ancy to arise from the fact that a very small part of the powder is
ignited immediately at the beginning of the explosion, that then a
large part is set on fire when the shot is just close to the orifice of the
cannon, and that finally the largest part is ejected not yet ignited;
perhaps this single reason causes the elastic force of the blast pro-
pelling the shot to be a hundred times as large as that which results
 HYDRODYNAMICS, CHAPTER X

from the experiment when no account has been taken of this matter;
this seems very probable to me from the fact that, with 4 ounces of
powder used in a 7. 7-foot-long gun, the shot could ascend in a
vacuum by this very thrust to a height of 58,750 feet, while for the
same quantity of powder and the same cannon, but shortened by
1. 7 feet, the thrust would correspond to a height of 6604 feet in a
vacuum, which height hardly exceeds the ninth part of the former.
From comparison between the two experiments I conclude that the
largest quantity of powder was ignited in the longer cannon while the
shot was close to the orifice, in fact not further away from the latter
than 1. 7 feet.
The thrust of the shot is also diminished by the size of the touch-
hole as well as by the opening which is left between the sphere and the
inner surface of the barrel, through either of which a noticeable,
useless part of the blast vanishes; however, not as large a decrease
arises thence as that which I had presumed before the calculation
had been performed; nevertheless, let me add the calculation in the
following, in order that a method be available for stating the very
outer limits for the force of gunpowder, which it will certainly
exceed in any case.
IV. The one which displays the greatest elasticity of the blast is the
third experiment, performed with the gun not yet shortened, which
shows that the shot could have risen by the impact received to a
height of a = 58,750 English feet. But the length of the barrel AG
was a = 7. 7; the length AC ( as much as I conclude from the area of
the barrel and the gravity of the gunpowder) was 0.08. Finally, the
value 26.8 is found for P itself (or for the weight of a mercury column,
the base of which is a great circle of the shot and the height of which
is 30 English inches, in proportion to the weight of the iron shot,
designated by unity), the specific gravity between mercury and
iron having been taken as 17 to Io; since, according to §3,
n= a/ (bP In i) approximately, there results n= 6004. Whence
it follows, if the blast of ignited gunpowder has an elasticity propor-
tional to its density, that the maximum elasticity of the former is at
least six thousand times as great as the elasticity of ordinary air.
V. But if we now consider the useless part of the blast which es-
capes through the touch-hole and the aperture left by the shot, we
will find a greater elasticity. Since the calculation which is required
for solving this question is not a little lengthy and is intricate, I did not
hesitate to apply slightly more liberal hypotheses, by which it be-
comes much easier; although the hypotheses themselves are not true
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 269

in all rigor, they can nevertheless not produce any noticeable error.
First, let me assume that either aperture through which the blast can
escape is practically infinitely small in proportion to the area of the
barrel; with this assumption the velocity at which the blast escapes
can be estimated at any instant directly from the pressure alone; but a
hypothesis of this kind can be formed without any noticeable error for
the entire fluid, even when the openings are not very small at all,
as we have deduced the corollary from our theory, and anyone will
see quickly that it can be assumed much more easily in a very elastic
fluid from the fact that the increment of the potential ascent in relation
to the internal motion is much less in proportion to the potential ascent
of a particle springing forth from the orifice in a fluid which is ex-
pelled by its own elasticity than that which is ejected by force of
gravity; for in the former case the local internal motion is less than
in the latter. Second, [it can be assumed] that the elastic force of a
blast of ignited gunpowder is such that it is not worthwhile to con-
sider the counteracting pressure of the atmosphere; third, that the
velocity of the shot in the gun, although very large, can nevertheless be
considered very small in proportion to the velocity at which the blast
escapes through either aperture, because indeed the inertia of that
blast cannot be not very small in proportion to the inertia which the
shot possesses; by virtue of this hypothesis the blast will escape through
either aperture at the same velocity, since otherwise, the velocity in
the touch-hole having been set equal to VA and the velocity of the
shot equal to v, the velocity of the blast in the aperture left between
the shot and the surface of the barrel would have to be called VA - v.
I now come to the solution.
VI. First it is to be noted that if the elasticities of the blast are con-
sidered proportional to the densities, the blast will escape constantly
at the same velocity through either aperture, as we have seen in the
problem in §34, and that this velocity will be nominally the same as
that which is generated by the height of the homogeneous gas, the
weight of which can prevent the contained blast from expanding.
Hence, the aforementioned velocity will be determined in this manner:
let the gravity of the shot, be unity, the elasticity or the weight which
can keep the blast of powder ACDB just ignited in that state of com-
pression be P, and the weight of the powder used be p; then the
weight of the blast of powder just ignited will also be p; and if the
length AC is set equal to b, it is evident that the height of the homo-
geneous gas which has the weight P will be ~ b. Therefore, the
velocity at which the newly created blast escapes through the
 HYDRODYNAMICS, CHAPTER X

touchhole is J~ b, and it will be ejected at the same velocity during

the entire explosion, and this not only through the touch-hole, but
also, approximately, through the aperture left between the shot and
theoarrel.
VII. Now, in addition let the area of the barrel be F , the aperture
contained between the shot and the barrel be f, the area of the touch-
hole be cp, the length of the barrel be a, and the quantity of the blast
at the beginning of the explosion be g. Let it then be understood
that the shot has come from E to e, and let AC be called x, the residual
quantity of blast in the cannon at that instant be z, and the velocity
of the shot in that position be v; the remaining notation has already
been explained earlier.
Since the elasticity is by hypothesis directly [proportional] to the
quantity and inversely to the space, the elasticity of the blast remain-
ing in AcdB will be zb P; this, indeed, is not all expended for pro-
gx
pelling the shot, but only a part of it, which is to the total as F - f is
to [F]. Therefore, taking dt for an element of time, one has
F - f zb
dv = ----P-dt.
F gx
But by the method shown in §34, where the quantity of air flowing out
in a given time element was specifically defined, one finds

- dz = f + cp · ~ ·
F X
J!:.p b · X dt.

From a comparison of these two equations, one has

f+c/Jg
-dz= - - · -·---dv
vb
F - f b VPp
which integrated, with the addition of the proper constant, gives

Z =
f+c/Jg
g - - - ·-·--·V
vb
F - f b VPp .

Now if this value found for z is substituted in the first equation, and
simultaneously dx is entered for dt, then
V

F - f b f + cp V bP
vdv = - - ·-·P-dx - - - --vdx
F x F xVp '
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 2 7I

or
Fvdvvp dx
-,
(F - f ) -bPvp - (f + cp ) · v-vbp X

which equation after its proper integration, x = a having been

entered, transforms into this:

In i= [-F(f + cp )vvp - F(F - f)-pvPb

-In (1 - (j + cp )v
(F - f)vbPp
)]ju+ c/>)2·VPb.
VIII. Now if that value of v were known from experiment, the
value of P itself which denotes the elasticity of the cloud of gunpowder
not yet expanded could thence be deduced. In order to illustrate
this by an example, let us use the same experiment which we have
shown already in Art. IV, in order that it be apparent therefrom what
increase of elasticity arises from the escape of the blast. Therefore,
the calculation will be performed as follows.
Because we have designated the weight of the shot, which was three
pounds, by unity, the four ounces of powder used will have to be ex-
pressed by --f2 ; therefore, p = f 2 . The measurements of the open-
ings which we consider I have not taken; but usually the aperture
left by the shot in a gun of this kind constitutes approximately a
fifteenth part of the area of the barrel; the area of the touch-hole,
I believe, can be neglected altogether here; therefore, I set F = 15,
f = 1, <p = o. Further, again a= 7.7, b = 0.08; the height to
which the shot could ascend in a vacuum ½vv = 58,750, or v = 343;
therefore, the last equation of the previous article will be this:

-5251
In 96 =-----:;=- + 17.5-ln V-
vP ,
V p p - 300

which is satisfied approximately if one takes VP = 534, and thence

P = 285,156, which equals the weight of a mercury column of the
same area as the barrel on the gun, the height of which is more than
ro,ooo times as large as the height of the ordinary barometer; more-
over, we have found above in Art. IV that the number n (which
signified the same) equals 6004. Therefore, we shall now safely
affirm (for everywhere the things which we have neglected render a
larger force to the powder) that gunpowder possesses an elastic force
at least ten thousand times as large as the elastic force of ordinary
air. Moreover, it is simultaneously apparent from comparison of the
 HYDRODYNAMICS, CHAPTER X

numbers 10,000 and 6004 approximately how much of the force of the
powder is lost due to the often-mentioned openings. Indeed, I would
have believed this decrement to be greater; but I was confirmed by
that calculation in this matter, about which once a man, knowing of
such matters, wanted me [to be] more certain, [since] indeed he had
observed no noticeable decrement in cannons after the touch-hole
was amplified beyond normal by daily use in a siege.
IX. Indeed, in order that from our equation some corollaries can
be deduced that are simpler although only approximately true, we
shall change the logarithmic quantity into a series. This is, indeed,

-ln (1 - (f + cp)v ) = (f + cp )v
(F - f )vbPp (F - f )vbPp
(f + </> )2 vv
+ 2(F - f )2 -bPp
(f + </> )ava
+ , ;-
3(F - j )3 ·bPpv bPp
+

After this value has been substituted in the last equation of Art. VII,
there results

ln ~ =
3
Fvv + F(f + cp )v + ...
b 2(F-f ) -bP 3(F-f )2 -bPvbPp .

We will notice that here this equation agrees perfectly with the last
equation of Art. II if the apertures f and <p are set equal to zero; for
what is indicated here by -½ vv and nP is indicated there by a and P,
the remainder of the notation being identical.
X. This equation will serve in order that it be apparent approxi-
mately how much the height of the thrust is decreased by the aper-
tures if these openings are very small. Let a indicate the height
which the shot can reach in a vacuum if it is assumed that none of the
blast escapes through the openings, and the decrement of that height
about to arise from the eruption of the blast through those same
apertures will be approximately this:

(2a) 312 - (f + </> )/3FvbPp.

Whence in the same gun, with the same quantity of powder used,
and with the weight of the shot remaining the same, the decrements
of the thrusts will be proportional to the areas of the apertures.
The same decrements follow almost in proportion to the square
root of the quantities of powder used, other things equal; because,
 PROPERTIES AND MOTIONS OF ELASTIC FLUIDS 273

indeed, the logarithms of large numbers increase at a much lesser

rate than the numbers themselves, and because, furthermore,
a = bP ln i, one could have stated that, the rest being equal, a is

proportional to b itself, because P is not affected by b. But the

to the quantity or in proportion to the quantity

itself, which denotes the weight of the powder used, is proportional

:t
decrement under discussion, other things being equal, is in proportion
a3 12 /v'bp, butp

to b; therefore, the previously mentioned decrement is approximately

in proportion to v'b, which is the square root of the quantity of pow-
der used. Therefore, from the usual reasoning concerning thrusts,
the decrements are much larger in feeble thrusts than in more violent
ones, and the experiments described in Art. III seem to confirm this
also; I see indeed no other reason why in the first table of experiments
the thrust of a shot in a vacuum, with two ounces of powder having
been used, should have been more than twenty-six times as high as
when half an ounce was used,. and why, after the quantity of powder
was duplicated to 4 ounces, a thrust only four times as high as that
with a quantity of two ounces should result from the calculation.
XI. Whatever other inequalities may appear [by comparisonJ in
either table of the experiments, I derive, as I have said above, to the
largest part from the fact that not all powder is ignited, nor is all that
which is ignited burned at once at the beginning of the explosion.
And I shall certainly not be astonished, since we have investigated
this in Experiment 4, Table r, that the total time of explosion makes
up not even the hundredth part of one second. Therefore, since it is
certain that the largest part of the powder is ejected not ignited, and
since not a small part of the remainder is set on fire more slowly than
was assumed in the calculation, and since, furthermore, a noticeable
part of the powder is adulterated by vapors and earthen material
which does not burn, it follows that the burning parts possess a far
greater elasticity than that which was determined from calculation of
the experiment in Art. X; perhaps it is ten or a hundred times as
large.
But perhaps it is only as large as the experiment has shown, namely,
ten thousand times as large as the elasticity of ordinary air; it follows,
hence, either that the elastic blast which develops from ignited gun-
powder is not ordinary air, or that the elasticities increase in a greater
proportion than the densities; indeed, the density of air that arises
from powder just ignited cannot be more than a thousand times
larger than the density of ordinary air if almost all the powder is
 HYDRODYNAMICS, CHAPTER X

composed of compressed air, which I conclude from the specific

gravity of the powder with respect to the air.
Meanwhile, the question has been treated for a long time whether
the elastic blast which is deduced from bodies is ordinary air or not,
which question I will not decide.
If, nevertheless, one assumes that gunpowder is a thousand times
as dense as natural air and ten thousand times as elastic, then it will
follow from §4 that the air, compressed by an infinite force, cannot be
compressed more than r 33 r times, and according to the same rule
the elasticity of air four times as dense as natural would be to the
elasticity of natural air as 4¼ is to r.
But whether experiments performed by others which make the
ratio of these elasticities exactly 4 to r have been done with sufficient
accuracy, and whether the temperature of the air remained the same
while it was compressed, I do not know. It is also quite likely that
the same blast which lies hidden in the pores of gunpowder is the
cause of the elasticity of elastic bodies or of resilient wool; indeed, as
long as it is present in the small cavities, the elastic blast is compressed
if bodies are driven by some force into an unusual shape, and as long
as it is restoring the most capacious form to the small cavities, the body
is returning to its original shape and length.
 ELEVENTH CHAPTER
Concerning Fluids acting in a Vortex,
while also Concerning Those which
are Contained in Moving Vessels

§1. From the time at which Kepler and Descartes were employing
vortices for explaining various phenomena of nature, many people,
reckoning that they were expending their effort wisely, reworked that
argument eagerly; but, unless I am wrong, Huygens first penetrated
the nature of it correctly in his Traite de la pesanteur; let me add certain
things which pertain to my purpose, perhaps not sufficiently examined
by others.
However, as is customary, one assumes that vortices are reduced to
the state of permanence or of persistence, so that the fluid, subjected to
no change, is moved constantly according to the same law.
§2. Let the cylinder ABCD (Figs. 65 and 66), the axis of which is
GH, be placed vertically and let it be filled to a certain height; let the
water be considered as having been formed into a vortex, and let all
.A D .A G D

FIGURE 65 FIGURE 66
 HYDRODYNAMICS, CHAPTER XI

things be already reduced to the state of permanence. Thus the

surface of the water will be depressed toward the axis and elevated
toward the sides. We will represent a section through the axis
terminated by the surface of the water by the curve EOF, and the
nature of this curve we will now give from the relation which the
velocities have mutually at certain distances from the axis.
Let ga andfn be drawn infinitely close and horizontal, and let am
be made vertical. Let Og = x, gf or am = dx, ga = y , mn = dy.
Moreover, it is evident that any volume element whatever located at
the surface presses perpendicularly to the surface by its own acting
force, composed of the horizontal centrifugal force and the vertical
force of gravity, because, if it would press obliquely, there would be
nothing which would keep the elemental volume in its place.
Therefore, if the centrifugal force of a volume element located at a
is expressed by the horizontal ba and the force of gravity by the ver-
tical ca, and the rectangle abec is completed, the diagonal ae will be
perpendicular to the curve; hence the triangle eca is similar to the
triangle amn, and thus dx: dy = ec: ca = ba: ca, or, as the centrifugal
force is to the force of gravity at the point a.
Moreover, Huygens showed that the centrifugal force of a body
driven in a circular course at the speed which it could acquire by
free fall through a height of half the radius is equal to its own force of
gravity: thus, if accordingly the height corresponding to the gyratory
velocity of the volume element is called V and the force of gravity g,
2gV 2gV
the centrifugal force will be --, from which dx: dy: : - - : g, or,
y y
dx = 2Vdy_
y
§3. If one sets V = ½Y, x will become equal toy, and accordingly the
line EO will be straight, forming a half right angle with the axis CH,
and the cavity will have the shape of a cone. But if the proportion of
the velocities is kept the same, namely, that they are everywhere
proportional to the [square] roots of the distances from the axis, and
the water is driven around more quickly or more slowly, the angle
EOG will be more acute, the more quickly [the water] is moved, so
that, if the velocity would be infinite, then the water would have to
stand perpendicularly to the base, as if it were a wall, and form a
cylindrical cavity in the interior, if only there would be a cover at
AD which would prevent all the water from being ejected.
§4. If it is assumed, a little more generally, that 2 V = jy", dx will
become_be- 1 dy, or, x = (~f- Hence it follows that the curve will
 FLUIDS IN VORTICES OR IN MOVING VESSELS 277

always be concave toward the axis, as in Fig. 65, if e is greater than

unity, and convex, as in Fig. 66, ifit is less. In the first case the angle
EOG is always right, in the other it is always null; only in the case in
which e = I can that angle have any value whatever.
§5. These things can serve for determining to some degree the scale
of velocities in an artificially produced vortex: for, if one sees that the
surface is concave, one will judge correctly that the velocities increase
in a greater ratio than the distances from the axis increase; if it is
convex, one will deduce the contrary. If the curve does not seem to
be of the parabolic type, this will be proof that the velocities cannot
be compared with some force determined from the distances. The
greater the observed line EM terminated by the horizontal OM, the
greater the absolute velocity of the particles, or the letter f
§6. But I think that in a homogeneous fluid a vortex cannot remain
in the proper state through some notable time if the centrifugal forces
of equal portions increase from the axis toward the periphery: for if
this were so, since there is nothing which would sufficiently repress the
centrifugal force of the portions nearer the axis, thus it would occur
that those nearer portions would perpetually recede from the axis,
and those more remote would be propelled toward it, and equilib-
rium or the state of permanence could never be attained under this
2
condition. Hence it appears that this quantity g V (which certainly
y
expresses the centrifugal force of equal portions in homogeneous
fluids) either increases together with y or at least does not decrease,
and thus if we go again to the special hypothesis formed previously
(2 V = ff ), e cannot be less than unity. Therefore, in all vortices
discussed here that have been reduced to a state of permanence, the
surface will never be convex, as in Fig. 66, but always either concave,
as in Fig. 65, or conic; and because e is either greater than unity or
equal to the same, it cannot occur otherwise than that the velocities
increase in proportion either equal to or greater than the roots of the
distances from the axis. When I consider these things accordingly,
I do not understand in what way Newton could assume that two vor-
tices of a fluid everywhere homogeneous are reduced to a state of
perpetual persistence in which in the one the periodic times of the
portions are as their distances from the axis of the cylinder, but in the other as
the squares of the distances from the center of the sphere. For in the first of
these vortices the velocities would be equal everywhere, and in the
other they would clearly decrease from the axis to the periphery.
It is more probable in the majority of vortices which have already
attained the state of absolute persistence that the periodic times of the
 HYDRODYNAMICS, CHAPTER XI

individual portions of either a homogeneous or heterogeneous fluid

will be the same as if the entire cylinder were solid, moreover that the
portions which are specifically heavier will be nearer to the circum-
ference. In this case v becomes proportional toy itself and V pro-
portional to the square of the same, and the curve EOF will be an
Appolonian parabola, the vertex of which is at O and the axis of which
is OG.
I presume especially that things will be approximately so if the
vortex is generated by the rotation of a cylindrical vessel about the
axis HG, or even by uniform agitation with a stick near the sides of a
vessel, the phenomena of which sort of vortices Mr. Saulmon showed
in the Commentaries of the Royal Academy of Science of Paris for the year
1716.
§7. The pressures which the different portions of the cylinder
ABCD sustain from the fluid are proportional to the heights of the
vertical columns corresponding to the same portions; indeed, it is not
required that we add to this weight the impulse of the fluid develop-
ing from the centrifugal force, because that impulse has already
obtained its effect in elevating the water. And if the vessel were not
cylindrical but of some irregular shape, it would be permissible to
assume a cylinder, the axis of which coincides with the axis ofrotation,
full of fluid such that the point O in the proposed vessel as well as in
the fictitious cylinder is located in the same place; for thus, at any
point in the cylinder the pressure will be as great as it is at the same
point in the proposed vessel. It appears from this very thing that the
surfaces of the vortices can be defined from a principle other than that
which we used above: indeed, after the horizontal line OM has been
drawn, and the vertical Na [has been drawn] with pn infinitely close
to it, it follows that the height Na or Og is proportional to the centrif-
ugal force of all the particles which are at ON, and that the difference
of the two neighboring heights, namely, am and gJ, is proportional to
the centrifugal force of the particle Np. Thence the final equation
. §2 1s
. h we presente d 1n . again
. d enve
. d , name 1y, dx 2
w h 1c = -V-dy·
y
§8. Now let us see what must happen to bodies floating on a vortex;
but in order that the matter may be made the more clear and simple,
in place of the body we will consider a small globule of the same
specific gravity as the eddying fluid.
Such a globule united with the fluid is driven immediately by two
forces, the one tangential, drawing its origin from the impetus of the
fluid, the other centripetal, which develops from the centrifugal force
of the fluid. These forces maintain a constant ratio to each other,
 FLUIDS IN VORTICES OR IN MOVING VESSELS 279

namely, the square of the respective velocity of the fluid, whether the
body is at rest or is carried in a circular motion.
Moreover, it deserves to be noted by those who adhere to Car-
tesian principles in explaining the phenomena of gravity that the
tangential force is incomparably greater than the centripetal force:
indeed, the former is to the latter as the distance of the body from the
axis of the vortex is to eight-thirds of the diameter of the globe; the
proof can be seen in the Commentaries of the Imperial Acaderrry of Science
of St. Petersburg, Book II, pp. 318 and 319.
§9. Although I know that many things have been alleged by various
people in order that they might show that a delicate material driven
very suddenly into a vortex can indeed dislodge bodies toward the
axis, on the other hand it does not follow thence that at the same time
those bodies are transferred by the vortex; nevertheless, I was not
able to remove this doubt after I learned that the tangential force is
almost infinitely greater than the centripetal force. Perhaps this
difficulty is obviated no better if we state that there are two vortices,
contrary and of equal strength, about the same axis. For it seems that
most phenomena of nature cannot be conciliated with the hypothesis
of vortices unless we assume that two or more vortices can cross over
one another very freely in any direction whatever: for the common
gravitation alone of all the celestial bodies toward one another, which
cannot be doubted, shows well enough either that one should bid
farewell to the hypothesis of vortices, or that the very free crossing of
several vortices in all directions should be concluded. Therefore, if
two vortices of equal strength were assumed to be contrary and about
the same axis, then the contrary impetuses would destroy the tan-
gential forces of each vortex; but at the same time each vortex would
join in depressing a body toward the common axis.
§rn. Another difficulty occurs in that the gravity of bodies cannot
be sought from the effect of two contrary vortices moving about the
same axis. For thus the bodies would not gravitate toward a com-
mon point or quasi point, but toward the axis, and they would glide
toward the same in a perpendicular motion, which conflicts with the
vertical descent of bodies and the roundness or quasi roundness of the
earth and of celestial bodies.
This other difficulty is also overcome if two axes are assumed per-
pendicular to each other or approximately so, about each of which
two contrary vortices of equal strength are driven. For the force com-
posed of all the vortices can be understood to be so constituted that
a body moves approximately toward the point at which both axes
intersect one another; nevertheless, the earth would always be
 HYDRODYNAMICS, CHAPTER XI

compressed somewhat toward the plane crossing through both axes.

But certainly it will be possible to contend with this inconvenience, if
only it is an inconvenience, by increasing the number of vortices
very much: for if the vortices are considered as almost infinitely many,
they all can cross over themselves with the same facility as rays of
light, which do not impede each other at all.
I wanted to add these things here for the sake of those who are
pleased by vortices, in order that they may see whether this motion
can be conceived more easily than that which Huygens assumed: for
the phenomena of nature can be explained equally by each. I
showed this idea a little more accurately in the dissertation which,
having been rewarded with a prize in the year 1 734, the Royal
Academy of Science of Paris chose to print.
§11. Because no one can doubt that all planets gravitate toward the
sun and satellites toward their own planets according to Newton's
thinking, and that the cause of this gravity is connected with that by
which terrestrial bodies tend toward the center of the earth, the
hypothesis of vortices will have to be extended to the entire system of
the world ifit is applied for explaining the gravity of terrestrial bodies.
Thus, indeed, planets floating in a fine material would be moved in a
resistant medium, and, gradually losing some of their own motion,
they should have to drift toward the center of the sun in the manner of
a spiral: but since this is not apparent from the most ancient observa-
tions, the hypothesis of vortices postulates that an eddying fluid is
assumed, rare and delicate beyond all measure, that is moving at a
velocity which the human mind can barely comprehend: for, the rarer
the fluid, the quicker one must assume the motion to be. Perhaps
the perpetuity of the motions will be explained more by a certain
reciprocal communication of motion, which is such that a celestial
body which has just propelled certain particles is propelled in turn
by them with a similar force.
§12. I now come to the remaining properties of gravitating bodies
which follow from the hypothesis of vortices. Let us consider that a
body which transmits no particles of fluid through its pores is resting
in an eddying fluid; thus the body tends toward the center of the vor-
tex, and its centripetal force will be precisely equal to the centrifugal
force of the eddying fluid which is located within the same volume at
the same distance from the center. Therefore, any bodies whatever
located in the same position in the vortex have the same centripetal
force if they have the same volume, even if the quantities of material
in any one body are unequal in any way whatever, and if bodies of
this sort can be moved freely toward the center of the vortex, they will
 FLUIDS IN VORTICES OR IN MOVING VESSELS 28 I

be carried at unequal velocities, reciprocally proportional, of course,

to the square roots of the quantities of material, if the measured
distances are equal.
§13. Those things which were mentioned in the preceding para-
graph are easily applied to the gravity of bodies, if only the origin of
gravity is the centrifugal force of some delicate material driven very
quickly in a vortex. But because experience shows that all terrestrial
bodies descend at the same velocity in a vacuum, and all bodies sus-
pended by an equal thread make tautochronous vibrations, we thence
conclude that ultimate heavy particles, through which, indeed, heavy
fluid cannot penetrate, are of equal specific density in all terrestrial
bodies, that is, they contain equal quantities of solid matter in equal
volumes, and this no less in the heavy particles which compose gold
than in those which compose feathers. But lest these things be under-
stood other than as I wish, I will have to explain what I mean by
ultimate heavy particles and by the solid matter innate to them.
§14. Therefore, heavy particles are those, having been appropriately
named such, which are impenetrable to a fine eddying material:
indeed, particles of this sort act the same way as bodies placed in a
vortex, which we discussed in§ I 2; nevertheless, although they may be
impenetrable to the fine material just mentioned, I would not believe
that they are perfectly solid, which Huygens seems to have presumed
in his treatise De Gravitate, that is, such that their entire space is
filled with material, without pores or inter-flowing fluid; I think,
rather, that these heavy particles do have their own pores again, and in
them there is some far more delicate fluid which traverses the heavy
particles with the same freedom with which a heavy fluid flows
through observable bodies: but the remainder, which coheres to
itself in the heavy particles, I call the solid material pertaining to the
same particles.
§15. From these [considerations] it is clear that the different speci-
fic gravities of bodies should be sought by no means from the different
density of the heavy particles, but from the fact that these particles can
be unequal in number, or even in magnitude, in different bodies
within the same volume, such that in the more compact bodies, or
those of greater specific gravity, the heavy particles either are positioned
with less interstices or are greater in volume.
But even if the heavy particles should have different specific densities
in different bodies, on that account the bodies would not have different
specific gravities, the remaining things having been set equal; more-
over, such bodies, having fallen from above, should descend toward
the center of the earth at different velocities from one another.
 HYDRODYNAMICS, CHAPTER XI

Therefore, it should be possible that bodies of equal specific gravity

should descend in the vacuum so commonly mentioned at an unequal
velocity no less than we see bodies of different specific gravity de-
scending at equal velocity. Moreover, in bodies of this sort the laws
of motions would be greatly different from what they are nowadays
when masses are estimated from weights alone.
§16, Finally, because all terrestrial bodies, as far as is known from
experience, have their heavy particles of equal specific density, as was
indicated in §13, I may indeed be easily induced to believe that the
same thing occurs in all planets considered separately. But it is
altogether probable to me that the planets compared to each other
have their individual heavy particles of different specific density,
because I do not see any reason why these particles should be similar
in all planets. But in any planet whatever, its centrifugal force or its
attempt to recede from the sun depends upon the density of its heavy
particles. Therefore, it is not yet allowable to infer that the centrifi1gal
forces of the planets are in an inverse square ratio of their distances from the sun
from the fact that the periodic times follow a ratio of the three-halves power of
the distances: for such a conclusion supposes the same density of the
heavy particles on all planets.
§17, The centrifugal forces of planets are certainly equal to the
contrary forces by which they are drawn toward the sun. But
because, as I mentioned in the paragraph above, it is not yet certain
in what ratio with respect to the distances from the sun the centrifugal
forces of the planets are changed, therefore it is not permissible to say
anything definite about their forces of gravity toward the sun. And
indeed there are many things in the hypothesis of vortices which
constitute and determine the forces of gravity at different distances.
For when the force of gravity is equal to the centrifugal force of a fine
material which cannot penetrate the heavy particles of a body, it
follows that the force of gravity is greater, the greater is the quantity
of fine material to which transit is denied; but because we know that
a body, impenetrable to one fluid, often offers the freest transfer to
another finer fluid, it can occur, if only we assume that the eddying
material is unequally fine at different distances from the center of the
vortex, that one and the same planet, at unequal distances from the
sun, is driven unequally toward the sun, which same thing can apply
more easily in different planets because it happens that the structure
of the heavy particles can be different.
In addition to these things are the different density of the eddying
material, the velocity, and the distance from the center, which [all]
contribute to forming the force of gravity. But if they are taken into
 FLUIDS IN VORTICES OR IN MOVING VESSELS 283

consideration, it will certainly appear that the forces of gravity can

decrease with increasing distances from the center of forces. Never-
theless, on account of the fact that the centrifugal forces of equal
volumes of eddying material do not decrease equally, I consider that
the latter [situationJ cannot occur on account of the reasoning shown
in §6.
But let these things suffice which we have discussed generally and
incidentally concerning the nature of vortices and their application
to the Phenomena of gravity: the intention was not to commend the
hypothesis of vortices, but only to draw certain conclusions from it
without which I should believe this hypothesis cannot subsist.
I come now to the other part of the chapter in which we will con-
sider briefly the state of fluids which are contained within moving
vessels; the subject is very fertile and variable in infinite ways. But
we will treat a few matters, or examples, to which many others can
be referred.
§18. If water is contained in a perforated vessel and the same vessel
falls freely, it is self-evident that no water will flow out during the fall
of the vessel, because, certainly, the upper particles do not gravitate
toward the lower. If the vessel indeed descends in an accelerated
motion, but slower than that by which bodies are accelerated natur-
ally in a vacuum, the water will flow out, but at a lesser velocity than
if the vessel were at rest; the contrary will be the case if the vessel is
drawn upward by an accelerated motion. Finally, if the vessel is
borne horizontally by an accelerated motion (for now we will not
attend to the remaining directions), it can happen that the velocity
of the water flowing out is either greater or less than the ordinary
velocity in relation to the position of the orifice. But the velocities of
the water will thus be determined.
§19. For example, let the cylinder ACDB (Fig. 67 ) be full of water
right up to AB, the base CD of which has a very small orifice at E
through which water flows, while in the meantime the entire vessel
is drawn upward by a descending weight P by means of a string
running over two pulleys H and G. Finally it is assumed that as
much water is constantly supplied from above as flows out through the
orifice E; but let the weight of the cylinder and the water contained in
it be indicated by p. Thus it appears that any volume element of
water standing in the vessel is stimulated to ascend by a force which
is in proportion to the natural force of gravity as pp - p is to 1. But
+p
because the reaction of a volume element against the base is equal to
the force by which any volume element is stimulated to ascend, it
 HYDRODYNAMICS, CHAPTER XI

exerts another pressure on the base, in addition to the natural pres-

sure, which will be expressed by pp - P_ But both pressures taken
+p
2
together will be to the natural pressure alone as p P is to 1, so that
+p

FIGURE 67

the base is pressed not at all otherwise by the water lying above than
2
if the cylinder were at rest and the height of the water were p P -AC,
+p
and from this itself it follows that the height due to the velocity of
2
the water flowing out uniformly is p p . AC.
+p
Therefore, if P = o, no water will flow out with the vessel falling in
a naturally accelerated motion; if P = p, the water will flow out at
the ordinary velocity, because then the vessel is at rest; and if P = oo,
the velocity of the water flowing out will be to the ordinary velocity
as V2 is to I.
 FLUIDS IN VORTICES OR IN MOVING VESSELS 285

§20. Now one seeks what must happen to a fluid which is contained
in a vessel to which a uniformly accelerated horizontal motion is
imparted. But it is very easy to see from this alone that now the
inertia of the particles is horizontal or opposite to the direction in
which the vessel is moved, while that of their gravity is vertical. But
each remains constantly the same.
Therefore, after the fluid arrives at the state of persistence or
permanence, its surface will be plane but inclined toward the direction
of motion. Moreover, the angle of inclination will be determined
as follows.
Let there be a cylindrical vessel ACDL (Fig. 68), positioned verti-
cally, which is moved in a uniformly accelerated motion over the

ct~....--

FIGURE 68

horizontal plane CDH by means of the weight P attached to the vessel

at S with the help of the pulley G, and let the weight of the vessel and
the water contained in it be to the weight P as p is to P; let the
natural gravitation be unity; the force of any volume element in the
direction GS, with respect to its own gravitation, will be - p
p.
+p
Therefore, if AB is in the same plane as SG and as the surface of the
water, and if AL is drawn, it is evident that the action of the natural
gravity will be to the reaction arising from the weight P as BL is to
AL, or as I is to - p
p ; and with the entire sine designated as 1, the
+p
sine of the angle LAB will be
p
 HYDRODYNAMICS, CHAPTER XI

Hence it is understood also that the base CD experiences a greater

pressure at C than at D from the water lying above, and this in pro-
portion to the heights AC and BD; and, if the same base is perforated
by a very small orifice, that the water will be ejected at a velocity
which corresponds to the height of the vertical column lying above.
Thus, indeed, this will be after everything has reached the state of
permanence; if the weight P is variable, the surface AB will never re-
main in the same position; moreover, the velocity at which the vessel
is moved in individual locations depends on that weight. Therefore,
if the total weight is removed after the vessel already has acquired
motion, the vessel will continue to be moved at its own velocity, but
the surface of the water will lose its slope and again be composed in a
horizontal position, just as if the vessel were at rest; therefore, in these
cases it is not the motion of the vessel which changes the state of fluids,
but the variation in motion.
§21. That which we indicated in the preceding paragraph about a
vessel positioned vertically is easily extended to a vessel of any shape
whatever: for, whatever is the inclination to the horizon of the
aqueous surface AB in the cylindrical vessel, it will be the same in all
other vessels; moreover, the pressure of the water on the walls of a
vessel is defined everywhere if the vertical column is considered from
that point for which the pressure of the water is to be defined up to the
surface of the water, which will have to be imagined if that should be
necessary. If in place of the vessel there is assumed, for example, a
pipe curved at each end, just as ACDL (Fig. 69 ), and this is moved in

A--,
...:'1:· .

C 'D

FIGURE 69

the direction CD, then each surface Mand N will change position to
A and B, until the straight line AB obtains the required inclination
defined previously; also, it can occur that part of the water flows out
through A before equilibrium is present; if the leg DL is directed
downward as in Fig. 70, the water will remain as if suspended;
indeed, in each case the inclination of the line AB will be the same,
the remaining things being equal.
However, in Fig. 69 the line MA will be greater, the longer is the
 FLUIDS IN VORTICES OR IN MOVING VESSELS 287

horizontal leg CD, such that the smallest accelerations or retardations

can be observed; this often can be useful for other things such as
measuring accelerations of ships and the pressures which rowers exert
in individual immersions of the oars; however, in these cases the entire

- - -- - ~ -- ·-- .

FIGURE 70

motion of the fluid which is developed in individual repetitions should

be investigated, because the state of persistence or permanence cannot be
assumed.
Because of this same reasoning, it is not yet allowable to determine
wholly from the preceding what must happen when vessels contain-
ing a fluid are pierced.
However, the rules of percussion can be deduced from the ordinary
laws of pressure, since indeed a percussion is nothing other than an
immense pressure lasting for a very short time.
§22. For example, let the cylindrical pipe ABCD (Fig. 71 ), placed
horizontally, be full of water, and let the sphere P impinge on the

FIGURE 71

extension AP of the pipe; then the water will suddenly press the base
BA violently toward P; in order to understand this pressure properly,
we will assume first that the pipe has no weight; thus it appears from
the equality between action and reaction that during the impulse of
the sphere the base is not impelled differently by the water than it
would be impelled in the opposite direction by the sphere if the latter
impinged directly against the base. But if the weights of the water
 HYDRODYNAMICS, CHAPTER XI

and the pipe are assumed to be in proportion asp is to 7T, the impulse
of the water against the base will be to the residual impulse asp + 7T
is top; for the impulse is distributed equally over all the material of
the water as well as the pipe, and only the fluid reacts on the base.
But now let us assume a very small orifice m in the base BA;
through this, nevertheless, water is considered to flow very freely;
thus we understand that a particle of water will be ejected through the
small orifice m during the impulse; however, the quantity of that
water cannot be determined, for it depends upon the rigidity of the
material AP receiving the impulse: indeed, if that material is very
rigid, a greater pressure is to be substituted for the impetus, but lasting
for less time; for example, let the same impetus be considered for two
different cases: moreover, in one let the pressure be quadrupled, in
the other let the duration of the pressure be quadrupled, which can
happen when the material is more rigid in the former case than in the
latter; thus, approximately double the quantity will flow out in the
impulse of the lesser pressure and greater duration than in the other
case. In this way the rigidities of materials can be explored: but
they can be found as well from sound.
 TWELFTH CHAPTER
Which shows the Statics of Moving Fluids,
which I call Hydraulico-Statics

§1. Among those who gave measurements of the pressure of fluids

existing within vessels, few have gone beyond the common rules of
Hydrostatics which we showed in Chapter II; nevertheless, there are
many other rules which pertain to the appropriately named Hydro-
statics, such as whenever a centrifugal force or the force of inertia is
united with the action of gravity, each of which we discussed in the pre-
ceding chapter; dead forces of this type can be devised and combined in
infinitely many other ways. But these are not the things which seem
to me to be most desirable, since it is not difficult to give general rules
for this procedure. I desire, rather, [to treat] the statics of fluids
which are moved within vessels in a progressive motion, such as of
water flowing through conduits to leaping fountains: indeed, this is
of multiple use, and it has not been treated by anyone, or, if some
people can be said to have made mention of it, it was not at all
properly explained by them; indeed, those who have spoken about
the pressure of water flowing through aqueducts and the strength
required of the latter for sustaining that pressure did not hand down
any laws other than those for extended fluids with no motion.
§2. It is singular in this hydraulico-statics that the pressure of water
cannot be defined unless the motion has been known correctly, which
is the reason that this doctrine escaped notice for so long; indeed, up
to now Authors were hardly anxious to investigate the motion of
water, and they estimated velocities almost everywhere from the
height of the water alone; however, although the motion often tends
so quickly toward this velocity that the accelerations clearly cannot
be distinguished by observation, and all the motion seems to be
generated in an instant; nevertheless, it is of interest to understand
these accelerations correctly, because otherwise the pressures of the
 HYDRODYNAMICS, CHAPTER XII

flowing water often cannot be defined, and on that account I esti-

mated that it is a matter of greatest moment to consider those changes,
however instantaneous, from the beginning of motion up to a given
limit with all care, and to confirm them by experiments, which I did
at different places in this treatise, but especially in Chapter III.
§3. If the motion could be defined everywhere, it would be easy
to develop the most general statics in moving fluids: indeed, if one
assumes an orifice which is infinitely small in that very place at which
the pressure of the water is desired, one will seek to learn first at what
velocity the water would erupt through that tiny orifice and to what
height that velocity would be due; moreover, one understands that
the pressure which is sought is proportional to this very height.
From this principle the pressure is to be sought which the hori-
zontal plate LQ in Fig. 43 sustains if it has not been perforated. In-
deed, since it has been shown by us in the second corollary of §3 I,
Chapter VIII, that, if the orifice His infinitely small in proportion
to the orifices M and N, and the ratio of these orifices Mand N is
indicated by a and y, then the height due to the velocity of the water
erupting through H will be aa (LB) - yy (NQ ), we will thence judge
aa + yy
that the pressure of the water against the nonperforated plate LQ is
proportional to this very height. We gave the same proof in another
way in §rg of the cited chapter. Hence it follows that it can occur
that the section LQ experiences no pressure, however great the height
of the water above it may be, as for example when y = aV LB/NQ;
indeed, the pressure can even be changed into suction.
§4. Similarly, the pressure of the water against the section LQ is
obtained if, for instance, the latter is perforated by an orifice of finite
size in proportion to the two remaining [orifices]. For if the section
is perforated by an infinitely small orifice with respect to that which
exists at H, the water cannot but erupt at a common velocity through
either one. And since this velocity is known (from §30, Chapter
VIII ) for the orifice H, the velocity is also obtained at which the water
must erupt through the tiny orifice which we conceive, and thus we
know the pressure of the water. For example, let the orifices M,
H, and N be equal to one another, and also let the height BL have a
ratio to the height LQ as ro is to 3, and the pressure against the plate
LQ will be one-tenth of what it is with the orifices Hand N closed off.
Finally, if one should desire the pressure of the water in another
location, he will simply add the height by which the section LQ
exceeds that point to the height of the thrust through the orifice H.
The same method serves for determining water pressures in the rest
 STATICS OF MOVING FLUIDS

of the vessels which we treated in Chapter VIII. But all these

matters differ from those which pertain to the motion of fluids through
conduits, because the water, on account of the infinite size of the
vessels assumed by us, is as if at rest in cavities, and nevertheless it
exerts a far different pressure from what is otherwise customary.
Moreover, in conduits the water changes its pressure more, the greater
the velocity at which it flows through, and it exerts almost all its
customary pressure if that velocity is very small.
This is so whenever the velocities of fluids can be determined by the
methods presented by us just above. But the matter must be handled
by a singular method when the water flows through conduits, and I
comprehend this doctrine especially under the title of hydraulico-
statics. Here, not so much can the pressure be defined from the
velocity as, reciprocally, the velocity from the pressure, if a small
orifice is made in the walls of the conduit. And in the present
chapter I decided to treat especially that hydraulico-statics, the appli-
cation of which is very broad.

