R336 60< A Fawcntt Premier h^Beek

sSfe
m- ''-^m m \.:

Wi. - _ V , - 'c -"tSffiK^SSii ft.--' 1

-

\ ->.

•S:
mk Si.
«

an easy
introduction
to the
slide ruie

rij
'W***^'''*""

iiAto-Ai!w,vi«^VA',';*rrf^A<a"-.'?.^^.
p» X
o

P* —4

m
t •

en
en
o
rsj
PROPERTY of:
ALEX/INDER G. GONZALES
Other Fawcett Premier Books by
Isaac Asimov:

Realm of Algebra
Realm of Measure
Realm of Numbers
 iiiii|iiii|iMmiii|iiii|iiii|iiii|iMi|iiiij

3 It I Isd.3.c Asiixiov
j,'?ir.'.'i|i'.vi|lv.viv,l|iv.'i|i?,.|! .M|M
3 it
I I I I I I I |lllt|llll|llll|llll|llll|llll|llll|llll|IIM|llll|
I I I I I I

'
L '
it 3
4 3 2

8.9,1
1 1
ihll I III I I I I I I I I I I I I I I I I I IM 11 I I M I I I I I I I I I I I I

7
ilililililihlilililililililililil
(,'/'""
,',l|(|'.',l|l,'.',l
I'P'I
1

An
I
^-0
tllllllllllllllllllllllllllllllllllMllllill

Easy
III |llll|llll|llll|llll|llll|llll|llll|llllj

6 7 8 9 1
Introduction
3 TC
to
I IJ 1 1
1 1 j 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i 11 1 III 1

liiiiiliil
the
|i

iii|iiii|iiii|

3
TC
Slide
„1

III
Rule
iiiiiiiiiiii{iiiii

I
^0 with Diagrams by
nil llllllll.iilllllllllltllllllllllllllll

William Barss
III {illl|llli|llll|llll|llll|llll|llll|llll| A Fawcett Premier Book
i 7 8 9 1 Fawcett Publications, Inc.
Greenwich, Conn.
Member of American
3 % Book Publishers Council
V,NlVn'.'.Tr.T,','(IVlli'.t.VII
(I I i|lili|liil|iiil|llli|liii|llli|lill|
I
'

^ 3 it
1
1 1 1 1 1 1 1 I I I
j
I |iiii|iiii|iiii|iiii|iiii|iiii|iiii|iiii|ii ii|ii ii|
1 1 1 1 1 1

n
1

' '

5 4 3
5 4 3 2 1
III III ll I I I I I I I I I I I I I I I I I
• I I M I I I I I I I I I I II I I I I I I I I I I I I I I I 1

,
8 9 I 1
ilihlilililililililil
l| |i|l|i|l|i|i|i|ii|i |i|l|i|l|

9 1

I
10
nil iihiiiliiiiliiiiliiiiliiiiliiiiliiiil
A Fawcett Premier book reprinted by arrangement with
Houghton Mifflin Company, Boston. This book contains the
complete text of the original hardcover edition.
Copyright © 1965 by Isaac Asimov. All rights reserved including
the right to reproduce this book or parts thereof in any form.

Library of Congress catalog card number: 65:19294

PRINTING HISTORY
First Houghton Mifflin Co., printing, September 1965
Second printing, June 1966

First Fawcett Premier printing, March 1967

Published by Fawcett World Library,

67 West 44th Street, New York, N.Y. 10036
PRINTED IN THE UNITED STATES OF AMERICA
 Contents

1 Arithmetic by Distance 7

2 The Addition Rule 19

3 Logarithms 39

4 Logarithms on Wood 58

5 Muhiplication 82

6 Division 107

7 Powers 131

8 Roots 152

Index 159
To a friend

who has accompanied

and served me

loyally for

twenty years

My Slide Rule
 1

Arithmetic by Distance

To Begin With —
We have all heard, these days, of electronic computers.
These marvelous instruments, which came into use during
World War II, are capable of performing in a few
seconds work that might take years if all we could use
were pen and paper.
There are times when arithmetical problems come our
way and we might wish that we ourselves owned such a
computer to do the work for us. Such a situation would
have its disadvantages, however. Electronic computers
are bulky, expensive, complicated, and can be handled
only by people with special training.
Besides, electronic computers aren't at theii* best
when used for everyday problems. That would be like

trying to shoot a fly with naval artillery.

For a fly, an ordinary swatter is much better, and for
ordinary mathematical problems, we could best use a
really simple computer.

There happens to be a simple computer, just suitable

for everyday computations, that was invented about

350 years ago. It isn't electronic; there are no electric
currents involved. In fact, it is no more than a piece of
wood with some marks on it. It looks like a ruler except
that it has a middle piece that can slide back and forth,
so that it is called a slide rule.
1
 A slide rule doesn't seem as impressive as a giant
electronic computer, but it has many advantages. It is

small enough to put in your pocket, it need not cost

more than a couple of dollars, it can't go out of order, and,
best of all, it can solve almost any numerical problem
that you meet up with under ordinary circumstances.
To add to all that, it is simple to operate. If you know
grade-school arithmetic, you can use a slide rule, even
though you may not quite see why it works! If you have
taken some high-school mathematics also, the reason for
its success can be explained with very little trouble.
In this book, I shall start from the beginning and try
to show you both how and why it works.
To be sure, merely reading this book will not make you
an expert at handling the slide rule. For that you will

need practice and I am not including practice exercises

as part of this book.

However, once you thoroughly understand the prin-

ciples of the slide rule — once you know what you are
doing and why — then it will be simple to set yourself
problems. It will be simple for you to use the slide rule
on problems that arise from day to day.
As you practice you will become expert, and you will

be amazed to see how, by merely sliding one piece of

marked wood against another, computations that seem
very complicated can be completed in a few moments.
In fact, if the time comes when you have a job in
which numerical computations have to be carried
through frequently, you will want your slide rule with
you at all times. Without it, you would feel like a doctor

without his stethoscope or a painter without his brush.

Let's consider, then, how a piece of marked wood can

8
 .

help us in our calculations.

One Ruler

One of the earliest tools used by civilized men engaged

in fine and accurate work — as in making plans for build-

ing temples or tombs — must have been a piece of wood

or ivory or metal which had a straight edge. By sliding

a pencil, or any marking device, down this straight edge,

a straight line could be drawn.
You understand how important it is, in drawing up
any accurate plan, to produce perfectly straight lines,

but to do so some guide is absolutely necessary. Try

drawing straight lines freehand, that is, without using
some straight edge to guide you, and see what a sloppy
appearance it makes.
An instrument possessing a straight edge and nothing
more is called just that, a straightedge.
It is simple, however, to put the straightedge to
another use, and make it a means for determining the
length of a straight line.
For this purpose it is only necessary to make small
marks upon it — marks spaced a fixed difference apart.
Such a marked straightedge is a rule or a ruler.
The ruler with which we are most familiar has the
marks upon it spaced an inch apart, and is 12 inches long
all together. Since 12 inches make 1 foot, such a ruler is a
foot-rule.

The foot-rule has its inches subdivided into halves

and quarters, usually eighths as well, and sometimes
even into sixteenths ( see Figure 1 )

Such a ruler can be used to determine the length of a

 Figure I

line or to draw a line of a particular length, and we are

all familiar with its use.

Can we, however, use it for some other purpose than

measuring length? Suppose we think about that a bit . . .

When we say that 1 + 1 = 2, we are talking about

"pure numbers." The and the 2 in that sum do not
I's

represent one of anything, or two of anything. Still the

sum can be made to apply to numbers that are not pure
— to count objects.
For instance, 1 apple + 1 apple = 2 apples; 1 chair
+ 1 chair =2 chairs; 1 star + 1 star = 2 stars; and 1

inch + = 2 inches. We can say this not only for a

1 inch

sum like 1 + 1 = 2, but for any arithmetical problem

that involves addition or subtraction. We know that
72 + 28 = 100 and that 35 - 20 = 15, and we can be
sure, therefore, that 72 inches + 28 inches = 100 inches,

and that 35 inches — 20 inches = 15 inches.

In other words, if we add inches to inches, or subtract
inches from inches, the numerical portion of the answer
we get would be the same as the numerical portion we
would have obtained if we had used apples, chairs, stars,

or anything else. It would be the same, indeed, as

answers we would have gotten in using pure numbers.

10
 This means that if we could somehow use some device
to prove that 45 inches and 32 inches taken together
make up a length of 77 inches, then we can be sure that
45 + 32 = 77. The device we would use to put lengths
together in this manner would, in eflFect, do our addition
for us.

Let's take a very simple example. Look at a ruler ( or,

if you don't have one handy, look at Figure 1) and,

starting at the left edge, count off 2 inches with your
finger. Your finger is now pointing to the number 2.

From there, count off 3 more inches. Now your finger is

pointing to the 5. What you have done is shown yourself

that 2 inches + 3 inches = 5 inches, and you can deduce
from that that 2 + 3 = 5.
Try again. Count off 3 inches and, starting from the
place you reached, count off 5 more, and you will find
that 3 + 5 = 8. In the same way you can use an ordinary
foot-rule to prove to yourself that 4 + 7= 11, 1 + 8 = 9,
or even that 6 + 3 = 3 + 6, for both these sums come
out to 9.

Subtraction works just as well. Count off 10 inches

and you are at the number 10. If you want to subtract 3
inches from- that, count them backward, chopping each
inch off the total length. When you finish you find your-

self at the number 7. This means 10 inches — 3 inches =

7 inches, or 10 - 3 = 7.
What we are doing is manipulating lengths and
using those lengths to tell us something about numbers.
Lengths are not numbers, of course, but in certain ways
they follow the same rules that numbers do. Lengths
have properties that are analogous to those of numbers in

some ways. If we then use lengths to work out, or

11
 )

"compute," the answers to number problems, we are

using the foot rule as an analog computer.
An example of a very elaborate analog computer is

Univac, which is used by television studios to work out

the course of elections in progress and to predict the
winner. ( Univac stands for Universal Analog Computer.
Univac makes use of flashing electric currents that

alter the position of many thousands of tiny switches in

millionths of a second. The properties of the changing
switch positions are analogous to those of numbers, which
is why Univac can be used to solve problems. However,
an ordinary ruler, although much, much simpler than
Univac, can also be used as an analog computer, and
we've just done it. In fact, it is easier to determine that

2 + 3 = 5 on a foot-rule than it is on Univac.

Two Rulers

Next, let's see how we can make a ruler more useful

for this new purpose — not of measuring lengths, but of

helping us work out simple additions and swfa'actions.

In the first place, a 12-inch rule is inconvenient. It only
exists because 12 inches make 1 foot, but we are not inter-
ested in that just now. We are interested in arithmetic
and in arithmetic we base everything on the number 10.

Suppose you start with nothing (zero) and begin to :

count: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9. Counting the zero, there

are ten different symbols (or digits) for the first ten
numbers. The next number, 10, makes use of two digits,

with a 1 on the left. The 1 on the left is kept there in ;

the next few numbers while the digit on the right goes
through the same series over again: 10, 11, 12, 13, 14, 15, ;

16, 17, 18, 19. With a 2 on the left, we again go through ,

12
 the same series; then we do the same with a 3 on the left,
*
with a 4 on the left and so on.
If, therefore, we learn to work with the numbers from
to 9 to begin mth, it will be easy to apply the knowledge
to the higher numbers which, in a way, are merely repeti-
tions of the first set. Consequently, a 10-inch rule will
be suflBcient.

Let's use c»j«r 10-inch rule to deal with the addition

problem 2 + 3 = 5. To begin with, we count off 2 inches
from the left, but we don't really have to do any actual
counting. If we count off 1 inch from the left, we end at
the figure 1; if we count off 2, we end at 2; if we count

off 3, we end at 3 and so on. Therefore, for the first

number in the problem, 2 + 3 = 5, we move straight to 2,

without counting.
we next have to count off 3, but why
Starting at 2,
count it off? Since we are actually dealing with distances
rather than with numbers, why not measure the 3 inches
with another ruler?
Imagine a second ruler, with the inches marked off

at the bottom, rather than at the top. Imagine the second

ruler, edge to edge, on top of the first. The I's will match
on the two rulers, and so the 2's, the 3's, and so on
(Figure 2).
Slide the top ruler to the right until its left fend is over
the 2 on the bottom ruler. Now, you see, the top ruler

is measuring off the inches starting at the bottom 2

(Figure 3). If you look at the 3 on the top ruler, that

marks a distance of 3 inches and immediately iniderneath

te it is the 5 of the bottom ruler. This is the method of

• For a fairly detailed discussion of how our number system

works, you might refer to another book in this series: Realm of
Numbers
13
 )

showing that 2 +3 - 5 by measuring distances.*

If you keep the top ruler in the position shown in

Figure 3, with its left end over the bottom 2, then

you can see that 2+1 = 3, because the top 1 is over
the bottom 3, and, for similar reasons, 2 + 2 = 4; and
2 + 4 = 6. The single position of the ruler gives us the
answers to a whole family of sums. It is easy to see how
this system can be used to give answers to still other
sums: 3 + 5 = 8; 6 + 1 = 7, and so on.
When two rulers are placed edge to edge, with one
sliding along the other, the result is a slide rule. (You
can also think of it as two sticks, one slipping along
another. People who use a slide rule often, sometimes
call it a "slipstick," but this is considered slang.
A slide rule consisting merely of two ordinary rulers
is anything but convenient, however. The ordinary ruler
• It may seem silly to you roundabout a method for
to use so
finding out something as simple as 2 -f 3 = 5. Right now, how-
ever, we are only working out the technique. Things will re-
main very simple for quite a few pages, but before long we will
be tackling more difficult problems —
and solving them with
just as little trouble as 2 -|- 3 = 5 gives us now.

14

J
 gure 3

tapers to a narrow edge at the inch markings, in order

to make it easier to draw a straight Hne and measure its

length. Sometimes, the narrow edge even has a thin

metal strip down its length to make it still easier to

draw lines.

We don't use a slide rule, however, for drawing

straight lines. We want to design it, instead, in such a

way as to make it easy to slide one ruler against the

other. With thin edges, it is diflBcult to do this, and with
metal strips it is practically impossible. It would be
[better to use two blunt rulers that are of even thickness
all the way across, and that meet, therefore, at thick
edges. (In fact, slide rules are usually a quarter of an
inch thick.)

I
Even meeting at a thick edge has its problems. If the

edges are quite smooth, the top ruler can easily slide
sideways ofiF the bottom one. For that reason, the bottom
•uler is generally made with a groove down its length
md the top ruler has a small tongue of wood that fits
 ,

into the groove. Now the top ruler can sUde back and
forth easily, without any danger of slipping off sideways.
To be sure, the top ruler can still move upward easily.
And if the slide rule happens to be turned upside down
at any time, the top ruler, tongue and all, will fall out

of the groove.
To prevent tipping, the top ruler can be made with
a tongue of wood on top as well at on the bottom, and
the upper tongue can be made to fit into a third ruler
with a groove. What was the top ruler becomes a middle
ruler. The next step is to bolt the uppermost ruler toi

the lowermost ruler, holding them firmly in place, while

the now middle ruler can slide back and forth easily
between them (Figure 4). The middle ruler cannot
tip, and the slide rule cannot fall apart.

You might suppose that if you hold such a slide rulel

by one end and let it dangle, the middle ruler will simply
slip out. However, the fit is usually made tight enough j

so that friction will keep it in place. To move thatj

middle ruler, you actually have to push it.

Making the Rulers

Ordinarily, you could follow the directions I will give

you for using a slide rule by following them on an actual
instrument. That would serve as an additional guide to

that of the illustrations that will be presented.

For a while, however, we are going to concern our-
*
selves only with addition and subtraction, operations

which are not conducted on the slide rules that are

actually manufactured. ( This is for the very good reason

that addition and subtraction are too simple to require

a slide rule.)
16
 Figure 4

In fact, in describing the use of rulers in performing

addition and subtraction, I hesitate to call the device
a slide rule lest you confuse it with the slide rules actually
manufactured for use in other operations. Let us there-
fore call the addition-and-subtraction device an addition
*

rule and use only that term for the purpose.

In order to follow the workings of the addition rule as I

describe them, you may have to rely very heavily on the

illustrations.You cannot be guided by the use of an
actual addition rule, unless you choose to make one for
yourself. Fortunately, it is not difficult to make an addi-

tion rule.

To do this, get a sheet of flexible cardboard, of the

type used in the manufacture of folders, and cut out two
pieces ( A and B ) according to the measurements shown
in Figure 5.

Fold Piece A along the lines a andThe end pieces,

h.

which will overlap, are pasted together with mending

tape, so that we now have something that looks like a
flattened cylinder with a long rectangular hole in it.

17
 13'

.
1" 1"
I'A" I'/J"

2" PIECE A

I' I"
•/j" .-1
r

I' I"

2" PIECE B

14"

Figure 5

Figure 6
Piece B will fit into Piece A snugly and can be moved
back and forth within it. Piece A will now serve as the
top and bottom ruler, "bolted" together with tape, while
Piece B will be the middle ruler. Where Piece A and
the lower portion of Piece B meet mark off inches from
to 10 on both as shown in Figure 6.

you cannot find a piece of cardboard 14 inches

If

long, you can make a smaller addition rule by working

with measurements half the size of those indicated in

Figure 5 and then marking off the to 10 divisions in
half-inch units.
The result, in either case, will be an addition rule
which, however crude, will give you the feel of the
instrument.
18
 2

The Addition Rule

Beyond Ten

We have reached the point now where we should

use proper terms, and no longer speak of top rulers,
bottom rulers, and middle rulers. The part of the addi-
tion rule which slides back and forth is the slide. The
rest of it is the body.
We can identify the range of numbers on the slide

as the "S-numbers," and the range of numbers on the

body as the "B-numbers." If I speak of S-7 and B-5,
I mean the 7 marked on the slide and the 5 marked on
the body.
What was the left end of an ordinary ruler — an end
not marked by a number, as you can see if you look at

one — is now represented by the number on the addi-

tion rule.

Instead of saying, then, "Move the middle ruler to

the right until its left end is over the 5 on the bottom
ruler," I will say, much more simply, "Place S-0 over B-5."
In order to add 5 and 2 on the addition rule, the
directions would be: Move S-0 over B-5, and under-
neath S-2, you will find B-7 (Figure 7), indicating that
5 + 2 = 7.
By now, though, you have undoubtedly realized that
19
 Rgur* 7 S + 2-7

our addition rule will give us the answer to only the very
simplest addition problems. Suppose we wanted the
answer to 7 + 5, a problem which is still quite simple
and which we know has the answer 12. Yet, at first

glance, our addition rule can't help us.

If we place S-0 over B-7, then we would expect to
find the answer under S-5. The trouble is, however, that
S-5 is not making contact with the B-numbers at all

(Figure 8). Now what?

One solution is to make the body of the addition rule
longer, 20 units long, rather than 10. That would mean
we can mark off numbers from to 20. Let us remember,
however, that the numbers 10 and beyond are repetitions,
at least in their right-hand digits, of the numbers under
10. We can make this plainer if, in the numbers from
10 to 20, we place the digit to the left in parentheses,
so that we can concentrate on the digit to the right
(Figure 9).
Using this extended addition rule, we can add 7 and
5 by placing S-0 over B-7 and finding S-5 over B-(l)2.
This shows us (Figure 10) that 7 + 5 = 12.
The trouble is that the longer an addition rule is, the

1 a 3 4 S «> •
s
B
(> 1 'i
I 3 *1 S 4k 7 •

1
 — r — r r

9
1
\ — © 10
I 1

7 + 5=7
Figure 8

less convenient it is. No one would want to use one

in which the bottom half of the body stuck out like an
elephant's tusk. Is there any other way out?
To find one, look at Figure 10 again. Notice that S-0
is over B-7 and that S-(1)0over B-(l)7. Ignore the is

parentheses and you'll see that the situation has re-

peated itself. What's more, this would happen every
time. If S-0 were over B-5, then S-(1)0 would be over
B-( 1 )5; if S-0 were over B-2, then S-( 1 )0 would be over
B-(l)2, and so on.
There's no puzzle about why this should be. In
moving from S-0 to S-(1)0, we are adding 10 and we
know that 2+ 10= 12, 5 + 10= 15, 7 + 10 = 17 and so
on. The right-hand digit of 10 is 0, and when we add
10 to any number, the right-hand digit of that number
(to which we are adding zero) naturally remains un-
changed.
In that case, it doesn't really matter whether we place
S-0 or S-( 1 )0 over a B-number. In either case, the right-
hand digit of the answer will be the same, and we must
only remember to adjust the left digit. If we use S-0,
the left digit of the answer is and is omitted. If we use

— ——
Figure 9

1
oil
\

()|2
I

(1)3
\

(1)4 (l|«
\

HI* (II'
1

(11*
—— I

(II* (2|0

21
 Figure 10

I
(b

7 + 5=12

S-10, the left digit of the answer is 1.

For instance, let's go back to our ordinary 10-unit

addition rule and consider 7 +5 again. This time, in-
stead of putting S-0 over B-7, we place S-10 over B-7
(Figure 11).
If you compare Figures 10 and 11, you will see how
they duplicate each other. In Figure 10, on the 20-unit
addition rule, S-4 is over B-( 1 ) 1, S-6 is over B-( 1 )3, and
S-8 is over B-(l)5, while in Figure 11, on the 10-unit
addition rule, S-4 is over B-1, S-6 is over B-3 and S-8 is

over B-5.
On the 10-unit addition rule, if we place S-10 over
some B-number, we need only add the left-hand digit 1

to our answer to get what we would have obtained with

the much less convenient 20-unit addition rule. In Figure
11, with S-10 over B-7, we find that S-5 is over B-2. We

add the left-hand 1 and we know that 7 + 5 = 12.

We can also tell by Figure 11 that 7 + 4= 11, 7 + 7 =
14, 7 + 9 = 16.

Thus, by making use of both S-0 and S-10, we can

carry our additions on the 10-unit addition rule up to a
sum of 20. And yet 20 isn't the highest possible sum
there is either. What if we wanted to get the answer to

17 + 14?
In order to tackle additions of this sort, let's take a
detour in what seems the opposite direction — numbers
smaller than 1 rather than larger than 20.
22
 '
4 k ( 1 1 (1)0
(j
p
1 i 1 I 1

> |l ii (mi) (1)3 (1)4 (1)5 (l)« (1)7 (l)t (2)0

~-

Ft actic>ns

So far, our addition rule only has digits on it, and we

have been dealing only with a few small whole numbers.
But you know that there are fractions — numbers that are
smaller than one, or that are intermediate in value
between neighboring whole numbers. In fact, if you
look at an ordinary ruler again, you will find fractions
there, for the inches are divided into halves, quarters,

and eighths, and often sixteenths, too.

Suppose we divide each unit distance on our addi-
tion rule into halves, quarters, and eighths (Figure 12).
YouVe undoubtedly had practice reading a ruler, so
that you will probably have no difficulty in reading a
particular mark in Figure 12 as 3/4, or 4%, or 7H. And,
of course, you will find the simple fraction % between
and 1.

Fractions work as well as whole numbers do on the

addition rule. Why shouldn't they? If we add VA inches

Figure II

I"
 10

Hf ++ h ' [i h i|t
[
i|t 4h|h4 H+M+ HHH* h4h+ t++ H+ H4H+ Ht +++

10

3'/2 4y, 7'/. Figure 12

and 2^/4 inches, we would get 3% inches, and that tells

us iy2 + 2H = 3%. Try it on the addition rule. Place

S-0 over B-VA and S-23^ is found to be over B'S%
(Figure 13).
By proper manipulation, you can find in the same
way that 2% + 47$ = 7%; that 5% + VA = TH, and so
on. (The addition rule should impress you a little more
now. It adds fractions as easily as it adds digits, whereas
with pen and paper, it is considerably harder to add frac-
tions than digits.)

For addition of fractions yielding sums over 10, we

use S-10. To add 5% and 7!4, for instance, we place S-10
over B-5% and find S-VA over B-2% (Figure 14). We
insert that left-hand digit 1, and find that 5% + 7H =
12%.
There is no question that this system of adding
fractions is very handy, but we can do still better. Re-
member that we changed the 12-inch ruler to a 10-inch
ruler because in our number system the right-hand digits
2'/4
Figure 13

4 5 6 7 8 9 10

+^ 4+ 4 +f+ f+j+H 4+ 4+ H* 4+ 444 4+ 4^ 4^ 4 4+ jJLi J.I.I. hi

I 9 10

l'/2 + 2«/4-^3y4
I'/J 3y4
 7'/4

I 2 3 8 ^ 10

s
ill iiliii Imliiilmlii "
I I III I

'lllllllll l ll H l
|
l ll|lll|ll l llil« TTTTTTTTTTTT^rnTTTTm^
|

2% - 5%
Figure 14
5H + 7l/4 = l2y,

repeat after intervals of 10 (see page 12). This makes

10 a particularly useful number in calculations of all sorts.
Following this system of emphasizing 10, we can divide
each unit on the addition rule into tenths, instead of
into halves, quarters, and eighths.

The result is shown in Figure 15. The %o mark (which

s equal to H made longer than the others so that it
) is

tands out. You may be less accustomed to this system

)f subdividing units, but you should have no trouble
kicking out SKo, or THo, or 2%o.
It may seem at first that we have simply changed one
;et of fractions for another and you may wonder what
ve have gained. But tenths, you see, are special fractions

because of our 10-based number system. With tenths,

kve can use decimal fractions. Thus IVio can be written

l.l; 3%o can be written 3.6; 9%o can be written 9.9, and

JO on. By using tenths, then, we have switched our ad-

dition rule to the decimal system.

We can use the addition rule as easily for decimals as
or ordinary fractions or for simple digits. If we want the
um of 1.7 and 1.4, we place S-0 over B-1.7 and find
5-1.4 over B-3.1 (Figure 16). We can tell, then, that
1.7 + 1.4 = 3.1. It is just as easy to discover that 2.3 +
r.l = 9.4, or that 5.8 + 1.6 = 7.4.

For sums over 10, we use S-10. If we place S-10 over

25
 )

4 S 7 t 9 10

mm mm mm mm m i n il n il n il \mm mm mm
: J : r. 10

2 l/IO 3 3/10 7 l/IO

Figure 15

B-8.3 and find S-5.6 over B-3.9 (Figure 17), we need

only add the left-hand digit 1 to conclude that 8.3 +
5.6 = 13.9. J

Moving the Decimdl Point

We are now ready to see a great advantage in the use

of decimals over ordinary fractions. The decimals make
it possible for us to handle additions that reach sums
well over 20.
We have already determined by means of the addition
rule that 1.7 + 1.4 = 3.1 (Figure 16). Actually, though,
we have discovered more than this. Changing the posi-
tion of the decimal point doesn't aflFect the actual digits
in the sum, provided the decimal point is moved in the

same way in each number involved in the addition. If

we moved each decimal point one place to the left, we

would have 0.17 + 0.14 = 0.31; and if we moved each
one place to the right, we would have 17 + 14 = 31.
(This answers the question I asked on page 21 as to

how the addition rule could be used to add 17 and 14.

Suppose we want to We have already

add 83 and 56.

determined (Figure 17) that 8.3 + 5.6=13.9. We can

tell at once, therefore, that 83 + 56 = 139.

You see, then, that the addition rule, subdivided in

decimal fashion, can be used to add, with one shift of
26
jii
 1.4

10

B
|"M|llti|m.| U k H i h i |i lii|i

*
h i|il i i

7
i i ii[iii i |ilii|ilii iiii
|

9
|
iiii

10
|
' iX^^^^^^^

I i I

1.7 3.1 Figure 16

1.7+1.4 = 3.1

the slide, any two-digit number to any other two-digit

number to give sums up to 200. This is better than merely
reaching 20, but 200 isn't the highest number there is,

either.

Suppose we divide each tenth on the scale into ten

still smaller subdivisions. We would have tenths of
tenths, or hundredths. We could then mark off numbers
like 5.23 and 1.81 and we could find that 5.23 + 1.81 =
7.04. By shifting decimal points mentally we would find

that 52.3 + 18.1 = 70.4 and that 523 + 181 = 704. We

could now add any three-digit number to any other three-
digit number and get sums up to 2000.

But there's a catch, and that is that there's a limit to

how finely you can divide your units. If you try to divide

the space between the units into hundreds, then you will
have tiny markings spaced a hundredth of an inch apart,
and these would be far too difficult to read without close
peering.
Actually, the smallest divisions into which an inch-

5.6 Ftqw 17

01 2 3 4S:6 7t 9 10

"' ''' " ' 'l l

'''| l l l l|l|ll[ Hl l |n l l [ll ll |ll lM ll L TjfmjTTTTJTTTTj-
| ^

9 10

3.9 8.3
8.3-1-5.6=13.9
 2 3 4 5 <k 7 8 9 M)

^ miliiii
MM
nil III!
flit Mil MM Nil Mil lilt nil lllilllll nil nil Mil nil iniliiiLnrNliiii
nil Mil Ml 1 mMiMV».inn Mil Mil
B
2 3 4 5 / 8 '•• 9 1 )

Figure 18
^ H^
6-..
8 8.1, 8.1
8.12 8.2

wide unit can be divided conveniently are 0.02 inch units.

Each tenth would have five subdivisions (Figure 18),

so that you could read a number like 8.12 or 8.16.
Using such an addition rule, you can find, without too
much difficulty, that 0.24 + 8.12 = 8.36 or that 8.12 +
6.72 = 14.84. This will tell you that 2.4 + 81.2 = 83.6
and that 81.2 + 67.2 = 148.4. It will also tell you that
24 + 812 = 836 and that 812 + 672 = 1484.
It may seem to you that now we can only handle
numbers with right-hand digits that are even, for the

smallest divisions shown in Figure 18 represent 0.02, 0.04,

0.06 and 0.08. There is no division equivalent to 0.03,
for instance, so that a number such as 7.33 could not
be read directly.

And yet why not? You have no trouble finding 7.32

and 7.34 on the addition-rule Betweenas we now have it.

them is an empty space with no markings and we can

imagine a fine line splitting that empty space in two.

This imaginary line (shown in Figure 19 as a dotted

line )would be 7.33. You might even imagine lines drawn
closer to 7.32 than to 7.34, or vice versa and representing
7.325 or 7.335. These lines, which wouldn't be shown
directly on even the most carefully manufactured 10-inch
 Figure 19

7.3 7.32 7:34

addition rule (if any were manufactured at all) can

nevertheless be estimated.
To be sure, no matter what we do, there are limits to
how finely we can read subdivisions. Our estimates are
bound to be uncertain and we can't push them too far.
In addition, the most carefully made markings may be
very slightly off position and even if they aren't, the
markings have a perceptible thickness (they must have)
and ideally they should have no thickness at all.

We are, in short, condemned to inexactness. To see

why that is, let's go back for a moment to ordinary
fractions.

Inexactness

The decimal addition rule can handle ordinary frac-

tions, too, provided those ordinary fractions are first con-
verted to decimals. We can add SYi and 2% without
trouble if we convert the numbers to 5.50 and 2.75. »

Using the addition rule, we find that 5.50 + 2.75 = 8.25.

If we wish, we can change 8.25 back to 8!4 and then we
have determined that 5y2 + 2% = 8%.
The conversion of fractions to decimals, however, can
present the addition rule with more than it can really
29
handle. Suppose, for instance you wanted to add 3%
andSH.
The decimal equivalent of SVs is 3.125, and you have
to estimate the position of 3.125 as an imaginary marking
just one quarter of the way from 3.12 to 3.14.

As for SVs, that, in decimals, is 5.3333333 . . . and so on

for any number of 3's you care to write. How do you
locate that on the addition rule? You have the markings
for 5.32 and 5.34, and the imaginary line halfway be-
tween is 5.33. Well, one-third of the way from 5.33 to
5.34 is the number you want.
Very well, then, you place S-0 on the B-3.125 and look
under the S-5.333333 . . . and find yourself at a position
just short of B-8.46. The position is somewhere between
that mark and the imaginary mark that represents 8.45.
You might estimate that it is three-quarters of the way
over from 8.45 to 8.46 and therefore decide that the
answer ought to be, well, 8.458.

But you have been estimating in three places now.

You Ve estimated for 3.125, and for 5.33333, and now for

8.458. What is the real answer? If you work out the sum
of SVs and SVi with pen and paper, you will find that the

answer is 8^M»4 and if you change that into decimals it

comes out to 8.458333333 . . .

The addition rule did not give you the right answer.)
Look how close it came, though. It was off by only;
0.0003333. You may feel hardhearted and say, "A missi
is as good as a mile and a wrong answer is a wrong'

answer." But is it? I

In the first place, in using the addition rule, you could j

find the almost-right answer in a moment, in a fraction

of the time it would take you to get the exact answer by

pen and paper.
 )

To be sure, if you simply must have the exact answer

and nothing else, the addition rule will have failed you,

but often you can make do with less. It frequently

happens in science, engineering, architecture, or in any
field where numerical calculations are much used, that
it is not necessary to get the absolutely exact answer. A
very close answer will be fine. In that case, the addition

ule with its close answer is what you need.