PROBLEM

§5. The very wide vessel ACEE (Fig. 72), with the cylindrical and
horizontal pipe ED attached, is to be kept constantly full of water;

A B

FIGURE 72

and at the extremity of the pipe let there be the orifice o emitting
water at a uniform velocity; the pressure of the water against the
walls of the pipe ED is sought.
 HYDRODYNAMICS, CHAPTER XII

SOLUTION. Let the height of the aqueous surface AB above the

orifice o be a; the velocity of the water flowing out at o, if one excludes
the first instants of flow, will have to be considered uniform and equal
to va, because we assume the vessel to be kept constantly full; and,
with the ratio of the areas of the pipe and its orifice assumed equal to

~, the velocity of the water in the pipe will be

I
Va_
n
But if the entire
base FD were missing, the ultimate velocity of the water in the pipe

itself would be Va, which is greater than va_

n
Therefore, the water
in the pipe tends to greater motion, but its pressure is impeded by the
added base FD. By this pressure and repressure the water is com-
pressed, which very compression is confined by the walls of the pipe,
and hence these sustain a like pressure. Thus it appears that the
pressure of the walls is proportional to the acceleration, or the incre-
ment of velocity which the water would receive if the entire obstacle
to motion would vanish in an instant so that [the water] might be
ejected immediately into the air.
Therefore, the problem is now changed into this: if during the
flow of water through o the pipe ED were broken at cd at an instant,
one seeks the magnitude of the acceleration the volume element acbd
would thence be about to obtain; indeed, the particle ac taken at the
walls of the pipe will sense that much pressure from the water flowing
through. To this end the vessel ABEcdC is to be considered, and with
regard to it the acceleration of an aqueous particle close to efflux is to

be found, if this would have the velocity Va_

n
We handled that
matter very generally in §3, Chapter V . Nevertheless, because the
calculation is short in this particular case, we will here again subject
the motion in the shortened vessel ABEcdC to evaluation.
Let the velocity in the pipe Ed, which [velocity] is now to be con-
sidered as variable, be v; let the area of the pipe, as before, be n, the
length Ee = c; let the length of the aqueous particle ac, infinitely
small and about to flow out, be indicated by dx. There will be an
equal volume element at E entering the pipe at the same instant that
the other, acdb, is ejected; moreover, while the volume element at
E, the mass of which is n dx, enters the pipe, it acquires the velocity v
and the live force nvv dx, which entire live force was generated anew;
indeed, the volume element at E, not yet having entered the pipe,
had no motion on account of the infinite size of the vessel AE; to this
live force, nvv dx, is to be added the increment of live force which the
water at Eb receives while the volume element ad flows out, namely,
 STATICS OF MOVING FLUIDS 293
wcv dv; the sum is due to the actual descent of the volume element
n dx through the height BE or a; therefore, one obtains nvv dx +
v dv a - vv
2ncv dv = na dx, or dx = - -c-·
2
Moreover, in all motion the increment of velocity dv is proportional
to the pressure multiplied by the differential time, which here is dx;
V
therefore, in our case the pressure which the volume element ad ex-
. proport10na
periences 1s . 1 to t h e quantity
. dx'
V dv
t h at 1s,
. to t h e quantity
.

a - vv
2C
.
B ut at t h at mstant at w h'1ch t h e pipe
. 1s v'a, or vv = -a ;
. b ro k en, v = -
n nn
therefore, this value is to be substituted in the expression a - vv,
2C
1
which thus is transformed into nn - a. And this is the quantity to
2nnc
which the pressure of the water against the portion ac of the pipe is
proportional, whatever area the pipe may have, or by whatever
orifice its base may be perforated. Therefore, if in a particular case
the pressure of the water would be known, it would be understood at
the same time in all remaining [casesJ : but, indeed, we have this
[pressureJ when the orifice is infinitely small or n is infinitely large
with respect to unity: for then it is evident from itself that the water
exerts its entire pressure, which conforms to the total height a, and
this pressure we will designate by a; but when n is infinite, unity
vanishes with respect to the number nn, and the quantity to which the
pressure is proportional becomes !!:... Therefore, if we wish to know
2C
in general how great the pressure is when n is any number whatever,
the following analogy must be used. If the pressure a conforms
to the quantity !!:.., what then will be the pressure for the quantity
2C
nn - r
- - - a? And thus the desired pressure is found equal to
2nnc
nn - r
- - a . Q.E.I.
nn
§6. COROLLARY r. Because the letter c vanishes from the calcula-
tion, it follows that all portions of the pipe, those which are nearer to
the vessel AG as well as those which are more remote, are pressed
equally by the water flowing through, and certainly less than the
elements of the base CG, and the difference is the greater, the larger is
2 94 HYDRODYNAMICS, CHAPTER XII

the orifice o; and, further, the walls of the pipe do not sustain any
pressure if in the latter the entire barrier FD is missing, so that the
water flows out from a full orifice.
§7. COROLLARY 2. If the pipe is perforated somewhere by a very
small orifice that is necessarily in some ratio to the orifice o, the
water will spring forth at the velocity by which it could ascend to the
height nna - a if only no foreign hindrances were interfering. Indeed,
nn
this will be the height of the thrust in Fig. 73, or In = nna - a_ But
nn

FIGURE 73

if the small tube gm is attached, vertical or even inclined in some way,

connecting with the horizontal pipe, but so, nevertheless, that the ex-
tremity of the inserted tube does not project into the cavity of the
horizontal pipe lest the water flowing past strike against that extremity,
the vertical height gh of the water standing in the inserted tube will
also be equal to nna - a; and it is not necessary in this latter case
nn
that the tube gm be very narrow.

ScHOLIUM

§8. Therefore, this theory can be confirmed very easily by experi-

ment, and this will be of more importance because up to this time no
one has defined equilibria of this sort, the use of which is very widely
evident, because by the same method the pressure of water flowing
through conduits can be obtained very generally for aqueducts in-
 STATICS OF MOVING FLUIDS 295
dined in any way whatever, curved, of varied area, and at any
velocity of water whatever; then, as well, because not only this
[theory] of pressures, but the entire theory of motions besides, which
we gave above, is confirmed by experiments of this sort, because they
prove that the accelerations of the water were defined correctly by us.
But one must take care in the experiment that the horizontal pipe is
very smooth on the interior, perfectly cylindrical and horizontal, and
that it is wide enough so that no noticeable decrement of motion can
arise from the adhesion of the water to the walls of the pipe; let the
vessel itself be very wide and be kept full continuously. Also one
must observe how great is the characteristic of elevating standing
water in the glass tube gm, which characteristic pertains to all capillary
or rather narrow tubes; for this elevation is to be subtracted from the
height gh; or, rather, a pipe of equal thickness is to be assumed with
the orifice o blocked off, the point mis to be noted, and then, with the
water allowed to flow, the point h is also to be noted; moreover,
according to the theory the descent will be mh = _..:_ a = _..:_ (EB ).
nn nn
Finally, one must pay attention as well to the stream of water
flowing out at o, for its contraction also causes the water in the hori-

zontal pipe to flow through at a velocity less than Va_

n
I treated that
contraction and the method of preventing it in Chapter IV. But
although it can happen with these inconveniences that no noticeable
error remains in the experiment, nevertheless, if we wish to apply
greater accuracy, the quantity of water flowing out in a given time
will have to be discovered by experiment, which [quantity], compared
with the area of the pipe, will give very correctly the velocity of the
water flowing within the pipe, which in the calculation we have set
Va
equal to - . But if in the experiment it will be found to be less,
n
such, for example, as is due to the height b, then the pressure of the
water flowing by will be approximately a - b.
§9. COROLLARY 3. If the orifice at o is blocked off at first by a
finger, and afterwards the water is allowed to flow, the pressure a at
the first moment of flow is changed into the pressure nna - a, but that
nn
change of pressures does not occur in an instant; if, indeed, one is to
speak accurately, it occurs at last after an infinite time, because, as
we saw in Chapter V, the entire velocity of the water, which was
assumed by us in the calculation to correspond to the whole height a,
is never present exactly; nevertheless, it tends toward this velocity
 HYDRODYNAMICS, CHAPTER XII

with an incredible acceleration immediately after the first drops have

been ejected, so that it seems to have acquired the total [velocity], as
much as can be judged by observation, without any noticeable delay,
unless the aqueducts are very long, for then the accelerations of the
water can be discerned clearly by eye, an example of which I gave in
§13, Chapter V. Therefore, in those conduits bearing water to a
leaping fountain from a reservoir located very far away, if the pres-
sures are investigated at some point by experiment in the manner that
I mentioned above, it is found that the pressure is diminished quickly
indeed, nevertheless not in an instant, and it will be possible to
distinguish the differences of the pressures.
But in order to define the force of the water generally, one must
assume for v that velocity which the water has at that same place and
that same instant at which the force is desired, and if this velocity is
known to conform to the height b, the force of the water will be a - b.
Hence, since those things which were offered in Chapter V have
agreed with the present, it will be possible to define what the pressure
will be at any moment.
For these [statements] it is not difficult to anticipate the laws of this
hydraulico-statics if both the shape of the vessel and the velocity of the
water flowing through the conduits are assumed at will as anything
whatever. Indeed, the pressure of the water will always be a - b,
where by a is understood the height due to the velocity at which water
will flow out of an abrupt conduit and vessel kept constantly full after
an infinite time, and by b the height due to the velocity at which the
water actually flows through. It is clearly amazing that this very
simple rule, which nature affects, could remain unknown up to this
time. Therefore, I will now show it more expressly.

PROBLEM

§rn. To find the pressure of water flowing at any uniform velocity

whatever through a conduit arbitrarily formed and inclined.
SOLUTION. Let there be a conduit ACD (Fig. 74), through the ori-
fice o of which water is considered to flow at a uniform velocity due
to the vertical height oS; let the line SN be drawn, and let the in-
finitely wide vessel NMQ,P be assumed full of water right up to NP,
from which the conduit draws its water perpetually and equally; I
assume these things accordingly ii:i order that a cause be present, or a
uniform propelling force, which propels the water at a given velocity
or maintains an equal flow of water. And without this hypothesis
our problem would be indeterminate, because the same velocity in
the same conduit pertaining to any instant can be generated in in-
 STATICS OF MOVING FLUIDS

finitely many ways, and therefore, in order that a measure of the cause
propelling the water be obtained, uniformity must be assumed in the
motion of the water.

FIGURE 74

Now the pressure of the water is to be defined at CF (or if) ; and to

this end we will consider again that the conduit is broken at the
section CE (or ce) perpendicular to the conduit, and we will examine
what acceleration or retardation the volume element CEGF (or cegf)
will receive after the first instant of rupture; for this reason we have
to define generally the instantaneous motion through the shortened
vessel NMECAQP (or NMceAQP). Therefore, let the velocity of the
infinitely small volume element CEGF (or cegf) at that very point of
cutting off be v, and let its mass be dx; the live.force of the water moving
in the shortened vessel will be proportional to the quantity vv; hence
we will set it equal to a.vu, understanding by the letter a. some constant
quantity which depends upon the areas of the suddenly broken con-
duit; however, its precise determination is not required here. Let it be
 HYDRODYNAMICS, CHAPTER XII

noted that the live force of the water in the fictitious vessel N MQ,P
is neglected on account of its infinite area; nevertheless, even if
it would not be of infinite area, no variation would hence arise in the
calculation. Now we have the increment of live force of the water
moving in the shortened vessel equal to 20:v dv, to which if there is
added the live force generated at the same time in the ejected volume
element, there arises 20:v dv + vv dx, which is the total increment of
live force due to the actual descent of the volume element dx through the
vertical height of the water above the point C (or c), which we will
designate by a; therefore, that total increment of live force is hence to
be made equal to a dx, such that
20:v dv + vv dx = a dx
or
v dv a - vv
Tx = ----;;-·
If the remaining things occur as in §5, and the velocity v is assumed
as ifit were due to the height b, it will be found that the pressure of the
water at CF (or cf) is as great as in water standing at the height a - b.
Here it can be noted that the height b is to the height oS, if there are
no alien hindrances to the motion and the stream flowing out at o is
not contracted, in proportion as the square of the orifice o and the
section CE (or ce).
§11. COROLLARY. When b is greater than a, the quantity a - b
becomes negative, and thus the pressure is changed into suction, that
is, the walls of the conduit are pressed inward; moreover, then the
situation is to be considered as if, in place of the aqueous column CT
lying above and set in equilibrium with the water flowing by, the
aqueous column ct were attached, the tendency of which to descend
is prevented by the attraction of the water flowing by, just as if, for
example, the area ce of the conduit were equal to the orifice o, where-
upon b = oS, not considering the accidental hindrances to the motion;
hence, if the tube er descends from the conduit, and ifit is full of water
from its origin c right to the point t placed on a level with the orifice o,
the water ct will remain suspended without motion; but if the point t
is placed below o, the water will descend through the tube er, and it
will flow perpetually at r, and, nevertheless, as anyone can now
estimate after this theory has been considered, the velocity of the
water flowing out at r will be that which is due to the height of NP
above r, and even if all hindrances are removed, this velocity will
correspond rather to the height tr, if only the tube is very narrow in
proportion to the conduit. If the point t is placed higher than the
point o, the water will ascend on its own, and after it will all have

.
 STATICS OF MOVING FLUIDS 2 99
entered the conduit, air will be drawn through the tube, and soon the
aqueous stream flowing out at o will be disturbed by the admixed air,
its clearness and solidity having been spoiled. Therefore, one sees
when the pressure will be positive and when negative: indeed, the
pressure is the greater in a tube, the larger it is [in area] and the lower
it is placed. Certainly in theory the height b = !._ (oS), if~ denotes
nn n
the ratio between the area of the orifice and of that section of the pipe
for which the pressure is to be defined. But when hindrances dim-
inish the motion notably, it will be agreed upon in estimating pres-
sures rather that the velocity of the water, as it actually is, be found by
experiment and the height required for that velocity be substituted
for b; similarly, the pressure will be estimated more accurately if for
a not the height of the aqueous surface NP above the place of effiux is
substituted but rather the height of the velocity at which the water
actually flows out from the conduit broken in the same place. Never-
theless, these corrections are not always important. But I will now
illustrate that general theory by certain examples.
§12. EXAMPLE I. Let there be a vessel ABFG (Fig. 75), from the
middle of the base of which the pipe DE descends, having the shape
of a truncated cone diverging toward the lower regions. Let water
be supplied perpetually at AG, so that the vessel is thus kept full.
Moreover, let the height of the aqueous surface above the orifice
Ebe a, and above D (which is the point at which the pressure of the
water is desired) be c, the area of the orifice at Ebe m, and the area
of the horizontal section at D be n. The pressure of the water at D
will be c - mm a, which quantity is negative by virtue of the hypo-
nn
theses, so that the walls of the conduit are pressed inward by an
aqueous column of height mm a - c.
nn
Therefore, if the curved pipe DLN is understood to be inserted in
the other pipe DE, the water flowing past D will be in equilibrium
with the water DLN when the height of D above N is mm a - c. If
nn
this height is less, the water will ascend on its own, and it will not stop
ascending as long as the orifice N is submerged in water, so that thus
water can be elevated from a lower place to a higher without any
external force, if it flows in at AG in sufficient quantity. But, indeed,
when the vertical height of D above N is greater than mm a - c, the
nn
water will ascend in the leg LN until it will be equal to the other.
300 HYDRODYNAMICS, CHAPTER XII

Finally, it must be recalled here that in passing I indicated that

experience shows that water is indeed far from flowing out through
pipes diverging from the vessel to which they are attached at its total
velocity which it should obtain by virtue of the theory; I indicate
reasons for this in §26, Chapter III.

I
IIN

FIGURE 75

Hence it occurs that the height of D above N is somewhat less than

that which should be found by theory. The error may be corrected
if in place of mm a the height of the velocity is used which the water
nn
has at D; this height is obtained from an experiment performed on the
quantity of water flowing out in a given time.
 STATICS OF MOVING FLUIDS

§13. EXAMPLE 2. If to a similar vessel a vertical pipe is attached,

which is represented in Fig. 76 by CE, in which the areas everywhere

)"

FIGURE 76

have an inverse ratio to the square roots of the heights of the water
lying above, that pipe is not affected by the water flowing through,
and it does not sustain either pressure or suction anywhere.
Hence it follows that the natural shape of a vertical aqueous fila-
ment, as long as it is continuous, is the same as that of the pipe CFE,
which both reasoning and experience confirm; moreover, the fila-
ment will be attenuated more quickly, the less the height of the
aqueous surface above the orifice C, or the slower the water flows out;
it appears that the aqueous filament is of this nature in order that the
same quantity of water may flow across the individual sections and
that the velocity is not changed anywhere, wherever the filament is
cut off, which same property occurs as well in the pipe CFE, so that
these things agree with each other very well.
§14. EXAMPLE 3. Let water discharge from a reservoir through a
conduit in the base of which there is an orifice through which water
springs forth vertically just as in a leaping fountain; I say that the
pressure of the water at individual points in the conduit is everywhere

equal if its areas are respectively as J~ x b' where a expresses the

height of the water in the reservoir above the orifice of efflux, x the
height of the same water above a point chosen at will in the conduit,
and b an arbitrary constant height, and then the pressure of the
flowing water everywhere will be to the pressure of standing water as
302 HYDRODYNAMICS, CHAPTER XII

b is to a. But because, with the remaining things being equal, wider

conduits are less resistant to rupture than narrower, and this indeed
in ratio of the radii, or because the attempt of water in rupturing a
conduit, with the remaining things being equal, follows a square root
ratio of the areas, it is evident that the conduit will be subject to the
same danger of rupture at any location if the area (y ) in proportion
to the orifice ejecting water (unity) everywhere follows the law of the
following equation:

(x - ~) vj = b, or xxy 4 - bby 3 - 2axyy + aa = o.

In a conduit of equal area throughout its entire length, the pressing

force of the water for rupturing the conduit will be everywhere
proportional to the strength of the conduit, if the thickness of the walls
of the conduit follows the ratio of x - _!:__, the area of the conduit in
mm
proportion to [that of] the orifice (unity) having been understood
bym.
§15. EXAMPLE 4. It can happen that the height of the aqueous
surface with respect to the place at which the pressure is to be investi-
gated is negative, such as in curved siphons drawing water from one
vessel to another placed lower. Then the pressure becomes negative
on two accounts, namely, equal to - a - b, with a denoting the height
of the point above the surface of the water and b the height due to the
velocity of the water at that point.
Truly these things will suffice, as I believe, for correctly under-
standing the statics of moving fluids. Now I come to certain other
phenomena, the solution of which depends on those rules which we
have just presented.
§16. In Chapter III, §25, I made mention of the cohesion of water
flowing through pipes; however, to define the true measures of that
cohesion everywhere is a matter which cannot be explained without
that previously mentioned hydraulico-statics: for it does not suffice to
have considered the vertical heights above the orifice of efflux, as it
is commonly thought, but it is necessary to know also the velocities
conforming to the water, and these are understood from the areas.
But in order that the general law may appear at once in defining the
force of cohesion or the inclination with which fluids tend toward
mutual separation, I say that that force of cohesion is equal to the
force by which the walls of a conduit are pressed inward, which we
defined in §ro. This proposition does not seem to me to need any
other proof; for just as the compression of water, or the force by
 STATICS OF MOVING FLUIDS

which its portions are pressed mutually against one another, is equal
to the standing aqueous column lying above, so in turn the tendency
to separate the fluids is to be reckoned equal to the [effect of] the
attached standing vertical aqueous column which is in equilibrium
with the water flowing by. In place of examples we will accept the
same things which we used above for indicating the negative pressures
of water.
I. In Fig. 75, explained in §12, if in the tube DLN the height of D
above N is such that the water standing in it is in equilibrium with
the water flowing past at D, then, in order that the water not be torn
apart at that place, the force of cohesion at D must be as great as that
which the weight of an aqueous column of similar base and vertical
height DN has. Hence that which I mentioned in §25, Chapter III,
is understood: that the length of a pipe can be increased so that finally the
water stops being continuous in the pipe, but rather it is divided into columns,
and this happens in cylindrical pipes when they descend beyond 32 feet; more-
over, in diverging pipes a lesser descent is required, so that, for example, if the
lower orifice were twice as large as the upper orifice attached to the reservoir,
pipes could not descend below eight feet without the danger of the dissolution of
the water being present. However, in these examples considered
theoretically, the water is assumed to flow at its full velocity without
diminution of motion.
II. From the same reasoning it is evident that if pipes converge
toward the lower regions, then they admit a descent greater than
32 feet; and finally, in the case of Fig. 76, explained in §13, the pipe
can be continued without end, as also in infinitely many other ways.
III. But if the height of the aqueous surface in a reservoir is nega-
tive with respect to the proposed point, as occurs when water is to be
carried across a mountain, never, no matter how the problem is
attacked, can the height exceed 32 feet, which is evident from §15.
For even if the water is to flow through at an infinitely small velocity,
a force of cohesion is already required which is equal to the entire
aqueous column, and a greater force is required if it flows through at
an appreciable velocity. Hence I consider the remedies employed
by some Writers as useless: certainly I know that without other artifice
water often remains suspended beyond a height of 32 feet, and Mer-
cury beyond 30 inches; but this effect is uncertain and not consistent.
Certain people also affirm that the flow of water through curved
siphons occurs in a vacuum; but whether the vacuum is such that not
even a sixtieth part of the air is left in the receptacle, and whether the
height of the pipe exceeds by more than half a foot the surface of the
water to be drawn, I do not know. Thus, therefore, I wish that those
 HYDRODYNAMICS, CHAPTER XII

things which I mentioned about the subsequent dissolution of the

water should not be considered other than hypothetically spoken.
It will suffice that I have determined accurately by what force the
water is urged to mutual separation.
§17. Further, there are other phenomena of nature, the true ex-
planation of which depends on that hydraulico-statics theory: for ex-
ample, that smoke ascending through a chimney draws air after itself
with great impetus through an orifice made in the chimney; that the
wind blowing from a rather narrow place into a more open one loses
some of its elasticity, just as it is gathered from this that opened
windows are closed by air attempting exit from a room on account of
its greater elasticity; and others of this sort, which individual cases it
is not permitted to study.
The pressures of moving fluids can be varied, indeed, in infinite
ways; nevertheless, I believe that all can be reduced to our principles;
we have examined two forms of that theory: I deduced the first from
the known motion which the fluid will have, if at the point where the
pressures are to be determined the vessel is perforated with an in-
finitely small orifice; the other I deduced, as they say, a priori from our
general theory; often they both pertain at the same time, as one
requires the help of the other, and then another estimation of pres-
sures arises which I will indicate by a single example.
§18. Let us consider, in the vessel which Fig. 72 shows, that the
horizontal pipe has, not only at its extremity but also at its insertion
EC, a section in a vertical place perforated in the middle, the remain-
ing positions indicated in §5 being maintained; the walls of the pipe
ED will endure a different pressure from the water flowing through
than if there were no section EC added, and certainly a lesser one,
although [the water] flows through at a lesser velocity. In order
that this pressure be accurately defined, the path to be followed is the
same as cited in §5: namely, first of all the velocity is to be sought at
which the water flows in the pipe ED after it has already been made
uniform. Then one should inquire as well into the value of v:v
if the
pipe is assumed to be broken off somewhere. x
But how this can be found is a matter which pertains especially to
Chapter VIII, with the precautions of§r4, Chapter VII, having been
heeded at the same time. In Chapter VIII the motion of fluids
flowing through many orifices is shown generally, and in §14, Chapter
VII, it is demonstrated in particular how the potential ascent which is
generated in volume elements is to be estimated when these [ele-
ments] flow through the orifice, not into practically standing water
but into water carried by a motion which cannot be neglected.
 STATICS OF MOVING FLUIDS

If one proceeds properly along these indicated paths, he will dis-

cover that the velocity with which the water flows uniformly through
the pipe ED conforms to the height mmppa , where by
mmnn + nnpp - mmpp
m, p, and n are indicated, respectively, the areas of the orifices made in
the sections EG and FD, and as well of the pipe ED; moreover, by a
one understands the height of the water above the horizontally posi-
tioned pipe ED.
Further, if one assumes that the pipe is broken off at cd and that the
volume element ad is being moved at the velocity v, or that the height
due to this velocity is vv, and if at the same time one indicates the
length Ee by c and the very small length ac by dx, one will encounter
the following equation:
nn
2cv dv + -mm vv dx = a dx
or
v dv mma - nnvv
dx 2mmc
mmppa
Now for vv let the value just indicated, -------=-.c=------ be
mmnn + nnpp - mmpp'
substituted, and there will be

V dv mmnn
___ _ _ -_ _
mmpp
....::..::___ a
dx 2c(mmnn + nnpp - mmpp) '
to which the sought pressure is proportional. But if the area of the
final orifice, indicated by p, is as if infinitely small, the pressure
becomes a. Generally, therefore, the pressure sought, by virtue of
§5, is equal to
mmnn - mmpp
-------=-=--------a.
mmnn + nnpp - mmpp
§19. If the area n of the pipe is as if infinite in proportion to the
areas of the orifices in the sections, the pressure becomes mma
mm+ pp
and so great also is the height to which the water flowing out at o
can ascend by its own velocity; therefore, this conforms with §4,
Chapter VIII, because the shape of the vessel, or [its being] of in-
finite area everywhere, docs not cause the velocity of the water
springing forth to differ.
When there is no plate at F, it happens that p = n, and the entire
pressure vanishes. This deserves to be noted because it shows the
 HYDRODYNAMICS, CHAPTER XII

reason why in diverging pipes the suction is not as great as it should be

according to the hypothesis in which all the live force is assumed to be
conserved. Indeed, in the present case we took into consideration a
live force which is continually diminished. And so also the sides of
the pipe experience no pressure when the section which is at EC has
an orifice infinitely smaller than that which exists at FD. Finally, it
is worth noting also that, although fluids being moved through con-
duits constructed without any cross-sectional plates generally effect
a pressure which corresponds to the difference of the heights due to
those velocities at which the fluid flows after an infinite time through
a cut off conduit and at which it flows actually through the unin-
terrupted conduit, in the present case this law is nevertheless least
valid, to which I wish that those would pay attention who want to
show the general proposition of §ro synthetically by our observed
hydraulico-statics theory. For perhaps there will be some to whom this
matter will seem so intrinsically obvious that it hardly need be proven;
but particular laws of this type which occur in hydraulico-statics show
that those, if there are any, deceive themselves by a certain false
resemblance to the truth.
§20. It will be to the point to undertake experiments also concern-
ing these things which were mentioned in § r 8, for [determining] the
velocity of the water flowing out at o as well as the pressure; for
hence in addition to the laws of pressures that theory of accelerations
will also be confirmed which obtains when a certain portion of the
live force is continually used up uselessly, which problem we treated
especially in Chapter VIII. Moreover, in undertaking an experi-
ment, as much as it can be done, let those hindrances be avoided of
which we have already often made mention.
§21. Let me inject here a question which certainly does not pertain
to the statics of fluids, but to the hydraulics or motion of fluids, but
which cannot be solved without those previously given hydraulico-
statics rules. In Fig. 72 (here I no longer consider any plate at EC)
one seeks, if the pipe is perforated by an orifice at ac having a finite
ratio to the area of the pipe as well as to the area of the orifice o, and
if the motion of the water has already been made uniform, one seeks,
I say, at what velocity the water will erupt through each aperture.
At this time let the height BE again be a, the area of the pipe be n,
the area of the orifice at o be p, the area of the orifice ac be m, and the
velocity of the water flowing out through o be v. The velocity of the
water which flows across the orifice ac will be f!.. v. Therefore, at that
n
same place it exerts a pressure on the walls of the pipe which is
 -
STATICS OF MOVING FLUIDS

a - ppvv (according to §5), and on that account I assume that the

nn
height which can generate the velocity at which the water springs
forth through the orifice ac will also be approximately as great, but

that this velocity itself is Ja - ppvv_ With this having been estab-
nn
lished, the velocities at the orifices a and ac will be as vis to Ja - ppvv;
nn
and thus any volume element whatever entering the pipe at GE, when
it arrives at the region of the first orifice, is separated into two portions,
one of which flows out through ac, the other through o; and thes~
portions are, respectively, proportional to the velocities at which the
efflux occurs on either hand multiplied by the areas of the orifices.
Therefore, if the mass of the entire volume element GE is called g,
the portion of it flowing out through ac will be equal to

J - ppvv;
gma nn [ pv J
+ ma nn '
- ppvv]

and the other portion flowing out through a equals

If these portions are multiplied, respectively, by the squares of their

velocities, their live forces will be obtained, the sum of which is to be
equated to g · a, that is, to the actual descent of the volume element g
through the height a. Thus, if it is reduced, the following equation
is obtained:

n3 vv - n3a = mpvv nna - ppvv

or
2n + mmnnpp + nnmpv4n 4 + mmpp - 4nnpp
6
vv = -----------:::-------,----------- a
6 2n + 2mmp
4 '

and this quantity expresses the height for the velocity of the water
flowing out at a, by which knowledge also is obtained the similar
height for the other orifice ac, which indeed is a - ppvv_
nn
§22. If p = n, it happens that vv = a; therefore, the water then
springs forth at the total customary velocity through the orifice a, and
nothing flows out through the other orifice ac. Further, in either ori-
fice the velocity corresponds to the entire height a, if p is as if infinitely
small. But if mis infinitely small, it certainly occurs that vv = a, but
 HYDRODYNAMICS, CHAPTER XII

the height of the velocity pertaining to the small orifice ac is a - pp a,

nn
as was already indicated in §7. If m = p, it occurs that
n4 a ppvv (nn - pp )2 a
v v -4- - - - - ~4· and a - - - = ~4' - - . . . . C . . . C . - ~4
- n - nnpp + p ' nn n - nnpp + p
Finally, it can be observed that the water is always ejected through
the orifice o at a greater velocity than that which corresponds to the
height a, which certainly occurs because the water at Ed makes some-
what of an impetus against the water at dF.
Meanwhile, although all these Corollaries agree splendidly with
the nature of the argument, nevertheless, the solution of that problem
cannot be considered other than approximately true.

HYDRAULICO-STATIC EXPERIMENTS FOR

CHAPTER XII

PERTAINING TO §§3 AND 4

The pressures which have been shown in the aforementioned para-

graphs can be confirmed by a simple experiment, if the vessel which
Fig. 43 shows and which is described in §30, Chapter VIII, is care-
fully prepared, and if in its cross section LQ a glass tube is inserted
vertically, either end of which is open; thus it will be observed, with
the orifices H and .N blocked off and the entire system filled with
water, that the water in the glass tube ascends to the level AB, or it
exceeds it according to the nature of capillary tubes. Then also, if
the finger is removed from the orifice .N, it will be observed that the
water in the glass tube descends, and after the measurements have
been taken, it will be found, unless I am mistaken, that the residual
height of the water in the glass tube (after the height due to the
effect of capillary tubes has been subtracted) is aa(LB) - yy(.NQ),
aa + yy
just as it was mentioned in §3, where the denominations of these letters
are explained.
Further, if from each orifice Hand.Na finger is removed, then the
residual height of the water in the glass tube will be just that which is
indicated in §4. Similarly, a glass tube can be inserted in the section
Q.N, and this then bent [upward], so that it can be learned whether the
pressures at the section Q.N have also been defined correctly.
But th - experiments which pertain to the pressures of water carried
through pipes I myself undertook in the presence of our Society, and
 STATICS OF MOVING FLUIDS

they are described in Vol. IV of the [St. Petersburg] Commentaries,

p. 194. Therefore, I will present those things here as they are
described there.
I used a wooden box, the width of"" hich was one foot, the length
three feet, the height r 4 inches. I filled this with water, and I
implanted horizontally in its final portion a cylindrical tube accu-
rately made from iron. But that iron pipe was made as follows:
namely, it had a length AB (Fig. 77) of 4 English inches, 2 lines, a
.e
I

A [D] I

lb]
FIGURE 77 FIGURE 78
diameter BC of 7 lines; the pipe was perforated in the middle by a
small orifice m, and at the same place the tube DE, likewise of iron,
having six lines in length and one and one-half lines in diameter,
was welded so that the small orifice m would lie in the middle of the
base. A little later I attached to this small tube a glass tube of
equal area, as it appears in Fig. 79, which shows the method of the
whole experiment. Further, I took care that three covers be made,
[ each] attached to the iron pipe and perforated by an orifice of
different size; such a cover is represented in Fig. 78.
With all these things brought together in that way, which Fig. 79
shows, and having insured that the water did not flow through
openings other than the aperture at BC, I blocked off the orifice at
BC, and then I observed, in the vertically placed glass tube, the

FIGURE 79.
310 HYDRODYNAMICS, CHAPTER XII

point n to which the water ascended, and this I marked by wrap-

ping a silk thread around it; but first I had determined the capillary
effect of that glass tube, and this I had found to be five lines, so
that, with the tube vertically immersed in the water, the difference
between both surfaces of water was five lines; accordingly, the point
n was elevated above the surface EF by the same number of lines,
and hence in the calculation any height Dn, Dg, is to be considered
diminished by five lines.
In the individual experiments the box was kept full of water so
that the height AF was g inches, 7 lines, but the height Dn was 10
inches. With all these things thus prepared for the experiment,
then by reason of the orifice having been opened at BC, efflux was
granted to the water, and directly the water descended in the glass
tube, as from n tog, which point g I marked again with another silk
thread wrapped around the tube beforehand. And thus at last we
performed the following experiments which correspond to §5 and
following.
EXPERIMENT 1. When the diameter of the orifice in the cover BC
was 2¼ lines, the descent ng was a little greater than one line, so that
no difference could be observed between the theory and the result of
the experiment.
EXPERIMENT 2. With another cover applied, in which the dia-
meter of the orifice was 3-} lines, or a little greater, the observed
descent ng was six lines and two-thirds, again clearly as the theory
indicates.
EXPERIMENT 3. With the third cover applied, in which the dia-
meter of the orifice was 5 lines, or somewhat less, we observed a
descent ng of 28 lines. According to the theory it should have been
about 29 lines, and, indeed, the orifice was seen to have not quite
five lines in diameter. The very small difference is to be attributed
to the hindrances which the water experiences in flow through the
tube which are greater than in the preceding experiments on account
of the increased motion within the tube.
EXPERIMENT 4. Finally, with no cover attached, we allowed the
water to flow out through the full orifice, and then almost all the
water had gone out from the glass tube; nevertheless, some portion
remained which we discovered to be eight lines high. But five of
them are to be attributed to the effect of the capillary tube, the re-
maining three are due to the hindrances which the water en-
counters in flow from D to B.
Thus, therefore, the experiments agree with theory correctly.
Moreover, hence it is not difficult to see in advance that it can
happen that the walls of the pipe not only are not pressed toward
the exterior, but also that they are compressed inward toward the
axis of the pipe (see §11). Moreover, I was shown this by the fol-
lowing additional experiment.
EXPERIMENT 5. In place of the cylindrical pipe AB I applied a
conical one, the external orifice of which was greater than the in-
 ST A TICS OF MOVING FLUIDS 31 I

ternal orifice, and at the same time I made use of a curved glass tube
such as Fig. 80 shows. And, while before flow the water stood at n
in the glass tube, the water descended in the same tube right to g
when water flowed through the conic pipe; and the point g was

I I
D I .J
I

FIGURE 80
below D, [serving as] proof that the conical pipe was under pressure
during flow. But in these cases there are significant hindrances to
the motion which make the velocities of the water at the external
orifice much less than those which correspond to the height of the
water; and for this reason the height of the point D above g was
not as great as it would have been otherwise, although there was
some [height]. I obtained the same but altogether more notable
effect in another way (see§r2 ). This other experiment I performed
in the following year in the presence of the Academicians, the Most
Serene Prince Emanuel of Portugal being present.
EXPERIMENT 6. In Fig. Sr ACFB represents a cylinder, in the base
of which was implanted the conic pipe DCHE; and the latter had a
small tube at the side at l which was joined by the extremity of the
curved glass tube lmn; the height CA was 3 inches ro lines; El, 4 lines;
lH, 2 inches g½ lines; the area of the conic pipe at l was to the area
of the orifice CH as ro is to r 6; ln was 5 inches 6 lines, and its orifice
n was submerged in water in the small vessel M.
With a finger placed over the orifice CH and the vessel filled, the
water trickled through the glass tube lmn into the vessel M; but
with the finger removed and the water now flowing out through
CH, the water ascended of its own in a reciprocal motion from the
small vessel M through the tube nml, and together with the re-
mainder flowed out through CH, during which time the entire
small vessel M would have been emptied. But water was supplied
continuously from above, so that the vessel was kept full. If a por-
tion of the orifice CH was blocked off by a finger, it was easy to
cause the water in the glass tube lmn to move up or down at will.
312 HYDRODYNAMICS, CHAPTER XII

If anyone should wish also to find out by experiment whether the

theory agrees with the problem of§r8, he will not have organized his
work badly, since indeed he will thus illustrate not only this new
hydraulico-statics of ours, but also the theory of Chapter VIII, equally
new and treated by no one, by a splendid example, and this very
easily.