Then, too, even if you want the exact answer, the
Addition rule can still be useful as a check. Suppose you
dded 3% and 5%, got the correct and exact answer of
W24 and wanted to change that into decimals. You
night perhaps get the answer 8.43842 through some
arithmetical mistake involved in long division. (It is

easy to make arithmetical mistakes in long division.

If you then perform the same addition on the addi-
tion rule and find the answer is "just short of 8.46," you
know something is wrong. If the correct answer were
3.43842, then the addition rule would say the answer
was "just short of 8.44." The addition rule, if properly
bandied, cannot mistake 8.46 for 8.44, and you know
you have made an arithmetical error in your pen-and-
paper calculations and start checking it. So you see, the

iddition rule can be extremely useful even when it gives

)nly approximate answers and not exact ones.
And, of course, it is important to realize that all through
this book, I will be giving approximate answers. When
[ say that aVs + SVa == 8.458, it is not intended to be an
3xact answer but an "addition-rule answer." It is suffi-
cient that it is a very close answer.

Reversing the Process

And what about subtraction? Since, in arithmetic.

subtraction is the opposite of addition, it would seem that
any device that is capable of working out additions can
work out subtractions, too, if it is run backward.
On page 11, for instance, I pointed out that you could
solve 10 —3 on an ordinary ruler by counting off 10
inches and then moving back 3 inches to the figure 7.

Can't that be done on the addition slide rule?

Your first impulse might be to place S-0 over B-10,,
but what good would that do us? We can't count back-
ward (that is, to the left) on the slide under such con-
ditions since there is nothing to the left of S-0.

One way out would be to move some other part ol

the slide over the B-10. We might move S-7 over it and
count back to S-4, or move S-5 over it and count back
to S-2. Best of all, we might move S-3 over it and count
back to S-0. j

You can see that the last alternative is best, because il

is autqmatic. The 3 is pait of the problem, 10 — 3, an(

you are, in effect, subhacting the 3 from itself, leaving 0.

Whenever you subtract a number from itself, you leave 0,

-«dd!tIon'

B — —r^
1
I

(b

5 +1= 7

32
 that if you follow this procedure, you will always find
e answer under S-0.
This fits in neatly with the process of addition, if we
^member that subtraction is the reverse of addition,
onsider two problems: 5 + 2 and 5 — 2. In the former,
e place S-0 over B-5 and look forward to S-2, under
hich we find B-7. In the latter, we place S-2 over 6-5
jid look backward to S-0, under which we find B-3
igure 20 ) Thus we find that 5 + 2 = 7 and 5 - 2 = 3.
Here we run the danger of falling into a difficulty. In
dition, it doesn't matter in which order we add our
imbers. Faced with the problem of 5 + 2, it doesn't
atter whether we treat it as 5 + 2 or as 2 + 5. The
swer will be 7 in either case. This lends the addition
lie a certain flexibility. If you place S-0 over B-5, the
iswer will appear under S-2. If you place S-0 over B-2,
le answer will appear under S-5. In either case, you will

id B-7 as the answer.

In subtraction, matters are less flexible. The answer to
—2 is not the same as the answer to 2 — 5. In the

8 9 10

1
1

Figure 20

8 9 10
) c 1 4[ !; ( : r

P '
I 1

'
< ^ : 1 ^' 1
) (11) '

5-2 = 3 =
 1

i
former case, the answer is 3, in the latter it is —3. * We
must, therefore, be careful about the manner in which wc
manipulate the addition rule if we are to solve subtractior
problems correctly.
The best system I know of is to remember that the B
numbers are immovable while the S-numbers can b<

pushed back and forth and can therefore be viewed ai

changeable. In the problem 5 — 2, the first number, J

(the minuend), is the number you start with; a numbei

which, if untouched, will remain 5 forever. It, therefore^
belongs among the unchanging B-numbers. The seconci
number, 2 (the subtrahend), introduces change; it-

presence alters the value of 5. The subtrahend, therefore

• It is possible to design an addition rule to deal with
negativ
numbers. To do this we would have the B numbers and S num
bers extend leftward from to give a series of ten negativj
numbers. You can then have readings of B-(— 5) or S-( — 3),
for instance. However, such a negative extension will have nt'
application to the real slide rule and its uses,which I will b
dealing with shortly. We will, therefore, pay no attention t
negative numbers. t

4— subtraction —

10

I I I

Figure 2

34
I L
* longs among the changeable S-numbers.

In the general subtraction problem, a — b = c, then,

u place S-b over B-a and then find B-c under S-0.

Following this system, you will see that to solve 5.84 —

8, you place S-2.28 over B-5.84 (never vice versa),
d under S-0, you will find B-3.56. You know, therefore,

lat 5.84 - 2.28 = 3.56 or that (just as surely) 584 - 228

356.
There's another way of handling subtraction. Suppose
le numbers on the slide ran in reverse, with the on the
treme right, and then read leftward, 1, 2, 3, and so on,

I the way to the 10 on the extreme left. This would

J an inverted scale, in place of the ordinary direct scale,

you place the of such an inverted scale over a B-

nmber, you could then find room to look leftward and
]
jrform a subtraction.
We can call such a range of inverted numbers on the
ide,
"^^^y Sl-numbers (for
^ "slide, inverted"). If we wanted

1 1

9 10

addition ^

I 1
5 i
'
'010 SI
1

d) db
1
1
;: :
1

' '- • 9 10

5 +2 7
Figure 22

to solve 5 —2 using Sl-numbers, we would place SI

over B-5 and, looking under SI-2, find B-3. If we wante
to add with the Sl-numbers, we could solve 5 + 2 b
placing SI-2 over B-5 and finding B-7 under SI-0 ( Figd

21) ^

If you compare Figures 20 and 21, you will see thij

whether you use S-numbers or Sl-numbers, you move t

the right (forward) in addition and to the left (bad

ward ) in subtraction. The difference lies in the use of tli

0. In addition, you place the S-0 over the number t

which you are adding, but not SI-0. In subtraction, yo'

place SI-0 over the number from which you are subtrac
'1

1 3 4 !. i 1
'
1 • i II

UB l'

i-
1

SI

-
1

— addition
1
*
} 1 1 ]

:
'
C
s p 1 !

f
• 1 1 1 1 1

1 2 3 4 i
dD '
f
1
'
c '1

*
>

6 + 3-9
i
s.
1

10 9

s
_1

Figure 23

36

I .
 , but not S-0. If, then, you want to make use of in

is fashion for both addition and subtraction, you must

le the S-numbers for the first and the Sl-numbers for the

«cond.
This does not mean we must have an addition rule and
jiubtraction rule, as two separate devices. There is room
Ir both sets of numbers, the S and the SI, on the addition
lie we have been using.

After all, there are two parts to the body of the addition
r le, an upper part and a lower part, and so far we have
ben using only the lower part. There's nothing to pre-
V nt us from using the upper part, too.

|Let*s place a set of numbers on the upper part of the

Idy, exactly like that on the lower part of the body.
\e can refer to the upper UB-numbers ( "upper
set as the

rdy") to distinguish it from the ordinary B-numbers we

ve been using till now.
Along the upper edge of the slide, adjacent to the UB-

1 1 1

3 2 1

7 8 9 10

1 1 1

10
' (^

Q) • subtraction

10

10

6 3 3
 )

numbers, are the Sl-numbers. Thus we have two pairs ol

scales on the same addition rule, one for addition with S-

numbers and one for subtraction with Sl-numbers ( Fig

ure 22. %
Let's use our double addition rule to work out 6 + {

and 6 — 3 just to see how it works.

For addition, place S-0 over B-6 and look under S-3
where you will find B-9, so that you see that 6 +3=9
For subtraction, place SI-0 under UB-6 and look over SIj

3, where you will find UB-3, so that 6-3 = 3 ( Figun]

23).
This brings to an end what I have to say about the

addition rule. There is considerably more that could hi\

written, but I now have all I need for the main businesj
of this book.
As I said in the first chapter, addition rules are noi^

manufactured, because addition and subtraction are sucl

easy operations with pen and paper (except for occa
sional fractions ) that no one bothers to take the trouble,
I-

to manipulate an addition rule instead.

Please don't feel cheated at this. Don't feel you have
learned various manipulations "for nothing." All th(

principles I have described in connection with the addi

tion rule can also be used on actual slide rules whicl

deal with operations more complicated than addition anc

'I

subtraction. [

On actual slide rules, however, you might have ha(^

trouble seeing why those principles worked, because th<.

scales used are tricky ones. As it is, you learned th<|

principles on the simplest possible scales, those made uj,

of evenly spaced digits. I hope that it will now be eas;
to transfer the principles from the simple scales of th<
»

addition rule to the complex scales of the true slide rules

38
 3

Logarithms

ultiples of Two

Actually, the whole key to any sHde rule, however

incy and complicated it may seem, is only addition and
btraction. The point is, however, that you must add
"jid subtract special kinds of numbers. In this chapter,
e will track down those special kinds.
Let us look, for instance, at the following set of num-
rs: 2, 4, 8, 16, 32, 64, 128, 256, 512 ..You can con-
.

fine the set as far as you wish, for, as you see, each num-
r is just double the one before.
An odd thing about such a list of numbers is this: If

")u multiply any two of them, the product will be

lother number on the list. Thus, 2 X 8 = 16; 16 X 32 =
12; 4 X 64 = 256. Then, too, if we want to get into
rger numbers, 512 X 512 = 262,144; and if you con-
ue to work out the list of numbers by doubling each
w value, you will find that the eighteenth number on
le list is indeed 262,144.
This list (or "set") of numbers is, in other words,
[Josed to multiplication."

Let's look at the list in another way. Since we get each

imber by doubling the one before— that by multi- is,

ying the previous number by two — we can write the

39
first number as 2, the second as 2X2, the third aS|

2 X 2 X 2, the fourth as 2 X 2 X 2 X 2, and so on. The.

hst becomes a set of multiples of 2. f|
Of course, it is clumsy to list the numbers as more and
more 2's multiplied together, and it is natural to search
for a simpler means of indicating The system com-
this.

monly used is to write 2, standing by itself, as 2\ The'

number 4, which is 2 X 2 (two 2's multiplied together),
is 2^. The number 8, which is 2 X 2 X 2 (three 2's multi-'

plied together), is 2'^. By this system, you would expect'

2-^0 be 2 X 2 X 2 X 2 X 2 (five 2's multiplied together)
and if you work out the product you will find it to be 32 '

Therefore, 2' = 32. '^

The small ''

( in the expression 2 '
) is an exponent antf
2"'
is an exponential number. Suppose, then, we try tc
•I
express our set of numbers in exponential form. Insteac
of writing 2, 4, 8, 16, 32 ... we can 2\ 2-, 2'\ 2^ 2
write it

. . . The latter method is clearly the simpler and neater o:

the two.
Suppose we multiply these numbers in their expo
nential forms. Instead of writing 2 X 8 = 16, we car
write 2' X 2' = 2\ Instead of 16 X 32 = 512, we hav<
2' = 2\ Instead
X 2''
of 512 X 512 = 262,144, we havfif.
2-'X2'' = 2^\
If you will examine these multiplications, you will se<

that, in every case, the exponents have been added. Ii

the case of 2^ X = 2^ 1 + 3 = 4.
2'^
In the other twcjtif

cases 4 +5=9 and 9 + 9 = 18.

|
This is not mysterious. In the case of 2^ X 2^, you areiic

multiplying a set of four 2's multipUed together (2^ =hj|ii

2X2X2X2) by a set of five 2's multiplied togethei |

(2^ =2X 2 X 2 X 2 X 2). If you multiply the first sejij

by the second set, you end with a set of nine 2*s multi^if,|
40
J
 [lied together:

„2 X 2 X2X 2) X (2 X 2 X 2 X 2 X 2) =
(2X2X2X2X2X2X2X2X2)
In multiplying sets of 2's in this fashion, you add the
imber of 2's in the various sets, and this is why expo-
;nts are added when exponential numbers are multi-
lied. We can make this general by saying that 2' X 2*^ =
(X + y)

I
Next, we'll try something else. Take a number of our
i5t of multiples of by a smaller number
two and divide it

Pthe set. The quotient is also a number of the set. Thus

56 -i- 32 = 8. Turn that into exponential numbers and it
28 ^ 2'^ = 2\ As you might guess, 8-5 = 3.
This also is not surprising. If we divide 2® by 2^, we are
u-rying through the following division:

2X2X2X2X2X2X2X2
2X2X2X2X2

The five 2's in the denominator cancel five of the eight

s in the numerator, leaving three 2's multiplied together

i the answer. Therefore, 2^ ^ 2^ = 2'^

and, in general
;C|
^ 2*' Zii 2^"^"^^ *

'

We have thus discovered a method of converting cer-

tin multiplications into additions, and certain divisions
ito subtractions. Since it is the exponents which are
rAded and subtracted in these cases, we are going to
« ncentrate on these exponents from now on.

we extract the exponent from an exponential number

If

ilid set it down all by itself, it is customary to give it a

w name and call it a logarithm. For instance, if we are

ti

41
(
considering the expression 2*^ = 8, then 3, taken by itself, i

is the logarithm of 8. Of course, we don't want to forgeti

the 2, which rests under the exponent as though it werej
a base. Consequently we speak of 3 as the logarithm o
8 to the base 2. 1
In the same way, the logarithm of 16 to the base 2 i

4, for 16 = 2^ and the logarithm of 32 to the base 2 is 53

for 2^ = 32.
We can abbreviate "logarithm to the base 2" as /og2. In
that case, we can say: log2 8 == 3, lo g2 16 = 4, log2 32 = 5,
and so on.
Suppose we want to work it backward now. We have
the logarithm to the base 2, and we want to write down
the number it represents. If we have the logarithm 4 to
to the base 2, then the number it represents is 16, so that

16 is the antilogarithm of 4. In the same way if we have

the logarithm 9 to the base 2 then its antilogarithm is 512,
sinc^ 2^ = 512. The antilogarithm in these cases is to the

base 2, and the expression can be abbreviated antilog2.

Thus, we can say that antilogi 4 = 16, and antilo g2 9 =

512.
These new terms are confusing at first and will cease

to be confusing only with practice. However, to be as

clear as possible right now at the start, consider the ex-

pression x° = b. In this general exponential expression, x

is the base, a is the logarithm, and b is the antilogarithm.
(It might help you to keep in mind the fact that log-

arithms are exponents, while antilogarithms are the "ordi-

• This relationship between multipUcation and division of ex-

ponential numbers —
where exponents are added in the first case
and subtracted in the second —
is not surprising. Division is an
operation which is the inverse of multiplication, just as subtrac-
tion is the inverse of addition. Keep this in mind for it will be
handy later.

J.
 )

numbers" you use in everyday computations.

It is easy to make a small table of logarithms and anti-
garithms to the base 2, as follows:

:?g2 antilogy logi antilogy

111

I 2 11 2,048
I 4 12 4,096
3 8 13 8,192
i 16 14 16,384
> 32 15 32,768
5 64 16 65,536
7 128 17 131,072
256 18 262,144
2 512 19 524,288

P 1,024 20 1,048,576

You can continue the list as long as you like, but we

ave enough now to make our point. Remember that
henever we multiply two antilogarithms, we can achieve
le same result by adding the corresponding logarithms,
iippose, for instance, we wanted to multiply 128 and
)96. These "ordinary numbers" are antilogarithms and
ay be found in the column headed antilogi. The loga-
thms corresponding to them are 7 and 12 respectively.
/e add the logarithms and find that 7 + 12 = 19, so that

) is our logarithm sum. We find 19 in the log2 column,

id see that 524,288 is the corresponding antilogarithm.
he logarithm sum is the antilogarithm product. We
lerefore conclude that 128 X 4096 = 524,288, a fact you
in check by long multiplication.
+ 8=18 logarithms we can see at
Again, since 10 ( ) ,

ice that 1024 X 256 = 262,144 (antilogarithms.)

'
We can try subtracting logarithms, too. Since 20 — 15
 .

= 5 (logarithms), we conclude that 1,048,576 ^ 32,768

= 32. Try this by long division and see if not right. it is

You will agree, I think, that it is much easier to add and

subtract logarithms and make use of columns such as
those given above, than it is to multiply and divide or- j

dinary numbers (that is, antilogarithms )

ff

Multiples of Ten

There are, however, flaws to this pretty picture. Let's

begin with the biggest flaw of all. What I have described,
so far, will only work for a few numbers which happen to
i
be built up through the multiplication of 2's. You can {

deal with 32 X 64, but suppose you want to deal with

31 X 63. You are stuck. You can't produce either 31 or
63 by multiplying 2's.

One possible help might be to produce numbers which

are multiples of other integers. For instance, you can't

produce either 27 or 81 by multiplying 2's, but you can

produce them by multiplying 3's. Thus, 27 = 3X3X3 =
3'^
and 81 = 3X3X3X3 = 3^ If you want the answer
to 27 X 81, try it in exponential form. Since 3 + 4 = 7,
3'^ + 3^ = 37 xi^e number 3^ represents the product of

seven 3's multiplied together, and that works out to

2187. So you can say that 27 X 81 = 2187.
In short, you can prepare columns that list logarithms
n
to the base 3 and their corresponding antilogarithms to

the base 3. You can then solve problems which you can't !

solve by using logarithms to the base 2 in the fashion !

described above.
In the same way, we can work with logarithms to the
base 5, or to the base 7, or to any base we may care to
choose and in each case work with a new set of numbers.

44
 .

However, this does not solve our problem. Every

umber can be found among the list of antilogarithms to
ne base or another, but what if you want to multiply a
umber from one list by a number from another list?
Consider, for instance, the problem 8 X 9 = 72. You
an see that 8 = 2'* and 9 = 3-. This means that /ogj
( = 3, and log:\ 9 = 2. These are logarithms to different
|iases. Does that matter? If they were logarithms to the

ame base, we would add them and achieve the same

esult that we would by multiplying the antilogarithms
\ and 9. Let's try adding them anyway. Since 2 + 3 = 5,
ve decide that 5 is the logarithm of the product — but
he logarithm to which base?
2'*
If 5 is a logarithm to the base 2, then the answer is

3"'
>r 32. If it is a logarithm to the base 3, the answer is

)r243. But the actual answer, 72, is neither. In short, we

imply can't add logarithms to one base and logarithms
another base, any more than we can add boys to cows.

If we are going to make logarithms useful, we must

tick to one particular base and work out a way for
nding logarithms to that base for all possible numbers.
Jut if we are to do this, which particular base are we
use?
Perhaps at first, you might think that the most con-
enient possible base is 2, and in some ways it is. After
1, doubling is so simple that it is much easier to prepare
list of multiples of 2 (2, 4, 8, 16, 32 . . .), than of 3's

3, 9, 27, 81, 243 .

.), or of 5's (5, 25, 125, 625, 3125 . .
.)

r of almost any other number. Furthermore, the num-

ers in the list of multiples of 2 are more closely spaced
3gether than the numbers in the list of multiples of any
[ther numbers, so there are fewer numbers not on the
45
 »

list for which we must find logarithms.

However, here as in many other cases, the fact that
our number system is upon the number 10 over
built
rides everything else. Because our number system isjj

built upon 10, it is easier to prepare a list of multiples of

10 than of any other number.
We begin with 10 itself. Then we have 10 X 10 = lOOJ
After that is 10 X 10 X 10 = 1000; 10 X 10 X 10 X 10 =i
10,000, and so on. Our list is 10, 100, 1000, 10,000 . .

each number possessing one more zero than the numberM

before. There is nothing easier than adding one morel
zero for each number on a list. It

We can express such a list in exponential form, too, as|i

10\ 10^, 10'\ 10^ . . . The connection between the ordi-l

nary 10-multiple and its exponential form is a simple one.
Since 10 = lOS 100 = 10 X 10 =10^, 1000 = 10 X 10 X^l

10 = 10"**,
and so on, you can see that the exponent isii

equal to the number of zeros in the number itself. Youil

can see without having to go into great feats of computa-

tion that 10,000,000 = 10^ and 100,000,000,000 = lO^^jl

The simplicity of this connection extends to logarithm?

to the base 10. Remember that the logarithm is the ex-'

ponent. Therefore since 100 = 10^, logio 100 = 2. AgainJ

since 10,000,000 = 10^ logio 10,000,000 = 7. Again, thel

logarithm to the base 10 is equal to the number of zerosj

in a number of this sort. ||i

Such logarithms are so easy to work out that loga-ji|

rithms to the base 10 are used far more often than loga4ll

rithms of any other kind.* Consequently, when people!

• There is a set of logarithms to another base very frequently, i

used in higher mathematics. Since this book does not involve i

higher mathematics, we will not need to consider this other set,} I

sometimes referred to as "natural logarithms."

46
L
 ! eak of "logarithms" without specifying the base, they
i e almost certain to mean logarithms to the base 10.
For the rest of the book, logarithms to the base 10 are
Te only ones I shall use and I shall refer to them merely
i logarithms, with the simple abbreviation log.
Logarithms of this sort are so easily obtained that one
esn't even need a set of columns of logarithms and
tilogarithms to carry through multiplication and divi-

im of multiples of ten. Suppose we consider 10,000 X

;0. By counting zeros, we know at once that log
1,000 = 4 while log 100 = 2. Since 4 + 2 = 6, we knoW
ie logarithm of the product to be 6. That means there
te 6 zeros in the corresponding antilogarithm, so that
I'OOO X 100 = 1,000,000.
Again, if we consider 10,000 -^ 100, we must subtract
karithms. Since 4 — 2 = 2, the logarithm of the quo-
*nt is 2 and the corresponding antilogarithm is 100.
jerefore 10,000 -MOO = 100.

I 'tween the Multiples

Of course, we can't be completely enthusiastic about

1 this, for so far we can only make use of the 10-
ultiples, such as 100 and 1000, and these are far fewer
tan the 2-multiples. If you inspect the columns on
J ge 45, you will see that there are twenty 2-multiples
I) to the neighborhood of a million. There are only six

•-multiples in that same range.

Furthermore, the 10-multiples are particularly easy to
die even without logarithms. It is no problem at all

decide, without logarithms, that 100 X 1000 = 100,000.

^hat we need is some easy method of carrying through
multiplication such as 72 X 263.
47
 We must therefore extend the notion of logarithms
1
tc

numbers that don't end with zeros. *i

Let's begin by considering the following: 10 -^ 10 = 1

We can express 10 in exponential form as 10^ and the^

express the division as follows (remembering to subtrac
exponents): 10^-^10^ = 10^. If we solve a problem
correctly, in two different fashions and get an answer ii

two different forms, those forms ought to represent th(

same number. Here we divide 10 by 10 and get twc

answers: 1 and 10^. It is reasonable then to suppose tha

10^ = 1. If we bring the exponent down, it become
a logarithm and we can say that log 1 = 0.
We know now that log 1 = and log 10 = 1. Wha
about the logarithms of numbers between 1 and 10? I

would seem that those logarithms ought to lie betweei

and 1 in value. They ought, in other words, to h\

fractions.
But what meaning could a fractional logarithm have
Suppose there were a number which had a logarithn
equal to K». What kind of number would fit such j

logarithm?
Let's consider a number x, such that log x = M». Re
member that a logarithm is simply an exponent brough
down from its position above. If log x = M», then x = 10''
We are certainly entitled to wonder what a number lik<

10'^ can possibly mean. Suppose we consider the produc

of 10'^ XIO'^ If we add exponents we can see tha
10'^ X 10*^^ = 10^ = 10.

We can therefore say that 10'^- is that number whicl

when multiplied by itself, gives 10 for an answer. Sue
a number is the "square root of 10" and is usually writte
ViO.
It is not difficult to work out an approximate value (

48 I
the square root of 10. Actually, it turns out to be an un-
ending decimal expression, but it is possible to work it

out to as many decimal places as are desired, and the

more decimal places that are worked out, the closer the

number is to the actual value of the square root of 10.

Worked out to six decimal places, the square root of
10 is 3.162120. It turns out that 3.162120 X 3.162120 =
9.9990028944, which is almost 10 as you see. In fact, it

would be pretty useful to speak of the square root of

10 as 3.162. This number, multiplied by itself, yields the

product 9.998244, which is still pretty close to 10.
Since 3.162 X 3.162 is just about equal to 10, we can
say that 3.162 is just about equal to 10'^^ We can there-
fore say, with a good approximation to the truth, that

antilog \k = 3.162 and that log 3.162 = ¥2.

Next let's consider that 10'^ X 10'^ X 10'^» = 10^ = 10.

Therefore, 10'^* is a number which yields 10 when mul-

tiplied by itself twice. We can, therefore, say that 10*^
is equal to the "cube root of 10" or v^lO. It is possible
to work out the cube root of 10 and this turns out to be

(to three decimal places) 2.154. Therefore we can say,

as an approximation, that antilog Vs = 2.154 and that

fog 2.154= Vs.
By adding exponents, we can see that 10^* X 10^ =
lO'^. We already know that 10'^ is equal to approximately

2.154. If we multiply 2.154 by itself we get about 4.64.

Therefore, lO'^ is equal to about 4.64, which means we
can say that ( approximately, at least ) antilog % = 4.64
and fog 4.64= %.
You can, if you choose, express such fractional loga-
rithms in decimal form; and, indeed, they usually are so
expressed. Instead of saying fog 3.162 = H, you would
49
say log 3.162 = 0.5. If you worked out the fact that log

1.585 = %, you could express it as log 1.585 = 0.2, and

so on.
The cases I've presented above are examples of ways
in which simple fractional logarithms can be obtained.
Those simple fractions have as their antilogarithms
rather complicated decimals. Mathematicians, however,
have worked out methods for obtaining the logarithm

of any number at all, including logarithms for all the

simple digits. Thus, the logarithm of 2 turns out to be
about 0.301 (to three decimal places), so we can say
/og 2 = 0.301.
Actually, the logarithm of a niunber, in almost every
case, is an unending decimal, but mathematicians can
work out as many places as they choose. They can then
prepare logarithm tables, in which the logarithms of a
set of consecutive numbers are presented. Such loga-

rithms are commonly presented to five decimal places.*

By using such a table, we can find that:

antilog log

1 0.00000
2 0.30103
3 0.47712
4 0.60206
5 0.69897
6 0.77815
7 0.84510
8 0.90309
9 0.95424
10 1.00000

This is equivalent to saying that we can place each

I
'of the digits from 1 to 10 into exponential form. We can
ay that 3 = 10^^^^^^^ 7 = 10^ ^^sio ^^^ ^^ ^^
The usual five-place logarithm table will, of course,

Iso give the logarithms of decimal numbers. For in-

tance, such a table will tell you that log 4.354 = 0.63889
nd that log 2.189 = 0.34025. That means that 4.354 =
^Qo that 2.189 = 10^
r>.->>889 "^^^^s
and
In multiplying exponential numbers, we add expo-
nents — which is to say, we add logarithms. Suppose, for
instance, you wanted the answer to 4.354 X 2.189. In-

stead of multiplying the numbers themselves, you add the

logarithms and find that 0.63889 + 0.34025 = 0.97914.
The sum 0.97914 is the logarithm of the product of 4.354
X 2.189. In the logarithm table, we find that 0.97914 is

tj|ie logarithm of 9.531. Therefore, we conclude that

4.354X2.189 = 9.531.
This is not the exact answer. The manipulation of
logarithms can't give you the exact answer unless the
exact values of the logarithms are used, and logarithnis
are virtually never known exactly. However, five-place
logarithms are accurate enough for most purposes. For
instance, if you work out 4.354 X 2.189 with pencil and
paper, you find the exact answer is 9.530906 and surely
9.531 is quite close.
In the same way, you can subtract logarithms instead
of dividing numbers. Suppose you wanted the answer to

4.354 -^ 2.189. Taking the logarithms again, we find that

0.63889 - 0.34025 = 0.29864. The logatithm-difference
0.29864 is the logarithm of the quotient of 4.354 ^ 2.189.
• If this were a book on logarithms, I would include such a
table and give complete instructions for its use. However, I am
using logarithms only to explain the workings of the slide rule.
I will ask you, therefore, to accept, more or less on faith, the

values of the logarithms as I present them in the meanwhile.

51
 I
The logarithm table tells us that 0.29864 is the logarithm 1
of 1.989. Therefore, we conclude that 4.354 -^ 2.189 =
1.989. If you work out the problem by long division, you
find the answer is actually about 1.98858, but again 1.989
is reasonably close.
With practice, one could learn to use logarithm tables
so quickly and would be no question of
easily that there

working out complicated multiplications and divisions in

the ordinary way —
unless one simply had to have

absolutely exact answers. Even then, the logarithmic

answer would be a convenient check against the ever-
present possibility of arithmetical eiTor.

Changing the Characteristic

So far, we have been working with the logarithms of

the numbers from 1 to 10. There is, however, a vast
array of numbers greater than 10 and smaller than 1.

What about those?

What might the logarithm of 52.38 be, for instance?
We can solve that problem by considering 52.38 asi

5.238 X 10. The logarithm of 5.238 is, according to the

logarithm table, 0.71917. The logarithm of 10 is, of;

course, 1. If we are multiplying two numbers to get a

product, we can just as well add the logarithms of those i

two numbers to get the logarithm of the product. Since

5.238 X 10 ^ 52.38, and 0.71917 + 1 = 1.71917, then
1.71917 is the logarithm of 52.38. We can express this)
more simply by saying log 52.38 = 1.71917.
It is simple to find the logarithm of 523.8 now. We see
at once that 5.238 X 100 = 523.8, and we know that the
logarithm of 100 is 2. Therefore using logarithms, 0.71917

52
 :

- 2 = 2.71917 and log 523.8 == 2.71917. In the same

fay, we can show without trouble that log 5238 =
.71917, that log 52,380 = 4.71917, and so on.
Let's make a small table to make sure we see the point
learly

antilog log

5.238 0.71917
52.38 1.71917
523.8 2.71917
5,238 3.71917
52,380 4.71917

We can divide the logarithm into two parts : the num-

er to the left of the decimal point, which is called the
aracteristic, and the number to the right of the decimal
int, which is called the mantissa. If we say that log
^2.38 = 1.71917, the characteristic of that logarithm is 1,

he mantissa is 71917.
When two numbers' differ only in the position of the
ecimal point, the mantissa of their logarithms is iden-
^cal, as you see in the example given in the table above.

he characteristic, on the other hand, changes with the

hifting decimal point in a number.
The change in the characteristic presents no problem.
\y inspecting the table above you can see that in the
ase of those numbers, at least, the value of the character-
5tic is one less than the number of digits to the left of the
lecimal point in the antilogarithm; and this it turns out is

rue in all cases.

This rule is so simple that we need no logarithm table

^ determine characteristics. By simply looking at the
lumber 35.62, we see that there are two- digits to the left

53
 of the decimal point and that the characteristic is there-
fore 1. In the same way we know that 282,100,000, which
has nine digits to the left of the decimal point, has aj
characteristic of 8.
All we need the logarithm table for, then, is to de-l

termine the mantissa, and it should, therefore, really be

called a "mantissa table." Since the mantissa is the same ^

for any particular combination of digits, regardless of f

the position of the decimal point, we can ignore the

decimal point altogether in finding the mantissa. t

Suppose we want the logarithm of 35.62. We look up i

the mantissa for the digit-combination 3562 and find it to '

be 55169. Since 35.62 has two digits to the left of the

decimal point, the characteristic of its logarithm is 1; and i

therefore log 35.62 = 1.55169.

In the same way, to get the logarithm of 282,100,000,
we up the mantissa for the digit-combination 2821
look
and find it to be 45040. Since 282,100,000 has nine digits
to the left of the decimal point, the characteristic of its

logarithm is 8. Therefore, log 282,100,000 = 8.45040.

Suppose, then, we wanted to multiply 123.1 and 35.2. |

We know at a glance that the characteristic of the loga-

rithms of these two numbers are 2 and 1 respectively.

We find the mantissas of the two digit-combinations 1231

and 3520 in the log tables and find them to be, re-

spectively, 09026 and 54654. Therefore log 123.1 =

2.09026 and log 35.2 = 1.54654. If we add these two
logarithms, 2.09026 + 1.54654, we get 3.63680, which
is the logarithm of the product of 123.1 X 35.2.

The problem, now, is to find the antilogarithm of p

3.63680. It is the mantissa that will give us the necessary
digit-combination, and the logarithm table tells us that
the digit-combination of the antilogarithm represented by
54
 )

e mantissa 63680 is 4333. The characteristic, 3, tells us

at we must place the decimal point so that there are

ur digits to its The answer is, therefore, 4333.
left.