FIGURE 81

After these things now have been collected in wntmg, I myself

undertook the experiments of which I just made mention. I used
the same mechanism for this that I just described and which is
represented in Fig. 79; but in addition, as the nature of the matter
requires, I placed another cover on the pipe at A; and the height of
water AF was 8 London inches; the diameter of the iron pipe AC was
again 7 lines. Also I used the same covers as before. But in every
experiment I observed the descent which the surface n made when the
finger was removed from the cover BC; moreover, at the same time,
after the measurement of the vertical height of the orifice C above the
floor had been taken, I observed the distance of that vertical line from
the place at which the aqueous stream struck. This distance I shall
call the amplitude of the thrust; but this vertical height was r 9 inches in
 STATICS OF MOVING FLUIDS

the individual experiments. With these things thus prepared I

performed the following experiments.
EXPERIMENT 7. When the diameter of the orifice of the interior
cover was 2½ lines and the diameter of the orifice of the exterior cover
3} lines, the descent ng was a little less than 7 inches, and the amplitude
of the thrust was g inches. However, in the theory shown in §18, the
descent ng is indicated as 6 inches ro lines and the amplitude of the
thrust as g½ inches.
EXPERIMENT 8. Next the diameter of the internal orifice was
5 lines and the diameter of the other orifice 3§ lines; the descent ng
was almost r 7 lines and the amplitude of the thrust 24 inches. In theory
ng is r 7£ lines and the amplitude of the thrust 23 inches.
EXPERIMENT g. Further, when the diameter of the internal orifice
was 3¾ lines and the diameter of the exterior orifice 5 lines, the de-
scent ng was almost the same as in Experiment 7, namely, about
7 inches. But the amplitude of the thrust was greater, that is, r r inches.
In theory ng is 6 inches r r lines and the amplitude of the thrust almost
r r inches.
EXPERIMENT ro. Finally, with the diameter of the interior orifice
being 3% lines and the diameter of the exterior orifice 2½ lines, the
descent ng was about one inch and the amplitude of the thrust 23 inches.
In theory ng is 14 lines and the amplitude of the thrust is 22½ inches.
Actually all these experiments agree splendidly with the theory;
perhaps a greater agreement would have resulted ifit would have been
possible to obtain the measurements of the orifices with greater
accuracy; nevertheless, no one, as I believe, is displeased by those
minimal differences in numbers. Moreover, they arise for the most
part from the compression of the water at AC which is produced while
the volume elements entering the conduit through the interior
orifice lose part of their motion; hence the amplitude of the thrust is
slightly greater and the descent ng is less in theory than in the experi-
ments; I did not wish to add the measure of this matter, although it
would have been within my power, lest the calculation become more
intricate.
 THIRTEENTH CHAPTER
Concerning the Reaction of Fluids flowing out of
Vessels and the Impetus of the Same) after
They have Flowed out) on the Planes
against which They Strike

§1. Water, while it is being ejected from a vessel, acts in the same
way against the vessel out of which it is flowing as a shot against the
cannon or rifle from which it is expelled: it certainly repels the vessel.
And this, indeed, Newton already noted in Principia Mathematica
Philosophiae Naturalis, first edition, p. 332, and from this he correctly
deduces the ascent of mortar shells which are filled with gunpowder
properly mixed with charcoal. For after the material has been
ignited, it projects the mortar shells upward while it expires slowly
through the orifice.
But neither did the cited author (since it was not in accordance
with his purpose) handle the argument generally enough for the
importance of the matter, nor did he give the true measurement of it.
Finally, in the two later editions he ignored it altogether. However,
he considered that that force of repulsion is equal to the weight of an aqueous
cylinder the base of which is the orifice transmitting the water and the height of
which is equal to the height of the aqueous surface above the orifice. Indeed,
this quantity is deduced correctly from the opinion which Newton
favored at that time about the velocity of the water flowing out of a
vessel, when he stated that the water can ascend to one half the
height of the surface by its own velocity.
But just as now the falsity of the latter proposition is unknown to
no one any longer, so also the defect of the other anyone hence easily
gathers, although at first glance it seems true enough.
§2. At first we will consider the matter in the very simple case in
which, certainly, we assume the water to flow horizontally out of a
vessel of infinite area. Moreover, I have demonstrated that the total
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 3I 5

force of repulsion is not present immediately at the beginning of flow

unless insofar as the total velocity itself is present in the water flowing
out, such that if the vessel is not of infinite area, the force of repulsion
together with the velocity of the water flowing out increases little by
little, or even decreases, according to the nature of the circumstances.
However, at first let us disregard these instantaneous changes by
assuming that the flow from an infinite vessel becomes constant.
And thus the force of repulsion is best defined if whatever force is
required for producing the motion is sought. Indeed, to this end one
has to look not only for the velocity of the water flowing out, but also
for the quantity of it; but the quantity depends partly on the magni-
tude of the orifice and partly on the contraction of the stream, which
latter is variable; indeed, we saw in Chapter IV that it can be
entirely avoided. If some contraction exists, nevertheless, the section
of the most greatly contracted or attenuated stream is to be considered
instead of the orifice, and then I say that the force of repulsion will be
equal to the weight of an aqueous cylinder the base ~f which is the orifice trans-
mitting the water ( that is, the section of the most greatly contracted
horizontal stream) and the height of which is equal to double the height of the
aqueous surface above the orifice, or, more accurately, to double the height
appropriate to the velocity of the water flowing out. Therefore, if there is no
contraction of the stream, just as there is none when water flows out
through a short pipe, the repulsion will be twice or almost twice as
great as that which was defined by Newton.
§3. In order that we may show this proposition, a certain Mechan-
ical principle will have to be considered here, for which I have often
found use in solving other questions. The principle is this:
lj a body has acquired the same velocity from rest through direct motive
pressures, variable in arry way whatever, and if the individual pressures are
multiplied by their proper differential times, the sum of all the products will
always be the same; that is, if the pressure is p and the differential time is dt,
f
then p dt will be constant. I showed this matter more clearly in the Com-
mentaries of the Imperial Academy of Science of St. Petersburg, Book I, p. 132.
§4. Let us assume now a cylinder of practically infinite area from
which water flows out horizontally at uniform velocity; let us dis-
regard the influence which gravity exerts on the particles after they
have flowed out, so that the individual ones continue to be moved
horizontally and uniformly. But the particles are accelerated and
they experience pressure as long as the maximum value of the velocity
is not yet present, and they obtain this value when they have arrived
at the place of the greatest contraction of the stream. For this reason
I said that the section of the stream formed at that place is to be
 HYDRODYNAMICS, CHAPTER XIII

considered instead of the orifice of efflux. Let the area of that section
be 1, and let the water there have a velocity which is due to the height
A. Let it be assumed that the cylinder of water has flowed out which
has I for its base and L for its length. If the time is expressed by a
length divided by a velocity, the velocity appropriate to the height
will have to be expressed by V 2A, and the time of flow by _ ; .
v2A
With these things set forth in advance, we will investigate the motive
pressure which can impart the velocity V 2A to the cylinder Lin the
time _ ; . Let the pressure be p, and let it be considered, for the
v2A
sake of a shorter calculation, to have acted during the time t and to
p dt pt
have given the velocity v to the cylinder. Then dv = Land v = I,'
Lv _ 1- L
whence p = ,:· Now let v 2A be substituted for v and V A for t,
2
and thus p = (LV 2A) I(v:A) = 2A. Therefore, the pressure con-

stantly exciting the water to efflux is equal to the weight of the aqueous
cylinder of which the base is the above-defined orifice transmitting
water, the height of which is equal to double the height appropriate
to the velocity of the water flowing out; and also just as great is the
reaction which repels the vessel. Q.E.D.
§5. The proof is the same if the water flows out not through an
orifice but through a horizontal cylindrical pipe at a constant velo-
city, or even through a pipe of size varying in any way. This latter
can also be proven directly if the pressure required in the individual
particles is expressed correctly so that these [particles] receive the
required increments or decrements of velocities.
§6. The height which we called A differs very little indeed in ex-
periments from the height of the water above the orifice of efflux,
especially if the water flows out from a very large vessel through a
simple orifice which is not very small. But the orifice of efflux more
often differs notably from the minimum section of the stream, which
we consider as the orifice transmitting the water; the quantity of water
flowing out in a given time, compared with its velocity, indicates this
in experiments.
Hence it occurs that our proposition of §3, after it has been chal-
lenged by experiment, ordinarily does not differ much from the pro-
position of Newton shown in §1. But if everything is carefully avoided
which can produce a contraction of the stream and which can dimin-
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 3I 7
ish the velocity, the repelling force according to our theory will
become almost double that which was defined by Newton, and
accordingly, such a value is also confirmed by experiments.
But in order that we may bring the matter clearly to light and treat
it now rather generally, we will handle it so that we determine the
repelling force from the beginning of flow, while the velocities are
being changed continuously; and, indeed, our first theory does not
have meaning other than when the velocity remains unchanged. In
order that we may be more intelligent in handling this slightly more
intricate question, it will help here to set forth certain rather general
things in advance.
§7. Momentum is the product of the velocity and the mass. If the
velocities are unequal, the absolute momentum will result if the indivi-
dual particles are multiplied by their own velocities, respectively, and
the sum of the products is taken. The momentum is generated by the
motive pressures acting for a given time, and the effect is to be con-
sidered equal to the cause. Therefore, the sum of the motive pres-
sures multiplied by their proper differential times is to be evaluated
from the momentum generated. And because any motive pressure
reacts on the vessel from which the water flows out, the total repelling
force for any instant whatever will be equal to the new momentum
divided by the differential time in which it is generated. With these
things having been set forth, I proceed to the question itself.
§8. Therefore, let the vessel ACDB (Fig. 82) be of infinite size, and
let the tube EHID, the areas of which are assumed unequal in some
way, be fastened horizontally to it. Let the area of the orifice HI
be I, and the length of the tube be m. The velocity at HI, variable
B

I
i

--..::::~~_;;~=::::=:::;~H I1
===~r Jr

FIGURE 82
 HYDRODYNAMICS, CHAPTER XIII

in any way, equals V 2v, or that which is due to the height v. I say at
first that the absolute momentum of the water contained in the tube
will be equal to mV 2v, that is, as if the tube were cylindrical and
equal in area to the orifice HI, because, of course, the velocity of any
sectionFGg.fis reciprocally proportional to the mass [sic.]
Now, indeed, let us consider that in a given infinitely small differen-
tial time there flows out through the orifice HI the small column
HLMI, the length HL or IM of which we consider equal to ex. The
mass of this column will be ex, and it will have the momentum exV 2v,
but in the same time the mass of water contained in the tube acquired
mdv . 1-
the momentum . 1_ (for it had mv 2V). Therefore, the absolute
·y 2V

. h e given
momentum generate d mt . d"ff, . I time
i erentla . . exv
is . ;-2v + .m1 _;
dv
V 2V
but if this is divided by the same differential time (which is to be ex-
pressed by ;
2
J, as we saw in §7 the required pressure repelling the

vessel will result, which, therefore, if it is called p, will be

• 1- m dv)/ m dv
p= ( exv 2 v + V 2V
ex
V 2V' or p= 2V + -ex•
(ex) It appears from this that the last definition of the question
depends on the ratio which exists between dv and ex; this, in fact, we
defined generally in Chapter III: however, no attention was paid to
the hindrances which are due to this case. Therefore, the shape of
the tube also contributes something here.
(/3) Further, it follows that, if the flow is considered uniform, p is
constantly equal to 2v, because then dv = o. In fact, this conforms
with what we showed in §5. But, while the flow is being increased
(which certainly it does noticeably, and this for long enough time if
the conduit EI is rather long), the vessel experiences a continuously
different repelling force.
(y) At all times, dv has a real ratio to ex. Therefore, the repelling
force is never null, so that from the first instant of flow the vessel is
repelled, even if hardly any water then flows out on account of its
trifling velocity. Truly, in order that the general use of our rule be
clear to everyone, we will now apply it to a special case by attributing
a cylindrical shape of area I to the tube EHID.
§9. Therefore, if the tube is assumed cylindrical, entirely open at
HI, with the other assumptions and designations having been re-
tained, the live force of the water contained in the tube will be mv; the
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 3I 9

increment of this is m dv, to which is to be added the live force of the

little column HLMI, or av, and their sum is to be made equal to the
product of the height of the surface AB of the water above the orifice
HI, which we will call a, and the differential mass a. Therefore,
. dv a - v
m dv + av = aa, and from this there results - = - - . But with
a m
that value substituted in the equation of the above paragraph, there
results p = a + v from which I deduce the following conclusions:
(a ) The length of the tube contributes nothing to the repelling
force which the vessel sustains if the velocity is assumed to be the same,
because the letter m vanished from the calculation. However, this
length (just as we showed in the above more than well enough)
causes the velocities to assume faster or slower increments, for the
longer the tube the more slowly the water will be accelerated, and
vice versa, so that it acquires in an instant from rest its maximum rate
of speed if the length of the tube is null. But if this same tube is of
infinite length, the water can acquire a noticeable degree of speed
only after an infinite time.
(f3 ) Therefore, it can occur, when the height of the water has not
been changed, the expenditure of water being howsoever small, that
the repelling force is notable and lasts arbitrarily long. And, indeed,
this can be obtained in a double manner, either by prolonging the
tube or by closing off the orifice rather often before the water has
attained a notable velocity. However, the former method assumes a
free flow of the water through the tube; indeed, when the flow of
water has been retarded by external hindrances, never to be avoided
in overly long tubes, the repelling force is also diminished.
(y) Let me be allowed here to mention in a few words a certain
proposition from Principia Mathematica Philosophiae Natura/is, 2nd
edition, of Newton. After he had changed his thinking shown in the
first edition of the cited work about the velocity of water flowing out
of a vessel, and after he had recognized in the second edition that, if it
is ~jected vertically upward, it ascends to the full height of the surface
of the water, the Author prest>nted the following words in the second
book, proposition 36, corollary 2 : The force by which the entire motion ef
the water flowing out can be generated is equal to the weight of a small cylin-
drical column ef water, the base ef which is the orifice EF (see Newton's
figure ), and the height ef which is 2GI or 2CK. That thinking was once
opposed by me and by some, and again confirmed by others. But
now, after I have thought about this theory of moving water, it seems
to me that the dispute is to be settled thus: when the water has arrived
at a uniform motion, which, certainly, is Newton's hypothesis, then
320 HYDRODYNAMICS, CHAPTER XIII

that force is defined correctly by the height 2GI, but at the beginning
of flow, when the velocity is still null, the force corresponds to the
simple height GI, and soon, with the velocity increasing, the force
animating the water to efflux increases simultaneously, and finally it
rises to that magnitude which Newton assigned. Now these things
are obvious to anyone, because the force generating the motion of the
water about which Newton speaks cannot but be equal to the repel-
ling force, which we saw to be equal to a + v. Also the Illustrious
Ricatti, with whom I had a discussion concerning this argument, when
asked whence that force corresponding to twice the height of the water could
arise, whereas it is apparently manifest that, with the orifice blocked off, the
volume element adjacent to the latter is pressed by the force corresponding to
the simple height, answered that one must distinguish the state of rest from
the state of motion.
§10. If the tube attached to the vessel is not cylindrical, the calcu-
lation will have to be performed thus:
Let the area of the conduit at FG or Jg bey, the distance of the sec-
tion FGgf from the orifice ED be x, and let the other designations be
retained. The live force of the water contained in the tube will be
vJ~' and its increment will be dvJ~' to which, as it was done in the
preceding paragraph, is added the live force of the small column
HLMI, or av, whereupon dv J~ + av = aa, from which it thus appears
that
dv = (a - v)
-;; /Jdxy ·
With this value having been substituted in the equation of §8, there
results

jJ = 2v + m(a - v) / J~-
Therefore, since in the uniform flow of water v = a, it follows that
p= 2a. In addition, as long as the flow of water is accelerated, the
motion of the water in the vessel ACDB near the orifice DE, which we
have disregarded in all this work, is not to be neglected here. But
that motion cannot be determined correctly, and, therefore, the ex-
pression which I gave for the repelling force does not apply accurately
if the water has not yet been understood to flow uniformly; but when
the water flows steadily, the expression prevails very accurately.
§11. After we have thus proven for the uniform efflux of water that
the repelling force is always equal to the weight of an aqueous cylinder
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 32 I

constructed above the orifice and rising up to double the height of the
water, it is pleasing to show it also indirectly by reduction to the absurd,
so that also those not knowing the rules of mechanics may perceive
the truth of this somewhat paradoxical proposition.
To this end we will consider water flowing vertically down from a
cylinder, disregarding the hindrances taking something from the
velocity of the water and [disregarding] the contraction of the stream,
which can be avoided. The vertical pipe which is seen in Fig. 76
corresponds to the orifice, and all the things behave as stated in Chap-
ter XII, §13: the water has constant flow; the walls of the vessel and
conduit are understood to be free from gravity; the height of the
cylinder is assumed equal to a, and the height of the small tube equal
to b; the height cF = x; the area at E equals 1. The area at F will be
v'a+b
v' , and at C 1t. will
. v'a+b
be v' Finally, the area of the cylinder
a+ x a
is set equal to M. After these things have been assumed, we will seek
the weight of all the water ABCE. We will express the weight of the
water ABC by Ma, and thus the weight of the water CE will be
2a + 2b - 2v' aa + ab; therefore, the weight of all the water ABCE
will be Ma+ 2a + 2b -2v'aa + ab. Therefore, under the assump-
tion that the water is at rest in the vessel and the tube, the force
required for suspending the water is Ma + 2a + 2b - 2v' aa + ab.
But now we will investigate a similar force when the water flows out
through E at its full velocity (by which certainly it can ascend to the
height a + b). But this will be obtained if the repelling force is
subtracted from the former force. If, therefore, this repelling force
is assumed, as we stated, equal to 2a + 2b, the force suspending the
water during flow will be Ma - 2v'aa + ab.
But, indeed, assume that the pipe CE is not present, and through
our same rules the suspending force while the water is discharging
through the orifice C will again be Ma - 2v'aa + ab, indeed because
the weight of the water ABC is Ma and because the area of the orifice

C is v'T, which, multiplied by double the height a, gives

2v' aa + ab. Therefore, our estimate of the repelling forces shows

that the suspending force during the effiux of the water is the same
whether or not there is a small tube present, and the tube may have
any length whatever as long as it has the shape described in §13,
Chapter XII; and the necessity of this agreement and identity ap-
pears also without calculation from the very nature of the matter,
322 HYDRODYNAMICS, CHAPTER XIII

because the tube thus formed makes no change in the water flowing
through, since the stream of water assumes of its own accord the same
shape that the tube has, as long as the water coheres. But ifwe esti-
mate the repelling force differently, we will generally never obtain
that agreement between suspending forces. Thus, for example, if
according to common sense we say that the repelling force is equal to
the weight of the often mentioned simple cylinder, then, while the
water is assumed to flow out of the vessel ABC through the conduit
CE, it will be a + b; and if this force is subtracted from the weight of
the entire water ABCE, or Ma + 2a + 2b - 2-Vaa + ab, there
remains Ma+ a+ b - 2-Vaa + ab, which is the force required for
suspending the system ABCE while the water flows. Moreover, we
saw that this force must be the same if the conduit CE is absent: but
then the suspending force is Ma - Vaa + ab, because the weight of
the water ABC is Ma and the repelling force by hypothesis is the simple
cylinder erected above the orifice C to the height a. Therefore,
according to this hypothesis it should always occur that

Ma +a+ b - 2-Vaa + ab = Ma - v aa + ab,

or
a+ b = Vaa + ab,
which is absurd. A similar absurdity can be shown if the stream is
considered to ascend vertically upward, and here in vain it would be
stipulated for confirming common sense that the stream flowing out
of CE cannot be assumed continuous unless some viscosity of the water
is assumed at the same time (for otherwise the stream will be broken
off in little drops directly in front of the orifice) and that the viscosity
changes the state of the situation; for, certainly, neither are the velo-
cities of the water changed by the mutual cohesion of the water at CE
nor do the sides of the conduit CE sense any pressure, just as I demon-
strated in Chapter XII, §13, so that I may pass in silence over the fact
that the cohesion of the water does not arise from viscosity but from
some other magnetic property or from mutual attraction, by virtue of
which the center of gravity of no system can acquire either a greater
or a lesser velocity. But, clearly, this exception for vertically ascend-
ing streams [taken] by opponents has no significance when the water
remains there continuously, even if the water has no viscosity or
mutual attraction.
But I could confirm our thinking in infinite other ways and by
particular examples if I wished to pause here for a longer time.
Thus, for example, in Fig. 29, described in Chapter V, §4, if the height
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 323

NS = 1, the orifice LM = 1, and the orifice RS = 2, then PB = ½;

the repelling force which develops from the flow of water through RS
is equal to 2 · ¾= ½, and I can show that the repelling force which
results from the effiux of water from the simple cylinder RN through
LM is also½, and thus that the total repelling force is ¾, which con-
stitutes precisely double the aqueous cylinder above the orifice LM
standing to the height NS + PB. Moreover, such an agreement in
no way appears in other theories erroneously conceived, so that there
can be no further doubt concerning ours, except by those utterly
unskilled in these matters. But if I wished to prove that what I said
-that the repelling force of the water flowing out of the simple cylin-
der RN through LM is ½-it is required that the repelling force be
defined when water flows from a non-infinite vessel at some given
nonvaried velocity. But lest I become too involved in this matter,
I leave this to be accomplished by others, and it should now no
longer be a great task. I proceed to other things.
§12. The proofs which we gave up to now are valid only for straight
tubes in which certainly the motive force of any of the volume ele-
ments and the repelling force arising from it are in accordance with
all the others, and they have a common direction. But when the
tubes attached to the vessel through which the water flows are
curved, another method of proof is to be employed. In order that
we omit nothing further in this new argument, we will also show this
case. And it will not happen that one regrets the work, since from
this will appear the true laws of pressures which nature follows, not
only in these cases but also in many others.
§13. And so let us consider a tube attached to an infinite vessel,
certainly of uniform area, but curved according to any curvature AS
whatever (Fig. 83), so that A is the point of insertion and S the point
of effiux. Let the tangents at A and S be drawn, namely AR and SB,
and let AB be perpendicular to SB. The velocity of the water flow-
ing through the tube will be uniform and that which is due to the
height A. The area of the tube everywhere is r. I Sl!J' that the total
repellingforce taken in the direction SB will again be 2A, and this alone will
be present.
For the sake of the proof, the infinitely close lines nq and ep are drawn
perpendicular to SB, and nm is drawn parallel to the same SB; let
Sq .= x, qp = dx, qn = y, and em = dy. The radius of the curve at en
will be -}~:Y, the elements en, which I will call ds, having been con-
sidered as constants. Moreover, the little column of water inter-
cepted between e and n has a centrifugal force to be determined thus:
 HYDRODYNAMICS, CHAPTER XIII

the gravity of the column is ds (because its base is I and its height is
ds) and if the radius of the curve were 2A, there would result, by the
theorem of Huygens, the centrifugal force of the particle equal to its

,---~------l}'
f

FIGURE 83

gravity, and the centrifugal forces are, other things equal, in recipro:
cal proportion to the radii. Therefore, the centrifugal force of the
. 1e co1umn 1s
11tt · - 2dy
A ddx ; t h"1s centn"fiuga 1 10rce
r · expresse d b y ec
1s
perpendicular to the curve, and co is drawn parallel to BS itself.
The force ec is resolved into oc and eo. There will be (on account of
the similarity of the triangles eoc and nme) the force

-2A ddx
oc = ds
the force
-2A dx ddx 2Addy
eo =
dyds ~

(on account of ds being constant).

But the elementary force oc alone acts in the direction SB, while the
other eo is to be neglected with respect to this direction. Let the
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 325

integral of the elementary force oc be taken with such a constant that

the integral vanishes together with the abscissa. This integral is
2A dx .
2A - ~ ' because at S 1t occurs that dx = ds. Now, in order that
the force be obtained in the direction of the tangent SB for the entire
tube,~ is to be substituted for!; therefore, the entire force follow-

ing the tangent SB is 2A -

2
Ai~). This certainly develops from
the centrifugal force of any volume element whatever, but another
force remains to be considered: namely, when the water flows con-
tinuously from the infinitely large vessel into the tube at a uniform
velocity corresponding to the height A, the vessel is repelled along the
direction RA by a force 2A (by §4); if this is resolved into a tangential
2
along SB and a perpendicular along BA, the prior quantity Al~B)
alone will have to be considered. And because it has a direction
2
common with the force 2A - Ai~B) developing from the centri-
fugal force and just defined, it will have to be added to the same, and
2A(RB) 2A(RB) .
thus the sum 2A - RA + RA , or 2A, expresses the repellmg
force in the direction of SB.
In order to show further that the vessel is repelled in no other
direction, we will return to the elementary force eo, which we saw to be
2 2
Ad~dy' the integral of which is Aj!B), which is cancelled precisely
by the force 2A repelling the vessel in the direction RA after the latter
has been resolved in the proper manner. Q.E.D.
§14. This simplicity of the most general theorem, by which cer-
tainly the repelling force in the direction contrary to the uniformly
flowing water is indicated constantly by 2A, can be the argumentum ad
hominem, as it is called, for its excellence, against those who either do
not understand our reasoning or who do not desire to examine it with
sufficient attention. If, in truth, one states that the repelling force
of the water flowing into the tube in the direction AR from the in-
finite vessel is A, one sees that the system is repelled in the direction
SB by a force which is 2A - AC::), which is absurd, as even the
formula itself seems to indicate to me. And in this opinion the force
in the direction perpendicular to the former would not be zero. For
the vessel should be pressed back in the direction BA by the force
 HYDRODYNAMICS, CHAPTER XIII

A ~A:), which again to me is absurd, and the falsity of this I learned

from experiment in the case in which ARS was a right angle and
AB= AR.
Many other theorems in favor of this argument, taken in the full
extension which it can have, could be elicited and demonstrated for
the flow of water not yet uniform through a tube irregular in any way
whatever, if only at the same time attention is paid to what was
pointed out in §8. But because there is not space [enough] to go
through the individual ones, I progress to examining another force
equal to the prior but in the opposite direction, that indeed which a
stream flowing out exerts on a plane when it impinges perpendicu-
larly on it.
§15. Concerning the impetus of an aqueous stream impinging on a
plane, many have written and performed very many experiments.
I also contributed something to this matter in the Commentaries of the
Imperial Academy of Sciences of St. Petersburg, Book II. Experiments are
conspicuous in the works of Mariotte in his traite du mouvement des eaux,
in the History of the Academy of Science [Paris] contributed by Mr.
Duhamel, p. 48, and elsewhere. Indeed, they do not all agree very
closely; nevertheless, most of them seem to indicate at first glance that
the pressure of the aqueous stream flowing uniformly is equal to the
weight of an aqueous cylinder, the base of which is the orifice through
which the water flows, and the height of which is equal to the height of
the water above the orifice. To this thinking the majority, in fact all,
adhered and do adhere up to this time, because it agrees wonderfully
with other experiments also, especially those which are customarily
performed on spheres moved in a resisting medium. Therefore,
I myself followed the same [thinking] in the cited St. Petersburg
Commentaries, although many made my mind uncertain, and I indeed
did not hesitate in this work itself, which I have at hand, to make use
of that as an example in Chapter IX, §31 and §32. However, after
I had thought over the matter more attentively, had applied new
principles, and at the same time had undertaken other experiments of
a new type, at last I saw clearly that that common opinion about the
impetus of a stream of water had to be changed in the same manner
as that of Newton about the repelling force, so that in place of the
orifice the section of the contracted stream should be considered, and
in place of the height of the water twice the height corresponding to
the actual velocity of the water should be applied. For I have proof
that the force of repulsion shown in §2 is entirely equal to the impetus
of the stream if all of it strikes the plane perpendicularly; it follows,
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 327

hence, that the impetus of the stream is greater, the smaller the con-
centration of the stream and, with the latter simply vanishing and the
water erupting at the same time at the full velocity which it can have
in theory, then the impetus is twice as great as is commonly stated.
Indeed, because the velocity always lacks something and the stream
is seldom not contracted to almost one half, it is a fact that most
experiments have seemed to support the simple height in the Cylinder
in estimating that impetus. Moreover, I would wish that it be noted
properly that I discuss here only solitary streams which the planes
receive entirely, but not fluids surrounding bodies and making an
impetus on the same, such as Winds or rivers. Indeed, I say that
these two types of impetus which authors have confused up to this
time are to be distinguished properly from one another on account of
reasons to be'explained briefly below.
§16. With respect to the aqueous stream I think as follows: I assume
that water flows out horizontally at a uniform velocity from the in-
finitely wide vertical cylinder ABM (Fig. 84) through the lateral ori-

fice CM, and that the stream impinges perpendicularly against the
plate EF; thus I see easily, since the following particles hinder the
prior ones so that they cannot rebound, that the individual particles
will be deflected to the sides, and this in a motion parallel, or almost
so, to the plate EF (if only the latter is large enough so that the entire
stream, however dispersed, is intercepted). And because all things
are in a state ofpermanence, it is permissible to assume that the plate EF
is fixed to the vessel and that the stream is surrounded by the lateral
 HYDRODYNAMICS, CHAPTER XIII

surfaces CHDGLM, so that the water can be assumed to flow out from
the vessel ABCHDEFGLM through the circular opening DEGF. If
this were so, we have shown in §13 that the volume elements flowing
out at DE would certainly produce a repelling force in the direction
of EF; but at the same time it appears that the repelling force at CF
is opposite to the former, so that here no attention has to be paid to
this class of repelling forces. But as far as the direction perpendicular
to the layer EF or to the cylinder BC is concerned, we showed at the
end of the same §r 3 that in this direction clearly no repulsion occurs.
Therefore, the plate EF is propelled just as much as the cylinder is
repelled. And this is what I wished to show. And hence it follows
now that the total pressure ef the aqueous stream which strikes the plate is the
same as the weight of the aqueous cylinder which has as base the cross section of
the stream (after the latter has reached a uniform area) and as height twice that
required for the velocity of the water (after this has similarly been made
uniform).
§17. I do not doubt that there will be many to whom this wholly
new proposition seems suspect and contrary to experiments. Indeed,
I would wish those to consider that the experiments performed up to
now by no means correspond accurately to the common rule, and in
most cases our Rule differs little from the common, although in theory
they are greatly different; then also I want those to have been in-
formed beforehand that I have undertaken other experiments which
individually confirm my thinking exactly, and clearly reject the old
one! At the end of the Chapter I will review the experiments per-
formed by me. Perhaps also the method of proof which I used will
seem insufficiently accurate to some, but I have another direct proof
which is supported by a new Mechanical property once observed by
me, and which I will communicate here, both because anyone can
deduce the said proof very easily, and also because he can apply the
same to other uses. And thus it is presented.
If a body is moved at a uniform velocity but changes its direction continuousry
by any cause whatever acting in any way whatever until it has acquired a
direction perpendicular to the first, and if the individual pressures deflecting the
body are resolved into two groups, the one parallel to the first direction, the other
perpendicular to it, and, finalry, if the individual parallel pressures are multi-
plied by their proper times, I say that the sum of the products will be constantly
the same and indeed equal to that which can generate the entire motion from rest
or absorb the entire generated motion.
In this dynamical relation, if we use it in our present problem, the
plate EF is to be considered, which by its reaction on the water
changes the direction of the latter until it has become perpendicular
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 329

to the first. Therefore, the proposition of the preceding paragraph,

with the help of this relation, will be proven in the same manner which
we used in §4 for determining the repelling force with the help of the
principle shown in §3. Therefore, that idea, which we must under-
stand about the impetus of water, seems true; however, it assumes
that the individual particles of water rebound to the sides in the direc-
tion of the plate, by which pattern I observed the water not always
to recede; indeed, I even saw that some particles, though few, spring
backwards; however, the latter produce a greater pressure than those
which are deflected to the sides. And from this itself I am convinced
firmly that if an aqueous stream with a great impetus impinges
obliquely against a plane, for example at an angle of thirty degrees,
a pressure will thence arise which is more than half of that which arises
from the same stream impinging directly, while according to the ordi-
nary rules it should exert exactly half the force; the reason for this
matter is that in an oblique impulse more particles can rebound than
in a direct one, in fact almost all [can rebound] if the velocity should
be great.
However, if all particles are assumed to rebound so that the angle of
incidence is equal to the angle of reflection, then each impulse will
have to be considered the same. The best method of estimating the
pressures of the water here is that in which the reasoning is supported
a posteriori.
§18. It follows further from the previously discussed well-known
relationship that the same effect arises from the pressures whether the
plate deflects the water to the sides or a cause is assumed which
absorbs all the motion which the aqueous particles having flowed out
of the cylinder have acquired. Hence it is understood what would
happen if the orifice CM (Fig. 85), through which water flows out of
the cylinder ABM, were submerged in other water standing in the
vessel PQFE. Certainly the cylinder ABM would be repelled against
PQ within the vessel PQFE if the latter were not connected to the
cylinder; but if the vessels were fixed to one another, the system would
undergo no prevailing pressure; for as great a pressure as there is
against PQfrom the water flowing out, also as great an opposite pres-
sure develops against EF from the continual destruction of the motion
which the particles having flowed out of the cylinder have acquired.
§19. I spoke about the pressure of a stream which, even if expanded,
is intercepted completely by a plate. I come to the other type of
impetus of water which indeed plates submerged in fluid on all sides
sustain; however, I consider that this cannot be defined absolute!:)!,
because the individual particles impinging on the plate are deflected
33° HYDRODYNAMICS, CHAPTER XIII

differently. But if the deviation of any particle whatever is assumed

to be known, the solution of this question will no longer be difficult
after the theorem which we used in §17 has been changed a little, and

FIGURE 85

this, given more generally, is as follows: if the angle of the change ef

direction in a moving body is not right, but less than right, then the sum of the
products ( which was the subject ef the discussion before) will also be less in
proportion as the sine ef the changed direction is to the entire sine.
Therefore, for any particle whatever it should be investigated how
much it is forced to change its direction of motion by the obstacle or
by the plate placed across its path. But this sort of definition can
hardly be shown accurately in theory; experience does not prove the
theorems customarily brought forth in this matter: such as that the
force of a stream impinging directly against a disk is twice as great as
the force of the same stream against a sphere of the same diameter,
and others which are similar. However, the fact that the quantity
of pressure for a sphere which is given customarily by authors agrees
accurately enough with experiments made by Newton and others and
recounted in Principia Mathematica Philosophiae Naturalis is, I consider,
after thinking over everything well, to be attributed to a fortuitous case.
In the Commentaries ef the Imperial Academy ef Science ef St. Petersburg,
Book II and following, I gave theorems which were developed for
motion in resisting media, considered theoretically, and also several
physical observations. Therefore, I do not wish to repeat those here,
although they pertain to our purpose; there is not space to delay any
longer on these hydrodynamical meditations. Hence, I hasten to the
end. I pursued this new theory about the reaction and the impetus of
fluids, which upsets the opinion accepted by all authors up to now in a
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 33 I

matter of great importance, in a singular Dissertation which should be

inserted in its own time in the Commentaries of the Imperial Academy of
Science of St. Petersburg, and I confirmed the same with indubitable
experiments. I come now to another argument, not at all unworthy
of the attention of Geometers.
§20. It entered my mind at one time that these things which I had
pondered about the repelling force of fluids while they are ejected,
which I exposed here for the most part, can be applied usefully to
instituting a new method of navigation. For I do not see what would
hinder very large ships from being moved without sails and oars by
this method: the water is elevated continually to a height and then
flows out through orifices in the lowest part of the ship, it being ar-
ranged so that the direction of the water flowing out faces towards the
stern. But lest someone at the very outset laugh at this opinion as
being too absurd, it will be to our purpose to investigate this argument
more accurately and to submit it to calculation, for it can be useful,
and it is very fertile for many geometric investigations.
Let me begin with this, for which then it will appear under what
circumstances the maximum success should be expected from that
new navigation.
§21. Therefore, it is to be noted that a ship is retarded continuously
by water drawn in on account of the inertia of the latter when the
same velocity is communicated to it at which the ship is borne, and
while it is communicated, the ship is forced backwards by the reaction
of the water, but at the same time it is pressed forward by the effiux
of the same. That meeting of the contrary actions places limits to
the force propelling ships to be obtained from a given absolute poten-
tial. For, if the prior action were not present (which, to tell the
truth, I did not consider for a long time), a force, however great, for
propelling ships could be obtained by the work, however little, of men, which I
thus demonstrate.
In Chapter IX (see especially §26) I showed that the work of men
expended in elevating water, which I designate by the term absolute
potential, is to be estimated from the product of the quantity of water
multiplied by the height of elevation, so that, for instance, by the
same labor, according to all measures, both four cubic feet can be
elevated to a height of sixteen feet, and sixteen cubic feet to a height
of four feet. Now, I say further that a uniform pressure is present,
propelling ships forward, as long as the fluids flow out at an equal
velocity; this pressure is to be estimated from the quantity of water
flowing out and from the root of the height of the water placed in the
vessel above the orifice; let the quantity of water flowing out in a given
332 HYDRODYNAMICS, CHAPTER XIII

time be Q and its height be A, then the magnitude of the orifice

spewing forth the water will have to be considered proportional to the
quantity ,h for that time; but the repelling force which in this case

promotes the ship is indeed equal to the magnitude of the orifice

multiplied by twice the height of the water (by§4), that is, equal to the
quantity ,h 2A, or 2Q VA. From a comparison of both proposi-

tions it follows that the labor of men engaged in elevating water for
thence obtaining the force for propelling ships is as QA is to 2Q VA,
or as VA is to some constant quantity. Therefore, the less the height
to which the water is elevated, the greater a force propelling vessels is
obtained from the same labor, so that by work of men, however little, an
arbitrarily great force for propelling ships can be obtained. But also the
inertia of the water which is taken in (about which we spoke at the
beginning of this paragraph), retarding the ships, obtains a greater
proportion to the force propelling the ships, the less the height A
is taken, to which proper attention is to be paid here.
§22. It is clear from the preceding paragraph that the height to
which the water is to be elevated is of the class of those which some-
where have a maximum. But in order that the height most bene-
ficial to our purpose be determined, other questions present themselves
to us for being examined first.

PROBLEM

Let a ship be assumed to progress at the uniform velocity which is

generated by a free fall through the height B, and let it be assumed
that water flows continuously into the ship, such as in the form of
rain, and certainly at that quantity which a cylinder constantly full
to a height A would supply through an orifice of magnitude M, with
all alien hindrances removed. There is sought how much resistance
the ship experiences from that perpetual and uniform inflow of water
and its inertia.
SOLUTION. Let any time t whatever be assumed; if this is established
from the distance which the fluid flowing in travels at its own velocity
divided by the same velocity, then the velocity is to be expressed by
V 2A, and the quantity of water flowing in during the time twill be
equal to the cylinder constructed above the base M of length tV 2A.
But that quantity receives in the time t, while it is discharged from the
ship, the velocity due to the height B, to be expressed by V2B;
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 333

therefore, the uniform force is to be sought which can in the time t

communicate the velocity V 2B to the aqueous cylinder Mtv 2A, and
that force will, on account of the reaction which acts on the ship,
have to be considered equal to the resistance sought. Let the pre-
viously mentioned force be p, and let it be assumed to have given the
velocity v to the aqueous cylinder Mt'V 2A in the time 0, and there will
result dv = p:;2A,and v
Mt
=
Mt
pJ-·
2A
Now let V 2B be substituted
for v, and t for 0, and it will occur that V 2B = ~ , or p =
M 2A
2MvAB.
Therefore, the resistance sought is equal to the weight of an
aqueous cylinder of which the base should be equal to the orifice M
and of which the length should be equal to double the mean propor-
tional between the heights A and B.

PROBLEM

§23. Let there be a cylinder in the ship of height A above the surface
of the sea, through the orifice of which, placed at the same surface,
of area M, water flows out toward the stern without any impediment,
and let the cylinder be kept constantly full of water. Determine the
force propelling the ship continuously.

SOLUTION

The force propelling the ship is equal to the reaction of the water
while it flows out, or to the repelling force diminished by the force
defined in the preceding paragraph developing from the intertia of
the water which is continuously drawn in. The repelling force is
equal, through §4 of this chapter, to 2MA, and this advances the ship;
the other force, which retards the ship, is, through the preceding
paragraph, 2MYAB. Therefore, the absolute force advancing the
ship is 2MA - 2MvAB.
§24. COROLLARY. If the ship has no velocity, the force urging it
will be 2MA; but if the ship is moved at the same velocity at which
the water flows out in the opposite direction, it occurs that B = A,
and then the ship will be propelled by no force. If, then, the ship
were moved very freely across the sea, it would nevertheless not
acquire from the action of the water which is taken in continuously
and flows out below a velocity greater than that at which the water
flows out, not because the water flowing out from a uniformly moved
334 HYDRODYNAMICS, CHAPTER XIII

vessel repels the vessel with a lesser force than from an unmoved
vessel, but because then the inertia of the water produces a resistance
equal to the repelling force.

PROBLEM

§25. For a given potential of the laborers who elevate the water
and a given height to which the water is elevated, find the size of the
orifice of effiux and the repelling force.
SOLUTION. Let the potential be such that by it a number N cubic
feet of water can be elevated in one second to a height of one foot,
which potential a number of laborers to be designated by 5/,µV can
develop according to the second experiment inserted following Chapter
IX. Let the height to which the water is continually raised be equal
to A, expressed in feet, and let the area of the orifice in square feet be
equal to M. The number of cubic feet of water which the laborers
can elevate to the height A in a single second by the given potential
will be equal to j (through §22, Chapter IX). Therefore, the orifice
will have to be constructed of an area so that in one second that num-
ber of cubic feet of water can flow out through it if the water flows very
freely. But let us assume instead of seconds the time which a body
takes while it falls freely through the height A. This time is to be
expressed here as ¼VA (it having been assumed for the sake of a
more simple calculation that a body falling freely from rest travels
r6 feet in one second), and in this time the number of cubic feet of
NI,;- N
water to be designated by -A· - v A or 1
.
_, must flow out. But
4 4vA
actually 2MA flows out, that is, the aqueous cylinder of which the
N
base is M and the length is double the height A; therefore, . ;-
4v A
is equal to 2MA, whence the area of the orifice is

M = 8A:A·
However, the repelling force becomes equal to 2MA or ~--
4v A

ScHOLIUM

§26. In any ship, water is to be elevated to a different height, so

that, by the same force which is expended in drawing in the water,
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 335

the maximum force advancing the ship is obtained, and two things
are required for defining that most useful height for a certain number
of laborers. First, it must be known what velocity the proposed ship
acquires from a given potential: with regard to this postulate we
assume that the ship receives a velocity which would be generated by
free fall through the height C from a pressure which is equal to the
weight of one cubic foot of water, or about 72 pounds; and since
from now on we will always express all measures in feet, the weight of
one cubic foot of water will have to be expressed by unity. Second,
the relation between the velocities of the ship and forces propelling
the ship is to be assumed as known; here it is commonly stated that
velocities are in proportion to the square roots of the propelling
forces; however, experiments do not confirm this hypothesis exactly
for slow motions; meanwhile, nevertheless, we consider this [hypoth-
esis] preferable to all the remaining. If someone wishes to explore
the matter under another hypothesis, he can perform the calculation
by the same method which we will now use.