The actual answer, obtained by multiplying 123.1 and

.2 in full, is 4333. 12.)
We can divide similarly. If we want to divide 123.1 by
.2, we subtract the logarithm of the latter from that of
e former. It turns out that 2.09026 - 1.54654 =
3.54372. The digit-combination of the antilogarithm of

e mantissa, 54372, is 3497. Since the characteristic is 0,

e decimal point must be so placed as to have one digit

its left. We can therefore say that 123.1^35.2 =

1.497. (If we work out 123.1 ^ 35.2 by long division, the
nswer turns out to be about 3.49715.
And what about numbers less than 1? Suppose, for
stance, we are faced with determining the logarithm of
1.481. We use the same strategy that worked for us be-

ore. We will consider 0.481 as 4.81 -^ 10. The logarithm

>f 4.81 is 0.68215 and the logarithm of 10 is 1. In dividing
lumbers we must subtract logarithms to get the logarithm

)f the quotient. Therefore, if 4.81 -MO = 0.481 and

1.68215 - 1 = -0.31785, then -0.31785 is the logarithm
»f 0.481.
This presents us with a slight problem. The mantissa
>f —0.31785 is 31785, which is different from the mantissa
>f the logarithm of 4.81. Furthermore the characteristic
; —0, whatever that means. Fewer problems are created
'
we leave the logarithm of 0.481 as 0.68215 — 1, without
rying to carry through the subtraction. That leaves the
lantissa the same as in the logarithm of 4.81 or 48.1 or
ny of that family. And we can call —1 the chaiacteristic.
In the same way, we find that 0.0481 can be written as
181 -T- 100, so that its logarithm is 0.68215 - 2, the
55
logarithm of 0.00481 is 0.68215 - 3, and so on.
If we write positive numbers less than 1, with one zero
to the left of the decimal point, we can then work out a
simple rule for determining the characteristic of their
logarithms. The characteristic is a negative number with
its value equal to the number of consecutive zeros that
begin the antilogarithm.
For instance, 0.235 begins with a single zero, and
the characteristic of its logarithm is —1. The number
0.000442 begins with four consecutive zeros and the
characteristic of its logarithm is —4.
It is only consecutive zeroes that count in this rule. The
number 0.02004, has four zeros all together, but it starts

with two consecutive zeros and its characteristic is —2.

We find the mantissa in the ordinary way. For 0.235,
we look up the mantissa of the digit combination 235 and
find it to be 37107. Since the characteristic is —1, log
0.235 = 0.37107 - 1.

Again, to find the logarithm of 0.000442, we look up

the mantissa of the digit-combination 442. This turns out
to be 64542, so that log 0.000442 = 0.64542 - 4. To get

the logarithm of 0.02004, we look up the mantissa of 2004

and find it to be 30190, so that log 0.02004 = 0.30190 - 2.
We can make use of these logarithms with negative
characteristics in quite the usual way. If we wish to

multiply 235 by 0.000442, we need but add the log-

arithms: 2.37107 + (0.64542 - 4) and this comes out to
3.01649 - 4, or, simplifying it, 0.01649 - 1. The man-
tissa 01649 corresponds to the antilogarithm digit-

combination 10387 and from the characteristic of —1,

we know the decimal point should be placed so as to I

make the answer 0.10387.

 If we were considering the problem 235 -^ 0.000442,
we would subtract the second logarithm from the first.

This means 2.37107 - (0.64542 - 4). As you probably

know, subtracting a negative number is the same as

adding a positive one, so that by removing parentheses,

we have: 2.37107 - 0.64542 + 4 = 5.72565. The man-
tissa 72565 gives us 5317 as an antilogarithmic digit-

combination. From the characteristic 5, we know that the

answer must have six digits to the left of the decimal
place and must therefore be 531,700.
Let us summarize, then. For any positive number, we
can work out the characteristic of the logarithm at a
glance. The mantissa can be found in a logarithm table
and will be the same for any particular digit-combination
*
regardless of the position of the decimal point.

• This is true, in our number system, only for logarithms to the

base 10 — another reason for preferring the base 10 to any other
in ordinary computations. Nor is this a coincidence. It is the
inevitable consequence of the fact that our number system is

built around 10 and multiples of 10.

57
 4
Logarithms on Wood

The L-Scale
•

Let now and

us stop consider the situation. In Chapter
2, I showed how we could add and subtract numbers,
automatically, on an addition rule. In Chapter 3, I

showed how we could multiply and divide numbers by

adding and subtracting their logarithms.
Could we not combine these two processes? Could we
not use an addition rule to add and subtract logarithms
and thus carry through multiplication and division auto-

matically?
Indeed we can. An addition rule used to add and sub-
tract logarithms is the true slide rule, and it is this slide

rule which this book is intended to explain.

The simplest slide rule we can imagine is, actually,

identical with an addition rule. Suppose, for instance, we

are faced with the necessity of performing a particular
multiplication. We use a logarithm table to look up the
logarithms of the numbers being multiplied. We then
add those logarithms on the slide rule, thus obtaining the

Prim«riM SccendariM

.0
lHHlMlllllMlMHllllllllHhlllllllllMllllllllllllllHlllHllHllllMlllllllllllllllllnilhlllllinlHlllllllllHlhMllllllllMlhlllhlll)lllllllllhllllllllhllllMlllllHhMllllMll»>>"i<llllllMll|lM

58
 .

logarithm of the product. We return to the logarithm

table to get the antilogarithm of that final logarithm, and

thus obtain the product of the multiplication.
This is clearly not very good. The time-consuming
part of working with logarithms is the finding of log-
arithms and antilogarithms in the table. Adding the
logarithms once they are found is very simple. We are
thus making the slide rule do the easy part and neglect
the tedious part.
But what if we could make the slide rule itself into a
kind of logarithm table? It might then be unnecessary to

lookup logarithms and antilogarithms at all. We could

make the slide rule look them up, so to speak, and, having
found them, add or subtract them, and then give us the
antilogarithm as the answer.
Can this be done? Or is it just a beautiful dream?
Delightfully enough, it can be done; and easily too. To
do it, we begin by marking off logarithms on the slide
rule.

For this all we need are the logarithms of the numbers

from 1 to 10. Once we have these, we can obtain the

logarithms of numbers less than 1 or more than 10 by

choosing the proper characteristic, which is easy enough
( see page 53 )
The logarithm of 1 is 0.0. As we progress through
numbers greater than 1, the logarithm increases through

|lHrtllMllMlllulllllllll'l|nilll^lhlMllMlllllllMHhMltllllllMlllllHllllhlllllllllHMlMlllllllhlllllllllMlllnilhMlllllHllHllllllMlllMlllllMhl^

59
higher and higher values — 0.1, 0.2, 0.3, etc. — until

finally the value of 1.0 is obtained, which is, of course,

the logarithm of 10.
Suppose, then, we represent these logarithms in the
form of a line on the body of the slide rule. The line

begins at 0.0 (the logarithm of 1) and ends at 1.0 (the

logarithm of 10). This line is divided into ten equal
divisions by marks we can call primaries. The primaries
represent logarithms of 0.1, 0.2, 0.3, and so on.
In order to simphfy matters and keep from blurring the
eye with unnecessary details, it is customary to omit the
zero to the left of the decimal point. The primaries on this
scale of logarithms are therefore marked oflF as .0, .1, .2, .3,

.4, .5, .6, .7, .8, .9, 1.0, from right to left. In some slide

rules the decimal point is also omitted.

This line actually deals with mantissas, rather than
with logarithms, just as a logarithm table does. If we
want to consider the line as dealing with mantissas to two
places, the primaries mark off the m.antissas 00, 10, 20,
30, etc. If we wish to consider the mantissas to three

places, the primaries are 000, 100, 200, 300, etc.; if to

four places, they are 0000, 1000, 2000, 3000, etc.

It will be most convenient to consider the mantissas to
three places, and we will therefore consider the primaries

as representing 000, 100, 200, 300, 400, 500, 600, 700,

800, 900, and 1000.
Between the primaries are ten equal divisions marked
off by secondaries, the middle secondary being longer

than the rest for convenience in reading mantissa values.

Between each neighboring pair of secondaries are five

still marked off by tertiaries.

smaller divisions
The result, with its numbered primaries, and its un-
numbered secondaries and tertiaries, is shown in Figure
60
24, in which the scale is marked "L" ( for logarithm ) . We
can, therefore, refer to it as the L-scale,
Notice that I have drawn the L-scale on the lower part
of the body of the slide rule in Figure 24. There is no
hard and fast rule about this. Whether the L-scale is

drawn on the upper body or elsewhere depends on the

particular slide rule.
Slide rules are manufactured in many different designs
which differ among themselves in details. The slide rule

I am using as a model for the illustrations in this book has

the L-scale on the lower body. Another slide rule in my
possession has it on the upper body. Still others have it on
the slide.
If your slide rule possesses scales in locations other
than those shown in the illustrations, do not feel dis-

turbed. The difference in location will not be important,

and you will be able, very easily, to adapt my instructions

to your slide rule.

If you have an inexpensive slide rule ( and some quite

decent slide rules can be obtained for not more than a
couple of dollars ) then some of the scales I will describe

in this l)ook may be missing altogether from your instru-

ment. You will be able to follow the use of the missing
scales from the diagrams in the book, however, and with
the scales you do possess you will still be able to perform
the essentials of multiplication and division.
If you measure the length of the L-scale on almost any
slide rule, you will find that it is nearly ten inches long.
It is this which allows people to refer to a 10-inch slide

rule even though the overall length of the usual slide rule
(which naturally extends some distance past both ends of
the L-scale) is about 12/1' inches long.
The fact that the L-scale has a total length of 10 inches
means that the primaries (numbered from 1 to 10) are
about 1 inch apart. This makes it tempting to look upon
the sHde rule as an "ordinary ruler/* This, however, is

quite wrong.

DO NOT make use of the L-scale to measure lengths.

The slide rule is a precision instrument and should be
used with the greatest care and only for the purpose for
which it is designed. (Would you hitch a race horse to <

a milk wagon?) Secondly, the slide rule is not adapted to >

the purpose and has a blunt end against which it is hard j

to measure lines. Thirdly, the L-scale is not quite 10

inches long. It is 25 centimeters long and that actually
comes to 9.84 inches. The primaries aie therefore only i

0.984 inches apart.

To summarize, the use of a slide rule as an ordinaiy
ruler is improper, inconvenient, and incorrect. Need I say
more?
It is not very difficult to read the L-scale. The reading
can be made quite easily to three places so that the pri-
mary readings, as I said before, are 000, 100, 200, 300 and
so on.
Concentrate for the moment on the primary markings

Figure 25
336 340
330 331
:
:

HllllnillMMllllllMMlinillllllllllllMlllllllllllllHlllllllMllllllJllMhlMllinlnHlMllllllllllltllinllHllllMMHlllMllllllllllllillllinillllllllulMII^

1
3S0 400
 3 and 4 which represent 300 and 400. The secondaries
lying between mark off tens: 310, 320, 330, 340, 350, and
so on. The long secondary in the middle is 350. This can
be told at a glance, and the shorter secondaries can be
counted off easily enough.
Between the secondaries are the tertiaries which mark
off by twos. Thus, between 330 and 340 are four tertiaries

which represent 332, 334, 336, and 338. To find 337 we

place an imaginary line midway between 336 and 338;
and we can find in similar fashion any odd three-number
mantissa.
With practice one can even estimate four place mantis-
sas. Suppose, having found 336 and 338, we consider
those two tertiaries to represent 3360 and 3380. Exactly
between lies the imaginary line marking 3370. Half way
between 3370 and 3380 would be 3375 and a shade to the
right of that would be 3376 (Figure 25).
Such an estimation of a fourth place is never accurate,
however, and it is quite impossible to locate a fifth place
on the L-scale. It is for this reason that the slide rule
cannot entirely replace a logarithm table. A decent
logarithm table gives logarithms to five places at least

!,•
imImiAiiMii.,I.,«|,,„|„„|„,,|,„,|„,,|„J„,,|,,„|,,,,|,„,,„,;|,,,,,,,„,,,,,,,,„,,,,J,,,,,,,,,,,„,,,,,,,,,_;|,,,,,,i,,,,ii,,,,i,,^ijJ^i^_,^i__^^

63
 .

.11 ttMit|lltl|IUI^t*1IMII<
»MM'»'\'iM'r\*l'»M'*M»^Ml (MI«|l«tt|tlWtltllUi

|lllltllM|l>llllMI|l><l,.M.,l..l,ll.,|. IMp.l .|l >. |M > . \ • •

J»1M'>>«\1">|""1,>"'\^"V'>M""\'"»«""1<" "IH>11H.1..J.,.*,
CvF I

CI litiiiiit<tiiiii«>ii\ii<iiiiii(iiiiii>ii\it<ti\a<Ki.t,M.\.i,t,i,i.\<i>><i<i>i<t>i>t.ii\U.t.i,i it iiiti ) . I
\ , t . I I > t Amijki
c
o

Figure 26

and thus offers an accuracy greater than that possible on

the shde rule. The slide rule makes up for its lesser ac-

curacy, however, as we shall see, by its far greater

convenience.

The D -scale

To make a logarithm table out of the slide rule, we

need more than just the mantissas of the L-scale. We?
need antilogarithms, too, and we need those antilog-
aiithms placed in such a way that we can in a moment
match up each logarithm with its corresponding anti-
logarithm ( or vice versa )
In order to do this we need a second scale above or I

below the L-scale. Each particular point on this second

scale will represent the antilogarithm corresponding to|

the logarithm at the point directly above or below on the.

L-scale.

To match a point on one scale with the corresponding!

point on another scale directly above or below is not easy

to do by eye, however. Slide rule manufacturers havej

therefore provided a mechanical device to help out.
This is an indicator assembly which consists (in most
slide rules) of two glass windows, one on either side of

I
 . y^><y^w^l«\^M« H ll^l l lll H^^»y»^V'^y"\
^

:he slide rule, fitted into a holder at the top and bottom
;o that they are held firmly in place (Figure 26).
The indicator assembly can slide back and forth easily
Jong the slide rule when it is pushed. If it is left alone, it

s held in place by the friction of a spring, so that it stays

^here it is put. The indicator assembly cannot be moved
dAF the slide rule, however, unless it is taken apart, for the
inetal bar that holds the two parts of the body firmly
cogether'acts to block the further movement of the indica-
:or assembly.
Down the precise middle of each glass window is a
ine vertical hairline. As the indicator assembly is moved
right and left, the hairline itself can fall over any part of
che L-scale from on the far left to 10 on the far right.

The hairline can act as a convenient guide for the eye.

If you want to locate 337 on the L-scale ( a reading we

can refer to as L-337), begin by adjusting the hairline so
that it falls between the tertiaries L-336 and L-338. It is
easier to put a real line between the two marks, than to
.try to imagine one. By adjusting the hairline one can
even try to estimate the position of L-3376 and mark its

^position until such time as the hairline must be moved

again. The hairline, in short, finds a point on the scale
and marks it.
65
 hairline set at L-337

•• .1 .2 .3
lllllllll>lllllllllllllll|j|llllllllllllll>llllll>lMlllllllllll>llll>llllllllMllllllril>lllllllllllllllllHlll>llllllllllllll(lll|lllllllllllllllHllllllllMlllllllll liiMliiiiliiiiliiiiliiiiliiiiliMilr

It does more, too, for once the hairline is placed ai

L-337, it also marks the corresponding points on variouj

scales above or below the L-scale. In particular, we car
find points on the scale of antflogarithms (which I an
about to discuss) that correspond to points on the L
scale.

The indicator assembly will play a part in virtually

every drawing in the remainder of the book. In order tc

simplify these drawings and in order not to obscure anj

markings, I will make no attempt to picture the entir<

indicator assembly. I will represent only the hairline as sj

hne crossing the slide rule vertically (Figure 27).

Where, now, shall we place our antilogarithm scaler!

There are four key scale positions on the slide rule: on thej

top and bottom of the slide, on the upper part of th<

body adjacent to the slide and on the lower part of thcl

body adjacent to the slide. (These are the position|

where I placed the four scales of my addition rule. Se<
Figure 22 on page 36)
On early slide rules, these positions were used for thj

basic scales and they were marked, from the top dowi^
A, B, C, and D (Figure 28). These letters are still apj

plied to certain scales in these positions, although a

66
 Figure 27

1.0
'iMMlMMliiMliinlMMliniliinliiMlMiiliiiiliiiilMnlinJniiliMilHiihiiiMMilinilnMliHilMiiliiiJiiMlnnlinHHiHiHiliHilNMliMiliiMliMilHiiliinliiiilhiiliniliHiliiiilnnliii^

Dther scales are lettered by means of initials of one sort or

another, as the L-scale is lettered L for "logarithm."

It is customaiy to place one scale of antilogarithms in

position D, and it therefore becomes the D-scale. By this

arrangement, the L-scale and the D-scale are both on the

iower body. Each maintains a fixed and permanent posi-
tion with respect to the other, since neither is disturbed
*
oy the to-and-fro motion of the slide.
We can begin matching the L-scale and D-scale at

^he extreme left. The leftmost reading on the L-scale is

300, and this mantissa has for its antilogarithm 1, or any

number obtained from 1 by moving the decimal point —
10, 100, 0.1, and so on. Naturally, it is simplest to con-
sider the antilogarithm as 1.

I If the hairline, therefore, is placed exactly on L-000, it

>hould intersect the left end of the D-scale and give a

I'eading of D-1.
'

(If, on your own slide rule, the hairline does not

Exactly intersect L-000 and D-1 simultaneously, then the

ilide rule is out of adjustment. It may be that the top and
bottom halves of the body are not lined up properly, or
• Where on the slide, as in some rule clesij;ns, the
the L-scale is

scale of antiloj^arithms can be placed in the C-position. The L-

scale and the C-scale will then have a fixed interrelationship.
Figure 28

that the indicator assembly glass is not seated properly

in its metal holder. It is best not to try the adjustment by
yourself —a matter of loosening screws, moving the
upper part of the body or the indicator glass, and tighten-
ing the screws again — but to allow someone familiar

with the slide rule to do so.)

Let's continue. If we refer back to the tabio on page
54, we will see that log 2 = 0.301. (We need only use
three decimal points here, for that is all we can read ac-
curately on the L-scale. ) That means antilog 0.301 = 2,
so that immediately over L-301, we place D-2. Continu-
ing to refer to the table on page 54, we must place D-3
directly over L-477, D-4 directly over L-602, and so on. .

The reading L-1000 represents a logarithm of 1.000

I

Figure 29 1

D
I I

1 2 3&SS769
,|iiii|illl|liil|illl|llll|MII|llll|llll|llll|llll|MII|IIM|IIM|llll|MII|nil|IIM|IMI|lll II I I
I
I I I l| I I M|llll|lll'

L I
llllllllllMlllllllllMllllMlllllllllllllllllinlllllllllllllllllllllllllllllllllllllllllllllllMllllllllllltllllllllllllllllllllllllhlllllMllllll iiiliiiiliiiiliiiiliiii[iiiiliiiitiMiliiiil:1i
I

L-000 L-301

68
and the mantissa here is once again 000. The antilog-
arithm is, once more, 1, so that the right end of the
D-scale, Hke the left end, is marked 1. You see the result
in Figure 29.

n
ompression at the Right

Tm sure that the first thing you notice about the D-

icale is that the ten primaries by the numbers
marked off

ire not equal in length. The distance between 1 and 2 is

onger than the distance between 2 and 3, which is in

!urn longer than the distance between 3 and 4. The

ihortest distance of all, that between 9 and 10, is less than
4i as long as the distance between 1 and 2. There is a

ll|llll|IMI|llll|llll|llll|llll|MII|llll I
I
I
I
I
I
I
I
I I
I
I |i|i| i|Mi|i|i|i|i|i|i|i i|i|i|i|i|i|i!l|i|.i llllllllllllllllll llllllllllllllllll
I

IiiiiIiiiiIiimImiiIiimIiiiiIiimIiiiiIiiiiIiii lllllllllllllllllllllMMIIllltlliMllllMllll iiiiIiiiiIimiImiiImiiIiiiiIiiiiIiu iiiiIiiiiIiiiiIiimIiiiiIiiiiIi liiiiliiiiliiiiliiiiliiiil iiliMiliiiiliiiiliiMM mIimiIiiiiImiiIii

L-602 L-699 L-778 L-845 L-903 L-954 LI 000

steady compression of the distances at the right in other
words.
To see why that is, let's consider the set of expressions:
2, 2 X 2, 2 X 2 X 2, 2 X 2 X 2 X 2, and so on. If we con-

sider the products of these expressions, we get the set ol

numbers: 2, 4, 8, 16, 32, etc., a set we have dealt witl

before.
Next let's consider the logarithms of the set of expres
sions we have just mentioned. The logarithm of 2 i*

0.301. If we want the logarithm of 2 X 2, we must take

the sum 0,301 + 0.301. Continuing in this way, we fine
that the logarithms of the set of expressions 2, 2 X 2
2X2X2, 2X2X2X2, etc., is 0.301, 0.301 + 0.301
0.301 + 0.301 + 0.301, 0.301 + 0.301 + 0.301 + 0.301, etc
Taking the sums, we get the set of logarithms: 0.301
0.602, 0.903, 1.204, and so on.
Now we
can prepare a table matching the products p
each expression in the set with the logarithm of tha
product, and we can begin with 1, which has the loga
rithm 0.000:
i

antilogarithm logarith;

1 0.000
2 0.301
4 0.602
8 0.903
16 1.204

and so on for as long as we care to continue.

Notice that the antilogarithms increase by doubling
Each antilogarithm must be multiplied by 2 to get the
next one in the set. A set of numbers in which each it
70
 )btained from the former through multiplication by a
^articular number (in this case 2) is a geometric pro-
iression.

The logarithms, on the other hand, increase by addi-

tion To each logarithm must be added 0.301 to get the
.

lext one in the set. A set of numbers in which each is

)btained from the former by the addition of a particular

lumber ( in this case 0.301 ) is an arithmetic progression.
It should be quite clear that the numbers in a geometric
)rogression increase in value much more rapidly than do
hose in an arithmetic progression. You can see it's so
n this case.

As we pass along the two sets of numbers in the table

ibove, you can see that a fixed change in the logarithm

orings about a greater and greater change in the anti-

ogarithm. To begin with an increase of 0.301 brings the

intilogarithm from 1 to 2; another such change of 0.301
)rings the antilogarithm from 2 to 4; then from 4 to 8;

hen from 8 to 16, and so on.

A broader and broader stretch of antilogarithms must

it into each fixed change of logarithm, which is why the

; lumbers on the D-scale are closer and closer together

is they get larger and larger. It is why there is com-
:)ression on the right.

1 We can show this graphically, too, as in Figure 30.

In this graph, we have a scale of logarithms from 0.0
:o 1.2 on the bottom, and a scale of antilogarithms from
) to 16 on the side. The dots mark the intersections of

ihe logarithm value and its corresponding antilogarithm

^alue, as given in the little table on page 70 . A smooth
^*urve is drawn through these dots and this represents the

ine of intersections of all logarithms with their corre-

71
 1 ' " ' ,

Antiloqarithmt ( Lo9«rithmic Scale)

16
n~
I 3 4\ 1
\ 6 1r

TT
8 9 1

-J
"1
1
"

15
— — —
14
f
\
/
y
13
/
—— - y
/
12
—
- h —— - - ...

--
y -

——
' --
, - —— ...
/ — —<
r
—
-

II
1

/
r 1

I
10 ~ U -- 1
/
/[
- - - — - — t — ....

£ 9 — — ^ - ^^ ^^ — .---

^
>
^ — -.
— - —— _ - (
e ;^
< f
7
/^
„ —— — "•
-- -- ... _ .. — — -- — — .. — —
— -H — -
/
7
r
J
J/
7
/
^r
^
iT
7
y
r
*^
/ <:

^ >

^^
/-

^ —
^ ^ ^— \

1
--
1

^^ ^ ^^^ ^^ __ 1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 I.I I.

Logarithms Figure

sponding antilogarithms.
As the logarithm increases at a steady rate (O.OC,

0.301, 0.602, 0.903, 1.204), while the antilogarithm i-

creases at a steadily increasing rate (1, 2, 4, 8, 16), th

curve bends upward, becoming ever more nearly verticl.

i
 2

If the horizontal hues, representing successive antiloga-

rithms from 1 to 10, are reflected, so to speak, upward
from the curve, we get a sequence of numbers from 1 to

10 at the top of the graph that are squeezed to the right

exactly as they are on the D-scale. The antilogarithms,

IS positioned at the top of the graph, are exactly over
heir corresponding logarithms at the bottom of the
graph. The top of the graph represents the D-scale,
therefore, and the bottom of the graph the L-scale.
Since the compression at the right in the D-scale is

jrought about by the adjustment of its numbers to a

jcale of logarithms increasing at a fixed rate, the D-scale
is an example of a logarithtnic scale.

Subdividing the D-scale

The spaces between the primaries on the D-scale can

be divided into finer subdivisions. Consider the space
between D-1 and D-2, for instance. D-1 is under L-000
and D-2 is under L-301 because log 1 = and log 2 =
0.301. Using the same system we can place the values
1.1, 1.2, 1.3, and so on, onto the D-scale directly over
appropriate values on the L-scale.
I
Thus, log 1.1 = 0.041; log 1.2 = 0.079; log 1.3 = 0.114,
etc. Therefore, D-1.1 can be placed over L-041; D-1.
«

over L-079; D-1.3 over L-114, and so on (Figure 31).

This gives us our secondary markings between D-1 and
D-2. Similarly, tertiaries can be placed between D-1.1
and D-1. 2, marking off 1.11, 1.12, 1.13, and so on. In
this way, too, secondaries and tertiaries can be set up
all along the length of the D-scale as finely and as ac-
curately as the length of the slide rule and the delicacv of
73
 1

•o
I t
O
I •

O O O

H
III I
l|lMI I I M I
I I
I I II I I
I
I I M lPi|i:;i iiPIII lllllllll lll|ini llllllll{ I I I I
I
1 1 1
1
1 I
1
1 iii|iiii{iiir
2

L .1
'iIMIuhIiiIiIiH niliHilniilid llltlniiliiMli iiiiiiiiiiiiii il(i<il>Mihi iIkmIiiiiI hull iiiliiiili IiimIii iiitlniiliiiiliiiiliiiiliiiiliiiiliiiiliniliiiiliiiilMiillllili }|

o
•o
1^ So

the instruments used will allow.*

The usual secondaries and tertiaries on the D-scale J
an ordinary slide rule are shown in Figure 32. You will
notice two things. First, the secondaries and tertiaries
get more crowded as one moves to the right. This is to be
expected in a logarithmic scale. The second is the con-

sequence of this, for the tertiaries represent different

values in different portions of the scale.

This should not be surprising. In an ordinary scale'
such as the L-scale, the spacing between the primaries
is equal throughout and secondaries and tertiaries can

be distributed evenly all along the line. In a logarithmic

I
• The difficulty of making an accurate logarithmic scale con[
tributes to the expense of a good slide rule.

Figure 32

i
 1

{|l|liM|iMI|ilil|iMi|liii|Mii|iiii|
'
4
I
I
I
I

'56 769
M M M MM I
I
•
I M '
I
I
M I
I
1
1

4MlliMiliiMliiiilniiliiiiliiiilMiiliiHliiiiliiMliiiiliiMliiiHMiilHiiliinliiiJiiiiliniluiihiiiliiiilinitiniltiiilinHii^
1 M M
>
I
>
M 1
1 1 M > 1
1 PI >IMM>l<l'l'IMi|'|i|'|i|i|i|'l'IMM'ii|'l'l'l'|)|>l'l'lfl<l'l'l'l

Figure 3

scale, however, the space between the primaries steadily

decreases as we move to the right so that room can be
found for fewer and fewer subdivisions.
As it happens, secondaries representing 0.1 each are
distributed all along the D-scale but get progressively

closer, and the space between them available for ter-

tiaries gets progressively skimpier. Consequently, there
are tertiaries of three diflFerent kinds present on the
D-scale.

j
Consider first, the space between D-1 and D-2, which
takes up nearly a third of the total length of the D-scale.
It can be divided up finely. The secondaries, represent-
ing tenths, are so relatively far apart that it is hard for
the eye to take them in at a glance. For that reason ( and
 Figure 33 -
UrtiariM 0.01 ««rticriM

because there is enough room for it), the secondaries

between D-1 and D-2 are marked with the numbers 1

through 9, numbers of smaller size than those marking

the primaries ( see Figure 32 ) . The small number 1

marks the 1.1 value, the 2 marks the 1.2 value, and so on
up to the 9, which marks the 1.9 value.
The spaces between these secondaries is divided into
ten smaller divisions. Thus, there are tertiaries, each
representing an increase in value of 0.1, running from
1.01 for the first tertiary after D-1 to 1.99 for the last

tertiary just before D-2. The middle tertiary between

neighboring secondaries, marking numbers such as 1.05,
1.15, 1.25, and so on, is longer than the others.
In addition, one can place the hairline midway be-
tween two tertiaries so that one can indicate 1.005 or
1.235 without trouble, or make rough estimates for 1.007
or 1.233.
Beyond D-2, the secondaries do not possess numbers
actually marked on the scale. If numbers were present |

they would be crowded too closely for clarity. On the|

other hand, the middle secondary in each case, repre-

senting 2.5, 3.5, 4.5, and so on, is made longer than the!
rest, and using these as guides, the value of any other |

secondary is quickly seen.

76
 S.70 S.75

|M. |MM|MH|MMM| M| | M 1 1
.
1
1
1
1 1
•
|
•
I M M M
•' I ' I ' I ' I ' I •
|
• MM HIM'MI'l<|i| <|i|i IHIMMMMIIMIIM'IIIMMIIIMI

t«rti«ri«t - 0.05

Since the stretch from D-2 to D-4 is just equal to that

from D-1 to D-2, there are twenty secondaries crowded
into the former stretch, where only ten exist in the latter.

[For that reason, there is no room in the D-2 to D-4

stretch for the secondaries to be conveniently divided into
ten smaller divisions. Five smaller divisions are set up
instead and each tertiary in this stretch therefore repre-

isents a value of 0.02. Thus, between D-2.1 and D-2.2

I are tertiaries representing 2.12, 2.14, 2.16, and 2.18. By
centering the hairline between such tertiaries, one can
I get 2.13, 2.15, and so on.*

The stretch from D-4 to D-10 is divided into sixty

secondaries that crowd closer and closer together. So
close are they that each space between secondaries in
this portion of the D-Scale is divided merely into two
smaller divisions, with a single tertiary representing a
value of 0.05. Thus, between D-4.3 and D-4.4 is a tertiary
indicating 4.35, and between 9.6 and 9.7 is a tertiary in-
dicating 9.65. By centering the indicator between a
tertiary and a secondary, one can obtain numbers such
as 4.325 or 9.675.

* Near the tertiary 3.14 is the special marking w. It is not

present in all slide rule designs. Its use will be discussed on
page 99 .

77
 To summarize, we can divide the D-scale into three
segments in which the value of the tertiaries differs ( see
Figure 33). From D-1 to D-2, each tertiary represents a
value of 0.01; from D-2 to D-4, a value of 0.02; and
from D-4 to D-10, a value of 0.05.

Naturally, when one is not accustomed to a logarithmic

scale, this change in the nature of the tertiaries can
be confusing. There will be some stumbling at first,

inevitably, but continued use of the slide rule will

eventually make it all seem natural and there will be no
trouble in reading the D-scale, no more than in reading
the markings of an ordinary ruler.
It might seem to you from all this that the left end of
the D-scale is more delicate than the right end, since
the left end has tertiaries that represent finer divisions.

Let's look into that situation.

Suppose you are trying to read the hairline position on

the left end of the scal^, and find it just about halfway
between the tertiaries 1.34 and 1.35. You conclude that

the reading, therefore, is 1.345. However, it may seem to

you that the position is just a shade to the right of the

midpoint, so that the reading might be 1.346. You can't
really tell exactly which of these it is so you have an un-
certainty of 0.001.
On the right-hand side of the scale you might make a
reading of 9.475, because the hairline is just halfway
between the tertiary marking 9.45 and the secondary
marking 9.50. However, again you cannot be sure that
it is exactly in the center. The space between 9.45 and
9.50 (a difference of 0.05) is only about two-thirds as
wide as that between 1.34 and 1.35 (a difference of only
0.01) and its exact midpoint is harder to judge. If the

78
 )

hairline seems a bit to the right of the midpoint, the

reading can easily be 9.485 rather than 9.475, an un-

certainty of 0.01.
Since the uncertainty at the left is 0.001 and the uncer-
tainty at the right is 0.01, you might think that the
accuracy at the left end is ten times as great as that at
the right end. That, however, is not so.

What counts is not the size of the uncertainty in itself,

but its size compared to the size of the number being

measured. What we ought to consider is the uncertainty
percentage. Thus, if you consider an uncertainty of 0.001
in a number such as 1.345, the uncertainty is about 0.075
percent of the number. An uncertainty of 0.01 in a
number like 9.47 represents an uncertainty that is about
0.105 percent of the number.
These percentages are not very different. We can
say, with reasonable confidence, that the accuracy of
a slide rule is the same all along its length and that it is

everywhere easily reliable to about a tenth of one per-

cent, or one part in a thousand. (A very careful user
could squeeze out an even better accuracy.