PROBLEM

§27. To find the height most useful to our purpose to which the
water is to be elevated continuously, namely, such that for the same
potential applied for elevating the water the force advancing the ship
becomes a maximum.
SOLUTION. Let all designations applied in the previous argument be
retained. First of all, one is to find the velocity of the ship or the
height required for this velocity, which we called B. But, because the
velocities of the ship are assumed proportional to the square roots of
the forces propelling the ship, the heights of the velocities will be
proportional to the forces themselves. Therefore, the following
analogy will have to be established.
Just as the weight of one cubic foot of water is to the height C (see
§26), so the pressure driving the ship or 2MA - 2MvAB (see §23)
is to the height corresponding to the velocity of the ship, which there-
fore will be 2MC(A - V AB). But this height we called B; therefore,
B = 2MC(A - v AB).
Hence the pressure driving the ship becomes equal to f, and there-
fore proportional to the height B, because C is a constant quantity;
therefore, both the pressure advancing the ship and the height corre-
sponding to the velocity of the ship become maximum at the same
 HYDRODYNAMICS, CHAPTER XIII

time. If, therefore, for the present purpose the quantity 2MA -
2Mv'Ai3, which expresses the pressure propelling the ship, is differen-
tiated, one can set dB = o. But before the differentiation is per-
formed, it is appropriate to substitute for M its value from §25, and
.
t h en t h e pressure a d vancmg t h e sh'1p b ecomes .N1_ - Nv'B
- A , 1n.
4vA 4·
which the letter N is a constant, but the letters B and A are variables.
Now let the differential of this be taken, and by making dB = o, one
sees that the former becomes equal to o, and thus it will be found that
A=~.
Therefore, the force advancing a ship is greatest when the height
to which the water is elevated is four times the height appropriate to
the velocity of the ship.
Let A =~be substituted in the equation B = 2MC(A - v' AB)
found above, and it will be seen that M = ~, and because (through
4
N
§25) M = . ;-, there then results
8AvA
A = (½NC)213, and B = H½NC)2/3,
§28. COROLLARY. If, according to the precept of the preceding
paragraph, the area ~ is attributed to the orifice through which
4
water flows out of the conduit from below towards the stern, that is,
one which is to an area of one square foot as a measure of one foot is to
four times the height appropriate to the velocity of the ship animated
by a force of 72 pounds, it will then occur that the ship is moved at
half the velocity at which the water flows out, and the repelling force
of the water flowing out will be

2MA = ~ (½NC) 213 .

2

But the force advancing the ship will be half of this, so that half the
effect is lost by the inertia of that water which is continuously drawn
1n.

ScHOLIUM

§29. After we have thus demonstrated how that method of naviga-

tion is to be undertaken most usefully and with the greatest success,
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 33 7

I believe that now this matter has to be illustrated by an example

which, I should think, does not agree poorly with the very nature of
the matter, in order that it be shown at the same time how the occur-
rence might take place, more or less.
Let us consider a trireme, commonly called a galley, with 260
rowers; let us assume that this galley, drawn by the weight of one
cubic foot of water, or 72 pounds, completes a length of two feet in
one second, the generating height of which velocity, indicated by C,
is / 6 , under the assumption that a heavy body falling freely from rest
travels 16 feet in the first second. Because, further, 260 laborers are
furnished, any one of which according to the second experiment per-
taining to Chapter IX, can elevate four-fifths ofa cubic foot to a height
of one foot in one second, there will be N = (4/5) · 260 = 208.
Therefore, let the orifice through which the water flows out be of an
area of 4 square feet; the laborers will be able to maintain the water
elevated in the conduit above the orifice at a height of approximately
3½ feet, which is indicated by the letter A, and if one takes the fourth
part of this height, one will have B = t foot, so that the ship will
progress by that navigation at the velocity that a weight acquires by
free fall through a height of i foot; thus, therefore, the ship completes
a length of 7½ feet in any one second and 27,000 feet in one hour, that
is, more than two Gallic miles; such a great velocity of a ship can indeed
barely, or not even barely, be obtained by rowing.
But now let me apply the calculation to another hypothesis, which
I trust those understanding nautical matters will not reject com-
pletely, for it agrees with many observations which I myself made at
sea: let me suppose that the sails of a trireme, expanded perpendicu-
larly to the keel, have a surface of 1600 square feet, and that a wind
which travels through a distance of 18 feet in one second strikes them,
impinging directly, but that the ship thus travels a distance of 6 feet
in one second in the same direction. So the wind strikes the sails at
a relative velocity of 12 feet; I estimate the force of that wind equal to
1600
the weight of 9 ' cubic feet of water, or almost 17 cubic feet of
8 50
water.
If these things are so, it follows that a ship can be propelled by the
elevating of water by 260 laborers at that velocity at which it travels
through a length of6½ feet in one second.
An estimate not very different from this follows from those things
which Mr. Chazelles has in the Commentaries of the Royal Academy of
Science of Paris for I702, p. 98, Paris edition. But in order that they
can be applied properly to our purpose, it will have to be noted that,
 HYDRODYNAMICS, CHAPTER XIII

in rowing, the force propelling the trireme is not to be estimated from

the pressure of the rowers against the oars, but from the pressure
which the extremities of the oars submerged in the water exert against
the water. In order that we may define this [pressure] approximately,
these things will have to be observed first: 260 rowers were furnished,
rowing with all their might; in the first minute 24 strokes (palades in
French) of the oars were made; the entire agitation of the oars occurs
in three motions, which I assumed to be of the same duration, and of
which only one advances the trireme. In this way the trireme was
carried forward at a velocity by which it traveled a length of 7½ feet
in any one second. The part of the oar inside the ship was 6 feet,
and outside the ship r 2 feet. But the surfaces of all the oars (Les pales
in French) which are impelled against the water, gathered into one,
make r 30 square feet according to Mr. Chazelles. He noted further
that the internal extremity of the oar describes a distance of 6 feet in
any one agitation, and because any one agitation is completed in a
time of ~ ~ of one second, and at the same time consists of three
motions, which I assume tautochronous, it appears that any retrac-
tion of the oar occurs in the time of~~' or i of one second, and in this
time the internal extremity of the oar completes a length of 6 feet.
Further, on account of the length of the surface of the oars which
is impelled against the water, not the entire [surface] is to be con-
sidered at a distance of r 2 feet. Therefore, I will assume that to be
at a distance of IO feet, just as if the part of the oar beyond the ship
projected a length of IO feet. The extremity of this part describes
IO feet in the time of i of one second; but because the trireme itself
has a velocity by which it travels 6 feet in the same time, it is to be
understood that the extremities of the oars are impelled against the
water at a relative velocity which describes 4 feet in a time of¾ of a
second. Therefore, the force propelling the trireme is equal to the
force which the water would exert against a surface of r 30 square feet
if it were to strike against it at that velocity by which it travels 4 feet
in i of one second. I find this force, according to common estima-
tion, to be more or less equal to the weight of 40 cubic feet of water;
however, that force is not applied continuously, but only during that
time in which the oars are drawn back: therefore, two thirds of that
force is to be removed, so that the force which propels the trireme
continuously is to be considered finally as equal to the weight of
r 3¼ cubic feet of water.
It follows thence, if the velocities of the ship are assumed to be in
proportion to the square roots of the propelling forces, that this same
trireme, if driven by the weight of one cubic foot of water, would
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 339

have had a velocity by which it could travel approximately two feet

in any one second. This hypothesis is the same as that which we
applied in the first place, so that it again follows therefrom that the
trireme will acquire from that navigation a velocity by which it can
travel 7½ feet in any one second, which velocity is a little greater than
that which was given to the trireme by the very strong rowing of
260 oarsmen.
With these things having been well considered, I am at a loss as to
which kind of navigation is to be preferred, rowing or elevation of
water, and I should believe that the success of either is almost equal,
and I dare to affirm for certain that if a ship is advanced less by the
elevation of water, the defect will be slight; but perhaps it will be
advanced more. Meanwhile, I do not doubt but that this new idea
of navigation appears to be groundless and ridiculous to those ignorant
of these things. But I feel otherwise, and I should wish that attention
be paid further to the following:
First. That water can be elevated easily in every kind of ship where
oars clearly cannot be provided, so that by that new navigation even
very heavy warships such as are used in naval battles can be driven
as it pleases in the absence of any wind.
Second. That thus in theory an example occurs in which the motive
or propelling forces are given, which can be called intrinsic. In-
genious minds will be incited by this example to devising other
principles for this kind of motion, to perfecting them further, and to
applying them to use in navigation.
Third. That in many ways the work of men in elevating water can
be assisted other than by the use of oars; there are indeed natural
things, extraordinary and furnished with almost incredible value as
compared to their moderate expense, by which the same can be
produced as by the work of men; the use of these things can serve
especially in establishing short passages during serene and tranquil
periods. I wrote in Chapter X, §40 and following, concerning the
innate value of natural things of that sort and concerning the effects to
be obtained therefrom and their measures, but especially I would wish
that attention be paid to §43, by which all to whom productive genius
was given by nature for devising machines should be incited to attempt
the perfection of that matter.
Fourth. That several other purely mechanical devices can be
applied similar to that which was given in §27, by help of which,
certainly, the effect of the same work in advancing ships increases not
a little; but it is not permitted now to treat all things according to the
true nature of the matter.
340 HYDRODYNAMICS, CHAPTER XIII

EXPERIMENTS WHICH PERTAIN TO CHAPTER XIII

In order that it be possible to understand the repelling force cor-

rectly by experiment, a vessel can be furnished which has the form of
a parallelopiped, and the weight can be taken empty as well as full of
water, and afterwards the ratio can be investigated between the area
of the vessel and the area of the orifice which must be in the side of the
vessel, just as also the ratio between the heights of the water above
the orifice and above the base. Then it will be possible to deduce the
ratio between the weight of the vessel full of water and that of
the aqueous cylinder lying vertically above the orifice. Further, from
the observed amplitude of the thrust, the velocity of the water will be
obtained; from this, if at the same time one applies in addition the
quantity of water flowing out in a given time, also to be observed, one
deduces the area of the contracted stream, which one would be able
to compare with the area of the orifice.
After all these things have been investigated, let a vessel be sus-
pended from a very long thread, with care having been taken at the
same time so that it cannot have any motion other than that which is
opposite to the direction of the water flowing out. Then at last let
effiux be granted to the water, and it will be observed that the thread
forsakes the vertical position, and from the angle of declination the
repelling force will be ascertained, and one will be able to compare
this with the measurements which we indicated.
EXPERIMENT I. At one time I myself did all the things that I just
indicated, and it was seen that our rule of §2 is properly confirmed;
nevertheless, I was not able then to perform the experiment with
sufficient accuracy with respect to time, nor did I repeat it later.
EXPERIMENT 2. At another time I tried the matter differently:
namely, I placed a vessel full of water, of which I had taken all re-
quired measurements, at the stern of a small boat; the boat was
floating on the water in a tub. Then, with the water flowing out of
the vessel (however, so that it did not strike against the boat), the boat
progressed in the opposite direction; I determined the velocity of the
boat very accurately from the space traveled through in a given time.
Then I inquired as to what elemental weight should be appended to
the boat [by means of string and a pulleyJ in order that, stimulated
by that weight, it would acquire the same velocity. Then, after the
comparison of that weight had been made with the weight of an
aqueous cylinder of given diameter, I saw that our theory was con-
firmed very accurately.
 REACTION OF FLUIDS FLOWING OUT OF VESSELS 341

EXPERIMENT 3. With the water from the vessel (superimposed on

the boat) flowing out into the boat, the latter remains wholly unmoved.
This indicates that the impetus of the aqueous stream is equal to the
repelling force, as I showed in §16 and §1 7. Then, as well, if the
aqueous stream was impinging directly on a plane affixed to the boat,
the latter similarly stayed unmoved, which again proves the equality
of impetus and repelling forc;e; but if the stream was striking obliquely
against the plane, the boat obtained a certain, but slower, motion.
Finally, if the water flowing out of the boat was intercepted so that
the orifice was submerged in the water standing within the boat,
similarly the boat remained without motion, as proof that the same
pressure arises from the stream, whether it happens that all its motion
is confined or that it is declined at a right angle, just as was demon-
strated in §18. I confirmed very precisely by several other methods
the equality between the repelling force and the force of the aqueous
stream striking perpendicularly against a plane. Moreover, I veri-
fied this force conforming to our theory and contrary to the opinion
common to everyone until now by an experiment overcoming all
objections; I conducted this in my home in the presence of Mr.
Emanuel Koenig, my paternal uncle icolaus Bernoulli, and my
Father, with such great confidence that, after all measurements had
been taken, I predicted with all precision how great the pressure of the
aqueous stream would be, although the experiment had never been
performed by me before. I communicated all these things found out
through new mechanics principles to the Academy of Science of

FIGURE 86
342 HYDRODYNAMICS, CHAPTER XIII

St. Petersburg, in the Commentaries of which they should be inserted

at some time.
EXPERIMENT 4. In order that I might also show the falsity of the
accepted rule concerning both the repelling force and the impetus
of the water, I furnished the vessel which Fig. 86 shows, connected
to the curved conduit AB of uniform area, of which the direction at
A was horizontal, at B vertical. I saw clearly that the vessel was not
repelled horizontally; therefore, by §14, the rule is false which
adheres to the simple cylinder defined in that place.

FINIS
HYDRAULICS
BY

JOHANN BERNOULLI
 JOHANNIS
BERNOULLI
HYDRAULICA
N unc prim um dctefra ac demonfhata direc\:e ex
fundamencis pure mechaniciso

ANNO I 73 2.

Photographic reproduction of the title page for the first

part of Johann Bernoulli's treatise, Hydraulica, as
published in the Memoirs of the Imperial Academy of
Science in St. Petersburg. It translates: "Johann
Bernoulli's Hydraulics, now first discovered and directly
shown from purely mechanical foundations. 1732."
The date is false, as Dr. Rouse explains in his Preface
to this volume, page x.
 LEONHARD EULER
Most Sagacious Mathematician

TO THE AUTHOR

P reviously indeed I highly praised Your Theory of flowing

water because of the true and genuine Method which You
alone, Most Excellent Sir, first revealed for fully treating Problems
of this type. But now, after I had examined another portion of Your
Studies, I was thoroughly astounded by the very fluent application
of Your principles to the solution of the most intricate Problems,
because of which most useful and also most profound finding Your
very distinguished Name will forever be revered among future
generations. But You also so distinctly and plainly explained the
most obscure and most abstruse question about the pressure which
the sides of vessels experience as a result of water flowing through
them that there remains nothing more to be desired concerning this
rather troublesome matter. Although indeed no one has undertaken
this matter except Your very renowned Son, who, however, defined
pressure in a rather indirect manner only so far as the entire motion
has already acquired the steady state, nevertheless, after the genuine
method had been brought to light, You at once determined most
accurately the pressure in every state of water, because of which
Your most praiseworthy discovery I congratulate You from my heart,
Most Excellent Sir, and for this communication I give You the
greatest of thanks.
 CONTENTS
Page

PREFACE 35 1

FIRST PART
Treating the motion of Water through Vessels and Cylindrical
Conduits which are Composed of Several Cylindrical Pipes
attached to one Another in Succession 356

SECOND PART
Containing the Direct and Universal Method for Solving all
Hydraulics Problems whatsoever which can be Formed and
Proposed Concerning Water Flowing through Conduits
of any Shape 39 I

HYDRAULICS PROBLEM 446

 PREFACE

Hydrostatics, v:hich deals with water standing in vessels which are

closed below the water surface, has its laws demonstrated and its
principles deduced from reason, whence the performance and the
phenomena are explained clearly and distinctly such that concerning
this Science there is scarcely any more to be desired. The situation
is different in Hydraulics, where not so much is done concerning the
gravitation of water and its pressure, but where, besides this, both the
motion which is thence produced if water can flow out of a given
aperture, or is forced to go from one pipe to another of different size,
and any other effects to be regarded which attend this motion must
be determined demonstratively. Surely this Science, commonly
called Hydraulics, is extremely difficult and up to this time has not
been subjected to the laws and rules of mechanics. Whatever
material Authors have written on this matter, they rely either on
experience alone or on theories that are wholly uncertain, having
insufficient foundation.
In the book Hydrodynamics which my Son published not long ago,*
he undertook that subject under luckier auspices, but he relied upon
an indirect foundation, namely the conservation of live forces, which
is most certainly true and was proven by me as well, but is still not
accepted by all Philosophers. It was I who first presented this
hypothesis in the Dynamics of solids (after Huygens used a similar
principle to determine the center of oscillation), and from that
hypothesis I firmly exhibited the same solution for a water-course
which is given by the ordinary principles of dynamics accepted by all
Geometers; t this clearly general conformity of the solutions elicited
by either procedure should by itself be sufficient for overcoming the
obstinacy of Skeptics. Thus far no one has given a direct method by
which, a priori and only through the principles of Dynamics, one can

* Daniel Bernoulli, Hydrodynamica, siue de uiribus & motibus jluidorum Commentarii.

Strassburg, I 738.
t See Nos. CXXXV, CXXXVI, CXL [Opera Omnia, Johann Bernoulli, Lau-
sanne et Geneve, 1743].
35 2 HYDRAULICS

investigate the nature of the motion of water issuing forth from vessels
through orifices or flowing through conduits of nonuniform size.
Having wondered from what source there is so much more difficulty
in successfully applying the principles of dynamics to fluids than to
solids, finally, turning the matter over more carefully in my mind,
I found the true origin of the difficulty; I discovered it to consist of
the fact that a certain part of the pressing forces important in forming
the throat (so called by me, not considered by others) was neglected,
and moreover regarded as if of no importance, for no other reason
than that the throat is composed of a very small, or even an infinitely
small, quantity of fluid, such as occurs whenever fluid passes from a
wider place to a narrower, or vice versa, from a narrower to a wider.
In the prior case, the throat is formed before the transition, in the
other, after the transition.
On the other hand, I will demonstrate that in the forming of the
throat, however small a size it may have, a pressing force is required,
nevertheless, which is not negligible and by no means infinitely small
but finite and determinate, and so far not at all to be disregarded but
wholly worthy of being taken into account. ow, that force re-
quired for the latter effect, which amazingly enough can be observed,
plainly does not depend upon the length of the throat, which can be
understood to be greater or lesser as long as it is considered extremely
small; it always consumes the same portion of the pressing forces in its
formation, if all other conditions are unchanged.
What the throat may be and in what manner it may be formed will
be understood from the very discussion of the matter, and at the same
time it will be evident that the formation of the throat is accomplished
without noticeable loss oflive forces with respect to the amount which
is present in the whole aqueous mass. Hence the reason is apparent
wherefore, safely and without error, the Theory of live forces can be
applied in Hydraulics, even if those who use this theory pay no atten-
tion to the throat, provided that they are not ignorant of the existence
of the throat and that they see that it detracts nothing from the con-
servation of live forces; for otherwise they cannot contend that they
themselves have arrived at the truth of the matter wholly and
scientifically.
I shall treat this investigation in two parts. In the first I will
consider the phenomena of flowing water and efflux from cylindrical
or prismatic vessels, be they either simple or composed of several
[sections], such as conduits composed of various pipes of different
size or of cylindrical pipes joined as syphons. In the other part I
shall examine completely all perforated vessels, whatever may be
 PREFACE 353
their shape, whether regular or irregular, and the conduits and pipes
attached to them.
In order to have a clearer understanding of things, I am setting
forth the following Definitions and Lemma ta, the validity of which is
manifest from Dynamics as well as Hydrostatics.
I. A uniform accelerative force [i.e., force per unit mass, or accelera-
tion] is that which impresses a given velocity on a given body in a
given time.
II. A motive force is that which, when it acts on a body at rest,
excites it into motion, or which can cause a body already moving to
accelerate, decelerate, or change its direction.
III. Motive forces are in proportion to the products of masses and
accelerative forces. Thus, for example, in order to move twice the
mass with the accelerative force tripled, or, which is the same thing,
in order to move three times the mass with the accelerative force
doubled, a six-fold motive force is required.
IV. The motive force divided by the mass gives the accelerative
force, but divided by the latter gives the mass.
V. The absolute gravity g, or the cause of gravity, whatever it may
be, is an accelerative force which, when it causes a prescribed mass m
of a body to move, produces in it a motive force gm. However, in our
thinking it will be permissible to separate it from the body and thus
to consider it in the same way as if it were acting externally upon the
body. We therefore consider that that same body, free from gravity,
will be accelerated by an external motive force gm according to the
same law by which it is accelerated naturally. However, it is con-
venient to call that same force gm, inasmuch as it exists beyond the
body, an immaterial motive force; therefore, if that force, translated
in another manner, acts on another mass M, the latter will be accele-
rated by an accelerative force gm /M.
VI. An immaterial and invariable motive force, acting without
impediment on a body, accelerates it in the same manner whether it
be at rest at this point or already in motion. Since this force always
follows the body, there is no relative motion between them, and thus
a motive force acts on a body in motion in the same way as if it were
completely at rest. This is the reason why heavy bodies, while
descending, are continuously and uniformly accelerated in accordance
with time, it having been supposed, certainly, that the intensity of the
accelerative force is not changed during the action, that is, neither
augmented nor diminished, just as in fact the force of gravity con-
tinually maintains the same intensity on a descending heavy body
from the beginning of the descent.
354 HYDRAULICS

VII. The intensity of an invariable motive force is the measure

according to which, on a body to be moved, there is produced a
greater or lesser accelerative force; thus gravity, on a body falling
vertically, has a greater intensity than it has on the same body sliding
on an inclined plane. In the first case, to be sure, a greater accelera-
tive force is produced than in the other; in either, however, gravity
is invariable.
VIII. A variable motive force is one of which the intensity is changed
while acting. Thus, for instance, the force of a stretched elastic has
a greater intensity, and as a consequence impresses on the body to be
propelled a greater accelerative force at the beginning than during
the progression of the relaxation. From the above, these Rules result:

Let the space traveled by a body = x,

the mass of the propelled body = m,
the motive force within the limit of the
region traveled = p,
the velocity acquired = v,
the time through x = t.

v
. be p
Hence dt = -dx ; there will - dt or p
m
J
- dx = dv, and therefore p dx =
mv
½mvv, which is very well known.
IX. The lower portions of the water contained in any vessel are
pressed upon by the aqueous mass lying above in accordance with the
depth alone, whatever shape the vessel may have. That is, if in one's
thinking the aqueous mass be divided into horizontal strata of in-
finitely small thickness, every one of these strata is pressed the same
amount as if an aqueous cylinder of water were lying over it having
the same altitude as that which corresponds to the depth of the stratum
itself in the vessel.
X. Hence the following is concluded directly: if the areas of the
strata, each having the same infinitely small thickness, are m, m', m"
m"', etc., and their corresponding weight elements are also in propor-
tion as m, m', m", m"', etc., their own gravitations can be imagined as
separable from the strata, so that their substance remains alone
without weight. But the same pressure will arise in the individual
strata as if they had remained in their natural state if, in place of the
gravitations which have been removed, just as many others are sub-
stituted, which together press the uppermost surface of the water;
[this is done,] certainly, by observing the following analogy at any
instant: as the area of an arbitrary stratum is to the area of the upper-
 PREFACE 355
most surface, so the proper gravitation of the stratum is to the gravi-
tation to be substituted.
XI. I call that mental substitution Translation. In order that I
might explain myself, let some stratum from the lower regions have
an area m, let its gravitation, or its own weight element, be rr, and let
the area of the uppermost surface be h. The translated gravitation to
the uppermost surface will be !!_ rr, which together with all the re-
m
maining translated in this way, constitutes the total immaterial motive
force by which all the water in the vessel is pressed downward, in the
same way as it happens naturally.

ADMONITION

It is now appropriate to say in advance that through all this

treatment of the motion of flowing water I avoid consideration of the
foreign and accidental hindrances which can alter the motion de-
termined through the rules. Such hindrances are the imperfect
fluidity of the water, also its adhesion, the friction at the walls of
vessels, excessive slenderness of pipes, narrowness of orifices or aper-
tures, tenacity of fluid particles on account of which they do not very
easily separate from one another, and others existing of this sort to
which I do not attend.
I would also wish it to be noted that it is not of absolute necessity
that the strata of water always be considered in a horizontal position.
They are more conveniently assumed perpendicular to the direction
of the motion of the water. Thus, for example, when water from a
larger vessel flows out into a narrower horizontal pipe through an
orifice or aperture lying in a vertical plane perpendicular to the wall
of the pipe, the water contained in the pipe is considered to be divided
very appropriately into vertical strata parallel to the plane of the
aperture, and more so in addition because by its very nature it tends
toward this sort of orientation. We see, for instance, that a column
of water in some pipe not greatly exceeding two lines in diameter
[ 1 Paris line = 0.226 centimeter] has both its extreme surfaces dis-
posed in positions perpendicular to the sides of the pipe, whether the
pipe itself is oblique to the horizontal or altogether horizontal. The
line joining the centers of gravity of the strata, whether it be straight
as in rectilinear pipes or curved as in curvilinear pipes, will be called
the center line, or simply the centric. Of course, individual strata
which have their substance concentrated at their own centers are
understood to have that motion which the [actual] strata themselves
have.
 FIRST PART
Treating the Motion of Water through Vessels and
Cylindrical Conduits which are Composed of
Several Cylindrical Pipes Attached
to one Another in Succession

SECTION I
First, let the conduit ABCFDE be given (Fig. 1), composed of two
cylindrical pipes of different size, AGDE and GBCF, of which the
former has a base GD open at the orifice CF through which it connects
to the narrower pipe BF. Now let the whole conduit BE be full of a
homogeneous liquid of no weight of its own but driven from a section
at the orifice AE by a given motive force [per unit density] = p,
which, by pressing equally, is expanded through the whole surface
AE of the liquid. The law of acceleration is sought according to
which the liquid flows through the conduit. Moreover, I consider
the conduit to be always full ofliquid, which occurs by understanding
that a new supply of liquid is provided freely from another connected
source, flowing into the pipe GE at any instant at the same velocity,
for the purpose of replacing that which is flowing out through the
other orifice CF into the pipe GC and from there is escaping into the
air through the opening BC.

E D
 MOTION OF WATER THROUGH VESSELS 357
SECTION II
From Hydrostatics I assumed that the immaterial motive force p
by which the surface of the liquid AE is being pressed is extended
instantaneously to the surface GF of the liquid contained in the pipe
BF, and this whether the liquid is standing still in the entire conduit
or is flowing, as long as it remains full.

SECTION III
While the liquid goes from one pipe to the other, in any case the
velocity will be changed in a manner reciprocal to the areas; on the
other hand, no change is sudden, but successive and gradual, pro-
ceeding through all possible intermediate values from the lesser to the
greater, or from the greater to the lesser.

SECTION IV
Hence when the liquid flows with a parallel motion in the direction
from AE toward GD so that at any moment the same velocity pertains
to the individual portions of the liquid, before the portions near to
GF itself arrive at the orifice GF, it is necessary that they begin to be
accelerated, at least through the small distance HG, and that they
continue accelerating until, at the entrance GF itself, they will have
acquired the velocity of the liquid flowing through the tube BF in a
motion uniformly parallel and common to the individual particles.

SECTION V
And, accordingly, there is formed along the indefinitely small
length HG something like a throat, IFGH, contracting from the wide
into the narrow, through which the liquid must pass, the acceleration
being continuous but nevertheless augmented gradually, with a
rather small portion of the liquid (which fills the small space IFD )
remaining at perpetual rest.

SECTION VI
Let the curve IMF defining the throat be of any nature whatever,
for it is not necessary to assume it of some prescribed shape. Directly,
indeed, I will show: that there is always the same motive force unique-
ly required for this purpose; that the liquid be driven through the
throat, whatever length HG it may have, as long as it be infinitely
small; and that the line IMF which connects the extremities I and F
may be of any nature whatever.
 HYDRAULICS, PART I

SECTION VII

Let no one consider that that motive force (which pushes a small,
in fact infinitely small, portion of liquid through the throat) must be
and is always very small and thus can be disregarded. Indeed, the
motive force is by all means one of finite quantity, because, although
the quantity of material moving is infinitely small, on the other hand
the accelerative force must be infinitely large compared to the former,
in order.that certainly in the infinitely short time in which the liquid
passes through the small space HG, a finite change in velocity can
nevertheless be created, since that which had been the velocity at H
is to that which now prevails at Gas GF is to HI.

SECTION VIII

The neglect of this motive force as if of little import has been the
reason why no one up to this day could have given from statical and
purely mechanical principles the laws ofliquids flowing through non-
uniform conduits. But those who undertook to determine those laws
exactly returned, by my example indeed, to the principle oflive forces,
the application of which to this problem and to others in solids as well
as in fluids they perhaps never would have considered if they had not
followed me, who by all means first showed how to derive these laws
from the conservation of live forces. But I myself, being dissatisfied
since this method was indirect and also founded on a theory of those
forces which is still not universally accepted, did not hesitate to
search for a direct method which would be supported solely by
dynamical principles denied by no one. Finally, after a rather long
meditation, I achieved my aim in the year 1 729, when I saw the crux
of the whole matter to lie in the contemplation of the throat, previously
considered by no one. And so now I am undertaking to share my
discoveries, already explained privately to certain friends, with the
public as well. Since the generation of the throat has now been
indicated, it is pleasing to pursue this task as far as I can with [any]
clarity.

SECTION IX

Let there be considered [Fig. 1] the abscissa HL = x, the ordinate

LM = y, and an element of the former Ll = dx. Let the area AE
or HI of the pipe HE be called h, the area BC or GF of the pipe GC be
called m, and the velocity of the liquid in the pipe GC be v. Accord-
 MOTION OF WATER THROUGH VESSELS 359

ingly, the velocity of the liquid in the pipe HE will be Xv, as the
velocities are reciprocally proportional to the areas. By the same
reasoning, the velocity of the liquid LMml at any place in the throat
will be:!! v, which may be set equal to u. Now therefore let there be
y
an accelerative force y which excites the section Lm of the liquid.
From the nature of the acceleration, it follows that y dx = u du, and
therefore yy dx = _yu du. That is, the motive force by which the
liquid section LMml is excited is equal to yu du. But this motive force,
according to §II, is generated by a single motive force existing in the
pipe HE and distributed over the entire area AE; in order that this
may be explained, yu du must be made into hu du in proportion as
LM is to HI, or as y is to h. The particular motive force in pipe HE
(translated certainly from yu du itself) will be hu du, which can pro-
duce the motive force yu du in the section LMml of the throat; and by
integrating through the whole throat, one has ½h( vv - : : vv), or

hh - hmm
2
vv, w h"1ch d es1gnates
· t h e motive
· fcorce m
• t h e pipe
· HE

required uniquely for creating the acceleration in the throat neces-

sary to change the lesser velocity to the greater, which must be done
in order that the liquid may flow into the narrower pipe GC.
COROLLARY 1. Hence it is evident that the nature of the curve
IMF, as well as the length of the throat HG, does not enter into the
determination of the motive force for generating the motion in the
throat. Thus, if the areas of the end sections HI and GF are given as
hand m and the velocity as v, one always has the motive force in the
. HE, generatmg
pipe . t h e motion
. . t h e t h roat, equa1 to hh - h mm vv.
m
2
COROLLARY 2. If now with a continuous flow of liquid the velocity
v in the pipe BF remains always constant, it is manifest that the
other velocity-in the pipe HE-remains constant as well, and
accordingly that the motive force, or the pressure p, adds nothing
more to accelerating the motion in either pipe; thus it is clear that the
entire force pis uniquely applied to forming the throat and maintain-
hh - mm
ing it in its proper state. Consequently, p = h vv.
2
COROLLARY 3. Let us consider the pipe HE or GE to be vertically
erect in the manner of a cylindrical vessel and to be connected to the
horizontal pipe GC [Fig. 2]. Also consider the force p to be the very
weight of the column ofliquid contained in GE, so that (withg having
 HYDRAULICS, PART I

been taken to designate the natural accelerative force of heavy bodies,

and also HA or GA = a) one hasp = gah equal to the weight [per
unit densityJ of the liquid contained in GE, from which gha =
hh - mm
h vv. But in order that v be determined by the vertical
2
altitude z through which some freely falling weight has acquired the
velocity v, g dz must equal v dv, from which gz = ½vv. Therefore by
substituting gz for ½vv, we will have gha = hh ~ mm gz, from which

emerges z = hh hh a, which gives the following hydraulics

-mm
Theorem.

SECTION X

THEOREM. Let the cylindrical vessel AGFE (Fig. 2) be vertically erect

and furnished at the base with a horizontal cylindrical pipe FB open at either

FIGURE 2

end. Likewise let both the vessel and the pipe be continuously full ef water, so
that, ef course, as much water as flows out through the opening BC is continu-
ously supplied through AE at the same velocity that the water has in the vessel.
I say that the velocity ef the water which flows out (if it starts from rest)
converges very rapidly to that which is acquired by a weight falling freely
through a height hh hh a.
- mm
The truth of this is evident from Corollary 3 preceding.
COROLLARY I. Whence if the opening BC be very small with re-
spect to the area AE of the vessel, such that m can be neglected with
respect to h, there will result z = a, that is, the velocity of the water
flowing from the pipe will be equal to that which a weight, having
 MOTION OF WATER THROUGH VESSELS 361

fallen freely from a height EF, acquires. This is a very well known
Theorem, but up to now not shown from dynamic principles, es-
pecially if the attached pipe BF were present, since previously the
Theorem was believed correct only for a small orifice placed at F.
COROLLARY 2. The greater is the opening BC with respect to the
area AE of the vessel, the greater becomes the maximum efflux
velocity of the water. Thus for m large, the value of the fraction
hh · increase
1s · d , untl·1 fcor m 1ncreas1ng
. . to h t h e maximum
.
hh -mm
velocity is infinite; and also from here it is evident that this is true
because in that case both the vessel and the pipe are of the same size
and they form one continuous bent pipe. And the force of the
weight of the water in the portion AF, always full, continuously
accelerates the entire aqueous mass such that finally its velocity,
generated in an infinite time, becomes also infinite itself. Now, with
the length of the pipe FC being called b, the mass [per unit density] of
all the water in the bent pipe AGC will be ha + hb; and this will not
be accelerated in any way other than as is some solid body which is

animated by an accelerative force equal to ha g1a hb = a ~ b' and

certainly such a body by falling for an infinite time would acquire
an infinite velocity.
COROLLARY 3. Surely if m were greater than h, that is, if the
horizontal pipe were larger than the vertical vessel, the maximum
velocity would certainly never be attained even in an infinite time,
since hh hh would be negative, with the result that, during flow
-mm
to eternity (if the flow lasts into eternity), the acceleration of the water
flowing out will not cease to be augmented. In this case, for instance,
an inverse throat will be formed in the pipe, looking back at the orifice
BC, which, as will be evident from the following, is of the nature that
it increases the motive force rather than decreases it, during which
time it disappears, so to speak, toward the pressure being produced
in the rear, whereby the water can descend rather freely in the vessel.

ScHOLIUM

So far we have considered the vessel and the pipe constantly full
of water and the water flowing out at its maximum and therefore
constant or uniform velocity, so that no additional motive force is
required for accelerating the water either through the vessel or through
the pipe, but that the total motive force p is used for controlling the
 HYDRAULICS, PART I

throat which is formed in front of the entrance from the wider region
to the narrower. Now we shall consider the velocity of the flow of
water as if it were increasing, starting from rest, such that its own
equally special portion of the motive force pis required for developing
the acceleration in the vessel as well as in the pipe. First we shall
examine the case in which a constantly full vessel is connected to a
pipe.

SECTION XI

Let x be the length of the region through which the water passes in
the pipe from rest. Then Xx will be the length through which it
passes in the same time in the vessel. Thus, similarly, with the exist-
ing velocity in the pipe equal to v, the velocity in the vessel will be
Ji v. From this the accelerative force in the pipe equals vd:v, and this
multiplied by the mass of the water mb will give the motive force
mb:/v, which, translated to the vessel (by§2), will give the equivalent

hbv dv f
--;fx' rom w h'ic h , certam
. 1y, t h at m
. t h e pipe,
. --rx-,
mbv dv
can b e pro-
duced. And so also the accelerative force in the vessel equals
mm v dv Im
Ith h dx = mv dv , w h'ic h , app1·ie d to t h e mass ha, gives
h dx . t h e motive
.

force m;x dv for propelling the water in the vessel; and thus the sum of
those three motive forces-through the throat, through the pipe, and
through the vessel-must equal the total motive force p. This gives
us the equation

hh - mm hbv dv mav dv
2h vv + --;Ix + ~ = p.

Therefore, as before, let p = gah be the weight itself of the aqueous

column, and let gz = ½vv be established as in Coroll. 3, §IX; with these
being substituted, this equation will result:

hh - mm hb dz ma dz
h z + -y;- + ~ = ha,
or
(hh - mm)z dx + (hhb + hma) dz = hha dx,
 MOTION OF WATER THROUGH VESSELS

from which dx = hh hh\: hma dz, which properly treated and

a - z + mmz
integrated logarithmically, will give

x = hhb + hma ln ( hha ),

hh - mm hha - hhz + mmz
from which, by progressing to numbers (assuming I = ln e), one has
z = ( hha
hh - mm
)(I _ i//hh-mm)x /(hhb+hmal).

But if the water in the vessel (which, for the sake of brevity, is con-
sidered to have only an orifice of area m, without the annexed pipe)
be animated by a gravity g' different from the natural gravity g, it
will be found that

z = g' hha (I _ i / /hh-mm)x /(hhb+hma)).

g(hh - mm)
COROLLARY. If x = co, which gives the case of the maximum
velocity to which the flow converges, it will be true that
I/e<hh-mm)x/(hhb+hma) = O,

and therefore z = hh hha for natural gravity g, which conforms

- mm
wholly to Coroll. 3, §IX; and if, in addition, mis infinitely small with
respect to h itself, then z = a, just as is developed in Coroll. 1, §X, all
of which confirms the method splendidly.

SECTION XII
Let us consider now the case where the vessel AF (Fig. 2) does not
remain full of water, but, as a measure of the water flowing out, it is
emptied gradually, and its surface AE descends continually.
Consider the water in the horizontal pipe to have passed through
the length x, hence that a quantity of water mx, that is, equal to an
aqueous cylinder of which the base is m and the length is x, has
flowed out from it (for I assume the vessel and the pipe to be full at
the beginning). But if, accordingly, in EF the portion EI is assumed
equal to Xx, it is clear that the horizontal HI is the location of the
uppermost surface to which the water descends in the vessel after the
portion of water mx has flowed out through the pipe. Therefore,
there will remain in the vessel the aqueous column GI= ha - mx, of
 HYDRAULICS, PART I

which the weight g(ha - mx) is directly that very thing which we
called p. Thus if, therefore, the accelerative force of the water re-
maining in the vessel (which in §XI is generally found equal to
:vd:v) is applied to the aqueous mass, which now is ha - mx, we will

have the motive force :vd~ (ha - mx), which is compatible with the
water flowing down through the vessel; from which now by collecting
the three forces-through the throat, through the pipe, and through
the vessel-and by equating the sum top itself, that is, to g (ha - mx),
we will acquire this equation:

hh - mm hbv dv mv dv
2
h vv + dx + h dx (ha - mx) = g(ha - mx).

By substituting g dz for v dv, and gz for ½vv, as we did in Coroll. 3,

§IX, we will change our equation to this other:

hh - mm hb dz m dz (h ) h
h Z +~ + h dx a - mx = a - mx;

and, by multiplying by h dx, into this:

(hh - mm)z dx + hhb dz + m dz (ha - mx) = (hha - hmx) dx.