Using the Wooden Log Table

Let's consider the L-scale and the D-scale taken

together. What can be done with them? Obviously, the
two scales have been constructed in such a way as to

give us a three-place logarithm table, with the logarithms

in the L-scale and the antilogarithms in the D-scale.

Suppose you want the logarithm of 3.5. You place
the hairline ovei* D-3.5 and find that it also marks L-544
( Figure 34 ) . You can conclude then that log 3.5 = 0.544.
79
 D
I

1
I I 11 I
I I I I

lasifSTiki
I
I I I I |lll l|llll|llll|lllllllll|llll|llll|llll|MII|IIU|IHI|llll|llll|llll|MII|llll|llll| I I I I
I
M I l| lll|llll|MIIII|||
'

liiiilMiiliHiliiMliiMliMiliMihiiilniiliiiiliiNliiiiliniliiiiliMiliiiihiiilMiiliiiiliiiiliiiiliiiiliiiiluiiliiiiliiiiliiiiliiiiliiiiliiiiliililiiMluiibinliniliii

Figure 34

If you were to look up the logarithm of 3.5 in a five-place

logarithm table, you would find that log 3.5 = 0.54407 so
that your slide rule answer is certainly satisfactorily close,
particularly since all you need to do to get it on the slide

rule is to shift the hairline. It is a much simpler task

to do that than to leaf through an ordinary logarithm
table and run your finger down columns of numbers.
Of course, the L-scale gives you only the mantissa,
but that is all you need. If you wanted the logarithm of
35, or 350, or of 0.35, you would still place the hairline
on D-3.5 to get L-544. You then merely adjust the char-
acteristic according to the principles described on page
54. We then see that log 35 = 1.544, log 350 = 2.544,
log 0.35 = 0.544 — 1, and so on.

It works the other way round, too. Suppose you are

given the logarithm 3.728 and want the antilogarithm.
To use the L-scale, which gives only the mantissa, you
drop the characteristic temporarily and search for 728
only. You place the hairline at L-728 and find that the
D-reading is just a shade to the left of 5.35 (see Figure
34 ) . You estimate the reading to be 5.345. The charac-
teristic tells you where the decimal ought to go. Since
80
 /Of. 3J 0.544

3J D.5.346
• •

IM|tHI|HII|IIM|MIIJ 1 1
l|l| IM|*I*I>IMM<|)MI |i|i|i|i|i|m|i|i|i|i|i|i|t|i|m|i|i|i|i|i|i|i|i|i|i|i|i|i|i|i|i|i|i|iii|i|i|i|i|i|i|i|i|i|

\* 1 -7

t i

44 L.728
Anfilog. 3.728 = 5345

5 characteristic is 3, the number of digits to the left of

5 decimal point must be 3 + 1, or 4. Consequently,
Hlog 3.728 = 5345. A five-place logarithm table will
I you that antilog 3.728 = 5345.4.
As you see, then, the D-reading is not fixed. What
) read as 3.5 can be 350; and what we read as 5.345 can
5345 or 0.0005345, for that matter. A reading on
J D-scale gives us a fixed set of digits, but it says
:hing about the position of the decimal point. That
ongs to us.
For this reason, I will no longer try to place a decimal
int in the D-reading. At first, it may seem very con-
lient to consider the secondary marks as representing
nths" and to read them as 1.1, or 2.4 or 9.6, but I
II consider such readings as D-11, D-24, or D-86 re-
jctively. In the same way tertiary marks will not

)resent 1.46 or 2.46 or 9.55. Thev will be D-146,

246, and D-955 respectively.
Ifou may feel lost without the decimal point just at

t, but you will get used to that too. And the loss is

ly temporary. As you will see, once a slide rule compu-

?on is completed, in will go that decimal point.
 5

Multiplication

The C'Scale

We have now converted the slide rule into a logarit

table and, having done so, w^e find that we don't res
need to l^bor in order to use it. The slide rule makes
use automatic. Whenever the haiiline is placed on gj

D-reading, it simultaneously marks off the logarithm!

that number as the L-reading. We don't have to 1

for the logarithm; it is there.

This helps us understand a crucial difference betwd

the evenly spaced scale on an ordinary ruler (such
the scales we used to work our addition rule in Cha
2) and the logarithmic D-scale.
In an ordinary scale, the numbers mark off lengj

that correspond in value to the numbers themselves,

number 2 is two inches from the left end; the numbe
is five inches from the left end and so on. By adding
subtracting these lengths, therefore, we add and subtn
numbers (see page 11).
The numbers on the D-scale, however, mark off leng
that correspond not to the numbers themselves, but
the lo^aritluns of the numbers. On the L-scale, le

•r
primaries are usually placed about an inch apart (e
page 62). That means that since D-1 is placed over Iftj

82
i D-2 is placed over L-301, D-2 is about 3 inches from
I left end of the scale. Similarly, since D-3 is under
177, D-3 is about 4% inches from the left end of the
lie. These distances, I repeat, are not equivalent to

J numbers themselves but to the logarithms of the

mbers.
Consider, then! When we used an addition rule
:h its scales representing lengths equivalent to the
mbers upon it, we could manipulate those lengths
as to add and subtract numbers. But if we use some-
ng like the D-scale in which the numbers represent
gths equivalent to the logarithms of those numbers,
can manipulate those lengths so as to add and sub-
ct logarithms. In adding and subtracting logarithms,
are, of course, multiplying and dividing the anti-
;arithms — that is, the numbers on the D-scale.
Fo perform such manipulations on the addition rule,

needed two scales of identical construction that could

moved against each other. We need the same now

two scales of identical construction that can be moved
linst each other.
rherefore, on the bottom of the slide, in the C-position
je page 66 ) , another scale, exactly like the D-scale, is

istructed. This new scale, the C-scale, can be placed

ictly over the D-scale, with every mailing on the
mer over the corresponding marking on the latter and
can then say the slide rule is in a neutral position

igure 35 ) . The C-scale can, however, be shifted out

neutral and moved along the D-scale in such a way
to add and subtract logarithms.
We can begin with a very simple: 2 X 3.

To do this, we follow a similar procedure to that used

83
 f

c 1 1 > )4SS7IS2
D 1 1 2 a4SC7l«2
|iiiijiiii|iiii|iiii|iiii|iiii|iiii|iiii|iiii|iiii|iiii|iiii|iiiiM^^^^^

Figure 35

in adding numbers on the addition rule ( see page V

We begin by moving C-1 over D-2.
And this at once brings up an important point. Th(
are two C-l's, one at the left end of the scale and one!

the right end. These are used so often in slide rij

manipulations that they are often referred to as the l\

index and the right index, respectively. In this bo(

how^ever, in order to avoid confusion, I will give

readings any special name. I will refer to the 1 at tj

left end of the C-scale ( the more frequently used ) as C

The 1 at the right end of the C-scale, I will refer to as O
(right). In the same way there is both a D-1 and a D
(right).

In placing C-1 over any D-reading, it is not necessi

to use the hairline as a guide. Standing as it does at tj

very beginning of the scale, C-1 is clearly marked d

and the eye can follow it without trouble. Markinj
along the interior of the scale are easy to confuse wij

their neighbors, however. Readings between markinj

are even easier to confuse. It is routine, then, to pla(

the hairline over any reading one wishes to make thai

does not involve the end-marking of a scale.

With all this in mind, we can return to the multiplicl
tion 2X3. Carrying it through by the system we us<
 .

I
A 5 6 7 i f 1

:he addition rule, we place C-1 over D-2, then shift the
rline over to C-3 ( Figure 36 )
Jince D-2 marks off a distance of log 2 from the end
the D-scale, and C-3 marks off a distance of log 3 from
J left end of the C-scale, we are adding log 2 and
; 3 in this manipulation. Since log 2 + log 3 = log 6

I
id you can check this in a log table ) , we would expect
find D-6 under C-3, for D-6 would be a distance of
i;
6 from the end of the D-scale.
We do not, however, bother reading the slide rule in

3 position shown in Figure 36 as an addition of loga-

hms. We do not say: log 2 + log 3 = log 6. The D-
de gives the antilogarithms and we read those anti-
jarithms directly, converting the addition of loga-
hms into a multiplication of antilogarithms. We say,

<3 = 6.

The same thing would happen if we wanted to mul-

•ly 3X2. We would place C-1 over D-3, then move
3 hairline to C-2, and find D-6 immediately under C-2.
lis is not surprising for log 2 + log 3 = log 3 + log 2,

d 2 X 3 =3X 2.

This is simple so far. We scarcely need a slide rule to

1 us that 2X3 = 6. However, what if it is 2.54 X 3.76
*t we are interested in. The multiplication is now
85
 i^mmm^KfivvnmiiHnmi

E "•

1 12S&ftT«ii2
"T lll||l'lll|llllt H II|II M |llllfl l l|M. nn|H..|Mii| .|iM inniiMin ill I ll|llll|llll|llll|lllllllll|||||i
J

'Off 2

/Off 2 + /Off 3

loff 2 + /Off 3 - li^l

2 X '3
- 4 t»

much more difficult on paper, but not a bit more difficu

Dn the sHde rule.

Move C-1 over D-254 (taking a Httle care to ma
sure you have the right tertiary ) , and then move the hai^j

Hne to C«376. You find the hairHne to be over D-955 (sq

Figure 37) and conclude that 2.54 X 3.76 = 9.55. Woi,

it out in full and you will find that the correct product
9.5504, but surely 9.55 is close enough — and think c|

the saving in time.

But what if it were 2.54 X 4.76 you wanted to solv

You begin, again, by moving. C-1 over D-254, but th

time you find you are stuck, for the C-476 which y
must reach with your hairline is off the D-scale and c

CI

Figur« 37 D-2i4

86
 5 i ? I
7
ihlilililililililiTilihlil lihliliMih'ihlt,
v/|i?/|iyi|iv/|ivi|iv/|iv/|iff|ivi|iv/|,ii|.ii|.|.i,|.iiU^

Figure 36

e you no D-reading.
ih, but we went through that in the addition rule
e page 21) and the same device will serve us here.
'
make use of the C-l( right) in place of the C-1 and
that over D-254. Now we can find C«476 over the
cale. The nearest tertiary is C-475 so we place the
rline just a bit to the right of that, about one fifth of
way over to C-480. means a kind of ( Naturally, this
iging by eye" but this can be done pretty well,
ecially with practice. ) The hairHne, on C-476, also

s just to the left of D-121 (which can also be read

D-1210). We might judge the mark to be D-1209
gure 38), and conclude that 2.54 X 4.76 = 12.09. If

C-374

iJilili iliililililiTthlilililihh
{/rvVvAHYi^tl1?TlJ^'i'i\\^\)\li\iil|^^

D-9S5
2M X 3.7* - f JS
 "

C-476 CI (ri)

3 It
,
I ,
il' ii|li'iiliinliililinil|i|llllMlllllllllill|ll I I I I I I I I I I I I I I

Ti
I 1
1 1 > 1 1
1 1
> ' '
' I > *

1 t

D-1209 D-W
2.54 X 4.76 = 12.09

we work the multiplication in full, the product tui^

out to be 12.0904.
11

The Decimal Point

This brings up the question of the decimal point. !

^

the example just completed ( 2.54 X 4.76 ) the slide ri IP

gave us the digit-combination 1209 as the answer. V^

might have decided that this meant 2.54 X 4.76 = 1.2C

or that 2.54 X 4.76 = 120.9, but we didn't. We sa*'^
that 2.54 X 4.76 = 12.09. Let's see why.
Consider the following multiplications:

X 1.5 3.23 = 4.845

15 X 32.3 = 484.5
0.015 X 0.323 = 0.004845
150,000 X 0.00323= 4845
to
150 X 323 = 48,450
Ea

This shows us once again that the decimal point is]

side issue. If we go by the slide rule we deal only w^

digit-combinations and all the examples above boil do^
 1

'

]
1 1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1 H I
1 1
1
1
1 n 1
1
1
1
1
1 u 1 1
l

y f f f >f>(>[if I f >i'i T i|iiiiiiiii ii>

|
w j
iif ifi)i(»|i w ^

8 '
d 1

Figure 38

15 X 323 = 4845. The decimal point — that side issue

- is for us to handle.

There is, however, no need to feel aggrieved, for,

lOugh placing the decimal point may be a tedious ne-

»ssity, it is not difficult. Occasionally, a bit of careless-

^ss will result in a misplaced decimal point, but this can

ippen even in pencil-and-paper calculations. The cor-

ct response to carelessness is a sober determination to

t careful, that's all!
*

The best way to place the decimal point is to consider

e problem and substitute, for the numbers involved,

nilar numbers that are particularly easy to handle,
ich similar numbers will give you an answer that is

rong, of course, but one that is close enough to the

ght answer to have the decimal point in the same place.

3u will have an answer that is of the same order of
agnitude.
Each shift of a decimal point by one place produces a
lange of one order of magnitude. For instance, 2.54 is

le order of magnitude smaller than 25.4 and two orders

'magnitude smaller than 254. Again, 2.54 is one order
89
of magnitude larger than 0.254 and two orders of magn
tude larger than 0.0254.
If two numbers differ by a factor of less than 5, the

can be lumped together as of the same order of magn

tude. Thus, 17 and 68 are of the same order of magnitud
since 68 is only four times as great as 17. On the oth
hand 17 is not of the same order of magnitude as 2.1
since 17 is eight times as great as 2.14. However, 17 ij

of the same order of magnitude as 21.4 which is, in turn]

one order of magnitude greater than 2.14. Therefore

we can say that 17 is one order of magnitude greate
than 2.14.
Now let's look at one of the problems in the list give

on page 88 , say, 15 X 32.3. We can convert 15 to 2i

fV

and 32.3 to 30. The numbers aie changed but not th

order of magnitude. It is easy to multiply 20 by 30 ii

our head. The answer is 600. That is not the answer wi

are looking for, but it is the same order of magnitude a
the answer. If the slide rule tells us that the digit-combi
nation of the product is 4845, then to make that the sam(
order of magnitude as 600, we must write it 484.5. W
know then that 15 X 32.3 == 484.5.
A similar treatment will give us the correct answer fo

any other example in the list.

And it will also give us the correct answer to th(

problem with which I started this section, 2.54 X 4.7

If we make the first number a little higher, changing it t

3, and the second a little lower, changing it to 4, w

retain the order of magnitude. Since we know tha
3 X 4 = 12, we know that the slide rule answer of 12
must be written 12.09 to keep the product in the correc

order of magnitude. Therefore, 2.54 X 4.76 = 12.09 anc]

90
 ) )

thing else.

Practice is all you need to learn to handle orders of

ignitude almost automatically, and once you've worked
at out in advance, you can manipulate the slide rule, get
ur digit-combination, and place your decimal point
thout any hesitation.
Let's take another example, and a slightly more difficult

e.

Suppose you wanted to carry through the following:

'1 X 7.23 X 1.15 X 0.86. This represents three multi-
cations to be carried out one after the other.

You can begin by placing C-1 over D-272 but then you
11 find that C-723 is off the D-scale. You therefore
vitch indices," bringing C-1 (right) over D-272, and
rrying the hairline to C-723. (To find C-723, you must
ice the hairline between the secondary representing
and the tertiary representing 725 — and place it a
:le closer to the 725 than to the 720.

Having done this, you will find that the hairline is also

uking out a reading just under D-1970. You don't,

wever, have to try to estimate what the answer is, or

en look at it. The hairline marks the product of
^2 X 7.23 on the D-scale and keeps marking it as long

you leave the hairline in place. This product must next

multiplied by 1.15.

You therefore bring the C-1 to the hairline (making

j:"e you don't move the hairline position in the process

,d this places C-1 over the D-reading that is the product

2.72 X 7.23. To multiply that product by 1.15, you
)ve the hairline (which you no longer need in its first

isition now that the C-1 marks it) to C-1 15. That marks
lew D-reading which represents the product of 2.72 X
91

;
 )

7.23 X 1.15.

This new product must be multiplied by 0.86. Y<

therefore bring C-1 to the hairhne marking that produc"

and find that C-86 is off the D-scale. You bring C-1
( right ) to the hairhne instead and then move the hairHn.^
to C-86.
Now the hairhne marks out a D-reading just about mi<
way between D-194 and D-195. Call it D-1945. That \
the digit-combination that represents the product
2.72 X 7.23 X 1.15 X 0.86 (Figure 39).
But where is the decimal point? If we look at th
multiplication problem again, we can change 2.72 to
7.23 to 7, 1.15 to 1, and 0.86 to 1. The problem become
3X7X1X1 which you can see at a glance equals 2i
To give 1945 the same order of magnitude as 21, w
must write 1945 as 19.45. Therefore 2.72 X 7.23 X 1.11

X 0.86 = 19.45.