This is the very equation from which, if the value of z itself is found,
the height will be obtained through which a heavy weight, having
fallen freely, will acquire the sought velocity, certainly equal to that
which the water in the pipe will have after the quantity mx has
flowed out.
The derived equation in which the unknowns are found to be
interrelated can also be integrated according to our rules with the
support of the Lemmata following shortly, and thus the value of z
itself may be known in finite terms. However, at this point one
should not tarry any longer on this matter: it suffices to me to have
reduced the Problem to a differential equation by using purely
mechanical principles, which might have been presented by someone
else before me, although I do not recall ever having seen it. Indeed,
it should be known that this very equation can be derived through
the method of live forces, so that their use and their validity hence are
substantiated against Adversaries.
COROLLARY r. In order to determine the maximum velocity of the
liquid flowing out and, furthermore, that of the liquid descending in
the vessel, one need merely set dz = o; after this has been done, our
 MOTION OF WATER THROUGH VESSELS

equation will furnish (hh - mm)z = hha - hmx, from which z =

hha
hh - hmx , w h"1ch , smce
. unt1·1 now 1t
. contams
. t h e un k nown x 1tse
. lf,
-mm
indeed determines nothing unless the value of z itself is also found at
the same time from the general equation.
COROLLARY 2. If m is very small with respect to h itself, the general
equation assumes this form: z dx + b dz = a dx; from which dx =
b dz , which gives z = a - a/extb_ Thus, in this case, in order that
a - z
z be a maximum, it is necessary that x be infinite, and then it will
occur that z = a, which certainly can be gathered at once from
dx = b dz , or from dx(a - z) = b dz; for by making dz = o, on
a - z
account of z itself being a maximum, one will have a - z = o, and
from that, z = a. From there, in turn, it is clear that in a very wide
vessel the water flowing out through a very narrow pipe immediately
acquires a maximum velocity, always constant thereafter and equal
to that which a weight falling freely from the height of the vessel
would acquire, as we saw above in Coroll. 1, §X. Obviously, in this
case the vessel can be considered as always full because, on account
of the comparatively infinite area of the vessel with respect to the
constricted [one] of the pipe, an almost infinite time as well would
certainly be required before the water descended noticeably in the
former.
SECTION XIII
Now take another case. Let the pipe (which is assumed full of
water right up to C at the beginning before the flow) be extended
indefinitely; thus it is certain that while the water is descending in the
vessel, nothing can flow out of the pipe, but always some of the liquid
descends from the vessel into the pipe, and this together with that
which is already assumed to be present there is forced to flow, pro-
pelled jointly within the pipe. The law of acceleration and the
velocity itself are sought after an arbitrary space has been flowed
through within the cavity of the pipe. The accelerative force in the
pipe, as was shown in §XI, here also will be v dv/dx, but the mass of
water to be propelled now is mb + mx, which, multiplied by the
accelerative force v dv/dx, gives the motive force in the pipe (mbv dv +
mxv dv )/dx; this, translated to the area of the vessel, gives the equiva-
lent motive force in the vessel equal to (hbv dv + hxv dv) /dx. And so
the three motive forces-through the throat, the pipe, and the vessel-
having been joined, and these having been equated to the total
 366 HYDRAULICS, PART I

motive force p, for the vessel always filled by new water flowing in,
this equation (see §XI) will result:
hh - mm hbv dv + hxv dv mav dv _ _ h .
2h vv + dx + dx - p - g a'
but for the vessel accepting no new liquid, this other will appear
(see §XII):
hh - mm hbv dv + hxv dv
2h vv + dx
mvdv
+ h dx (ha - mx) = p = g(ha - mx).

With gz substituted for ½vv, the prior equation gives this:

(hh - mm)z dx + (hhb + hma + hhx) dz = hha dx;
but the latter gives this:
(hh - mm)z dx + (hhb + hma + hhx - mmx) dz = (hha - hmx) dx.
However, either equation can be integrated through the Lemma
promised above, which I now show.

SECTION XIV
LEMMA. Let the equation to be integrated (and indeed without the
necessity of separating the unknowns) be
az dx + (f3 + yx) dz = (E + 0x)-dx.
I write y for f3 + yx, from which dx = dy/y, and the equation 1s
changed into this: ~ z dy +y dz= (E + 0x) dx; after this has been
y
multiplied byf fr -1, there will be obtained
a
- .zya/y -1 dy + ya fy dz= (€ + 0x) dx-ya fy -1
y
I
(E + 0x) ·-· (f3 + yx)afr - 1 y dx.
y
After integration there will result

f 1Yz = f (E + 0x)t(f3 + yxttr- 1 y dx

= ~(f3 + yx)afy·(E + 0x) -J..!!_ (f3 + yx)a fr ydx

a ya
I 0
= - (f3 + yx)afy(E + 0x) - - - (f3 + yx)afy+l
a aa + ya
 MOTION OF WATER THROUGH VESSELS

Let it be noted here that the last two terms given are added for the
purpose of rectifying the equation, as is customary; for, with x
vanishing, z certainly vanishes also. Now let the equation be divided
by ya 1r, that is, by (/3 + yx)a 1r, and the correct value of z itselfresults,
namely
I 0
z = - (E
IX
+ 0x) - - - - (/3
<XIX + ya
+ yx)

+ ( B 13a1r+l _ ~ 13a1r) · (/3 + yx)-alr.

<XIX+ ya IX

SECTION xv
Therefore, in order that the application of this might be made to
the prior equation, (hh - mm)z dx + (hhb + hma + hhx) dz = hha dx,
here there will be a = hh - mm, f3 = hhb + hma, y = hh, E = hha,
and 0 = o, by substitution of which there will be obtained

Z = hha - hha (hhb + hma) (hh - mm)/hh

hh-mm hh-mm
X (hhb + hma + hhx)<-hh+mml/h\
or, which is the same thing,
hha (
1
( hb + ma ) (hh - mm)/hh).
Z = hh - mm - hb + ma + hx
But if on the contrary it is applied to the latter, where a = hh - mm,
= hhb + hma, y = hh - mm, E = hha, and 0 = - hm, there results
f3
hha - hmx hm
Z = hh -mm + 2 (hh -mm ) 2 (hhb + hma + hhx - mmx)
-hm hha ]
+ [ 2 (hh _ mm) 2 (hhb + hma) 2 - hh _ mm (hhb + hma)
· (hhb + hma + hhx - mmx) - 1 ;
after these have been separated in order, by proceeding as usual, one
has at last
hhax - ½hmxx )
z = ( hhb + hma + hhx - mmx ·
COROLLARY I. If m is very small with respect to h, there will be,
for the case of the vessel always full, z = b ax ; similarly, for the
+x
ax
other case, there results z = -b- - , which certainly must happen
+x
thus generally, because, indeed, on account of m being infinitely
 HYDRAULICS, PART I

small, the water must flow out for an infinite time before its uppermost
surface descends noticeably in the very large vessel. In any case, it
is evident that it is the same as if the vessel remained always full, and
accordingly these two cases certainly reduce to the same.
COROLLARY 2. If b = o, that is, if no water is contained in the
indefinitely long horizontal pipe FB at the beginning of flow, for the
case of the vessel always full,
Z = hha (r - ( ma )<hh-mm)/hh);
hh-mm ma+hx
but for the other case in which no new liquid is received,
hhax - ½hmxx
z = -c---,-,-~---
hma + hhx - mmx
In this last case the following is also noteworthy: at that moment at
which the surface of the liquid will have descended all the way to the
bottom of the vessel, which is done by assuming x = !!_ a, one will have
m
z = ½a, that is, the velocity of the water in the pipe after total deple-
tion of the vessel will be that which a weight would acquire by falling
from half the height of the vessel.

CONCERNING A CONDUIT OF THREE OR MORE

PIPES

SECTION XVI

Let there now be (Fig. 3) the conduit AL consisting of three pipes,

AD, CC, and BL, all full of water. And let there be a motive force p

FIGURE 3

-- -

I
 MOTION OF WATER THROUGH VESSELS 369
which, expanded uniformly over the surface AE, drives or presses on
the same. The acceleration and the actual velocity with which the
water flows out of the pipe BL are sought.
First of all it should be noted here that two very short throats are
formed, one in the transition at GF, the other in the transition at BK,
which individually require their own motive forces which are to be
translated to the area AE, to which then are to be added the motive
forces of the aqueous columns contained in the individual pipes, after
translation of these forces to the area AE; after this has been done, the
sum of all these translated forces is to be equated to the total motive
p,
force from which the desired equation will result.

SECTION XVII
Consequently, let the lengths of the pipes be AG = a, GB = b,
BM = c, and their areas be AE = h, GF = m, and BK = n. Here
also let the velocity in the last pipe BL be designated as v, the velocity
in the second pipe GC as u = !: v. Thus there will be, through the
m
reasoning furnished in §IX, a motive force at the surface AE, re-
quired for forming the throat through GF, equal to
hh - mm hhnn - mmnn
2h uu = 2hmm vv
nn
(by substitution of the value of uu itself, which 1s - vv);
mm
likewise the motive force in the pipe GC required for the throat
through BK is mm - nn vv, which indeed, after having been trans-
2m
. mm - nn . hmm - hnn .
1ate d to t h e area AE, b y mak mg - - - - vv mto - - - - - vv, 1n
2m 2mm
proportion as m is to h, gives the motive force in the first pipe AD for
producing the throat through BK; and so both forces added together
hhnn - mmnn hmm - hnn h . hhmm - mmnn
give h vv + vv, t at 1s, h vv or
2 mm 2mm 2 mm
hh - nn
h vv, equal top. And ·thus is determined the velocity of flow
2
through the three pipes after the former has reached constancy.
COROLLARY. Hence it is evident that the water is moved through
the three pipes in the same manner as if, the second having been
removed, the third were immediately attached to the first, after it has
been stipulated, of course, that the flow has reached the greatest and
constant velocity; finally, it is now clear that henceforth, no matter
37° HYDRAULICS, PART I

how many pipes are considered, the motive forces at the individual
throats, translated to the first pipe and added together, are equivalent
to that unique motive force in the first pipe to be applied to the
unique throat which would be made by attaching the last pipe im-
mediately to the first pipe. And thus the same constant velocity to
which the flow converges is obtained in each case, whether the water
goes through the entire conduit composed of all the pipes, or, the
intermediate ones having been omitted, through the first and the
last connected to each other directly. Everything, therefore, which
we have shown above concerning constant velocity through two pipes
is to be applied to a conduit consisting of as many pipes as one might
wish.

SECTION XVIII
Now there comes up to be considered the acceleration in a conduit
of many pipes when indeed the flow of water begins from rest, with
the first pipe, however, remaining always full by means of the influx
of new water following the descending [ waterJ at the same velocity.
In this matter, nothing else is to be done than to translate the motive
force, considered to be in proportion to the aqueous mass to be driven
through the individual pipes, to the area of the first pipe. If the sum
of these translated motive forces is added to the motive force through
the throats, that is, through that single one which would be formed if
the last pipe were attached directly to the first, the force of all will
result, which is to be made equal top itself.

SECTION XIX
Thus let us apply this rule to a conduit of three pipes the lengths of
which are a, b, and c, and areas h, m, and n. Let x be the length of the
space through which the water, beginning from rest, travels in the
last or third pipe and v be the velocity acquired in this pipe. For
the purpose of imitating the process in §XI, !!:.. x will be the distance
m

which the water travels in the same time in the second pipe, and !!:.. v
m

its acquired velocity. Likewise ix will be the distance traveled

m the first pipe, and iv the acquired velocity. Hence the

accelerative force in the third pipe is vd!v, and this multiplied by the
 MOTION OF WATER THROUGH VESSELS 37 1
· w1·11 give
· t h'1s pipe
aqueous mass nc m · t h emotive ncv dv' w h'1c h ,
· fcorce ~

·
trans 1ate d to t h e fi rst pipe, w1·11 give
· t h e eqmva
· 1ent hcv dv
--z;:-· Th us a 1so

· fcorce m
t h e acce1erat1ve . t h e secon d pipe
. 1s
mm m
I
. -nn v dv -n dx = nv
-d
m x
dv ,

which, applied to the mass of water mb of the second pipe, gives the
motive nbv dv w h'1c h , trans 1ate d to t h e fi rst pipe,
. fc· orce -d-, . hnbv dv
. ld s --d-·
y1e
X m X

Thus the accelerative force in the first pipe, ~~ v dv/~ dx = 7::,

applied to the mass ha of the first pipe, gives the motive force of the
· h e fi rst pipe
water mt nav dv w h'1ch , smce
· -;fx' · · 1s
1t 1 d y mt
· area · h e fi rst pipe,
·

is not to be translated farther. Those three forces are therefore

hcv dv hnbv dv nav dv .
-d-, --d-, and -d-' the sum ofwh1ch, added to the force caused
X m X X

. hh - nn ( hnb ) v dv
by the throats, will be found to be h vv + he + -;; + na Tx'
2
equal to the total force p.

SECTION xx
Now let there be four pipes, of which the lengths are a, b, c, and e
and the areas are h, m, n, and q; let x be the distance traveled in the
last pipe and v the velocity acquired in the last pipe. In order to
observe the uniformity and the law of progression from one pipe to
another, I will begin at the first, in which the accelerative force is
hh v dv It dx, the velocity is %v, and the element of velocity is t dv.

And as the element of distance to be traveled is t dx, so there results

from the law of acceleration the accelerative force

qq /q qv dv.
hh v dv h dx = h dx '

and by multiplying this by the mass of water to be moved, [one

causes] this motive force hqav dv/h dx to appear, which, because
it is already in the first pipe, does not require further translation.
But in the second pipe, the accelerative force .!l!!_ v dv/i dx = qv ddv,
mm m m x
applied to the aqueous mass mb, gives the motive force in the second
37 2 HYDRAULICS, PART I

pipe mqbv dv/m dx, which, translated to the first pipe, gives the
equivalent hqbv dv/m dx. In the same manner, the motive force
translated from the third pipe to the first will be hqcv dv/n dx, and the
motive force translated from the fourth to the first will be hqev dv/q dx.
Therefore, all added together equal

hqav dv + hqbv dv + hqcv dv + hqev dv = (~ + ! + ~ + ~) hqv dv_

b~ m~ n~ q~ h m n q ~

Generally, therefore, for any number of pipes whatever, the lengths of

which are a, b, c, . .. , 7T and areas h, m, n, . .. , w, the sum of all the
motive forces translated to the first pipe will be equal to

(~ + l1_ + ~ + ~ ... + ~) hwv dv;

h m n q w dx

to which if the motive force hh ~h ww vv for all the throats is added,

there emerges the total motive force to be set equal top itself. This
equation results therefrom:

hh - ww
2h vv
(a b c
+ 1t + ;; + ri · · · + z:; --;Jx = P;
7T) hwv dv

or, by writing gz for ½vv, this other:

hh ~ ww z + (~ + l1_ + ~ ... + ~) hw dz = !.. p

h m n w dx g '
or,

(hh - ww) z dx + (~h + !m + ~n · · · + ~)hhw dz =

w
~ p dx.
g

COROLLARY I. If the lengths a and 7T of the first and last pipes and
the lengths of the intermediate ones as well remain invariable, the
first certainly through continual influx, the last through efflux, the
sum of the series _ha + !m + ~n · · · + ~w will be constant, which may be
called M, and p = gha, from which this equation appears:
hh - ww Mhw dz _ h
h Z + dx - a,
or,
(hh - ww)z dx + Mhhw dz= hha dx,
of which x is computed by logarithms given in z, z itself by numerals
given in x.
 MOTION OF WATER THROUGH VESSELS 373
COROLLARY 2. But if now, with no new water flowing in, the first
pipe be depleted by the flowing out through the last, of given length-
just as would be done if the first pipe in the form of a vertically
erected vessel were to contain liquid pressed by its own weight during
the time in which it would be expelled through the horizontal conduit
which the remaining pipes form-and if the distance traveled through
the last pipe from rest be called x, the height of the liquid remaining
in the cylindrical vessel will be a- :x; and therefore from the series
a b C 7T • WX I
- + - + - · · · + - there 1s now to be removed hh, and for - p there
h m n w g
must be written ha - wx, which gives this equation:
(hh - ww)z dx + Mhhw dz - wwx dz = (hha - hw x) dx,
which through the Lemma of§XIV, can be integrated.
COROLLARY 3. Furthermore, if the last pipe be prolonged indefi-
nitely, such that with the uppermost surface of the water descending
in the vessel, the water indeed does not flow out of the last pipe but is
continually thrust forward in it more and more, there is to be written
in the series not only a - :x for a, but also 7T + x for TT, and so for
this case we will acquire this other equation:
(hh - ww)z dx + Mhhw dz - wwx dz + hhx dz = (hha - hwx) dx.
This is integrable through the same Lemma.
COROLLARY 4. Ifby consideration of the vessel itself as the first pipe

there results _ha = !!_

= !:. · · · = :!., that is, if the lengths of the pipes of
m n w
which the number be N be everywhere proportional to their respec-
tive sizes, our general equation is changed to this:
I
(hh - ww)z dx + Nhaw dz = g- hp dx.
COROLLARY 5. But if, with exception of the vessel or the first pipe,
there results!!._ = !:. · · · = :!., and if the number of remaining pipes be
m n w
N, surely there will be
Nhhw dz I
(hh - ww)zdx + haw dz+ - - - = -hpdx.
m g
COROLLARY 6. There shall now be an infinite number of pipes, but
each one of them, except the first, of infinitely short length, such that
all of them combined represent a truncated conoidic conduit of which
374 HYDRAULICS, PART I

the anterior area equals m, and posterior equals w, such as RSTV

(Fig. 4). If this is conceived to be cut by two closely spaced planes,

FIGURE 4

sr and tv, parallel to SR and TV themselves, srvt will be one of those

short pipes, having for a length an element of the length RV of the
whole conduit, and for an area the plane sr. From this, in order to
. t he sum of t h e senes
ob tam . -b + -C · · · + -, - must b e mtegrate
ff M . d,
m n w sr
which in many examples can be done algebraically: for instance, if
ST be a straight line, that is, if SRVTbe an ordinary truncated cone;
likewise if ST be the arc of a hyperbola of any kind whatever asymp-
totic to RV.
SECTION XXI
Let us illustrate the very matter in the prior example. Let
SRVT be a truncated cone of which the anterior area SR = m, the
posterior area TV = w. Furthermore let their semidiameters be
proportional to Vm and Vw. Henceforth let its abscissa Vv be equal
to t, its element vr = dt, the semidiameter of the area tv = y, and the
total length of the pipe RV= L; y will be found proportional to
(tVm - tV~ + LVw )/L; but the area sr itself, which is proportional
toyy, is (tvm - tV~ + LV~) 2 /L2, because of which
vr L 2 dt
sr (tVm - tV w + LV w) 2 '
the integral of which, rectified in an appropriate manner, is
V Lt ; and therefore, through the whole conduit RSTV,
t mw - tw + Lw
taking Vv, or t, = VR = L, the required integral is found as./
vmw
!!_ + : · · · + :!.. And thus our general equation of§XX,
m n w

(hh - ww)z dx + (~h + !!_m + :n · · · + :!.)hhw dz

w
= g~ p dx
 MOTION OF WATER THROUGH VESSELS 375
will give the following equation for the conduit or conic pipe the length
of which is L and the two extreme areas of which are m and w, for the
height of the vessel being a and the area being h:
hhLw dz 1
(hh - ww)z dx + haw dz+ v' = - hp dx.
mw g

SECTION XXII

For the case of Coroll. 1, §XX, there will result as well

(hh - ww )z dx + Mhhw dz= hha dx,

where M = ~ + v'~w· For the case ofCoroll. 2 of the same section,

the same having been assumed for M, there will result

(hh - ww )z dx + Mhhw dz - wwx dz = (hha - hwx) dx.
For the case of Coroll. 3, it must be understood that the conic pipe
has a cylindrical pipe of indeterminate length and of area w attached
to it at its extremity so that the propelled water may always be con-
tained in it and travel through the distance x from the beginning of
motion; there will then result
(hh - ww )z dx + Mhhw dz - wwx dz+ hhx dz = (hha - hwx) dx.

SECTION XXIII: GENERAL ScHOLIUM

Many other corollaries could be derived from these, useful no less

than curious and elegant. Certainly those which pertain to this
matter have their entire basis in the ones already transmitted and
explained; it is clear that I did not indicate this in expressed words.
For example, we have supposed that some water or any other liquid
gravitates in the first pipe only, just as in a vessel, and from there is
forced through a conduit having a horizontal position, through which
the water, while it is being moved, is deserted, so to speak, by its
own gravity. Meanwhile if in this conduit, or in the pipes which
compose it, it also retains its own gravity, whether the total or only a
part, as would happen if the pipes were not horizontal but either
vertical or inclined to the horizontal in some diverse manner, this
causes no difficulty at all. For instance, the very weight of the water
from any pipe whatever, by means of §II, can be translated to the
vessel or the first pipe so that the water in the remaining pipes may be
considered as being without gravity. But the translated gravities,
 HYDRAULICS, PART I

added together with the weight of the water in the vessel or the first
pipe, can be considered in place of that which we called p or the fun-
damental motive force by which the flow of the total aqueous mass is
generated. Thus if, arbitrarily, the conduit EGEL (Fig. 5) were to

''
':'

HL...Bi
l
''
0 L............- - L
M
FIGURE 5

consist of three pipes, AD, GC, and BL, of different areas, the first of
which, AD, would have an area AE or GD, the second GC an area GF
or BC, the third BL an area BK or ML; and if the first were vertical,
the second were to make the angle GBH with the horizontal, the third
the angle BMO; and if the areas be AE = h, GF = m, and BK = n;
then the force of gravity or the natural accelerative force is g, and the
motive force in the pipe AD, full of water, is gh -AG. Likewise
GB/GH = g ;g·:;t is the accelerative force of the liquid in pipe GC.

Similarly BM/BO = g /g~e: is the accelerative force in the pipe

BL. Thus g -:/eH m. GB or gm· GH will give the motive force of the
water in the second pipe; similarly gn-BO gives the motive force of the
water in the third pipe. But the motive forces are now to be trans-
lated from the oblique pipes GC and BL to the vertical one by setting
m/h = gm-GH/gh-GH, and n/h = gn-BO /gh-BO. And so in this way
all the water in the pipes can be considered as being without gravity,
but in its place the first column AD can be considered pressed by the
 MOTION OF WATER THROUGH VESSELS 377
motive force expanded uniformly over the surface AE, which force
would be equal to gh(AG + GH + BO ) = gh · A = p (because
AG + GH + BO equals the total vertical height of the conduit,
which let be A). And so we have reduced this case and other similar
ones to our general method.
Note: If one or more of the oblique pipes is directed upward,
there will be, on account of it or them, a negative motive force trans-
lated to the first pipe, and there will have to be assumed for A an
excess by which the sum of the positives is greater than the sum of the
negatives, or vice versa. In a word, A will be the excess or defect
by which the surface of the water in the first pipe is higher or lower
than the horizontal which is the surface of the water in the last pipe.
This serves in the determination of the law according to which
liquids oscillate in pipes curved in any shape whatever. At this point
also refer to the following Problem, proposed to me by my Son six
or seven years ago but expressed slightly more generally.

HYDRAULICS PROBLEM

SECTION XXIV

ABCD [Fig. 6] is a vessel filled with water to EF. GI is a cylindrical pipe

the section KI of which is also full of water. With the thumb spread over the
orifice GO, the pipe is immersed in the water contained in the vessel, but just to
the extent that the section MI of the pipe, greater than KI, penetrates the external
water up to MN. The thumb now having been removed, the surface KL will
scend (on account of the prevailing pressure of the external water ), and on account
of the impulse received, it will reach above the surface EF up to PQ,. The
height MP or NQ, is sought to which without question the water in the pipe cer-
tainly can rise.
SOLUTION. Let the immersed part HM of the pipe be set equal to
a, its portion HK originally full of water equal to b (less than a), the
area EF of the vessel h, the area GO or HI of the pipe m. But the
water now surrounding the pipe and pressing down by its own weight
tries to enter through the opened orifice HI and to ascend by pro-
pelling the portion of water HL lying above. I understand that action
and effect in this way. Let another pipe, facing downward, having
an area h equal to the area EF of the vessel and having a height
HM= a, be attached to the orifice HI. Let this pipe be full of water,
but of water such that it would be raised, that is, would be pushed
upward, and certainly by a force precisely as large as that by which
the water gravitates downward in the vessel of height MH (in place
378 HYDRAULICS, PART I

of which that in the pipe is substituted through a fiction of the mind).

Accordingly, the motive force of the water in this fictitious pipe ex-
tending upward will be gha, and thus, with respect to this, the negative
motive force of the water HL in the pipe HO will be gmb, which
translated to the fictitious pipe gives ghb, which, of course, being
opposite to that gha, is to be subtracted from the same, and there
remains gha - ghb or gh(a - b) for the motive force which we
called p, to which, therefore, are to be equated the motive forces
which are generated by the flow through the throat to be formed at
the entrance HI in flowing through the pipe HO and rising in the
fictitious pipe. Thus if, beginning at KL, the distance traveled by the
water within the pipe HO is called x, and as well the distance through
which the surface of the water in the fictitious pipe travels in rising is
Xx, for the sake of an imitation of the reasoning of §XIII we shall

have the accelerative force in the pipe HO equal to vd:v, which, multi-
plied by the mass of the water to be pressed upward, mb + mx, gives
the motive force in this pipe as (mb + mx) vd:v, [which force is] to be
translated to the fictitious pipe in order that from there we might
V dv
obtain the equivalent motive force (hb + hx) Tx· And since in
addition in the fictitious pipe (in which the water ascends at the
velocity 7i v through the distance Xx) its own motive force, which is
. mv dv
not to be translated any further, 1s (ha - mx) h dx, therefore, after
that which is required for forming the throat has been added to these
two forces, we will obtain the total motive force
hh - mm v dv mv dv
2
h vv + (hb + hx) dx + (ha - mx) h dx ·

However, since here

p = gh( a - )i X - b - X) or g(ha - hb - mx - hx),

there will result the equation for determining the velocity v, namely
this:

hh - mm ( mm ) v dv
h vv
2
+ hb + hx + ma - h x dx = g(ha - hb - mx - hx);
 MOTION OF WATER THROUGH VESSELS 379
which having been reduced, and gz having been written for ½vv, will
yield
(hh - mm)z dx + (hhb + hhx + hma - mmx) dz
= (hha - hhb - hmx - hhx) dx,
which through the Lemma of §XIV is integrable. If the vessel AC
or the fictitious pipe is of exceedingly great size (which is the tacit
understanding of the Problem), there emerges, obviously, this much
simpler equation (after the terms in which m appears have been
neglected, and the remaining ones divided by hh)
z dx + (b + x) dz = (a - b) dx - x dx;
this, integrated, gives (b + x)z = (a - b)x - ½xx, from which, if
z = o, that is, if the surface KL ceases to rise, which happens when it
reaches the maximum height PQ_ to which it can ascend, it is neces-
sary then that also (a - b)x - ½xx becomes = o, wherefore a - b =
½x or x = 2a - 2b; therefore KP= 2KM.

SECTION XXV
The same Problem can be solved more easily if it is considered as
the case in §XIII. By understanding, of course, that the vessel AF
in Fig. 2, full of water at the beginning of flow, has a height a = MH
in Fig. 6, and the pipe FC, which is horizontal in Fig. 2, is now verti-

D
G ··· 0

FIGURE 6

cally erect and continued indefinitely, in the latter let the lowest part
of length b = HK be full of water at the beginning in Fig. 6. Now,
therefore, if because of the prevailing pressure of the aqueous column
in the vessel the water in the pipe ascends above b through the dis-
tance x, and likewise that in the vessel descends through the distance
 HYDRAULICS, PART I

Ji x, we will have the motive force about to arise in the vessel due to
the weight of the water lying above equal to g(ha - mx), and the
motive force opposite to the prior [forceJ in the vertical pipe coming
from the weight of all the water existing in the pipe equal to
g(mb + mx), which, translated to the vessel, gives g(hb + hx), to be
subtracted from the former g(ha - mx). And so there remains
p = g(ha - hb - hx - mx), to which must be equated the sum of the
three motive forces generated by the motion through the throat, the
pipe, and the vessel, as we found in §XIII; after that has been done,
the following equation is obtained:
hh - mm (hbv dv + hxv dv) mv dv (h )
2h vv + dx + hdx a-mx
= p = g(ha - hb - hx - mx);
which, after corresponding terms have been joined together, will have
this form:
hh - mm ( mmx) v dv
2
h vv + hb + hx + ma - h dx = g(ha - hb - hx - mx),
exactly that which we found just above.

SECTION XXVI
From our Theory set forth so far, the physical reason can be given
(which, certainly, neither Newton nor anybody else gave correctly
from purely dynamical principles) as to why obviously a solid cylin-
drical body which is moved uniformly in a continuous infinite fluid
of the same density as the body, with its own base directed forward,
suffers a resistance equal to the weight of the cylindrical body, with
the assumption, of course, that the velocity of the body is equal to that
which a heavy weight can acquire by falling freely from a height
equal to the side of the cylinder. From a number of proofs, which
are mine, it is pleasing to give the following support to our Hydraulics
Theory in this writing.
Let the cylinder RMNS (Fig. 7), which may be moved in the direc-
tion of side MN, be in a standing fluid, equally dense, continuous, and
infinite. Let the velocity of the cylinder be v, the side MN be a, and
the base, or the area NS, be h. Let us imagine that in place of the
solid cylinder there is the pipe MS full of the same fluid matter, and
through this stationary pipe (where beyond the boundary configura-
tion I consider nothing additional) there flows at a continuous and
constant velocity v an integral fluid cylinder, such that the pipe re-
 MOTION OF WATER THROUGH VESSELS

mains always full, and that as much as flows out through NS is brought
in through MR as a new supply at the same speed. To an observer
it becomes manifest at once that the fluid cylinder in efflux through

TF

FIGURE 7

NS meets, of course, the same resisting force by the approach to the

standing fluid external and opposed to the motion as a solid cylinder
itself would meet, because the fluid cylinder, while it is moved
through the pipe, can be considered as solid, and all the other cir-
cumstances are equal. Therefore, it is only to be seen how great the
resistance is which the fluid going out is experiencing at the moment
of egress itself. However, it is evident that this resistance develops
from the throat TNSV which is formed behind the orifice NS of the
pipe; the shape of this throat must be such that at an arbitrary small
distance it has the asymptote FG perpendicular to the direction of the
axis of the pipe for the reason that, on account of the very rapidly
decreasing and wholly vanishing motion of the fluid which has come
out, the areas of the throat must increase in turn, and in a very short
time they must be spread out practically to infinity. I assume, to be
sure, that the fluid coming out of the pipe is not miscible with the
other at rest outside. Thus through those matters which were ex-
plained in §IX, and because the ultimate velocity in the throat is v,
the force through the throat will be ½hvv, and as well, on account of
the constant velocity in the pipe, there will result, through Coroll. 2,
§IX, ½hvv = p = gha, that is, ½vv = ga. Thus by writing gz for ½vv,
there will result z = a. It is proper, therefore, in order that the resis-
tance become equal to the weight of the cylinder, that the required
 HYDRAULICS, PART I

velocity of the fluid in the pipe should be that which a heavy object
falling freely from the height a would acquire. Q.E.D.
COROLLARY. From the fundamental property shown (previously
not sufficiently accurately established), there follow all things beyond
this which are commonly transmitted concerning the resistances of
continuous and nonelastic fluids. Certainly the resistances in fluids
of this sort exerted perpendicularly on opposite planes of bodies are
composed in proportion to the square of the relative velocity and the
first power of the density of the fluid. From this at last the remaining
are deduced.

CONCERNING THE PRESSURE ON THE BASE OF A

CYLINDRICAL VESSEL (WITHOUT AN ATTACHED
PIPE) DUE TO A FLUID FLOWING OUT THROUGH
AN ORIFICE

SECTION XXVII
Let there be constantly filled by fluid the cylindrical vessel AF
(Fig. 8), of which the area AE = h, the length AG or EF = a, and the

FIGURE 8

area GB of the orifice = m. Let the discharging fluid, after [fluid]

has already been flowing out for some time, have a velocity v, so that
in the vessel itself it has the velocity 7i v. Let x be the length CL of
a cylinder of which the base is m, which cylinder may define the
quantity of fluid already having flowed out. Now, furthermore, let
the velocity which will exist afterwards be u, and the length of the
previously mentioned cylinder of fluid which will flow out further be
y, and also the total length of that which has flowed out and that which
will flow out be x + y. Let us also consider the fluid to be free from
all gravity, and likewise to have no other force for pressing the base
than that which stems from the motion. This force meets an equal
 MOTION OF WATER THROUGH VESSELS

resistance from the opposition of the base on account of the equality

between action and reaction. But the resistance is found if the re-
tardative force, which diminishes the velocity of the column of fluid,
is found in the usual manner, and that [force], multiplied by the mass
of the column, that is, by ha, will give the resistance of the pressure on
the base. I accomplish the matter as follows: the equation shown
in§XI,
hh - mm hbv dv mav dv
2h vv +~+ ~ = p,
is changed in the present case (where the length b of the pipe, since it
is absent, is o, and the weight p of the column of fluid in the vessel is
o) into this particular equation,
hh - mm mav dv
2h vv +~ = o,
and by replacing v by u, into this similar one,
hh - mm mau du
2h uu +~ = 0.

SECTION XXVIII
Through reduction, and after dy has been written for dx (for x is
now constant, while x + y is the indeterminate and variable length of
the cylinder of fluid flowing out), the equation appears in this form:
hh - mmd ma du
DI+ - - = o·
2h U '
and after integration,
hh - mm hh - mm
2
h (x +y) + ma In u = ma Inv + 2
h x.

I write this so by adding the last two constant terms for the sake of
rectification, to the end that, for y vanishing and u beginning at v, the
equation itself becomes an identity. And so there will result
ma In(;) = - (hh ~h mm)y, from which, by going to numbers and
by putting I = In e, one has uu = vve-<hh-mm)y/hma.

SECTION XXIX
After proper differentiation of this derived equation (after certainly
having assumed v as constant), there will result

u du = -vv dy ( hh - mm)/ e<hh-mm)y/hma_

--=---
2hma
 HYDRAULICS, PART I

However, the accelerative force in the vessel is negative, that is, it is

.
transforme d into t h e retar d ative wrce - mu
r du , w h'1ch accor d'mg1y
h dy

will be VV (hh - mm)/e<hh-mm)y/hma_

2hha

SECTION XXX
This force, which is useful to us at the very first moment after the
abolition or supposed cessation of the gravity which previously the
column of fluid in the vertically erect vessel had, will be vv = 2gz, as
certainly y = o; and so that force which was found will be
gz (hh ~amm) (see §XI, where

z = (hh ~amm). (I _ i//hh-mm)x/hma)),

and hence by multiplying by the mass ha of the fluid, the resistance or

the pressure on the base arising from the motion of the fluid alone is
found to begha(1 - 1/e<hh-mm)x/hma), to which ifin addition is added
the weight gha of the column of fluid, which in a vertical position acts
constantly on the base whether the fluid is moving or at rest, there
will result a total pressure gha + gha(1 - 1/e<hh-mm)x/hma).
COROLLARY. If x = oo, the total pressure will be gha + gha/e0 =
2gha, as e0 = 1. But if x = o, the pressure at the base will be
gha + gha ( 1 - 1) = gha, which is also obviously true from the fact
that initially only the weight of the fluid cylinder acts on the base,
and afterwards, with x increasing, the pressure also increases, and in
such a way that it never attains 2gha, much less exceeds it, although
it may approach this quantity rather closely.
ScHOLIUM. On the other hand, let no one believe that the heavy
liquid in the vessel perhaps presses on the base differently when it is
being moved from when it is at rest; notwithstanding that the con-
trary is readily evident to an observer of the nature of immaterial
forces; just as for example the cause of gravity was considered to be
apart from a body, so these forces act instantly throughout the total
mass to be set in motion, and thus they act in the same manner and
they exert the same pressure on an obstacle as if the heavy liquid
lying over it were at rest. Nevertheless, I will prove the truth of the
matter in our case through calculus. Certainly it is clear at once
that the heavy liquid is accelerated by descending in the vessel.
Moreover, its motive force, however great it may be, must have two
parts, of which one is given over to the retardative force at the base
 MOTION OF WATER THROUGH VESSELS

opposed to that to be considered, but the other part remaining is

employed in the acceleration of the actual descent. But this latter is
exactly that which results from the equation of §XI which is to be
solved, that is,

z hha
= ___ (I _ i /e<hh-mm)x/hma),
hh - mm
by differentiating which and multiplying it by g we will have

g dz = V dv = -h g dx je<hh-mm)x/hma.
m

Therefore, the remaining accelerative force in the vessel, or :v d~'

is equal to g/e<hh-mm>x/hma, to which if there is added the [accelera-
tive] force which the above-found retardative force diminishes,
g( 1 - 1 /e<hh- mm>xlhma), together they constitute an accelerative force
g, and therefore they constitute a pressure gha developing from the
weight, that is, equal to the weight itself. Q.E.D.
And indeed one will have to proceed thus in the remaining cases
where one or more pipes are attached to the vessel, so that, naturally,
first of all one may find the retardative force on the base of any one of
the given pipes by supposing that the fluid suddenly parts with its
gravity, and then one may add to the discovered retardative force the
pressure provided by the weight only of the fluid (by considering it as
being at rest and stagnant) and propagated either directly to the first
base or indirectly through the preceding pipes to whatever base we
wish.

APPENDIX:
OUTLINE OF THE CALCULATION TO BE EMPLOYED
FOR DETERMINING IN A SINGULAR WAY THE
VELOCITIES OF WATER FLOWING THROUGH
MANY PIPES FROM ONE TO ANOTHER, AND
ESPECIALLY IF IT SHOULD FLOW OUT THROUGH
SEPARATE INDIVIDUAL PIPES, AND HENCE TO
BE EMPLOYED FOR FINDING THE PRESSURES
EXERTED UPON THE BASES OF THE
INDIVIDUAL PIPES.

In advance it is necessary to point out that we assume this conduit

to be composed of several pipes attached to each other one by one,
 HYDRAULICS, PART I

having any position whatever, vertical, horizontal, or inclined. We

next assume that the conduit is constantly full of water and that the
flow has come to steadiness, during which time as much liquid flows
out of any pipe as is necessary for supplying the next smaller pipe,
such that therefore every one is constantly full, and that they may be
thus considered individually as if they were isolated and set apart.
Let the length of the first and largest pipe be a, that of the second
and next smaller be b, that of the third be c, etc.; the area of the first
be h, of the second be m, of the third be n, of the fourth be q, etc.; the
orifice of the first pipe, equal to the area of the second pipe, be m, the
orifice of the second be n, the orifice of the third be q, etc. The
natural gravity beg, the natural gravity in different oblique directions
bey', y", y'", etc. But for the vessel or first pipe let the gravity arising
from the mutual action in the attached pipes beg', for the second g",
for the third g"', etc. Let the length of the aqueous cylinder having
flowed out through the first orifice be x', that through the second x",
that through the third x'", etc. Let the height from which a weight
having fallen naturally acquires the velocity of the water flowing out
through the first orifice be z', that which pertains to the second z", to
the third z"', etc.
And so with these things having been set forth, and the remaining
things as they are in hydraulics Literature, certainly there results, for
pipes attached to each other in succession, through the translation of
forces, and indeed for a conduit of two pipes gha + y' hb = g' ha +
g"hborga + y'b = g'a + g"b;foraconduitofthreepipesga + y'b +
y"c = g'a + g"b + g"'c. I call these equations fundamental.
Next it is clear that there results z' = !!!!.. z" = .!l!L z'", etc. Like-
mm mm
wise x' = !!:.. x" = !l.. x"', etc. On the other hand, through §XI, the
m m
following equations result for the individual pipes being vertically
erect:
For the first pipe,
g(hh - mm)z' dx' + ghma dz' = g'hha dx'
For the second pipe,
g(mm - nn)z" dx" + gmnb dz" = g"mmb dx"
For the third pipe,
g(nn - qq)z"' dx"' + gnqc dz'" = g'"nnc dx"',
and so on in succession.
 MOTION OF WATER THROUGH VESSELS

Note that, if any one of the pipes were horizontal, the y pertaining
to it would vanish in the fundamental equation. Thus for example
if there were three pipes, of which the first were vertical but the re-
maining two were horizontal, the fundamental equation would be
this: ga = g' a + g 11 b + g111c; but if in fact all three were vertical, this
fundamental equation would result: g(a + b + c) = g'a + g b + g 'c.
11 11

Now since that is the prize achievement in this investigation, it is

necessary to define the forces of the gravities g', g 11 , g'", etc., resulting
from the mutual action of the natural gravity, from which later the
velocities as well as pressures on the bases of the pipes become known.
Moreover, I describe this so. By imitation of the operations applied
in §XI, that which follows is found for the individual pipes:
For the first pipe,

z' = !{ ( hha
g hh-mm
)(r _ r/e<hh-mm)x'/hma)

For the second pipe,

Z 11 -
_ g"
- (--
mmb
- - )( I _ I / e<mm - nn)x" /mnb)
g ,mm - nn
For the third pipe,
,,,
z=
g Ill ( nnc (
- - - -) I - r/e<nn-qq)x"' /nqc)
g nn - qq
And so on.
And accordingly, since z ' , ;:,
11
, z'", etc., as well as x', x 11 , x"', are given,
· 1y, m
respective · terms of eac h oth er-name1y z' = -nn z II
= -qq z,
111

mm mm
and also x' = !!.. x = !l x"'-let
11
the values of the individual z and x
m m
be expressed by one of them, and there will result as many equations,
less by one, as there are pipes, or as there are hypothetical gravities
g', g 11 , g'", etc. Certainly, for example, for three pipes, z' having been
retained to which the remaining z", z"' are to be reduced, and also
x 11 , x"' to the retained x', there will result these two equations:

z' g' ( hha ) ( _ 1 /e<hh - mm)x' /hma) = .!!!!_

g
z11
or hh - mm 1 mm
or
11
nng ( mmb )( 1 r/e<nlm)(mm-nn)x'/mnb),
mmg mm - nn -

and that first one is also the same as

qqg'" ( nnc
mmg nn - qq
)(r - rfe(q/m)(nn-qq)x' /nqc).
 HYDRAULICS, PART I

But since there are three hypothetical gravities g', g", g'", to be sought,
another equation is required at this point for the solution of the
Problem. Moreover, this must be sought from the fundamental
equation ga + y'b + y"c = g'a + g"b + g"'c, or (if indeed the two
pipes are assumed horizontal) from this there is merely ga = g' a +
g" b + g'"c, since y' and y" vanish.
Let us make an application, for the sake of brevity, to the very
simple case of two pipes constantly full of water, of which the first may
be vertical, the other horizontal, and let us stipulate that the flow has
reached uniformity, that is, x', x", etc., = oo. There will result one
equation drawn from z',
'( hha ) nng" ( mmb )
g hh-mm = mm mm-nn'
another from the fundamental ga = g'a + g"b, from which, by pro-
ceeding customarily, one has
, gnn(hh - mm) ,, ghha(mm - nn)
and
g = mm(hh - nn) g = mmb(hh - nn) ·
.. d"d . , hhnna
t e rema1nmg are enve , as certain1y z =
H ence a llh (hh )
mm - nn
and z" = hh hha , agreeing absolutely with those which we have
- nn
proven and given above. Likewise the pressures on the base of any
pipe are most easily determined. Since surely the individual pipes
can be considered as if they were solitary, the formula must be em-
ployed which we found above for the first and only pipe with the
writing of only the letters which are suitable for any other pipe con-
sidered as if alone or separate. Accordingly, since for that one alone
the total pressure was found as gha + gha[1 - 1//hh-mm>xlhma], here
the total pressure will have to be written:
For the first pipe, g'ha + gha[I - 1//hh-mm)x' Jhma],
For the second, g"mb + gmb[I - 1/e<mm-nn)x"/mnb],
For the third, g"'nc + gnc[ I - 1 je<nn-qq)x"'/nqc];
and so, after the values of g', g", and g"' themselves have been substi-
tuted--or, because we make the application to only two pipes, and
indeed where x = oo, merely the values of g' and g" must be substi-
tuted, which are g' = gnn(hh - mm) /mm(hh - nn), and g" =
ghha(mm - nn) /mmb(hh - nn)-the first pressure will result as
nn(hh - mm) 2ghha (mm - nn)
2gha mm (hh -nn )' and the second pressure as -m- hh -nn ·
If in addition h = oo, but m and n are finite, the first pressure will be
 MOTION OF WATER THROUGH VESSELS

2gha -M = oo, as 1s
. proper, b ut t h e secon d pressure w1"11 b e
mm
2
ga (mm - nn), which is finite.
m
This method having been observed properly, the gravities for
5 pipes, g', g", g'U, g1v, gv, are found as follows:
g' = ghhssa(hh - mm) /hhmma(hh - ss)
g" = ghhssa(mm - nn)/mmnnb(hh - ss)
g'" = ghhssa(nn - qq) /nnqqc(hh - ss)
g1v = ghhssa(qq - rr) /qqrrd(hh - ss)
gv = ghhssa(rr - ss) /rrsse(hh - ss).
From this foundation, the law of progression for any number of
pipes develops more satisfactorily. And so in the truncated conoidic
conduit FE (Fig. g) attached to the cylindrical vessel AF, which conduit

FIGURE 9

is considered as being made up of innumerable pipes of infinitely small

length, there will be found, for any area NO, the hypothetical gravity
by which the layer of water of infinitely small thickness is set into
motion when the flow will have reached constancy. By designating
the area NO by y and the thickness of the aqueous layer by dx, and by
employing the rest of the nomenclature which we have used so far,
[one finds that] the gravity setting this layer into motion will be
ghhwwa · 2_y dy ghhwwa · 2dy
Further, the weight itself of this
y 4 dx(hh - ww) y 3 dx(hh - ww)"
 HYDRAULICS, PART I

layer, or the pressure by which it is agitated, will result if one multi-

plies by the quantity of matter y dx; therefore this pressure will be
ghhwwa,2dy
To this then there are to be added the pressures of all
yy(hh - ww)'
the following layers from O right through to the limit B, but gathered
by translation to the position 0, as our method explained at the be-
ginning requires. In the meantime, to this end lety be assumed to be
constant and the other area, RS = t, to be variable; the pressure of the
. ghhwwa · 2dt .
latter layer t dx will be (hh )' which may be translated to the
tt - WW

. ghhwwa · 2dt . ghhwway · 2dt

invariable position NO b y mak ing ----,,,-,----..,. mto - - - - - as
tt(hh - WW) t3 (hh - WW)
t is to y, the properly correct integral of which gives
ghhwway ghhwway
ww(hh - ww) - tt(hh - ww)'
. ghha(J.i'I! - ww)
where now, by puttmg t - y one has "---,-,-'"--',__,~----,--~ as the
- ' y(hh - ww)
total pressure by which certainly the water at NO is compressed.
Therefore, in order that z, the altitude of the aqueous cylinder
of which the base is y and the weight is equal to this pressure,

may be found, gyz is to be set equal to g~:~y- -w:)), from which

hha(yy - ww)
"' - ---,-"--'---~~ And thus the water in the pipe inserted at the
,._ - yy(hh - ww) ·
position N will remain at this height NM. But it is to be noted that
the conduit FPBC is considered as having the diameters of the maxi-
mum area FP and minimum area CB small enough with respect to
the length PB so that certainly the tangent to the curve FNC at
any point N makes a very small angle with the horizontal PB. For
otherwise, from the impact of the water while moving toward the
overly curved side FNC of the conduit, there would arise a new
force of pressure (which is here being neglected as accidental)
which, augmenting the prior [force], would increase the height
NM, just as actually occurs if the conduit FB ends in a plate, per-
forated by an orifice of area w, on which the water, impinging
perpendicularly, can increase the compression in the regions near the
orifice. In the more remote regions that increase becomes less
noticeable, and it varies as the curvature of the throat postulates,
which moreover I consider to depend upon the peculiar nature of the
water or other fluid flowing through, and so to be generally indeter-
minable.
 SECOND PART
Containing the Direct and Universal Method for
Solving all Hydraulics Problems whatsoever
which can be Formed and Proposed
Concerning Water Flowing through
Conduits of any Shape
SECTION I
Consider any conduit whatever; let it be straight or curved, or let it be
continuous or composed of many cylindrical pipes, or, finally, let it be
vertical, or horizontal in part, or inclined differently in its different
parts. Let this conduit be full of water, or some other heavy liquid,
homogeneous and very fluid. Moreover, let it begin and continue to
flow by accelerating (as much and as long as it can), and indeed in
such a way that the conduit remains constantly full, with new water,
of course, entering from another source; this replaces that escaping
from the last orifice at any moment by flowing in through the first
orifice with that velocity with which the uppermost surface would
subside if the inflow were suddenly stopped. This condition is added
for the sake of easier calculation; the method is surely valuable if no
new liquid were to enter, right up to the complete depletion of a
vessel or conduit. First, the velocity of the liquid flowing out is
sought for any given quantity of liquid already having flowed out.
Then is sought how much the sides of a conduit are pressed at indivi-
dual points by the liquid flowing through, or, since it comes back to
the same thing, to what vertical height a liquid of the same kind as
that flowing through must remain elevated in a pipe inserted at any
point and erected vertically.

SECTION II
And so let the arbitrary conduit ECce be given (Fig. 10), the vertical
line AB considered as the axis of the abscissas, attached to which in
39 2 HYDRAULICS, PART II

order are AEe, PFJ, TNn, and BCc, the parts Ee, Ff, Nn, and Cc, of
which may define the areas or horizontal sections of the conduit. Let
the liquid contained in it be considered as divided into horizontal

FIGURE IO

layers of infinitely small thickness, FMmf, NLln, etc., the intermediate

points of which, or the centers of gravity G, H, V, I, etc., may form the
line GHVI, either straight or curved, which I shall call the center line,
or simply the centric; the latter certainly will be given, on account of
the given curves EFG, and efc, which are determined from the given
Shape of the conduit.
Let the first area Ee be h, the last area Cc be w, some intermediate
area Ff bey, another intermediate one Nn be r, and the thickness of
the individual layers PR or TS be dt. The layers themselves will be
Fm = y dt, Nl = r dt. Let, furthermore, some indeterminate straight
line ID, which is oflength x, touching the centric at I, be equal to the
length of the oblique cylinder of liquid flowing out in the direction of
ID itself, the base of which cylinder is Cc, and let this contain the
quantity of liquid already emitted. The velocity of the liquid
flowing out at that very moment is v. Let the gravity by which bodies
are naturally set in motion beg. Now by designating the elements
Hh, etc., of the centric line as ds, the gravities by which the layers are
set in motion in the directions Hh, Vu, etc., will beg it, and therefore
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 393

directions will be gyd~t

2
, gr ;;2, etc.; but the absolute motive forces
themselves [acting] in the vertical direction will be gy dt, gr dt, and
so on.

SECTION III
By translating these absolute forces (through the Hydrostatic
Principle as it was shown in the first Part) to the first area h, the former
will be gh dt for the individual areas. Therefore, by integrating
through all dt, that is, through the total height AB (let this be equal
to a), gha will be equal to the total pressure to be applied vertically
to Ee, equivalent to the sum of the absolute motive forces in all the
sections. And this total pressure gha to be applied to the first area is
that which is customarily called p by me.
Now let the tangent to the center line CHI at I be to its vertical
subtangent [i.e., vertical projection] as ex is to 1, and the tangent G to
its subtangent as f3 is to 1, but the tangent at any intermediate point
H to its subtangent as ds is to dt. Certainly through the resolution of
the motion, v, or the actual effiux velocity of the liquid at I, will also
be to its vertical subvelocity [i.e., vertical velocity component] as ex
is to 1, and therefore that subvelocity is !!..
Similarly, by calling u
ex
the actual velocity at Hin the direction of Hh, its subvelocity will be
u dt
In order that the actual velocity u may be found, however, it
is to be noted that the subvelocities of the layers are in reciprocal ratio
to their areas, to the end that they transmit equal quantities of liquid
in the same elementary time interval; and so, by setting y /w = ~;vw,
ex exy
vw will be equal to the subvelocity of the layer Fm. Furthermore, by
exy
. vw;vw ds vw ds .
settmg dt/ds = - -d-, -d- will be equal to u, or the actual
exy exy t exy t
velocity of the layer Fm in. the direction Hh . Hence the actual velo-
city of the first layer adjacent to the area Ee (wherey is inserted for h,
and ds : dt = f3 : 1) will be {3:v·
By no means is one to reason differently in finding the actual dis-
placement of the layer Fm in the direction Hh. For by calling dx
the instantaneous displacement of the last layer Cc in the direction
the weights or the motive forces of the layers themselves in those
 394 HYDRAULICS, PART II

ID, its subdisplacement in the vertical direction will be dx; more-

a
over, here the subdisplacement of the layers are also in reciprocal
. . dx;w dx w dx
rat10 to the areas, and so, by settmg y /w = - --, [then] - -
a ay ay
will be equal to the subdisplacement of the layer Fm. Therefore by
. w dx;w dx ds w dx ds . .
settmg dt/ds = - - --d-, --d- will be equal to the actual d1s-
ay ay t ay t
placement of the layer Fm in the direction of its motion, Hh.

SECTION IV

With the water already in motion, its layers act mutually on each
other in different ways by pushing and resisting, and certainly with
different forces, on account of the diversity of the surroundings with
respect to position as well as speed. And so, in the meantime, let the
indeterminate accelerative force which arises from the mutual action
be called y, and let the acquired velocity which some layer Fm has in
the direction Hh be called u. Therefore y ds = u du, from which
u du
y = ds . Let this be multiplied by the mass of the layer, y dt, and

its motive force yy dt = yu ~; dt will develop in the direction Hh.

However, in order that this may be obtained in the vertical direction,
from which the former can be produced, one has to set

yu du dtl
dt/ds = ----;J;- yu du.

This will be the motive force required in the vertical direction, which,
therefore, translated to the first area h, gives the equivalent hu du .
Let this be integrated so that ½huu is obtained, which, through the
necessary correction, is to be applied to all layers contained and added
together in the entire conduit ECce. Therefore (since the velocity of
the last layer is v, and that of the first is f3a:v), the correct integral is

h( f3f3ww ) vv (aahh - f3f3ww) h. . l h ·

- vv - -hh vv or h ; t 1s 1s equa to t e eqmva-
2 aa 2aa
lent vertical force to be applied at Ee by which, of course, the indivi-
dual layers obtain their own particular forces for pushing each other
mutually, to the effect that they merely conserve their own effort at
the moment at which the water is flowing out at the velocity v; thus,
since it occurs not continuously but in an indivisible instant and

--------- - -
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 395

[since it] depends upon the shape of the conduit only, this force,
having arisen from the translation, can be called the static force or
static potential, or, if it be more pleasing, the hydrostatic potential, since it
consists in the effort alone of making the transition from one layer to
the next lower, no attention having been given to the actual accelera-
tive force.

SECTION V

Further, the other force is to be sought which arises from the actual
acceleration of the liquid flowing through. To this end I set the
actual accelerative force of any arbitrary advancing layer Fm = y';
(on account of the actual displacement of the last layer Cc through
the small space dx) the displacement of the layer Fm will be w dxdds,
ay t
y'w dx ds , , wwv dv(ds) 2 .
and therefore (§III) d = u du = (d ):.i , from which
ay t aayy t
yI = WV dv ds
d d , an d t h e actua1 motive
. fcorce m
• t h e d"irection
. of t h e 1ayer
ay X t
uh,
n, or y 'y dt, is
. wv dv . 1 motive
d ds , an d so t h e vertica . fcorce firom w h"1ch
a X
2
the latter can be produced is wv ~(~) , which, translated to the first
a X t
. h . 1
area h, gives t e eqmva ent hwv dv(ds)2
d d ; m . ord er t h at t h"is may b e
ay X t
integrated through the total length AB of the axis corresponding to
the whole conduit, for any layer and for any acquired velocity v, herein
not only h and w but also v ddv must be considered as constant; and so,
a X

by integrating, we will have hwvddv

a X t
J(dsd)
y
2
equal to the other force
that is to develop from the actual acceleration of the liquid flowing out,
which it may please one to call the hydraulic force, as distinguished
from the hydrostatic force, which consists of effort alone, or pressure
employed instantaneously, and this at every instant, however the
liquid may be moved.

SECTION VI

These two forces, the hydrostatic and the hydraulic, compose the total
force which certainly is generated by the action of the primitive p,
which was found in §III to be gha. Accordingly, by equating this
 HYDRAULICS, PART II

to the sum of those two found in §§IV and V, we will obtain the
most general equation for the determination of the velocity with which
the liquid flows at any moment, which equation is this:

vv(aahh - f3{3ww)
2aah
+ hwv dv
a dx
J(ds)
y dt
2
= ha.
g

Here it is to be understood that J~1t 2

denotes the sum of all the
2
dsd which are contained not only between Cc and Ff but everywhere
y t
between the extremes, by including all from one [extreme] to the
other.

SECTION VII

If one now wishes to express the equation in z, or the height from

which a given body, by falling under natural gravity g, would acquire
the desired velocity v, then one should write, according to dynamic
principles, 2gz for vv and g dz for v dv, which will give this equation:

gz(aahh - f3{3ww) + ghwdzf(ds) 2 = ha·

aah a dx y dt g '
or, after a reduction has been made, this:

(aahh - f3{3ww) z dx + ahhw dz I 2

~1t = aahha dx;

or, because J~dsJt

2
' to be taken through the total length of the axis,
can be considered as constant, and thus as given, at least through
quadratures, let that be called M; and the equation will be reduced
to the following form:
(aahh - f3{3ww)z dx + aMhhw dz = aahha dx.
From the resolution of this equation, z will be found in quantities
given in x and the constants M, a, h, w, a, {3.
COROLLARY r. For an existing uniform velocity of efflux, to which
it reasonably arrives very quickly, and almost in one wink of the eye,
as will be shown in the proper place in this writing, dz vanishes;
therefore, after this has been neglected, the following algebraic
equation develops: (aahh - f3{3ww)z = aahha, from which the desired
aahha
Z = aahh - f3{3ww ·
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 397

COROLLARY 2. Hence let two vessels or two conduits have any

shapes whatever, even though very different from each other, provided
only that they have the same vertical height a, and that they have the
uppermost and the lowest areas, or the first and the last, h and w, in
the same ratio, and furthermore that they have a and f3 in either case
proportional to each other. The water will fl.ow out from either
conduit or vessel at the same velocity after the flows in both cases have
become uniform.
COROLLARY 3. If the centric line GHJis a straight line, whether it be
vertical or oblique, there will be f3 = a, and ds : dt = a : r. Hence,
2
ds = a dt, and J~1t ' or M, = aa J~' which changes the general
equation (aahh - f3f3ww)z dx + aMhhw dz= aahha dx to this:
dt
(hh - ww)z dx + ahhw dz
Jy = hha dx.

But, whatever might be the position of the straight center line, whether
vertical or oblique, in the case of uniform effiux z will always be
hha
hh - ww·
COROLLARY 4. Now, with the shape of the conduit or vessel being
maintained, and both the uppermost and lowermost areas Ee and Cc
as well, if a little change is made in the direction of the inflowing and
outflowing liquid, that change, even if it be hardly noticeable, can
produce a marked change in velocity. In Fig. 11, for example, if
borders or lips Emne and Cpqc, having very little height Em and cp, are
added to the vessel or conduit ECce such that the areas mn and pq
are kept the same as the previous values Ee and Cc, and such that the
total height of the vessel is not increased noticeably, no one will
easily believe how much the change in velocity is going to be due to
this operation. Whereas now certainly the water flows in and out no
longer obliquely but vertically, on account of the vertical direction of
the lips, which even gives a vertical orientation to the extreme tan-
gents of the center line, and so yields a = f3 = I, it is clear that the
general equation
(aahh - f3f3ww) z dx + aMhhw dz = aahha dx
now at once assumes this form:
(hh - ww)z dx + Mhhw dz= hha dx;

and, for a uniform velocity, there will be z = hh hha ; I consider it

- WW
 HYDRAULICS, PART II

worthwhile to show this, lest otherwise, if, in experiments yet to be

performed, one does not attend accurately enough to trifling circum-
stances which seem to be ofno importance, and then, what is produced

FIGURE II

seems, falsely, to agree less accurately with ours; lest, I say, our theory
be immediately suspected of being in error. And so it happened once
to the illustrious Poleni, otherwise a well-regarded and industrious
Man in experimental matters, who, about to see what different quan-
tities of water in a given time openings of different size would emit,
[the openings] having been affixed to the same vessel full of water,
had arranged for this purpose that there should be several thin plates
not altogether compact but that every one be pierced by an orifice of
some particular size, so that first one and then another [plate] would
cover the aperture at the bottom of the vessel. Moreover, it occurred,
unless I am perchance mistaken, that he repeated the experiment twice
with some of those plates, and diligently many times thereafter, where,
to his astonishment, he observed that that one and the same plate,
through its same orifice, had emitted a sometimes greater or sometimes
lesser amount of water in the same time, accordingly as one or another
face of that plate was looking outward. Finally the form of the orifice
was examined more carefully, and then it was observed that the shape
of the orifice, obviously having been cut in a thin plate, was not exactly

-----

1
 ------

METHOD FOR SOLVING HYDRAULICS PROBLEMS 399

cylindrical, but rather like a small truncated cone, having one base a
little larger than the other, which was already sufficient for revealing
the reason why, with the larger base of the orifice being open out-
wards, the water would flow out more fully than in the contrary
sense; and this on account of a double reason, for both the aqueous
stream springing forth was wider, and its velocity was greater, just as
is evident from our formula, z = hh hha , where it is clear that the
- WW

value of this fraction would be greater if w were greater, h and a

remaining fixed, and contrarily, it would be less if w were less.
COROLLARY 5. In the case in which ah = (3w, or where a : /3 =

w : h, it will be true that M- w · dz = aa dx, from which z = Maax,

·w
whence it is clear that, with the effiux x increasing to infinity, z also
increases to infinity, and thus that the velocity never converges to
uniformity. This certainly appears also from the very formula of
Coro11ary I,
c z
1or
-
/3/3 ww = ("1n t h"1s case )
= aahhaahha

aahha a
aahh - aahh 0
= 00.

SECTION VIII

ScHOLIUM I. It is to be noted that in conduits and pipes [that are]

not very wide and not sufficiently long, it is commonly observed, as I
already hinted in the Preface,* that the sections Fm (Fig. 10), being in
a state of flow, easily adjust themselves from a horizontal position to
a position perpendicular to the sides, or rather to the centric line
CHI, which is clearly evident, in any case, from the motion of the
uppermost surface Ee (if no other liquid follows), as for example in
barometric tubes and other syphons of that sort having diameters of
not more than one or two lines; whether this is accomplished on
account of the adhesion of the fluid to the sides, which must be uni-
form around the boundary in a circuit of the layers, in order that the
fluid may be moved as aptly as possible and without noticeable
friction, or it occurs on account of some other physical reason, is not
to be determined at this point. It is sufficient to insert here that this
circumstance does not oppose our Theory. For, because, through
the general rule, the center of gravity of bodies urged into motion by
* p. 355.
400 HYDRAULICS, PART II

any cause whatever is moved in the same manner and at the same
velocity in its initial direction as if their entire material were concen-
trated at the center of gravity itself, certainly the material of any
layer Fm can be considered as concentrated at the center of gravity
Hor h. As, therefore, in not very wide oblong conduits any arbitrary
small portion of them can be taken as quasi-cylindrical or prismatic,
it is evident that every layer Fm, for the slightest reason, can change
its horizontal position Ff to rs, perpendicular to Hh, during which Hh
retains the same length, and the quantity of the new layer rtos is
equal to [that of] the layer FMmf Therefore, let us consider how
individual remaining layers NL (without any other change either in
velocity or in direction with respect to Vu) may adjust themselves into
a position perpendicular to the sides of the conduit or, preferably, to
the centric line. If now we consider further what would happen if
the exit Cc were closed and in its place the orifice cd, of the same area
as Cc, were opened in the wall of the conduit, with no difficulty we
understand that the water must go out through the aperture cd under
the same inclination to cd under which it was going out through Cc,
and thus that its direction bg will be horizontal. Since in addition
the aperture cd is set equal to the area Cc, and the tendency of flowing
through Cc already is diverted towards de (through the common
hydrostatic law), it is necessary, certainly, that the velocity of the water
flowing out through cd will be the same as we determined for Cc.
Whence it is also realized that if to the orifice cd were attached a new
horizontal conduit, in which certainly the centric line would be hori-
zontal, the motion and the velocity of the water flowing through it
and flowing out will obtain in the same manner as they would obtain
if that same new conduit (cd being closed) were attached to Cc in the
direction ID, but in which the water flowing would have to be con-
sidered as deprived of its own gravity. Thus, in order that here as
well the weights of the layers translated to the area Ee yield the same
sum gha, just as if the new conduit were not present, nothing in the
expressed (§VII) general equation

(acxhh - fJfJww) z dx + aMhhw dz = aahha dx

should be changed other than that M, or J~jt 2

' now expresses the
2
sum of all the (dsd) which are contained in both conduits. But the
y t
uniform velocity on either hand will be the same, whether in the
simple or in the combined conduit (because the term in which Mis
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 401

found vanishes in the case of uniformity), as it is always that which is

mxhha
found through z = aahh - /3/3 ww ·

CONCERNING THE PRESSURES WHICH THE SIDES

OF A VESSEL SUSTAIN BECAUSE OF LIQUID
FLOWING THROUGH

SECTION IX
In order that we might comprehend clearly and correctly in what
that force consists which is exerted on the sides of a conduit while
liquid flows in it, it is to be understood that that force is nothing more
than that which takes its origin from a compression force by which,
certainly, consecutive portions of the fluid, for example EFfe and
CFfc, are driven one against the other; whence at Ff by this very con-
tact there arises through action and reaction an intermediate force
which I customarily call immaterial, because it is apart, so to speak,
from the portions pressing each other and yet intermediate between
the two, and it does not pertain to one more than to the other. It is
characteristic of this force to drive the preceding portion of liquid
forward, or in the direction in which it is going, but the following
portion backward, or in the direction from which it comes, and to make
the following portion ofliquid, which is propelled by translated forces,
and the preceding portion of liquid, against which some of the accel-
eration must press, acquire at this very contact an equality of acceler-
ative forces; just as we showed a short while ago,* the same effect
occurs in solid bodies, where, after they have been animated indivi-
dually by different accelerative forces, there arises in their contact,
when they begin to act on each other, an intermediate immaterial
force, appearing to be truly common to each body, which thus would
regulate the particular accelerative force of each, the one by diminish-
ing, the other by increasing, in order that thence in the total mass
combined from these two bodies one common accelerative force may
result.

SECTION X
This, however, is the distinction in the manner of acting: that in
solid bodies acting directly upon each other, that immaterial force
acts forward and backward like some elastic straight line which,
* Nos. CLXXVII, p. 262, and CLXXIX, pp. 333, 340. [Opera Omnia,
Book IV.]
402 HYDRAULICS, PART II

placed between the bodies, tries to expand itself; but in portions of

fluid acting mutually on each other, the immaterial force lying be-
tween must be considered just as elastic air, which extends itself not
only in opposite directions, but into all surrounding regions; from
which now it is easily understood that from this immaterial force
itself the pressure, which is the subject here, develops. This certainly
is exerted on the walls of a conduit, by which in turn it must be con-
fined while it acts freely forward and backward on the portions of the
liquid wherein it exists.

SECTION XI
It therefore remains that, according to this given idea concerning
the immaterial force, we determine its quantity or measure. Let
that [which is] to be sought, which we may say is 1r, be anywhere in
Ff [Fig. 10]. Now I proceed thus: for the time consider a part of
the conduit EFfe (during flow) to be removed suddenly, the remaining
CFfc staying in its place with all its environs, and [consider] at that
same moment that there is placed at the area Ff a new motive force
equal to 1r itself. One understands at any rate that in this way the
effiux of liquid flowing out of the truncated conduit is to be accelerated
(at least in the first instant of time) just as if the conduit had remained
whole. Therefore, I will already consider the residual conduit
CFfc as an integral conduit, the uppermost or first area of which is y,
or Ff, any variable intermediate area Nn is r, and the adjacent
section NL is r dt. Thus if (§IV) for h I substitute y, I will have
2
vv(aayy - ww(ds ) f( dt )2) as the hydrostatic force; indeed that which
2aay
at the first point G was called /3 is ds/dt at point H, the ratio, of course,
of the tangent to the subtangent, and (§V) a dx
J
ywv dv (ds) 2 • h
r dt rs t e

hydraulic force, where, in the integration, r is considered to be con-

tinuous from w right on toy.

SECTION XII
The sum of these two forces, the hydrostatic and the hydraulic,
should be equated to the original force p, which here should be
(§§ III and VI ) gyt, if indeed this alone were acting on the liquid
contained in the truncated conduit; but, because 1r acts together with
gyt, it is necessary by all means to establish this equation:
vv(aayy - ww(ds) 2 /(dt) 2 )
+
ywv dv J(ds)d 2
_
+ 7T.
2aay a dx r t - gyt
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 403

From this emerges at once the sought value of 7T itself: namely, for
gyt having been transposed, there results

2aay a dx
J
r dt
2
vv(aayy - ww(ds)2/(dt) 2 ) + ywv dv (ds) _
gy
t = 7T
'

where also it is to be warned that, in the integration J(:1t2

' r must be

taken as variable from B to P, whence, for any assumed y, J(:1t 2

will
be given. Therefore, let this be set equal to N, and there will result
vv(aayy - ww(ds) 2/(dt)2) Nywv dv
2aay
+ a dx - gyt = 7T.
Since, therefore, from the resolution of the general equation (§VII)
there evolves the value of vv itself, or 2gz, this substituted in the
latter will give the value of 7T itself in terms of g and purely linear
quantities. Thus, now,
gz(aayy - ww(ds) 2/(dt) 2) gNyw dz
aay
+ a dx - gyt = 7T.

SECTION XIII
If now, furthermore, it is desired to know, if some tube open at both
ends is introduced at some arbitrary place j in the conduit and is
erected to a vertical position, how far the liquid must ascend in it on
account of this pressure 7T which makes it ascend, it is agreed to con-
sider that 7T is equal to the weight of some cylinder formed from the
liquid, animated by the natural gravity g, which has for a base the
area Ff or y, and for a height that very [height] of the liquid standing
in the tube; whence 2 will be this height at which the suspended
gy
liquid will stand in the tube, invariable indeed, after the velocity of
the liquid flowing out will have reached reasonable uniformity; but
before this occurs (although it happens in an instant, more or less),
the liquid proceeds to ascend in the tube until it will have acquired
the appropriate stable location, when indeed the liquid flowing out is
no longer noticeably accelerated.

SECTION XIV
It happens in certain cases that the value of 7T itself might be nega-
. .h h 11 h . . . -vvww(ds)2
tive, e1t er w en, natura y, t e negative quantities (d ) - gyt
2aay t 2
 HYDRAULICS, PART II

prevai·1 over t h e positive

. . 1
2 vl!)I
ywv dv
+ ·--d-
ex x
J--d-'
(ds)2 or, wit
r t
. h t h eve1ocity
.
2
. um·rcorm, so t h at dv = o, w h enever vvww(ds
1 d y b eing
area
)
(d ) + gyt is
•
2exexy t 2
greater than ½Vl!JI; this can happen not only in those cases where exy
is w ds b ut a 1so mt
. 1ess t h an dt' . h ose w h ere exy is w ds i.f on1y
. greater t h an dt'

at the same time gyt be great enough that its excess over

ywv dv
ex dx
J(dsr dt)
2

be greater than the defect of the prior. However, in whatever

manner it may occur, it is plain that in cases of this sort the pressure is
changed into relaxation, which causes the walls of the conduit around
Ff not only not to be pressed outward, but to be contracted inward
everywhere (if the rigidity of the walls is not interfered with). Hence
it follows that the liquid can be elevated upward as ifby suction to the
height !!.... through a tube implanted in the conduit as before but
gy
directed vertically downward to a very low point where it opens into
a little vessel full of liquid.

SECTION xv
ScHOLIUM 2. Thus far we have not attended to certain particular
and accessory causes (not always having importance) which can
change either the pressures or the suctions TT determined by our
method. Among such causes, one occurs principally which acts
such that the water, being in motion and striking an immobile surface
in its path, impresses a force upon it during [its] approach which is
called the force offluid resistance, unquestionably proportional to both
the square of the velocity and the square of the sine of the obliquity
of the incidence, as is known. And so, for this very reason, that
force becomes unnoticeable in rather narrow oblong conduits; for in
these, on account of FM being almost parallel to hH itself, which is
the direction of the motion of the fluid when it reaches Ff, just as at
any other place Nn where the direction Vu is almost parallel to NL
and nl, and the sine of incidence can be regarded as negligible. In a
conduit made up of cylindrical pipes that sine is absolutely zero,
because the direction of the fluid is everywhere parallel to the sides
of the cylinders throughout the total length of the conduit. Besides,
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 405

another accessory cause which can confound the effect arising from
the pressure 7T is found in an exceedingly curved conduit, in which of
course swiftly flowing liquid acquires a centrifugal force (which we
discussed elsewhere); this centrifugal force would render a greater
pressure 7T than is correct in the convex part of the conduit, but a
lesser one in the concave part of the same. And therefore, if one
would like to conduct an experiment with the aid of a tube to be
implanted in this conduit, the insertion should be made neither in the
convexity nor in the concavity of the curve, but at the side, such that
the tube extends out from the conduit perpendicularly to the plane of
convexity and concavity, and therefore, if that plane be not hori-
zontal, the tube as well as the structure is to be turned until the tube
attains a vertical position.

GENERAL COROLLARIES ABOUT VELOCITIES AND

PRESSURES

SECTION XVI

In the uniform and constant efflux of liquid, the general equation

(§VI) is changed, on account of dv = o, to this:

vv( aahh - {3{3ww) _ h 2aaghha

2<X<X h
- g a, or, vv = --c-:----=--=---·
aahh - (3(3ww

Hence this elegant Theorem is deduced: If there are two vessels or

conduits having equal vertical heights and equal uppermost and
lowest areas, no matter what shapes the remaining portions of the
conduits may have and however different [they may be] from each
other, if only their centric lines be so comparative that the ratio be-
tween a and /3 in one be the same as between a and /3 in the other
vessel or conduit, I say that from either (understandably both being
full) the liquid, after the efflux will have reached equilibrium, will
flow out on either hand at the same velocity. This surely is evident
from the very value of vv itself, which is h:aag~; , inasmuch as
<X<X - WW

herein the intermediate areas y are not obtained at all. Let there be,
for example, two vessels of which the shapes are ABCD and EFGH
(Fig. 12), of which the centric lines are straight, and indeed it matters
not whether they be vertical or oblique, or one more or less oblique
than the other, since in all of these cases on either hand it will always
 HYDRAULICS, PART II

be that a = {3, provided only that those two vessels have equal
vertical heights a, likewise equal extreme areas AD = EH= h and
BC = FG = w, or, which suffices, only that AD : EH = BC: FG, and

B
FIGURE 12

that those vessels be constantly full of water or some other homo-

genous liquid in which on either hand there is, of course, the same
natural gravity g animating the sections; the maximum and uniform
velocity of the water flowing out through BC will be equal to the
maximum and uniform velocity of the water flowing out through FG.
Indeed in such a case, on either hand (§VI ), on account of dv = o,
and a= {3, there results, I say, (hh - ww )z = hha, and thus
z = hh hha , conforming to Corollary 2, §VII, and such others
- WW

as we found in this way in the first Part for a cylindrical vessel.

Meanwhile, it is manifest that the value hh hha for a given height a
- WW

of each vessel is the same if the ratio between h and w is the same for
each, that is, if AD: EH= BC: FG. But it is to be noted that here
we avoid the contraction of the aqueous stream which is customarily
observed somewhere beyond the orifice, particularly in those vessels
which suddenly terminate in an orifice opened in a rather wide base,
contrary to what occurs in those having the shape of EFGH, con-
verging, so to speak, into a cylindrical pipe in which no noticeable
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 407

contraction of the stream is apparent. Meanwhile, if the under-

standing of this is also to be obtained, the vessel should be considered
as if continued to the maximum contraction of the stream where it
stops drawing together, and then the area of the vena contracta should
be taken as the smaller orifice w itself, and its distance from the
greatest area as the true vertical height.

SECTION XVII

With the same conditions existing, as in the preceding, of two vessels

of equal height, ABCD and EFGH (Fig. 12), [being] equally large at
the extremities and having straight centric lines, if now in addition to
this they [each] have a third area LM and NO mutually equal to
each other and equally distant from the orifices BC and FG, then not
only will the maximum velocity of each (by the preceding) be equal,
but also the pressures at LM and NO will be equal, and therefore in
tubes inserted in these places and bent to a vertical position, the
water in each will stand at the same height. The truth of this
is evident from the equation (§XII) which, on account of
a = f3 = ~' in the case of uniform velocity results in this simpler one,
vv(yy - ww)
- ~ - - ~ - gyt = TT, or (by writing 2gz for vv) in this,
qy
gz(yy - ww)
- - - - - - gyt = TT,
y
in which, because for either vessel z, y, w, g, and t are the same, the
same value of TT itself must result by all means, and hence also of
TT z(yy - ww) _ t = hha(yy - ww) _ t.
gy yy yy(hh - ww)

SECTION XVIII

Thus if all the LM and NO at equal vertical distances from BC and

FG were equal-which would occur if those two vessels ABCD and
EFGH were, for example, truncated conoids of the same type, one
right, the other scalene-in this case not only would the uniform
velocities at which the water would flow out from each vessel be equal
but also the pressures at the individual equal heights would be
identical, and thus even the suspensions of water standing in the tubes
would have the same height in each vessel.
 HYDRAULICS, PART II

SECTION XIX

The equation found in §VII for determining velocity in general,

whether it be already steady or not yet steady, gives

dx = aMhhw dz .
mxhha - aahhz + (3{3wwz

If this value is substituted in the equation for the pressures 7T (§XII)

and 2gz for vv, we will have:

gz(aayy- ww(ds) 2 /(dt) 2 ) (aahha - aahhz + (3(3wwz)gNy

aay + aa Mhh - gyt = 7T.

Thus the height of the liquid in the tube, or .:!..., is

gy

z(aayy - ww(ds) 2 /(dt)2) N(aahha - aahhz + (3{3wwz)

--'----'-'-------'----'--'--'--'- + -'------~--'--'---'- - t,
aayy aaMhh

which, therefore, for any determined z whatever expresses generally

the height of the liquid in the tube. Next, which is interesting, that
initial height is found at once, that is, that which would be observed
in the tube at the instant that the orifice below is opened and the
liquid is about to flow. Since indeed at the first instant of time
z = o, it will surely occur that (z having been deleted) .:!... = N,Ma - t.
gy

SECTION xx
Previous to this proof someone might have doubted whether or not
perhaps at that instant that the orifice BC is opened and before the
liquid pours forth into actual motion, whether or not, I say, the
pressures at any location LM were still the same, at least for an in-
stant, as they had been previously when the orifice BC had still been
closed or blocked off. Truly, the often-used and common hydro-
static law accepted everywhere shows, for the case of the vessel closed
at BC, that the liquid in a tube inserted somewhere in the circum-
ference of some section LM, and vertically erect, will stand suspended
at some height a - t, that is, at the same horizontal as the uppermost
surface AD of the liquid contained in the vessel. But now we see that
the situation develops in one way in the case of the orifice BC closed
and in another in the case of the same opened, even if the liquid is not
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 4?9

yet actually flowing out. Since N, as a part, is certainly less than M,

Na . Na
the whole, M will be less than a, and therefore as well M - t less than
a - t. Hence it is evident that in the first instant in which the ori-
fice BC is opened some of the liquid is already lost, so to speak, or,
rather, emits because of its own gravity: this does not contribute to
pressing the sides but to propelling the liquid to such an extent that it
no longer presses the sides of the vessel as greatly as it had done before
the aperture was formed. Meanwhile, surely, it should be indicated
here that I avoid the accessory causes which can alter the determined
height in the tube, ":,; t. For example, something should be men-
tioned concerning the shape of the vessel; for instance, if it were very
wide, and suddenly it converged into a narrow orifice, then un-
doubtedly our Theory could be out of tune with what experience
would show. The reason is that the Theory supposes that the sec-
tions Fm and Nl (Fig. ro) aspire to this arrangement with respect to
the motion (although they are not yet moved by the impulse) : that
[on the one hand] the areas Ff, Nn, etc., through the whole traverse
retain a horizontal position, extended throughout to the sides, and
[on the other] that the centric line GHI everywhere passes through
the midpoints G, H, V, and I, which certainly must thus prevail
accurately enough, as the Theory shows it, in those vessels and con-
duits the sides of which gradually, not suddenly, converge or diverge
toward the orifice below. But in others, being very large and ending
in a large base which contains a narrow orifice for an exit, in these, as
one would expect, the layers do not spread throughout entire cross
sections of the vessel, but rather to various [radial] extents, in propor-
tion as the quality of the liquid, not perfectly fluid but more or less
viscous, requires this, that it may experience the minimum possible
~esistance from friction, with the remaining part of the liquid near the
sides of the vessel certainly being at rest or without sufficient disposi-
tion to motion. From this it can happen that in such a vessel a
continuous throat or some sort of cataract is formed, such as Newton
fully conceived, although not according to that law which he indicated
as necessary. It is understood from the foregoing remarks that in
vessels of this sort it is not the external and artificial shape of them that
is to be considered, but that internal [shape] of the continuous throat
formed by nature, not such as Newton conceived, but that which is
best suited to the quality or the constitution of the liquid. There-
fore, if in the enclosing surface of this throat a tube could be im-
planted, the height of the liquid suspended in it would always be
410 HYDRAULICS, PART II

observed entirely as our rule postulates, whether the liquid is already

in motion or is beginning to be moved in the vessel.

SECTION XXI

But since it cannot be easily noticed by observation just when a

throat or cataract is terminated in a very wide vessel or just where its
enclosing surface is located, it will be safer if the tube penetrates
perpendicularly within the vessel as far as the middle, that is, right to
the centric line, and then the other part of the tube, bent upwards
outside the vessel, is vertical. In this way, then, the height of the
liquid in the tube shows uniquely how great is the compression of
the liquid at the place on the centric line with which the orifice of the
tube is in contact, where those accessory causes altering the effect of
the compression, concerning which we have treated above (§XV),
have no effect. Nevertheless, care is to be taken that at least the
portion of the tube to be projected into the liquid in the vessel be
slender enough, lest otherwise its very great thickness present some
impediment to the free motion of the liquid. These precautions
having been properly taken, I do not doubt that the following will be
most accurately observed: r, that the general height of the liquid in
a tube, with the acceleration of the effiux continuing to this time, will
be (§§XII & XIII ),

z(aayy - ww(ds ) 2 / (dt ) 2 ) Nw dz

-71' = --'-~----'----'----'---'-'- +- - - t.
gy aayy a dx '

2, but with the noticeable acceleration ceasing, at which time

certainly the velocity will have reached reasonable uniformity, there
2 2
will be!!.... = z(aayy - ww(ds ) / (dt ) ) - t; 3, that the initial height
gy aayy
will be (§ XIX) !!.... = N,Ma - t.
gy

APPLICATION OF OUR THEORY TO EXAMPLES OF

VESSELS AND CONDUITS ALWAYS FULL

SECTION XXII

Let the center line of some vessel be in a vertical position. For

this rather simple case, where a = f3 = r, it is true that (§VII)
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 41 I

hha
z = ---- (r - r /e< hh-ww)x /Mhhw), which is found from the reduc-
hh - WW
tion of the equation (hh - ww)z dx + Mhhw dz = hha dx developed
there and applied to this case, after, of course, ln e = r has been
agreed upon; and thus for x = oo, that is, in the case of steady
efflux or uniform velocity, there will result z = hh hha , from which
- WW

follows the Theorem already demonstrated in §XVI. Therefore,

let the vessel ABCD (Fig. r 3) be of any shape whatever, its centric or

FIGURE 13

vertical height be a, and let [the vessel] have attached to it the con-
duit CK composed of many pipes, for example, of three cylindrical
pipes CG, FI, and HK, placed in a horizontal position. Let the upper-
most area AD (to which the vessel with the pipes is assumed [to be]
always full) be h, the areas CE= m, FG = n, and HI= q, and the
orifice of the last pipe be w. I say that z will always be hh hha ,
- WW

and thus that neither the shape of the vessel nor the number of pipes
nor their sizes come into consideration, provided only that the first
h and the last w be given, and the height a of the vessel be given.
And also it is not of concern to know whether the throat or cataract
extends throughout the whole internal region of the vessel or only
occupies some portion of it around the centric line. The matter is
clear by §XV, because the conduit CK is assumed horizontal, and
thus the uniform velocity is always the same as if the orifice w were
attached immediately to CE by applying some perforated section to
the aperture CE.
COROLLARY. If w is very small with respect to h, there will result
z = a, and accordingly the velocity of the water flowing out
412 HYDRAULICS, PART II

uniformly, which is the greatest velocity that it can acquire, will be

equal to that which a heavy body acquires by falling from a height a.

SECTION XXIII
For finding the height of the liquid in a tube to be implanted some-
where in the horizontal conduit CK, let it be noted that in this case
t = o, since t signifies the excess of vertical height of the place where
the tube is inserted above the height of the orifice through which the
liquid is flowing, or, which is the same, t signifies the height of the
place of insertion of the tube above the place of efflux. Here, more-
over, on account of the horizontal position of the conduit, the centric
itself is also considered as horizontal, especially if its pipes, of which
the conduit is composed, are not at all large enough that (§XXII)
the sections of liquid flowing through them may become vertical.
Therefore, the height of the liquid in the tube (§XIX) will be
:!.... _ z(yy - ww) .N(hha - hhz + wwz) . _ R _ <!:._ In
gy - yy + Mhh ' smce a - t-' - d(
the case of uniform velocity, where (preceding paragraph) one has
z = hh hha , after this value has been substituted for z, the latter
- WW

.N(hha - hhz + wwz) . z(yy - ww)

term Mhh vamshes, and the former yy
hha(yy - ww)
becomes yy (hh - ) . Therefore, the height in the tube will be
WW

1T hha(yy - WW)
Let the results which we have already found in
gy yy(hh - ww) ·
the Appendix* at the end of the first Part be compared, although
in another way. Indeed, for the first pipe CG, where y = m, that
height will be hha(
mm 0; - ww/; for the second pipe FI, where y
- ww
is n, the height will be hhag;
nn -- ww/ ; for the third pipe
ww
HK, where

y = q, the height in the tube will be hha(~: - ww/, and so on for

qq - WW

however many pipes there may be composing the conduit. All of

this corresponds most accurately to the experiments performed
concerning this matter.
COROLLARY. The initial height in the tube is 1;_;, as we found above
for vessels themselves without conduits.
* Above, p. 390.
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 413

SECTION XXIV
And thus this is related to the height of the liquid in the tube to be
implanted somewhere in the vessel itself (and if it is desired, to be
extended right to the center line, which for the future we always
assume as a vertical straight line, but with the conduit attached to the
vessel, we assume the other [center line] as horizontal) ; let the place
of insertion be at some point in the indeterminate section LM, the
distance of which from the lowest horizontal is t, and the area of the
throat (if it is not LM itself), whatever it may be in the experiment,
should be taken as r . The height of the liquid in the tube (by §XI,
applied to this) will be obtained as
z(rr - ww) N(hha - hhz + wwz)
rr + Mhh - t,

where N
J
= r&contamed
. between BE and LM, and M J&r' but =

contained between BE and AD (§VII). For the velocity of the

liquid flowing out uniformly, where z = hh hha , let this value be
- WW

substituted for z, and the height in the tube will appear as

hha(rr - ww) - t (smce
. t h e term m
. w h.1c h .;,~r an dM appear vams
. h es ) ,
rr (hh - ww )
and thus, after the height t of the place of insertion has been added,
the total height, above the lowest horizontal BE, of the whole of the
. .d stan d.mg m
11qm . t h e tu b e w1·11 b e hha(rr
(hh - ww)) . Thus 1.f w b e 1n-
.
rr - ww
finitely small with respect to h and r, that total height will be a, that
is, the whole of the liquid in the tube is at the same level as the upper-
most area AD; that this must so occur we certainly should be able
to understand also from the fact that the liquid in the vessel is, so to
speak, at rest. Finally, the initial total height in the tube, with
z = o certainly, here is also 1;;. All these results are in excellent
accord with each other.

SECTION XXV
Now let us assume that the horizontal conduit CK converges in a
truncated cone, or any conoidic whatever, and has its greater base
directed toward the vessel. For uniform velocity of the outflowing
water, the height in the tube implanted at any place whatever be-
tween F and H (by designating the area FG = y ) will be, I say,
 HYDRAULICS, PART II

. hha(yy - ww)
that height (as expressed m §XXIII) (hh ) . Hence if the
JJ - WW
smaller base were attached to the vessel, and the oblong conduit were
not suddenly to diverge too greatly, lest the water diffuse in it, but the
sections succeeding in order follow the preceding ones as the cus-
tomarily established Theory supposes, then, on account ofy being less
than w, the pressure on the sides will be negative, and therefore it is
changed to suction, by which it occurs that, with the water descending
vertically in the tube and discharging into the water contained in the
vessel below, [the water in the tube] is raised through suction to
. hha(ww - yy)
the height (hh ). But if also w be greater than h, then the
YJ - WW
numerator and denominator of the fraction become negative, and
thus its value is again positive, which indicates that pressure is
present. Whereby, the vessel ABCD being always full, so that the
uppermost area AD would be less than the orifice of the divergent
conoidic, through the larger base of which the water emits, it is to be
observed again that the water will continually and without end rise
in the upwardly erected tube. In fact, in such a case the acceleration
of the flowing water never ceases, hence it never reaches a constancy
of velocity, which is evident from the general equation (from Art. VII
applied here) (hh - ww)z dx - Mhhw dz = hha dx; or, more clearly,
from the equation shown in Art. XXII in finite terms,
hha
Z = ..,....,---
hh - WW
(I
-
i / /hh-ww)x /Mhhw),

which is equivalent to z = hha (e<ww-hh)x /Mhhw - 1), from which

WW - hh
it is clear at once that in the case in which w is greater than h, it turns
out that z = oo, and therefore that the velocity is infinite when x is
infinite, contrary to what occurs if his greater than w.

CONCERNING THE SHORTNESS OF TIME FROM

THE BEGINNING OF EFFLUX RIGHT UP TO THE
ESSENTIALLY CONSTANT OR UNIFORM VELOCITY

SECTION XXVI
Although, accurately speaking, an infinite time is required before
the flow of water springing forth from vessels through orifices would
arrive gradually to perfect and geometric uniformity, experience
nevertheless shows daily that water, especially from rather wide
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 415

vessels, even though hardly three or four feet high, converges with
such rapidity from the first moment of flow to its maximum and con-
stant velocity of flow, while it flows through an admittedly somewhat
narrow orifice, that the gradual increments of velocity through which
it goes from rest to the uniform and maximum possible velocity
which it can essentially attain cannot be perceived by observation. In
order that we might interpret the reason for this phenomenon from
our Theory, let us consider a cylindrical or prismatic vessel of suffi-
ciently large area h and of suitable height a, from which water rushes
forth in a horizontal direction through a narrow orifice w which may
be formed either immediately beyond the vessel itself or [beyond] an
intermediate conduit having the orifice w at its extremity. However,
let us first consider, for the sake of brevity, that, of course, the orifice
w is formed near the bottom in the very wall of the vessel.

SECTION XXVII
The general equation (§VI) for the determination of the velocity
increasing to this point was this:

vv(hh - ww)
2h
+ hhwv dv
dx
Jc!!=
y
ha·
g '

which, in our case, where Jt = ~' and ww as compared to hlz can be

neglected, is changed to this:

2awv dv = 2gha dx - hvv dx, or dx = 2awv dv .

2gha - hvv

And thus the element of time d0, or dx, will be equal to

V

2aw dv 2aw dv/h aw ( dv dv )

2gha - hvv = 2ga - vv = hv2ga V + v2ga + -v + V2ga .
By integrating one has

0 = __!!!!!_ In ( v + V2ga )
hv2ga -v + v2ga
= (since v =
-
V 2gz) hv
aw
, 1- In
( vzV-+ -,/ay1-)
2ga - z + a
aw
= (§XXII)
hv2ga
X In (1 + 'Vr - r/ehx/aw) /(1 - VI - r/ehx/aw).
 HYDRAULICS, PART II

Indeed in this case z = a( 1 - I / e"xfaw). Hence

hv'2ga y1---- y1----

--- 8 = In (1 + I - 1 / ehx/aw) /( 1 - I - 1/ehx/aw).
aw
By transferring from logarithms to numbers and proceeding in the
customary way, one finds that ehxt 2aw is equal to the fraction

SECTION XXVIII
Certainly from the dynamic principle for the free fall of heavy
bodies, by letting C equal the height through which a freely falling

weight travels in the given time 8, one finds that 8 = J~C· Let this
value be substituted in the fraction found above, and it will result that
ehxt 2aw equals this other fraction (e 2 h../acJaw + 1 )/2eh../ac/aw.

Now, because the area h of the vessel is assumed to be much greater

than the area w of the orifice, and the height C of free fall to be traveled
through in one second is 15 feet, and in addition, from the nature of
the logarithmic curve, e is greater than two, it is manifest that for any
ordinary height a of the vessel, the number e2h../actaw is immensely
greater than unity, such that the latter can be disregarded in the
numerator of our fraction. Therefore there will be, essentially,
e"xf2aw = e2h../ac/aw ; 2 e1t...Jac/aw = ½e"../ac/aw,
or

.
or, b y ta k mg t h e I ogant
. h ms, -hx- + I n 2 = -hv'-
aC, firom w h'1ch
2aw aw
-
x = 2 v' aC -
2aw In 2
h = 2 v'- aC (on account of w being incomparably

less than h), which is equal to 2v'60 feet (by putting a = 4 feet and
C = 15 feet), or roughly 16 feet. Thus if, therefore, in the equation
z = a( 1 - 1 /ehx/aw) which determines the velocity for any effiux of
water of length x, we substitute 4 for a, 16 for x, 2 for e (although,
which would prove the case better, e is greater than 2), and ifwe con-
sider the area h of the vessel to be to the area w of the orifice as I oo
is to 1, we will have z = 4( 1 - 1 / 2 400 ) , which, on account of the
extremely small value of the fraction 1/ 2 400 , is reckoned to be not
different from four feet, which defines the height of the vessel, and
likewise that height from which a weight having fallen acquires a
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 417

velocity equal to that which the effiux has when it will have come to
uniformity; from this it is evident that after one second of time has
elapsed, the water flowing out already has essentially that uniform
velocity.
But in order that it be more evident with what promptness the
velocity of effiux may converge to uniformity, let us see how insig-
nificantly the velocity of the outflowing water acquired after one
tenth of a second has passed should be out of accord with the maxi-
mum velocity which it could acquire if the effiux were to endure for
an infinite length of time. Let us reduce feet to inches, and we will
have a = 48 inches, and C is found to be about 2 inches, from which
x or 2VaC is about 20 inches, and t!'xlaw = e2000148 , for which I write
merely 2 40 • Thus there will be z = a( 1 - 1 /2 40 ), which still, on
account of the imperceptible smallness of the fraction 1 /2 40 , is to be
considered as differing not at all from that very a defining the uniform
velocity.
COROLLARY. The effiux of water from rather wide vessels through
narrow orifices can safely be considered as constant the instant after
the beginhing of motion.

GENERAL HYDRAULICS THEOREM DEDUCED

DIRECTLY FROM THE PRINCIPLES OF
HYDRODYNAMICS, PROVED THROUGH THE
INDIRECT METHOD OF LIVE FORCES

SECTION XXIX
For a more substantial confirmation of the validity of our direct and
universal method, it is pleasing now to propose an indirect solution,
to be derived from the Theory of the conservation of live forces, of the
principal Proposition concerning the velocity of water fl.owing out of a
vessel or conduit which is always full, just as we established through
the equation shown in §VII.
Let us consider that the water fl.owing out through Cc (Fig. rn) is
directed immediately to a horizontal position, so that it can be re-
garded as without ascent and without descent in its displacement.
Let x actually be the length along the oblique direction ID of the
aqueous cylinder having Cc for a base, which cylinder contains as
great a quantity of water as has already flowed out. That quantity
will be wx, the differential w dx of which defines the elementary
a a
particle of water about to flow out from Cc immediately after the
 HYDRAULICS, PART II

quantity wx has been emitted. Let z be the vertical height from

(X

which some heavy body having fallen freely may acquire the sought
velocity which of course that elementary particle of water w dx must
(X

have. Through the principle of its live forces, the velocity will be
Vz and its subvelocity will be .:. Vz, from which the subvelocity at G is
(X

w~z; but the actual velocity itself in the tangential direction at G is

/3w-Vz Similarly the subvelocity at any point H is w-Vz, and

~ ~

therefore the actual velocity itself at His /3 dzjz.

ay t

SECTION XXX

But since, through the individual cross sections in the entire con-
· at t h e same mstant,
d mt · t h e same quantity
· - - o f water must fl ow,
wdx
(X

such an elementary quantity w dx is to be considered as stationed

(X

above the uppermost surface Ee at a vertical height equal to 13 !:;z in

order that, beginning to fall by its own gravity at a suitable time, it
might arrive at the area Ee, and there replace, at the same moment
. /3w-Vz h . l wdx d
an d at t h e same ve 1ocity --h-' t e uppermost partlc e - - e-
a ex
scending in the conduit, and in this way the conduit will be kept
continually full, as the statement of the Problem requires.

SECTION XXXI

Thus, if therefore each individual w dx has been set at the proper

(X

height /3::h~Z above Ee, and if all have fallen successively and are
about to enter through Ee, maintaining the conduit always full, it is
evident that an equal quantity of water Jw dx or wx, which we under-
a ex
stand to be moving in the extended plane Cc, must have flowed out
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 419

through the orifice Cc, and certainly in such a way that each of its
individual particles w dx has the proper acquired velocity
a
Vz. For

. l es -
t h at reason any one of t h ose partlc w- . to b e cons1'd ered as
dx 1s
a
having fallen from the original place of rest all the way to the lower-
. {3{3wwz {3{3wwz
most level, BCc extended, from the height aahh + AB = aahh + a.
Accordingly, it is necessary to multiply the descents by the descending
. l . d b . {3{3w 3 z dx aw dx h' h . d .
part1c es m or er too tam a 3 hh + -a-, w 1c mtegrate gives

~;: f z dx + a:x as the live force acquired from the universal de-
scent of the heavy particles. This, moreover, must be equal to its
own effect, which consists of the sum of the products which are formed
by multiplying the individual particles by the squares of their re-
spective velocities.

SECTION XXXII

For this reason I multiply the particle w dx, which has already
a
flowed out of the conduit, by the square of its velocity, which is z, and
I will have wz dx, the integral
a
~
a
fz dx of which expresses the live
force arising from the velocities of the entire aqueous matter flowing
out of the conduit; at this point one must add to this [the live force]
that all the matter flowing within the conduit has, and which is de-
termined by multiplying the individual layers y dt by the squares of
. . l . l . . wwz(ds)2 h r l d
t h eir respective u t1mate ve oc1t1es (d ) , so t at 10r any ayer y t,
aayy t 2
. r . . fi . . wwz(ds) 2 d wwz(ds) 2
t h e 11ve 1orce ansmg rom its own motion, (d ) 2 y t = d ,
aayy t aay t
is found. Therefore, the live force of all the layers to be considered
. con d mt
t h rough t h e entire . 1s
. -wwz- -
(X(X
f
(ds)
y t
2
d = (on account o f -d fy t
2
ds ,

which may be called M, having been given throughout the entire

. Mwwz
condmt) - - - . Therefore, by considering the sum of both these
(X(X

live forces arising from the motion, we will have the live force of the
.
entire aqueous systems equal to - WI Mwwz
z dx + ---·
a aa
420 HYDRAULICS, PART II

SECTION XXXIII

Thus, by equating this force, determined from the motion, with that
which we determined immediately above from the descent of the
particles, the following equation will be yielded to us:

f3/3w
-
3
-·
a.3 hh
f awx w
zdx+-=-·
a a
f Mwwz
zdx+ --,
aa

which, differentiated and freed of fractions, furnishes this:

f3{3wwz dx - aahhz dx = Mahhw dz - aahha dx,

or finally (after the reduction has been completed), this:

(aahh - f3f3ww)z dx + aMhhw dz = aahha dx,

just as we found through the direct method (§VII).

COROLLARY. If h or Ee be of very great size with respect tow or Cc,
the equation found reduces to this: az dx + Mw dz= aa dx; and for
uniform effiux to this: z = a. If, however, Ee indeed be not of very
great size with respect to Cc, let us assume, nevertheless, that new
water follows by continuously descending within the conduit, not
with some acquired velocity, but that it begins to follow from rest,

FIGURE 14
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 42 I

such that also in this manner the conduit is kept always full. This
should be applicable to the result if to Ee (Fig. 14) is attached a very
wide vessel, but of very small height, which is full of water. Cer-
tainly the water flows out therefrom taking its motion from rest, and
yet, flowing into the conduit with the required velocity, it will con-
tinually maintain the fullness of it. The very matter is evident from
the figure, where the conduit ECce has the very wide cylindrical vessel
AQVK attached, the height AK or QV of which is assumed very small,
so that essentially it does not increase the vertical height AB of the
conduit Ee, so that KB can be taken for AB, and still the volume AV
of this cylindrical vessel may enclose a very great amount of water.
For this reason, in order that the velocity of the water already emitted
through Cc may be determined, no longer Ee but KV is to be taken for
the first area h, with the vertical height a of the conduit nevertheless
being maintained, because according to the hypothesis AB does not
differ noticeably from KB. This having been agreed upon, we will
always have z = a for uniform velocity, that is, that velocity which a
weight would acquire by falling freely from the height a = AB or KB.

A SINGULAR EXAMPLE OF DETERMINING THE

MOTION OF WATER DESCENDING VERTICALLY IN
A CONOIDIC CONDUIT WHERE NOTHING FLOWS
OUT AND NO NEW WATER FOLLOWS THE
DESCENDING [WATER]

SECTION XXXIV

Let the Hyperbola BEG (Fig. 15) exist between orthogonal asymp-
totes, the one AM vertical, the other AH horizontal, the ordinates
DI, EK, FL, GM, etc. of which let define the very areas of the conoidic
conduit, continued to infinity, which is known to be generated if
another hyperbola described between the same asymptotes (the ordi-
nates of which are in proportion to the roots of the first ordinary
hyperbola) is revolved about the vertical asymptote just as about an
axis. Let it be understood that at some place in the conduit desig-
nated by the hyperbolic area DK a portion of water is furnished,
beginning to descend from rest, and that by descending it has arrived
at some other place FM, so that, as a consequence, FM = DK. The
velocities at GM, FL, etc., and the velocity at any other intermediate
section POop are sought.
422 HYDRAULICS, PART II

------------------,A
H

FIGURE 15

SECTION XXXV

Let every individual rectangle of coordinates, that is, the product of

each area PO by the altitude AO of the conoid, be aa; likewise let
the abscissas be given as AI= b, AK= c, and any descent AL of the
upper area be assumed as x. There will result, from the nature of
the hyperbola, the area DI = aba' EK = aa, FL = aa, and in addition
C X
a . Mb
(on account of DK= FM) AM= -b, from which GM= - , and the
ex
very area, or rather the solid DK or FM, is aa(ln c - ln b). Further
(with the calculus proving it), the distance of the center of gravity of
C - b
the area DK from the horizontal AH will be b and the
1nc - 1n
distance of the center of gravity of the area FM (note that I always
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 423

understand the solid to be represented by the area) from AH

. x(c - b)
will be b(ln c _ In b)' Thence let the descent of the center of gravity

from the position DK to the position FM, (:(I: :)~c

1
: ~), be multi-
plied by the quantity of descending water, defined by aa(In c - In b),
aa(x - b)(c - b)
and the product _,___---'--'---'- will be equal to the live force
b
produced from the descent.

SECTION XXXVI

Now if the velocity at GM be set equal to Vz, the velocity at FL will

be ~C Vz. Also, if any AO equaly, [then] PO will be aa, the section
y
Po will be aa dy' and its velocity will be precisely by Vz. Therefore,
y ex
the square of this applied to the layer Po gives aabbzy dy' equal to the
CCXX
live force of the layer Po, the properly corrected integral of which (by
. a, b, c, z, an d x as constants ) 1s
tak1ng . -aabbz
-yy - - aabbz
- = ('1n case y,
2CCXX 2cc
.a ~~ ~ .
or AO, 1s -b, or AM) ½aaz - - - = - (cc - bb), equal to the hve
2cc 2cc
force arising from the motion of the entire aqueous mass. By com-
paring this with the preceding, we will have

aa(x - b)(c - b) = aaz (cc_ bb),

b 2cc

.. from which, by reduction, there is found z = (~c ~ bi~c, equal to the

square of the velocity at GM; correspondingly, the square of the
velocity at FL is (x - btb, and the square of the velocity at any
c+
intermediate section Po whatever (pertaining to the abscissa AO or y) is
(x - b)2byy
(c + b)xx ·
COROLLARY 1. The lowermost layer EK, descending from rest to
the position GM, acquires a greater velocity than a heavy body falling
 HYDRAULICS, PART II

(x - b)2cc .
freely from the height KM. For (c + b)b 1s greater than KM or
ex
than b - b.
COROLLARY 2. But the uppermost layer DI, descending to the
position FL, acquires a lesser velocity than a weight falling freely
from the height IL. For (x - btb is less than IL or than x - b.
c+
COROLLARY 3. Therefore, the lower parts of the aqueous mass are
accelerated more vigorously and the upper parts more sluggishly than
if they were to descend freely, animated only by natural gravity.
This could also have been foreseen before the calculation from the
fact that the portions of water in the narrower places are pressed upon
by those lying above and so are incited to greater acceleration. But
on the other hand, they resist those portions which occupy the wider
places, and thus the upper [portions] are retarded in relation to their
own natural acceleration.
COROLLARY 4. Hence, somewhere an intermediate layer Po is
given which is neither incited nor retarded but which is accelerated in
the same manner as if it were to descend freely. In order that this
may be determined, I make AL to AO, or x toy, in proportion as
AI or b is to Aw, which will be by; and 11w will be the original position
X

of the layer Po. Thus wO, ory - by' is the height through which the
X
layer Po descends. Therefore, in order that the acceleration of this
layer be equal to the natural [acceleration], it is necessary only to
makey - by= (( - bityy; now this will yieldy = (c +/)x, which
X C + XX 2
shows that the distance AO is the arithmetic mean between AL and
AM, just as Aw is the arithmetic mean between AI and AK. And
similarly, LO = OM and Iw = wK. And so in these places the inter-
mediate layer Po is pressed downward by the water FO lying above
just as much as it is pressed upward by the water pM lying below,
such that it descends by no means differently than if it were to d e-
scend freely, animated by natural gravity alone. Furthermore, this
is also to be observed, that in these same places the compression of the
water becomes a maximum, from which we conclude that if at the
position of the section Po a tube were inserted vertically, the water in
it would ascend to a greater height from the place of insertion than if
it were inserted in a section at any other place between FL and GM.
For truly the height in the tube depends upon the pressure of the
water alone, as is evident from that explained above.
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 425

COMPARISON OF THIS SOLUTION THROUGH LIVE

FORCES WITH THAT WHICH IS SHOWN THROUGH
OUR DIRECT METHOD OBTAINED MERELY FROM
DYNAMIC PRINCIPLES

SECTION XXXVII

In the second Part of this hydraulics Paper, §VII,* we gave

through dynamic principles the most general equation for the deter-
mination of velocity, namely this:
(aahh - fJ{Jww)z dx + aMhhw dz = aahha dx.
But there it is supposed that the vessel or conduit is always full, indeed
with new liquid following continuously at the same velocity [and]
adjoining to that which is already descending at the uppermost area,
and the conduit itself is of a given and determined vertical height.
But since in the present example this conduit is of indefinite height,
and of course, only the portion FM is filled by liquid which always
changes its position, and accordingly has its height LM variable at any
instant, it may not immediately be seen that the case of this example
is included by that formula found in §VII, especially since no new
liquid here succeeds that descending, but its one and the same quan-
tity DK or FM is always retained in the conduit, occupying one place
after another.

SECTION XXXVIII
Meanwhile, we will show by what means, through some fiction of
the mind, the present case can be resolved according to the hypo-
thetical laws established in §VII. Certainly the original space
DEKI [Fig. 15] is to be considered as a conduit of given height IK,
the upper area of which is DI, the lower EK, both given and deter-
minate. Now while liquid flows from EK, about to occupy the lower
positions in the prolong~d conduit, I consider some fluid to be flowing
in through the uppermost area DI, free from gravity as well as from
all inertia or resistance, which, although this may not exist in the
nature of things, nevertheless can be assumed such that it does nothing
other than fill the space which would be left empty by the descending
liquid. Thus, this having been presupposed, the actual liquid will
have descended from DK to the position FM. I translate the forces
of these individual layers PO, hydrostatic as well as hydraulic, to the

* Page 396 above.

 HYDRAULICS, PART II

uppermost area DI, where the fictitious fluid which occupies the space
DFLI presses and [ where it] must maintain the same effect that the
force arising from the real liquid FM, similarly translated to the upper-
most area DI, maintains by gravity, and this is that force which I
called p or gha, [determined] by applying all these things to the
reasoning of our Theory explained in this hydraulics Paper. And
thus no forces come into account other than those which result from
gravity and the motion of the real fluid, the fictitious fluid contributing
absolutely nothing and serving no other purpose than to transmit the
translated forces [required] for expelling the real liquid FM.

SECTION XXXIX

Therefore nothing remains other than that the required application

be made of the method explained in §§IV and V, to the end that
it may be adapted to the proposed example, where it is immediately
evident that a = f3 = I. But here the remaining letters define those
things which follow: surely h = DI; w = EK; a, or the height of the
real liquid, LM; M equals the sums of all f; contained in the height
LM; further, z is the height from which a weight having fallen
freely acquires the velocity v with which, immediately after the
beginning of the descent, the real liquid or, afterwards, the fictitious
fluid flows out through EK; and, finally, dq is the instantaneous
displacement from the area EK.

SECTION XL

These things having been properly attended to, I now proceed as

follows: after the velocity through EK has been set equal to v, the
velocity through FL will be :v; the velocity through GM is ~v; the in-
stantaneous displacement through GM is d(AM) = cix; the dis-

placement through PO is d(AO) = y dx; the displacement through

X

EK, or dq, is c dx; all of these displacements, since they must be

X
simultaneous, are therefore in reciprocal proportion to the areas,
just as are the velocities themselves, and thus in direct proportion to
the distances from the horizontal AH. Now there is to be sought,
by repetition of §IV (since a = f3 = I), the hydrostatic force which
 METHOD FOR SOL YING HYDRAULICS PROBLEMS 42 7

is expressed through half of the uppermost area DI multiplied by

the difference of the squares of the maximum velocity through GM
aa
and the minimum velocity through FL. Moreover, DI= b' the

velocity through GM is ~v, and the velocity through FL is :v, from

which the total hydrostatic force is

aa ( ~ _ xxvv) = aaxxvv (cc _ bb).

2b bb cc 2b 3cc

SECTION XLI

By repetition of §V, I determine the hydraulic force in the

following manner. I multiply the accelerative force of the indefinite
layer Po, which I call y', by its displacement y dx, and I will have,
X

through the dynamical principle, yy dx = u' du' = yyv dv, from which
X CC

the progressive accelerative force y' = xyvddv, and the motive force
CC X

itself of the layer Po, that is, y' Po, is aaxv ~v dy' which, translated to
CC X

aa fc aaxv dv y dy h. h . d .
t h e area DI, or to b' orms bee dx , w 1c integrate gives
aaxv dvyy
(by rectifying it, or by applying it to all yy that are con-
2bcc dx
. d m
tame . t h e mterva
. 1 LM) aax3 v dv
b3 (ccd - bb) equa1 to t h e h y d rau 1·1c
2 CC X
force. The sum of the hydrostatic and hydraulic forces will be

aaxxvv(cc - bb) aax 3 v dv(cc

+ ___ __,____ ....:,'
- bb)
2b 3cc 2b 3cc dx
or
aaxx ( xv dv)
2b3cc vv + dx (cc - bb),

which sum of forces must be equal to the original translated force

about to arise from the gravity of the layers. Moreover, the original
translated force of any layer Po whatever is found by changing
g(Po) or ga; dy into ga: dy just as PO is to DI, or as AI is to AO, that is,
 HYDRAULICS, PART II

just as b is toy, which properly integrated through the interval LM

gives gaax~b - b) for the total basic translated force.

SECTION XLII

Thus we obtain the equation between the sum of the hydrostatic

and hydraulic forces and the original translated force, which equation
aaxx ( xv dv) gaax(c - b)
thus appears as bcc vv + dx (cc - bb ) = bb , from
2
which, by dividing by aha; (cc - bb), there results:

=-g__
- x ( vv xvdv)
+--
2bcc dx C + b'
or, by reducing:
2gbcc dx
xvv dx + XXV dv = c+
b ;

and so, by integrating and organizing in the proper manner (in order
that v itself equals o when AL, or x, equals AI orb ) there will result
.l 2gbccx - b,.2gbbcc, or b y wntmg,.. accor d'mg to t h e dynam1c .
2 xxvv =
c+
. 'd'mg b y g, t h ere w1·11 b e xxz = 2bccx - b2bbcc ,
Iaw, 2gz fior vv, an db y d 1v1
c+
and thus z = (x(b )
XX C
2
+
t;', which determines the velocity through EK,
from which now the velocity through any other area whatever is
. . (x - b)2bcc . (x - b)2cc
determined. And m fact, by changmg xx(c + b) mto b(c + b)

just as (GM) 2 is to (AM )2, that is as cc is to c~t, or just as bb is to xx,

this will be equal to the height from which a weight falling freely
acquires the velocity which the liquid has at the lowermost point M.
. (x - b)2bcc . (x - b)2b . .
F urt h er, b y c h angmg ( b) mto b JUSt as cc 1s to xx,
XX C + C +
[this result will be] equal to the height from which a weight having
fallen freely will have that velocity which pertains to the liquid at the
uppermost point L. And finally, by changing (x( b)+ :;c
2
into

(x ( b)
XX C +
2
:r XX C

just as cc is to yy, this will indicate the height from which

 METHOD FOR SOLVING HYDRAULICS PROBLEMS 429

a weight must fall freely in order to acquire the appropriate velocity

which the liquid will have at any given intermediate point 0. All
this agrees marvelously with those [results] which we found through
the Theory of live forces.

SECTION XLIII

On the other hand we can now show that which cannot be shown
as easily through live forces, that is, how to find how much the liquid
at any section may be pressed during descending. We certainly saw
above in Caroll. IV that the liquid, after it descends from its initial or
original position DK to the position FM, undergoes different pres-
sures at its individual sections PO, and that the maximum of these
occurs where PO cuts LM in half; truly, to determine its magnitude
[there] and to compare it with some given weight-much less in
other places inasmuch as PO certainly divides LM in some other
proportion-would be a matter of more profound involvement if
someone would wish to show this from the nature of live forces.
Through our direct method, explained in the chapter on pressures, it
is by no means difficult to obtain what is desired, even for this special
example.

SECTION XLIV

And so let it remain to be investigated by how large a force the

mass FM of the liquid is pressed at any place PO whatever, which
force to be sought here I will now call 1r. I showed in §§XI and
XII that, if only the portion PM, the remainder FO having been
removed, were to proceed to descend so that it was driven not only
by its own gravity, but in addition by 1r as well, it will (except for the
first instant) be accelerated in the same manner as the total mass FM
must be accelerated by its own gravity alone. Let IK be cut at win
a similar ratio as LM has been cut at O, so that AI: Aw = AL : AO;
1rw will be the initial position of P itself. And 1rK = PM. Let AL
be to AO, or AI to Aw, as I is ton, from which AO = nx and Aw = nb.
And let us imagine that the liquid contained in 1rK descends by the
force of its own gravity and, in addition, by the force 1r which pertains
to it at any position of descent, in order that the acceleration occur
just as if the total mass DK were to descend by the force alone of its
own gravity. Therefore, that which was AL, or x, now is AO, or
nx, and that which was AI, or b, now is nb. And so the hydrostatic
and hydraulic forces are found, if one writes nx for x, n dx for dx, and
430 HYDRAULICS, PART II

. aannxxvv
likewise nb for b; and thus there will result b ( cc - nnbb) or
2n 3 3 cc
aaxxvv .
-b (cc - nnbb), equal to the hydrostatic force, and also
2n 3 cc

aan 3 x 3 v dv aax3 v dv
b d (cc - nnbb) or b3 d (cc - nnbb),
2n 3 3 ccn x 2n CC X

equal to the hydraulic force. Moreover, the force of gravity arising

. h" h gaax(c - b) . d d .
t h rough t h e trans 1at10n, w ic was bb , now m ee is
gaanx ( gaax (
nnbb c - nb) or nbb c - nb).

SECTION XLV
Now, by taking the sum of the hydrostatic and hydraulic forces,
and by equating that to the original translated force arising from
gravity, g:ba; (c - nb), to which must be added the force of pressure

translated from PO to TTw, which is obtained if we change rr into x;

just as PO is to rrw, or as AI is to AL, that is, just as b is to x, we get this
equation:
3
aaxxvv aax v dv ) ( bb) = gaax (c _ nb) xrr
(- - + 2nb
2nb 3 cc
- -
3 -
cc dx
cc-nn nbb + b'
which, adjusted, assumes this form:
aa gaa dx
-bb [d(½xxvv)(cc - nnbb)] = -b- (c - nb)
2n cc n
+ rrdx.
Since, moreover, the velocity of the diminutive mass PM, which,
however, is pressed by the force rr, must be the same as the velocity of
the whole mass FM, for which velocity we found just above that
I
-xxvv = 2gbccx - b2gbbcc, let us wnte. t h e d"iuerentia
rr .
1 of t h"is, w h"1ch is
.
2 C +
2
gbcc :x, for d(½xxvv), and this will yield
c+
gaa dx gaa dx
nb(c + b) (cc - nnbb) = ---,;r- (c - nb) + rr dx.
By dividing by dx and transposing, we will obtain
gaa[cc - nnbb - (c - nb)(c + b)] gaa(cn - c - nnb + nb)
7T = nb(c + b) n(c + b)
 METHOD FOR SOL YING HYDRAULICS PROBLEMS 43 I

Therefore, if at PO a tube, which should be erected vertically, is

inserted in order that it may be known to what height the liquid can
stand suspended in it (at least for a rather short time), 7T is to be
divided by g(PO), that is, by gaa, in order to obtain
nx
7TnX x(cn - c - nnb + nb)
gaa c+ b '
the height of the liquid in the tube, from which height the absolute
compression of the liquid at PO is to be evaluated. Q.E.I.

SECTION XLVI

Thus, if it be desired further to determine the point O in any given

LM where the intensity of the pressure is greatest, that is, where the
liquid in the tube will obtain a maximum height, the derived quantity
x(cn - C - nnb + nb) 1s . to b e d·cr · d , wit
1uerent1ate · h n h av1ng
· b een
c+ b
assumed as variable and the rest as constants; this having been done,
there will appear c +b- 2bn = o, from which results n = c ~ b,
and thus

nx or AO = (c + b)x = c + b AL = !_AL + !..AL= !_AM+ !..AL.

2b 2b 2b 2 2 2

From this it is evident that the point O of maximum pressure is half

way between M and L, clearly just as we foretold by conjecture in
Coroll. IV above.

SECTION XLVII

In addition, if the derived value of n itself, which is c :b b, is substi-

. h .
tute d m t e express10n
x(cn - c - nnb
b
+ nb) , h . h . h
t e maximum e1g t
c+
itself of the liquid in the tube, :i~c-+bi;, will result. Thus, because

½LM or LO= x(c ~ b), LO or the height of the liquid in the conduit
2
above the point O where the tube is inserted will be to the height of
. .d. h b x(c - b) . x(c - b) 2 ( b) .
t h e 11qm mt e tu e as b 1s to b(c + b)' or as 2 c + 1s to
2 4
432 HYDRAULICS, PART II

c - b, or as double the sum of the original distances AK + AI is to the

simple difference of the same, that is, as 4Aw is to IK.

SECTION XLVIII

Finally, sound reason alone dictates that the mass of the descending
liquid FM undergoes no pressure at the extremities FL and GM, and
therefore that the height in a tube inserted either at Lor at M must be
zero. And this very fact is certainly confirmed by the formula, for in
x(cn - c - nnb + nb)
the prior case where n = 1 the formula - - - - - - - - - 1s
C + b
changed to this:

x(c - c - b + b)
= o;
C + b

but in the latter, where n = i, the same is changed into this:

x(cc - cb - cc + cb)
= o also.
b(c + b)

SECTION XLIX

ScHOLIUM 3. This example of a liquid descending by its own

gravity in an indefinitely continued hyperbolic pipe, which I treated
rather extensively as an example, shows in what manner one is to
proceed in other cases of this type where a liquid descends within a
sufficiently long conduit, with the identical quantity always, so that
certainly nothing flows out from it and, equally, no new liquid flows
in. Besides, it reveals the access to the solution of Problems concern-
ing the determination of oscillatory motion of fluids in bent or reflex
pipes, whatsoever may be their shapes, and of area varying in any
way whatever. For, in fact, in such pipes or siphons, while a certain
portion of the fluid descends through one leg, through the other leg,
albeit very dissimilar, another portion of the fluid equal to the former
ascends, that is, descends negatively, so that, just as in pipes continu-
ously inclined downward, the same mass of fluid is always contained
within. Thus, if, with the signs in the calculation having been
changed for those quantities which require it, one proceeds by the
same method which we showed in the example presented, by no
means will it be difficult to perform the computation of the velocity
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 433

of the fluid flowing through in the individual places for any ascent or
descent of it whatever; from this all the remaining [relationships]
are derived.

CONCERNING VESSELS WHICH, DURING THE TIME

WHEN THEY ARE EMITTING LIQUID THROUGH AN
APERTURE MADE IN OR NEAR THE BASE, RECEIVE
NO NEW LIQUID FLOWING IN FROM ABOVE

SECTION L

For those cases of vessels which by emitting liquid, but by accepting

no other, are finally depleted and evacuated, that method could be
applied which I explained in §XLIV and following, where [the
text] is concerned with the determination of the motion of a given
aqueous mass falling continuously within a hyperbolic-conoidic con-
duit, certainly with the aid of a fictitious fluid which does nothing
other than fill the space left empty by the descending fluid. But,
moreover, after that fiction of the mind has been considered and
ignored, in the present situation our equation suffices [which was]
given in §VII, (aahh - f3f3ww)z:, dx + aMhhw dz:, = aahha dx, which
we say to be valid for vessels always remaining full. Of course, that
equation will be accommodated easily to these vessels also which are
depleted slowly on account of no new liquid flowing in, and in which,
hence, the uppermost surface Ee (Fig. 10) continuously descends.

SECTION LI

In repetition of that which has already been solved in the first

Part, §XII,* for the case of a cylindrical vessel, let us consider here
a vessel of any shape whatever into which no new liquid is flowing
from above, while that which is already within is escaping by con-
tinuously passing through the final orifice. And so let us prescribe
that the uppermost surface, by descending from the position Ee, has
arrived at the position Ff, at which moment the velocity of the
flowing liquid is to be determined. For this purpose the area Ff
itself, or y, obviously variable, is to be taken as the uppermost area to
which the variable height BP or t corresponds (indeed the nature of
the translation of motive forces in fluids permits this, as will be evident

* Pages 363-64 above.

434 HYDRAULICS, PART II

to anyone who considers it); but M itself, or Jydsd2t , which was con-
stant in vessels constantly full, is now variable; certainly it is to be
applied throughout the variable height BP.
Therefore, in our equation let one write yy for hh, t for a, and ~
for f3, and the following equation, which is the one desired, will be
produced:

[aayy - ww ~!;~:]
Z dx + ayyw dz
2
J
~Jt = aayyt dx.

But from this, since the element of falling water w dx equals the de-
a
scending layer -y dt (I put -y dt because t decreases with increas-
mg x \1, dx ·11 b e --w-.
WI - ay dt Therewre,
c b y su b st1tutmg
. . t h.IS vaIue
for dx, the equation will have this appearance:

(ds) ]
2
ds J 2

[ aayy - ww (dt) 2 z dt - wwy dz y dt = aayyt dt;

for vessels having a vertical centric where a = I and ds = dt, this

transforms into the rather simple equation
dt
(yy - ww)z dt - wwy dz
Jy = yyt dt.
SECTION LII
From this equation we will now pursue to the end the example
shown in the first Part (§XII), where the vessel was assumed
cylindrical, and we will add certain worthwhile notes. Here, there-
fore, h is to be used, contrarily, in place of the constant y, and J~
will be J~' or ~' and the whole term wwy dz J~ will be wwt dz,
whence the equation for the case of the straight cylinder will have this
form:
(hh - ww)z dt - wwt dz = hht dt;
then, by the method of integrating used formerly by me, this gives in
finite terms the required value of z itself as

hht [I _ (t/a)<hh-2ww)/ww].
hh - 2WW
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 435

Here at once it is evident that the velocity, whether initial or final,

is zero, that is, in the case in which t = o, as well as [in that] in
which t = a. Hence it is deduced that somewhere the velocity of the
water will be a maximum while flowing out of the vessel or while
descending in the vessel itself. In order that this [ maximum velocity]
be determined, the maximum z is to be found which would result
when the surface of the water in the cylinder descends to that distance
WW )WW/(hh-2ww)
t from the base which is a ( hh _ ww ; this may be found
in two ways: namely, either by differentiating the ascertained value
of a itself in the usual way, or, as is easier, by setting the second term
wwt dz in the preceding differential equation, (hh - ww)z dt -
wwt dz = hht dt, equal to zero, from which there results z = hh hht ,
- WW

which, compared with the ascertained general value

hht
hh - 2ww
[i _(~a)(hh-2ww)/ww],
will give, as I said,
WW ) ww/(hh - 2ww)
t=a ----
( hh - WW
·

SECTION LIII

Meanwhile, in the very special case where hh = 2ww, where the

area of the cylindrical vessel is to the area of the orifice as V2 is to 1,
this inconvenience occurs, that ~ - - - 1 - -
hht [ (t
)(hh -2ww)/ww]
hh - 2ww a
becomes hht ( 1 - 1), and indeed that the other,
0

WW )WW/(hh-2ww) (ww)WW/0 00
a ( hh , becomes - , or 1 ,
- WW WW

from either of which nothing can be concluded. However, this in-

convenience is taken up through the rule, applied with some dex-
terity, which I communicated some time ago to the illustrious
L'H6pital, as is to be seen in the Analysis of infinitesimals, Art. r 63. *
hht [ (t)<hh-2ww)tww]
First, then, the recently introduced hh _ 2ww I - a
is found for the present case as -2(a - t) ln (a - t), but the other,
* See No. LXXI, p. 401, [Opera Omnia] Book I.
 HYDRAULICS, PART II

WW ) ww/(hh - 2ww) I
a ( hh _ ww , as~' by assuming, naturally, a as unity and
In e = I; and so here either quantity, the subtangent of which,
a = I, is determined very easily through common Logarithms.

CONCERNING THE MAKING OF CLEPSYDRAS

SECTION LIV
Until now we treated the shapes of vessels passing water through an
orifice below exclusively as if the shapes had been given, so that we
might surely bring out the laws according to which the motion of the
water would proceed. But now I should like to inquire in reverse
order into the shape of the vessel required in order that the uppermost
surface of the water might subside according to some proposed law;
for example, that it might be lowered at uniform speed, whence from
the magnitude of the descent the duration of flow is known imme-
diately; this very frequent use of Clepsydras was instituted long ago
among the Ancients for measuring time. Moreover, this can be
obtained principally in two ways; one certainly is from the quantity
of water having flowed out, the other from the quantity of water still
remaining in the vessel; from either, judgment can be passed concern-
ing the interval of time. Let us treat each one separately.

SECTION LV
Let us consider the rather simple vessels which indeed have their
centrics vertical, and of which the equation (§LI) is this:
dt
(yy - ww)z dt - wwy dz
Jy = yyt dt,
where the unknown t andy have their origin at the lowest point, or at
the orifice w. Thus, if now we wish to consider that the water flows
out at a uniform velocity, z is to be put equal to the constant c, which,
this having been done, will give dz = o; and so the second term
wwy dz Jyvanishes, and the others, divided by dt, will give this
algebraic equation:
(yy - ww)c = yyt,

from which it 1s found that yy = cw'::._ct and y = jcw'::._ct or

v-y -- :J wwc C - f
0
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 437

COROLLARY 1. The nature and shape of the vessel are therefore

such that it can be generated from the revolution of a plane biquad-
ratic Hyperbola described between two asymptotes, of which one is
vertical, the other horizontal [and] distant from the orifice w by a
height C.
COROLLARY 2. Assume that the first height or the initial t equals c,
and that the initial areay is infinite; then indeed one has

y = Jwwc
c-t
= J wwc
c-c
= 00.

COROLLARY 3. The quantity of water remaining in the vessel for

f
any height t, or y dt, is found by integrating and properly organizing

Jjdt wwc
C - t
= 2wc - 2w'Vcc - ct. Thus the total quantity of water
from the beginning of flow is 2wc, that is, equal to the cylinder of
water of which the base is w and the height is 2c.
COROLLARY 4. Therefore, if a cylindrical receptacle of capacity
not less than 2wc is placed under the water about to flow out of the
orifice w, the water that has flowed out and has been collected in that
receptacle ascends equally in equal times, and so, after the height of
the receptacle has been divided into equal graduations, the Clepsydra
will result.

SECTION LVI

ScHOLIUM 4. It should not be concealed that this type of Clepsydra

can hardly have any use in practice on account of the immense
height which should be given to the hyperbolic vessel in order that the
efflux can endure through a period of time, even if it be rather short;
this can be understood satisfactorily from the fact that if the vessel
were 15 feet high, that is, if c = 15 feet, it would contain water
2wc = 30w, that is, an aqueous column the height of which is 30 feet
above the base w. Since, moreover, from the orifice w the water
flows out at the uniform velocity required for the height of 15 feet,
and since the time of fall through this height is less than one second,
it is certainly evident that within one second the aqueous cylinder of
area w and of length twice fifteen, or thirty, feet flows out of the
vessel, and therefore in such a short time the whole vessel will be
emptied. Let me say nothing yet about the impossibility of the
structure of the vessel, since one supposes that its uppermost area is
infinite; I would certainly like to correct for this inconvenience by
prolonging-the vessel near enough to the upper horizontal asymptote
 HYDRAULICS, PART II

that it may acquire an area much greater than that of the orifice w,
which area then could be enclosed by a lip rising up to the asymptote.

SECTION LVII

The Clepsydra can be made in a more convenient manner by means

of constant descent of the uppermost aqueous layer in the vessel itself;
a suitable shape of vessel is therefore to be investigated in order that,
with water escaping through the orifice w, the surface of the water
remaining in the vessel may descend equally in equal times, so that,
in a given interval of time, the water surface passes through a given
number of equally spaced divisions into which the vertical axis of the
vessel has been divided. But there are two cases in which the in-
tended [effect] can be obtained: namely, either the area w of the
opening is so small that it has no sensible ratio to the area y in the
vessel; or the area w is large enough that it is comparable with any y
whatever. We will now treat the first case, since it is easier and more
useful in practice; the other will be treated later.

SECTION LVIII

It is certainly evident that our equation

ydt = yyt dt
(yy - ww)z dt - wwy dz
J
is reduced, in the case in which w is very small in proportion toy, to
precisely these two terms: yyz dt = yyt dt, from which z = t; that is,
in any vessel whatever having an orifice w, water certainly emanates
from it at that velocity which is acquired by a heavy body falling from
the height t which the residual water in the vessel has; of the truth of
this matter (found scientifically by us) hydraulics Writers of previous
years had knowledge only through experiments. However, once this
had been supposed, it was then easy to discover the nature of that
conoidic vessel having the orifice w as if it were its vertex, facing
downward, because then it has this effect, that at any moment the
uppermost surface y of the residual water descends at a uniform or
constant speed. For since the velocities of the fluid flowing in the
same quantity and at the same time through two different areas are
· rec1proca
m · 1 rat10
· o f t h e areas, one must set y / w = v, z r;wvz
y ' and
there will result ~ Vz designating the speed of the descending surface
y
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 439

Y, which speed, since it must be constant, let be set as ~

y
-Vz = Ve,
from which there will be 0'Y = z = t. Moreover, in straight conoids
WW

the areas are nothing else but circles, of which the areas y are as the
squares of the radii, or of the ordinates of the generating curve which
by its revolution about the axis describes the desired conoid. There-
fore, by designating the ordinate in the generating curve as s, and
the radius of the circular orifice w as b, and likewise by saying that
the area of the circle is to the square of the radius as n is to r, one will
have, certainly, y = nss, yy = nns4, w = nbb, and ww = nnb 4; and so
.
fior t h e equat10n t = -0'Y t h ere w1·11 resu1t t = -b
cs4 or -b4t = s4 ; t h'1s
4
WW C

shows the generating curve of the desired conoid to be a biquadratic

Parabola, at the lowest point of which, or, at the vertex facing
downward, the orifice w emitting the aqueous stream must be
fashioned.
COROLLARY. The parameter of the derived Parabola is b{/bJc,
where c is arbitrary; and therefore c can be assumed so small that
b{I bJc can be made as great as desired, to the end that the areas of the
conoid become incomparably greater than the area of the orifice w.
Accordingly, therefore, for the pleasure of it, the capacity of the vessel
can be made so large and the area of the orifice w can be made so
small that the efflux of water may persist for a very long time before
the vessel is completely exhausted; attention is paid to this especially
in the design of Clepsydras.

SECTION LIX
Now consider the other case, where the orifice w is not assumed of
so insensible an area that in the universal equation

ydt
(yy - ww)z dt - wwy dz
J = yyt dt
those terms vanish in which w appears; it is surely necessary that all
2
terms remain, and then that 0'Y be substituted for z, and 0' dy for dz,
WW WW
to the end certainly that the uppermost surface of the water in the
vessel descend at uniform speed, which speed is due to the arbitrary
height c, whence the resulting equation to be solved is

ydt
(yy - ww )c dt - 2cww dy
J = wwt dt.
44° HYDRAULICS, PART II

There are certain indications from which I suspected at once that

there exists a certain algebraic curve which agrees with that equation
taken in the abstract. Thus after a short investigation, this appeared
directly to me:
ww (t + 3c )
yy = C ;

and for the generating curve, this:

which again is a biquadratic Parabola, but with abscissas t which take

their origin not from its very vertex, but indeed from below the same,
on the axis, at a distance 3c. Meanwhile, the derived equation
yy = ww(t + 3c), which certainly satisfies in the abstract, cannot for
C

that reason be accepted in this instance because it does not fulfill a

tacit condition, which condition consists in this: that for t = o, the
entire equation must go to zero, and therefore even Jt must vanish,

which does not occur in the case of the equation yy = ww (t + 3c).

C
Of course, it is nothing new that a certain proposition which is true
in general is not always correct in particular, especially when it is
required to satisfy additional conditions to which it is not necessary
to attend when taking the matter in general. Moreover, the true
generative curve for the desired shape of the conoidic vessel is ascer-
tained if by this art the equation

ydt
(yy - ww)c dt - 2cww dy
J = wwt dt

can be universally solved so that y is determined through t, or vice

versa, t throughy, either that it be done in finite algebraic terms or in
exponentials, or, indeed, through quadratures. But I leave this
matter, since it is not important here, to be resolved by others to whom
time is available.

SECTION LX

ScHOLIUM 5. An opportune occasion is now given for exammmg

the Newtonian cataract, which that Author describes in the second
edition of Principia Mathematica Philosophiae Naturalis, Prop. 36,
Book II, pp. 303 ff. There at once it is to be observed that that form
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 441

which Newton gave to his cataract ABNFEM (see Fig. 16, which is
Newton's from that selected place) certainly is the same as that which

p
,Q

I :L
K - ---------- ----- .
A~----1f,.,.H,.,.---711B

··········o·····-· N

FIGURE 16

I found above (§LV) for the shape of the Clepsydra which emits water
through the orifice w at a constant speed. Let us note now what
Newton himself recognizes, that in such a cataract any layer MN
whatever descends with that velocity which it would acquire if it
would fall freely from a given point I through the height JO, ani-
mated by no force other than its own natural gravity; from this it
follows certainly that the layers remain in contact with each other in
descending, but nevertheless such that they exert no force on each
other, either by pressing or by resisting, just as if individual particles
were descending by their own weight. And thus that pressure con-
cerning which I treated above in a particular section will be null
through the whole Newtonian cataract, and accordingly, concerning
the force of pressure which I called Tr, not even the least will be exerted
on the sides AME, BNF, which also will be evident from my very
formula which I gave for TT in §XII. If indeed that is applied to the
present case, it is obtained, as I said, that the value of Tr itself is null
through the whole height of the cataract. What therefore must be
concluded from this? Undoubtedly the following: that if the sides
AME and BNF of the cataract were rigid, resembling those of a
funnel, with which Newton compares it, and if at any place whatever,
an orifice having been made, a tube erected to a vertical position
were inserted, none of the flowing water would enter from the
cataract into the tube and ascend, as would happen if the sides were
pressed by the water flowing through. Meanwhile, the sides AME
44 2 HYDRAULICS, PART II

and BNF are pressed inwardly against the axis HG by the weight of the
water standing in AMEC and BNFD, through the common hydro-
static law which shows that pressures experienced at the individual
places M and N are proportional to the heights HO. Moreover,
since the sides of the cataract are not rigid, and those pressures of the
standing water have no pressures opposed to them from the water
flowing through, it is of course necessary that the water which is
considered to stand and which presses continuously obtain its own
effect, that is, that it pour into the cataract and mix together with the
flowing water itself. Therefore, the shape of the cataract will be
destroyed and will be thrown into disorder, and the water descends
differently than according to our explanation.
Therefore the Newtonian explanation, since it is adverse to the laws
of hydrostatics, cannot stand.

EPIMETRUM: CONCERNING THE FORCE BY WHICH

A VESSEL IS PUSHED BACKWARD WHEN WATER
FLOWS OUT OF IT IN A HORIZONTAL DIRECTION

It is a very well-received truth that Action is equal to reaction, that is,

that any external force whatever which acts on a certain body or any
other obstacle also acts backwards in the directly opposed direction on
whatever barrier it may have or find, just as we see it happen when,
for example, a stretched elastic placed between two bodies begins to
be relaxed at once upon release of the restraint; it propels each body
by equal forces, one forward, the other backward. We observe that
it does not occur otherwise when an iron shot is fired with great
violence from a cannon by the ignition of gunpowder, that similarly
the cannon itself is driven back in the opposite direction, however by
a much lesser impetus on account of the huge mass of the cannon with
respect to that of the shot itself.
In an equal manner, when we see that water is expelled from a
vessel in a horizontal direction, whether it occurs immediately
through an orifice opened in the side of the vessel or intermediately
through some conduit having a horizontal position, we must conclude
that some force or power exists which may produce this expulsion,
which then acts equally strongly in pressing backward at the wall of
the vessel directly opposed to the final orifice of the conduit. There-
fore, this effect is to be considered in order that the magnitude or
measure of that force driving backwards and forwards may be prop-
erly estimated. At once several ideas come to mind which at first
glance seem to reduce to the desired result, but, since the various ideas
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 443

show measures of that quantity differing from each other, 1t 1s un-

certain whether one or another of them is correct, or, rather, whether
or not all of them are out of tune with the truth. And, therefore, the
path which we shall follow will be safest and most certain ifwe deduce
the solution immediately from the very principles confirmed in this
hydraulics Theory of ours.
To this end let us consider what was stated at the beginning con-
cerning the translation of the motive forces of the individual aqueous
sections Fm, Nl, etc. (Fig. ro), to the common uppermost area Ee,
where it is shown that those forces, thus collected into one, furnish
the same effect in expelling the liquid, assumed free of gravity,
through the aperture Cc, and this at the same velocity at all times as
that which occurs in the natural manner through the gravitation of
the liquid descending without translation. Now indeed we clearly
understand that that translation to the uppermost area Ee is arbitrary,
inasmuch as it can easily be gathered from the same hydrostatics
principle which we assumed that the motive forces of the individual
layers can be transferred to any other assumed area whatever, for
example to Ff, which may be considered as given or constant; and it
is not so that one may say that the forces translated to that [area] act
only by pressing the layers below but not those above. For while the
continuous layers are thus joined to each other so that the one without
the other cannot be moved from a [particular] place, with no difficulty
at all do we understand that the propelled aqueous mass below
FCif carries, so to speak, that above, EFfe, along with it as well, and
that with that [mass] the effect of the pressing potential at Ff must be
shared to such a degree that the liquid will flow out through Cc at no
other velocity than as if the motive forces of the layers had been
translated to the uppermost area Ee.
With these things properly understood, let us now consider that the
forces of the layers are translated not to the uppermost area Ee, nor
to some intermediate one Ff, but to the lowermost Cc, so that the
resultant force from rhe collection of all the forces of the layers, while
it acts immediately only on the lowermost layer, nevertheless must
drive the entire aqueous mass ECce into motion precisely by the same
law as if the motive forces had been translated to the uppermost
surface Ee, or to any other whatever, Ff. Moreover, from the
common hydrostatic principle, the sum of the motive forces translated
to any area whatever is proportional to the area itself, just by making
gha, or the weight of the aqueous cylinder the base of which is Ee
and the vertical height is AB, into gwa, or the weight of the aqueous
cylinder which has the same height a or AB and the base w or Cc, just
444 HYDRAULICS, PART II

as Ee is to Cc, or h is to w; let this weight be designated p, and thus we

have the quantity of pressure p by which the water is expelled through
the orifice Cc, and which, in pushing forward, at the same time is
pushing backward in a common direction.
Let an application be made: Let the vessel ABD (Fig. 13) having
any shape whatever be attached to the not very large oblong conduit
CEK lying in a horizontal position. For the sake of brevity, let the
centric of this vessel be vertical, and it is not important what shape
the conduit may have, whether it be a truncated conoid or composed
of many pipes, of which [conduit] the extreme orifice at K through
which water is expelled equals w. Now when the vessel has been
assumed constantly full, we saw above that for the determination of
the velocity v of the water flowing out, this equation results:
vv(hh - ww) hMwv dv _ h
2h + dx - g a,

if, of course, the motive forces are translated to the uppermost area h,
or to AD; whence if we translate the same to the lowermost area w,
. vvw(hh - ww) Mwwv dv
there will result hh
2
+ dx = gwa = p, [ equal] to
the weight of the aqueous cylinder the base of which is w and the
height is a, the height of the water above the horizontal BK. There-
fore, imagine the small surface Bb in the wall BA of the vessel from
the region opposite and equal [in area] to the orifice K; this Bb will
endure a similar pressure on itself, equal top, from the retroaction
expelling the fluid.
Certainly this retropressure p exerted at Bb rises to its full intensity
immediately as it begins to be moved; that is, from the first instant of
motion when the velocity is as yet infinitely small right up until it
reaches the level of equilibrium, that force of retropressure is con-
stantly the same. Certainly whatever v is, there always results
vvw(hh - ww) Mwwv dv . .
gwa or p = hh
2
+ dx , whence at the begmnmg of
fl ow, w h en v =
. 1n
o or 1s . 1y sma11, one w1·11 h ave p
. fi n1te =
Mwwv
dx dv ;
but when v will have arrived at uniformity, so that dv = o, then,
vvw(hh - ww)
similarly, p = hh . From this it follows that
2

Mwwvdv vvw(hh - ww)

dx 2hh '
assuming, certainly, v for the initial velocity in the first term, but v for
 METHOD FOR SOLVING HYDRAULICS PROBLEMS 445

the uniform velocity in the other. Further, it is also evident from

here that our retropressure is equal to the weight itself of an aqueous
cylinder having a base w and that height which the water has in the
vessel above the horizontal BK, because a very well-known hydro-
static principle shows that the area CE through which the water enters
into the conduit CK is pressed by the weight of the aqueous column
lying above, of which the base is CE itself and the height is that of the
uppermost surface AD above EC. And so with the area CE having
been set equal to m and the height of the water in the vessel equal to a,
the pressure which drives the water in the conduit from CE against K
will be gma; with gma having been made into gwa in proportion as the
area CE is to the area K of the aperture, that is, as m is to w, the force
gwa will be, through the same hydrostatics principle, that very
[force] which drives the water to eruption through the orifice w;
therefore, the force directly opposed to this, arising from the reaction
and action at Bb, will likewise be gwa = p. Q.E.D.
COROLLARY. It turns out that in vessels constantly full the force of
retropressure is invariable from the first instant of efflux right up to
the constant velocity, inasmuch as it is always gwa. Therefore, they
err who state that it is variable: namely, less in slower efflux, greater
in faster, a maximum in constant efflux.
 HYDRAULICS PROBLEM

AB is a cylindrical pipe, uniformry wide and open at both ends A and B

[Fig. 17]. Let it be immersed perpendicularry anywhere in an infinite fluid the
surface of which is LR, and let the immersed part of the pipe be BC. Let it
be understood that the whole pipe AB is full of the same fluid, so that with the
thumb held over it at A nothing can.flow out. Now there is sought, iffreedom is
given to the descending fluid by the thumb having been removed, how far it will
descend below LR, and afterwards how far it will ascend again above LR; that
is, P having been taken for the termination of the descent and Ofor the termina-
tion of the subsequent ascent, the length GP and as well the length CO are sought.
A
El
1-· 0
M. I
L/:t,··- iN ~1
-_ -- --_- -- - - - - -
-- ~ _-_-_-_-_-_

.. p
F
B
FIGURE 17

SOLUTION

First it is to be considered that the portion of the fluid contained in

CB has no weight, or rather, that its attempt to descend developing
from gravity is eliminated by the opposing pressure of the fluid
surrounding the pipe; and thus at once some of the fluid while de-
scending from the projecting part AC descends into the submerged
 HYDRAULICS PROBLEM 447
part of the pipe; further, the former is to be considered to be separated
from its own gravity so that the total accelerative force at any time
depends only on the weight of that portion of the fluid which exists as
yet in the projecting part of the pipe. Thus at once it is evident that
the descent of the fluid must be accelerated as long as some remains in
AC and hence that the maximum velocity will occur when it will have
descended to the horizontal LR; that then again it must be retarded
on account of the prevalent pressure of the ambient fluid; that, when
the velocity has been completely eliminated at P, the fluid is com-
pelled to ascend again to O by the external ambient pressure, and
that it will have a maximum velocity again at C. Then from 0, by
another change, it falls, but not to as great a depth as previously, at
once about to ascend upward again, and thus it will continue oscilla-
ting successively. And so are sought the first descent to P and the
first subsequent ascent to 0, inasmuch as upon these all the remaining
depend.
As far as the determination of the descent is concerned, it is achieved
in a two-fold way, either directly from the principles of mechanics or
indirectly from the nature of live forces. Accordingly, I accomplish
the task by each method.
FIRST METHOD. The principles of mechanics show that the instan-
taneous increment of velocity is obtained if the element of time is
multiplied by the accelerative force; moreover, the element of time is
obtained by dividing the element of space to be traversed by the ac-
quired velocity. Thus let the total length of the pipe AB be a, the
immersed part CB be I, any arbitrary part AE passed through by the
descending fluid be x, and the natural gravity by which, certainly,
bodies are naturally animated or urged to descent be g. Let the
velocity acquired by the fluid having fallen from A to Ebe designated
by v. With these [terms] so defined, the weight of the fluid filling up
the whole pipe will be ga; the weight of the part CB, g(r) = g; and
the weight of the part CE, g(CE) = ga - g - gx. Now, because equal
to the weight of the portion CB is eliminated by the equivalent pres-
sure of the ambient fluid, only the weight of the fluid CE remains,
which must accelerate all the remaining fluid EB into descending.
Thus the accelerative force will be g(BCE'E) = ga - g - gx, whence
a-x
ga - g - gx (dx)
a _ x '"v = dv; accordingly

v dv = ga dx - g dx - gx dx = g dx _ g dx ,
a-x a-x
 HYDRAULICS

and by integration,

½vv = gx + g ln (a - x) - g ln a.

(Note: I add g ln a here for the sake of correction, so that certainly for
x equal to o, where the velocity is null, the value derived for it
vanishes. ) And thus, in order that it be known how far the liquid
must descend in the pipe so that v again becomes o, it is necessary
to make x + ln (a - x) - ln a = o, and, by taking AP as the root of
this equation, the point P will be the terminus at which the descent is
ended.
Indeed the root x is obtained with the help of a Logarithmic
[curve] in this way: let the Logarithmic be HCC [Fig. 18], the sub-
tangent of which equals r = BC, the immersed portion of the pipe.

L B
FIGURE 18

In BC extended, let BA be taken equal to the length of the entire pipe,

up to which [lengthJ certainly at the beginning it is full of
liquid; [the line] AH, parallel to the asymptote BL, is drawn
from A, meeting the Logarithmic at H, and from this point H
[the line] HG, which cuts the curve at G, is drawn parallel to the
tangent at C; whence, further, GP is drawn parallel to the asymptote
BL; the point P will be the terminus to which the fluid descends and
at which the descent will be completed. The proof is simple. Indeed
because AH= ln AB= ln a, and PG= -ln PB= -ln (AB - AP)
= - ln (a - x), and AH+ PG= AP, there will beln a - ln (a - x)
= x, from which x + ln (a - x) - ln a = o. This is the very same
equation [which is] to be developed. Q.E.D.
 HYDRAULICS PROBLEM 449
SECOND METHOD. From the Theory of live forces I find the same in
the following way, by indicating here again (Fig. 17) AB by a, CB
by r, and the indeterminate portion AE of the descent by x. Let z
be a certain vertical height through which a weight falling freely would
acquire a velocity equal to that which the fluid acquires after it has
fallen in the pipe from A to E. This velocity will therefore be Vz.
Moreover, by cutting off BF = AE, it is clear that through the de-
scent of the fluid from A to E the column AF is translated to EB and
that then its individual parts have a velocity v1z; wherefore all the
fluid substance constituting this column EB will have a live force
(a - X) Vz-VZ = az - xz , to which are to be added the particular
collectively taken live forces of the individual particles contained in
the space BF, which, with the fluid descending from A to E, have
flowed out successively from the orifice B, of which certainly the sum
f
is ,<; dx. Therefore, the magnitude of the live force of the entire
f
substance then in motion is az - xz + z dx.
Because, indeed, this live force must be the effect of the gravity of
the portion of the fluid lying above the horizontal LR, therefore it is
f
necessary that that quantity of the live force az - xz + z dx, as the
effect, be equal to its own adequate cause, that is, to the sum of the
products which are formed by multiplying the individual particles of
descending fluid by the appropriate height of each through which any
one of them is depressed by its own gravity. Therefore, CD = AE
having been assumed, DC will be the descent of the column AD, and
therefore AD(DC) expresses the live force arising from the gravity of
the substance of the fluid contained in the column AD after it de-
scends to the position EC, where the lower portions that will already
have fallen below the horizontal are beginning to be freed of their
own gravity. This, moreover, relates to the particles of fluid con-
tained in DC. Individual [particles] make individual descents before
they arrive at the horizontal LR according to their different distances
from LR; since these distances are expressed through the indeter-
minate x, and any particle pertaining to it through dx, the product
x dx will be the live force arising from gravity of any particle what-
f
ever, and thus x dx or ½xx is the live force of all the particles con-
tained in CD, taken collectively, which, therefore, having been added
to AD (DC), that is, to (a - r - x)x, or to ax - x - xx, will become
the total live force arising from gravity, ax - x - ½xx, consequently
equal to the derived live force gathered from the motion, az - xz +
f z dx. And thus, by differentiating, there is a dx - dx - x dx =
dx
a dz - x dz, from which dz = dx - - - · integrating again,
a - x'
45° HYDRAULICS

z = x + In (a - x) - ln a. But z is proportional to the square of

the velocity sought, vv. Therefore vv, or, if you prefer, ½vv, equals
x + In (a - x) - ln a, wholly as we found above by the direct
method, because certainly the square of the velocity of the fluid de-
scending through the height x must be proportional to the same
x + In (a - x) - ln a.
Now further, in order that one may determine the point O [Fig. 17]
to which, after the descent has been completed, it ascends again, it is
to be noticed first of all that the liquid will have ascended right up to
the uppermost point A if the same particles which had gone out of the
orifice B during the descent, their own acquired velocities having
been retained, were now again to enter individually in reverse order
through the orifice B, and thus any one of them together with the
preceding would be thrust on upward by the pressure of the ambient
fluid, just as it occurs in every type of oscillation which, when the
resistance has been removed, always does the same thing going and
returning. But since the particles of fluid flowing out fall within the
ambient liquid immediately after discharge and then diffuse from
here, it is manifest that those same particles no longer flow into the
pipe at their own acquired velocities, but other [particles] in their
place, which surround the orifice without motion, are driven upward
successively with the preceding column of fluid; these particles there-
fore, since as yet they have no motion at the beginning of the entering,
cause the column of fluid preceding together with those [particles]
attaching themselves to ascend more slowly than it would do if the
particles were to flow in with some velocity which would aid the
ascent. Hence the liquid cannot ascend to A but to some lower
point O. In order that the height CO of this point O above the hori-
zontal may be known, let us understand, meanwhile, that the par-
ticles which have fallen from the orifice B into the ambient liquid
within which they have no gravity are directed upward, and there-
fore, with no loss in their velocity having occurred, arrive at the
horizontal LR, from which afterward the individual [particles], with
their gravity recovered, spring upward as much as they can according
to the velocity of each, occupying in order a position according to
some curve LMN. Indeed now because the amount of live force
must be conserved, it is necessary that the sum of the products of the
individual particles contained in CO and their respective ascents above
the horizontal LR, together with the sum of the products which are
formed as well by multiplying the individual particles on the curve
LMN by their ascents and distances from LR, be equal to the sum of
the similar products of the entire column AC; certainly the liquid
 HYDRAULICS PROBLEM 45 1
possessed this at the beginning of descent. Therefore, if from the
sum of the products of the column AC there is subtracted the sum of
the products of the particles LMN, there will remain the sum of the
products in the column CO, and hence the height CO itself becomes
known. For the prior sum is ½AC(AC) = ½(AC) 2 = ½aa - a + ½;
the other is f
z dx = ax - x - ½xx - az + xz = (in the case in
which x, or the indeterminate AE, becomes AP, where z vanishes)
ax - x - ½xx, and therefore ½aa - a + ½- ax + x + ½xx is the
sum of the products in CO, or ½(CO)2; from this,

CO = V aa - 2a + I - 2ax + 2x + xx = I +x - a
= (in this same case where x = AP) GP.
From this it is evident that the fluid in the pipe ascends to a height
CO above the horizontal LR [which is] as great as the depth GP
to which it had descended below LR immediately before.
COROLLARY. With CO known, one can find, through the construc-
tion given above, the second descent and the subsequent second
ascent; and then from this the third descent and ascent; and so on.
In this way the extents of the individual oscillations can be
determined.
 INDEX
FOR
Hydrodynamica and Hydraulica

Acceleration (in contractions), 37off Cardano, Girolamo, 2 1 3

Acta Eruditorum, 5 Castelli, Benedetto, 2
Adhesion of water to pipe, 69, 72, Cataract, 1 1 o
155 ff Newtonian, 440 ff
Aereo-aethereal particles, 20 Catherine I, ix
Air, Celerity of sound, 253
homogeneous, height of, 252 Chazelles, Jean Mathieu de, 337
rare, 264 Chimneys, 55
Amontons, Guillaume, 229, 261 Clairaut, Alexis Claude, 70
Angle of refraction, 24 7 Clepsydra, 165, 436 ff
Aqueducts, 6 Cohesion of particles, 96
Archimedes, 1, g, 206 Cochlea, g
Ascent, Conduit composed of several pipes,
actual, 203 356 ff
potential, 13, 35 ff, 104, 124 ff Conservation of live forces, 13, 42
loss of, 56, 159 ff, 188 Contraction of a stream, 71 ff, 82,
see also: hindrances, impediments go, 92,406
Attraction, mutual, 21, 96 Cramer, Gabriel, 224
Ctesibius, 192
Barometer, 230 ff
Bernoulli, Daniel, 13, 26, 59, 266,
Delisle, Joseph Nicolas, 241
279, 280, 309, 314, 326, 330,
Density of compressed air, 229
35 1 Descartes, Rene du Perron, 275
Jakob, 5, 20
Descent, actual, 35 ff, 104, 127, 141,
Johann, 5, 8, 12, 128, 341
150, 307
Nicolaus, 341
Dilation of a stream, 82, 84, 95
Bilfinger, Georg Bernhard, 31
Discharge,
Bladders, 6, 23 ff
from a cylindrical pipe, 86
Borelli, Giovanni Alfonso, 5
from a dilating pipe, 53, 65, 84,
Boyle, Robert, 5
89, 96, 299
Calandrini, Giovanni Ludovico, 224 Discontinuity [vaporization] of liq-
Capillary effect, 31 ff, 109 uid in a vertical pipe, 54, 303
454 INDEX

Displacements, isochronous, r 34 Galilei, Galileo, 5, r 2

Duhamel, Jean Baptiste, 236, 326 s'Gravesande, Wilhelm Jacob van,
50, 64
Emanuel, Prince of Portugal, 3 I I Gravity,
Establishment of flow, 76 ff, 96 ff center of, 12, 20, ·130
Euler, Leonhard, ix, 173 ff, 414ff component of, 386
mean specific, 237
Fabretti, Rafaello, 4 Guglielmini, Domenico, 2, 4
Fahrenheit, Gabriel Daniel, 252
Hales, Stephen, 261
Feuillee, Louis, 241
Hawkbee, Francis, 24 7
Flow, within an arbitrary conduit,
Height, barometric, 230 ff
391 ff, a conoidic conduit,
Hermann, Jacob, 64, 75, 97
r 25, 373, 389, 42 r, an ob-
Hindrances to fluid motion, 42, 53,
liquely oriented conduit, I 27,
56,156,160,201,355
376 Hire, Philippe de la, 58, 70
Fluids,
Huygens, Christiaan, 12, 275, 276,
elastic, 29, 34, 226
heterogeneous, I 66 ff 281, 324, 351
Hydraulico-statics, 7, 10, 289 ff
within moving vessels, 283 ff
Hydraulics, 351
Fontana, Carlo, 55, 2 I 7
Hydrodynamics, r, 351,417
Force,
Hydrostatics, 35 I
accelerative, 353
hydraulic, 395 Impediments to motion, 42, 63
hydrostatic, 395 Impetus, 5, 69, 70, 102, I 15, I 19,
live, 129, 139, 149, 258, 351, 187, 202, 217, 314 ff
425 Impulse, 220
actual, 258 [Impulse-momentum law], 328
conservation of, 13, 42
contained in a compressed Jallabert, Louis, 225
elastic body, 257 Jet, 184, 293 ff, 312 ff
[Jet propulsion,
loss of, 139 ff, 149
potential, 258 ofa ship], 331 ff
of any vessel], 442
motive, 353, 354
immaterial, 353, 401 Kepler, Johann, 275
of fluid resistance, 404 Koenig, Emanuel, 24 I
of ignited gunpowder, 9, 264 ff Laborers, day, 185 ff, 335 ff
of repulsion, 314, 3 15 Leibniz, Gottfried Wilhelm von, ix
static, 395 l'Hopital, Guillaume Frarn;oisde, viii
Fountain, leaping, 7, 45, 47, 58, 70,
I 14, 122, 296, 301 Machines, hydraulic, 183 ff
Francini, 2 I 7 Maclaurin, Colin, ix
Freezing, I 5 Mariotte, Edme, 2, 3, 8, 9, 16, 33,
Friction, 42,45, 114,164, 182,201,326
in pipes, 42, 50, 63, 66 Maupertuis, Pierre Louis Moreau
mechanical, 203 ff de, 70
Fron tin us, Sextus Julius, I, 4, 53 Moment of a force, 210
 INDEX 455
Momentum, 317 (see also: force of Shape of vessel, effect on discharge,
fluid resistance, 404) 4o5
Motion, internal, 15, 72 Siphon, 119, 303, 339
Smoke, 55
Newton, Isaac, 2, 5, 8, 71, 76, 129,
Specific gravity, mean, 237
248, 314, 319, 330, 4o9, 44° Spiess, Otto, xiv, xv
Orifice size in proportion to pipe Straub, Hans, xiv, xv
size, 46 Thermometer, air-mercury, 230 ff
Oscillations, Throat (gurges), 352
center of, 12 Time of depletion, 77 ff, 363
in a curved prismatic tube, 128 ff Time of establishment of flow, 76 ff,
in a straight conical tube, 158, 96 ff, I 15 ff
in a straight prismatic tube, Time, periodic,
I 79 ff, 3 77 ff, 446 ff of a particle in a vortex, 277
isochronous, 8, 132 of a planet, 282
Papin, Denis, 5 Torricelli, Evangelista, 2, 5
Particles, ultimate heavy, 281 [Trajectory of a jet], 95 ff, 313
Pascal, Blaise, 5 Translation, 355
Pendulum, tautochronous, 133 ff Treadmill, 187, 204
Percussion, 287 Trireme (galley), 337 ff
Permanence, state of, 196, 275 Truesdell, Clifford Ambrose, viii,
Perrault, Claude, 194, 217 xiii
Phoronomia, 5
Vaporization of fluid in a vertical
Pipes, obliquely oriented, 49, 61 pipe, 303
Plane, inclined, 186, 203 Varignon, Pierre, 5
Poleni, Giovanni, 5, 71, 119 Velocity, height of fall correspond-
Potential,
ing to, 2, 4, 5, 7, 12, 41 ff,
absolute, 184
56 ff, 76, !02 ff, 142, 161,
animated, 217
289ff, 298, 315, 377, 385,
moving, 184
397,405,4 21
static or hydrostatic, 395 during depletion, 363 ff, 433 ff
Pressure, 22 ff
in a clepsydra, 436 ff
barometric, 237 ff in a vortex, 278
effect of moon on, 240 in composite (manifold) conduits,
in underground caverns, 239
50 ff, 64 ff, 165 ff, 368 ff,
of an aqueous stream, 328
385 ff
of flowing water, 29, 382, 401 tt for compressed air, 257 ff
Pumps, 187 ff Vena contracta, 407 (see also: contrac-
Refraction, angle of, 24 7 tion of a stream)
Ricatti, Jacopo Francesco, 320 Vessel,
Rive, Ami de la, 224 composite, 356 ff
Rowers (oarsmen), 337 manifold, 176 ff
submerged, 119, 139
Saulmon, 278 Vis viva, (see: force, live)
Scheuchzer, Johann Jacob, 240 viscosity, 42
 INDEX

Vitruvius Pollio, Marcus, 206 ff Waterwheel, 204

Viviani, Vincenzo, 5 Weidler,Johann Friedrich, 205,
Vlacq, Adriaan, 242 261
Vortex, 94, 275 ff Whorl, 92, 94
Windlass, 186
Waterscrew (cochlea) of Archi- Windmill, 221 ff
medes, 206 ff Wool, resilient, 274