( I have, by the way, just worked out the triple multipli

cation by pencil and paper as quickly as I could and
took me two full minutes to get the answer 19.4492784.
worked it out slowly and carefully by slide rule and il

took me just 15 seconds — one-eighth the time — to gefj

19.45, which is almost exactly the correct answer.

Folded Scales

Sometimes, as you saw in the case discussed just abov^

you must use C-1 and sometimes C-1 (right). Wit]
practice, you get the "feel" for which one to use.

Suppose, for instance, you consider 2.14 X 37.6. Yoj

can adjust these figures to their easy approximations of
and 40 and see, at a glance, that 2 X 40 = 80. In thj

92
se, you will recognize that it is safe to bring C-1 to
•214 for you expect C-376 to be somewhere near D-8,
kich would keep it safely on the D-scale. Sure enough,
you perform the manipulation, you will find C-376 over
-805 so that 2.14 X 37.6 = 80.5.

If, on the other hand, you were dealing with 4.14 X

76, which is close to 4 X 4, which in turn is equal to 16,

u would expect the answer to be beyond D-1 (right),

lich would represent a digit-combination of only 10.
)u can consider D-1 itself to be 10 also, of course, but
that case 16 would be near the left end of the D-scale
d must be sought there. Therefore you bring C-1-
ight) to D-414 and then find C-376 over D-1556 so that
L4 X 3.76 = 15.56.

However, it is always possible to be fooled. Suppose

u try 13 X 79 and decide to approximate it at 10 X
= 800. Therefore you move C-1 over to D-13 expect-
g to find C-79 somewhere near D-8 and safely on scale.

m find instead, to your horror, that C-79 is just beyond

e end of the D-scale. You must therefore switch indices
d bring C-1 (right) all the way over to D-13 and then
id C-79 over D-1027 (Figure 40) and decide that
; X 79 = 1027.
Whenever this happens, there is bound to be a certain
nount of muttering under the breath.
It would be nice if this could be avoided. Suppose, for
stance, that the D-scale were continued past D-1-
ight) so that there were a D-2( right), a D-3( right)
id so on. Then, if a C-reading moved off the ordinary
rscale, it could be picked up on the D-scale extension.
ae disadvantage of this is that the slide rule must be
iigthened, making it both more expensive and more

93
 )

C-723

M I I I I I I I I I ll >l I ihilllll lIllMlMdllMltnMllMllllllhllliliMilMiiliinl

.
ll H*I|IHiIiI|Ii|iIiIiIiIiI|Ii|iIi|Ji|iI|Ii|i|iI|Ji|iI|I|
I I

n
i|i II i|ii)i|ilii|liii|iiii|ilii|liii|iiii|ilii|liii|iiH|ilii lniliiW|Mn|l(ii|MilMl
I
,\i','l','iyh'itii'/i'/|i/wi'

D-(272 X 723)

C-l C-IIS

fiffffffWffWfffiTfffffffftft 1ii|iiii|m(|iiiimii|iiii|iiii|iHimM|MiiiiiM(Miiliii|i A////////fi(ll ili|iili|nfi)uiii|

D-(272 X 723) D-(272 x 723 x I

C-16 C-I(ri9iil

3
I . I
n
,, I.., I' .i,i..i».<i
^
.1 . I I I I

i
I I , I I I >
I
<

1
I
I
5,6,7,8,
I

2SiC«7«9
> h lil|lMMl||llllll|llllll| llllll||llllll||lll (llllllll lIlllllllJ!
1''/ll'lt,'iVi',ll llllil

D-I94S D-(272 x 1]

2.72 X 7.23 X LIS x OJ

unwieldy. (In some slide rule designs, however, paj

ticular scales are extended about half an inch or so pj

either end to take care of borderline cases.

94
 mmmtmmm

,....) i..njnn 'ii' r i

M |
i i'iii in i iiii
i i iiii |i

6 7 '8*91
l"MI'l"I"l"I'l"l"|'l'l"l'l'l"l"l"IM'l'l'l'l'l'IM'IM'l
1

'""""*' ' ^' ilihlih

m\Wn\\Wnvlfl\\M^^^^

ff ll lllljlfll|lllli l lll llll

|
|

Figure 39

Another way out is to move the D-scale bodily to the

eft so that D-l( right) is moved leftward, leaving room
eyond it for an extension without making the sUde rule
95
 .

C79 CI (right)

I , I . < . I I , I I Iil.hlilililihhlililililililililililililililil
lll|llll|llll|HII|IIM|IMI|ltll|IHI|llll|llll|llll|llll|ll
M t,'l',1'.'|','lt,r.l:'l',1'/l'/it.'l','

DI027 DI3
13 X 7<

any longer in that direction.

If you keep the rest of the D-scale intact, however, i

then begins to stick out at the left end like a tusk. Ii

order not to extend the slide rule in either direction, wi \

must arrange to have the left end of the D-scale disappea

as it moves leftward. The additional room at the right fc

an extension must then be made up for by a disap

pearance of the scale at the left ( Figure 41 )

As you move D-l( right) leftward — provided yotp"

keep the overall length of the scale fixed — the extensioi]

you gain on the right is exactly balanced by the loss o

the left. It follows then, as you can see in Figure 41, tha

the reading on the left end of the scale is always the sam
as on the right end of the scale. If the scale begins witl
D-2, it ends with D-2( right ) ; if it begins with D-4 or D-8
it ends with D-4 (right) or D-8 (right).
As you move D-l( right) leftward, you experience j

gain because of the extension on the right, and you als^

experience a loss because of the disappearance on th<

left. To increase the gain as much as possible you wanj

to push D-l( right) continually leftward. To minimize.

96
 ipwpnPHViPHMi^i^pm

i
i
i
im
TTTnpTiriiMmMim|im,.M.,....|..M,....,.Jlfi|l|l|l|l|rM'|ITM!|l|^
rnTTimii|iiii|iin|iininii|iiii|iiii|iMi|Miii i
I I I I I I I
• l ' l • l
] 'I'l'l'i'i'ri'ri'iM'i'i'i'i'ri'i'riTiM'i'i'

Figure 40

e loss you want to push D-l( right) continually right-

ard. The two impulses balance, as you might expect,

ecisely midway and the ideal situation is to have D-1

actly in the middle of the scale.

If D-1 were exactly in the middle of the scale what

Duld be the numbers on the left end and right end?
whatever number would be at the left end would, of

2 3 4 5 6 7 8 9 1

|ure4l

1 1 1 1 1 1 1 1

3 4 5 6 7 8 9 1 2

1 1 1

5 6 7 8 9 1 2 3 4

n
91 2
\ \

3 45678
\
 ).

OF
ill|llll|llll|ll!MiM^^^^
CF

C V 23iSS7IS2
D 123«Sfi7«92
}
jiH|iiii!iiii|iiii|iiii|iiii|iiii|iiii|iiii;iiii|iiii|iiii|iiii|iiiM^^^^^^^^
1

course, also be at the right end.

^
To answer this question, let us remember that th<

entire stretch of the D-scale represents the stretch of th

L-scale from 0.000 to 1.000 (see page 67 ). If D-1 is i|i

the middle of the scale, the logarithmic stretch from thapi

middle to the right end is, obviously, 0.500 or, in fractions;
/!». Earlier in the book, I pointed out that 10^ = V^, ol

3.16. Therefore, antilog 0.500 = 3.16.

The numbers on the D-scale are placed at the distano

from D-1 that corresponds to the length of the logarithm

of those numbers. Therefore the right end of a D-scale

that has D-1 in the middle must be 3.16, since the
distance of 3.16 from D-1 is then just half the length ol

the scale or 0.500 ( in logarithms ) and that is, indeed, the

logarithm of 3.16. Consequently the left end of the D-
scale with D-1 in the middle must be 3.16 also. Omitting
decimal points, a D-scale with D-1 in the middle should
run from D-316 to D-316 ( right ) !

Such a scale is called a folded scale, the notion being

that an ordinary scale, folded in half, has 316 appearing
at the fold, and the new scale staits and ends at that fold

In my slide rule, the folded scales are present in the A^j

and B-positions (see page 66), but they are not referre

to as A- and B-scales. Those names are reserved fo

other scales which I will discuss later in the book.
98
 12945S7«92 3ll'
;

i
12345S7e92
ilil|llll|llll|llll|llll|llll|llll|llll|llll|llll|llll|llll|llllM^^^^^^
'
3*

iiiiliiiiliiiiliiiM^^^^^^

}
'
L 5 6 7 8 9 1

Figure 42

Instead, the scale in the A-position is the DF-scale,

^hile the scale in the B-position is the CF-scale ( Figure
2 ) . The DF-scale is so named because, like the D-scale,
is on the body, while the CF-scale, like the C-scale, is

n the slide. The F, of course, stands for "folded."

'i

If you look at Figure 42 carefully ( or, better still, look

t a sHde rule) you will see that the DF-scale and the
]F-scale do not actually start and end with 3.16 but
ath 3.14 and that the two ends are marked thus: t:.

Behind that lies a story. In the case of every circle,

owever large or small, the length of the circumference is

bout 3V7 times that of the diameter. The actual ratio is

never-ending decimal, but, to five decimal points, it is

.14159 which, as you see, is not far removed from 3.14.

his ratio is customarily symbolized by the Greek letter

( Pi ).

It so happens that t: is not confined merely to questions

ivolving the length of the circumference of a circle in
?rms of its diameter. It crops up in numerous mathe-
natical equations and comes into play in almost every
acet of science.

Consequently, scientists and engineers frequently find

99
occasion to multiply r. by some number. That is the
reason for the special r. marking on the C- and D-scales.
It is also the reason why the DF- and CF-scales are de-
signed to start with t:. Although those scales should,
ideally, start with 3.16, it is not much of a departure from
the ideal to have them start just to the right of 3.14
instead. It is an alteration of not quite a single tertiary

and the insignificant bit of imperfection introduces such

a great deal of convenience that no one would object.

To see what this means, let's begin by noting that C-1 is

directly under CF-i:. This is a fixed position, for both the

C-scale and the CF-scale are on the slide and neither can
be moved relative to the other. This is also true for the

D-scale and DF-scale which are both on the body so that

one can't be moved relative to the other. Therefore, as-
suming both parts of the body to be in perfect adjustment,

D-1 is permanently under DF-ir.

Next, let's consider the hairline at some point on the
C-scale — any point. Let us say the reading is C-a. At

the same time, the hairline is indicating a reading on the

CF-scale; say, CF-b.
We already know that the C-a reading is at a distancel

log «

CF-7

.i|iMi,...i|iMiM.M|Mii|.iii| i>|i|i|i|i|i|>l<l>l>l<ii|i|i|i|i|i|i|i|>nMliMIMM'MMl1
II I

4
.
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I , I

'
7
'

5 S

I • I I I I I I I
1
I M 1 1 1 1 1 1 1
23(iS7l92
1 1 M 1 1 1 1 1 1 1 1 M iliintniiliiiiliM''*'iiliiMliiMtiiiilitiiliiiiliiiiliiiiliiiil ( I I I I I I 1 1 1 1 • 1 1 1 1 1 1 1) illll

CI
4- log •
from the left end of the C-scale equivalent to the log-

arithm of a. Since the left end of the C-scale is C-1, which

represents a logarithm of 0.000, the logarithm of the C-a
reading is 0.000 + log a, or simply log a.

The CF'b which is simultaneously marked off by the

bairline also represents a logarithm, which we can call log

b. But CF'b (marked out by the hairline simultaneously

with C-fl ) must be as far from the left end of the CF-scale
as C-a is from the C-scale. Hence CF-fc is a distance from
CF-TT that is equivalent to log a. But CF-t: is equivalent
to log r.. Therefore, log b = log u + log a. If we convert
those logarithms into antilogarithms (and remember to
change the addition into a multiplication in so doing ) , we
Snd that b = TzXa { Figure 43).
If this is not instantly clear to you, it will become clear
IS soon as you put the notion into actual practice.
Suppose you want the value of 2t:. You have only to
move the hairline to C-2, and you will find the hairline
.will simultaneously give a reading of CF-6.28, which is

equal to 27:.

In the same way you can find with a single setting

of the hairline that 4.37: = 13.51, since, when the hair-

CF-b log b = io9 ir + lo, .: b = IT X .

Figure 43

Mil II

2 3 45S7IS2
ll|lilll|llll|llll|llll|llll|llll|llll|llll|lltl|tMI|llll|llll|llll|llll|llll|llll|llll| I I ll|llll|llll|llll|llll|llll|llll|llll|llll|llll|llll|ll
'
it

4IIIII iiiiliMiliiiiliinliiiil 1 1 1 h lililililililiTilililililililililiTilililililMihlilililililililililililililihlililililililiTilililililililililil

Cm
jl/^ , log « = log I + log •

101
I
 1

OF liy^^gj^yg^yjux^iiLduiidiiiliiiiiiiiiTiiiiiiiiiliiiiiiiiili
ll//M illl)| ^t |l ll
W w|
lll )

CI

iirii |iMi|iiM|li |mT|nnff)ii|iiii|iiii|iiimiii|iiMMMt|iMiiiMi|i)ii| m I/'-''Aillf/|'i'i\l|l///if\ll|li

*•
i < 1 1 • 1
1 1 1 1 1 1 1 II 1 1

1 t

4
D-22*

line indicates C-43, it also indicates CF-1351. Asj|

other examples you can find that 75.2 r. = 236, thatj s

0.3887:= 1.22, and so on. The placing of the decimallj

point is no problem, for you have but to remember that i

multiplying by ir is not much different from multiplying

by 3.
^
Avoiding the Shift
y

But multiplication by t: is a side issue. When I started

discussing the folded scales, it was not with such a multi-i

plication in mind. The folded scales were introduced im

order to make it unnecessary to shift from C-1 to C-1-
( right ) in ordinary multiplications. ( Or at least to reduce!
the number of times the shift must be carried through,
if the necessity isn't eliminated altogether.)
In order to go back to this earlier problem, take another

Rgur* 46

f OF 1

iliyjiiiiliii
1 1 1 1 1 1 1 li III
lill.!jl!lijjj.l.l.!l.lllllllll!ttti<ijil'''''''^''''''''

ffipipipPPPpy^^ ||(
fl

102
 ' 1

CF-224 DF.294

!liiiV//iiii\'.^/ii|iv//i(iibiiij.!l\'iiir.f/ii^^ V.V|'llU\lV/l|l\V.V/lllV.V/MlIV,'llli,^...,.M,n n.,.,....|... |iiii|tiii|itii|iiii| 1 1

L
1
1
1 II 1
1
1

||p|,mryyl,.,|y.|.,.|T^,||||\ym^
lljJUUlllulllUl|l|lllUlJl|l|T. .•JtllluTl.U liiiiiiiiiTihiiiiiiiiiiiiiiiiTiiiiiiiiiiiiiiiiiiii'

Figure 44
r
look at the CF/DF scales as compared with the C/D
5cales. When the sHde rule is in its neutral position, C-1
is over D-1, and CF-1 is under DF-1, as shown in Figure
42.

If the slide now moves in such a way as to displace the

C-scale aganst the D-scale, it also displaces the CF-scale

against the DF-scale. And since the C-scale and the CF-
scale are on the same piece of wood, the amount of dis-

placement must be precisely the same for both.

Suppose, for instance, that you move C-1 over D-226.
^ou will find that CF-1 will move directly under DF-226
( Figure 44 ) . Furthermore, if you look at other portions
)f the scales, you will see that there are similar situations
ill along the line. Thus, C-13 is directly over D-294 and
CF-13 is directly under DF-294, and so on for any other
eadings you care to make.
Since the readings on the C- and D-scales are exactly

WAf4w*wi^^'i'T/ivvi'i'i'/i'i\'r 1 1 1
1
1 1 1 1
1
1 II ii 1 1 1
1
ii I
itii iii|i<iii'
1 2 i

. numbers ever 314 uubU-

numb«rt undar 314 utabU

' '

jaA|AI.V.I/|IAI|ll.l^l4Mi|MA'l\MVlr«iV|'IVI'('l7\V"'|''VI|T/l^^^

m
103
 1

OF jjji^^jiiJiminiiTilililililililililiTililililililihlililili ;\l|l|||||M*
wm iiiiiiiiiiiiiiiiii

U
C-l

iy>wwywww f iiiiiii ] i
i iiiiii i
iii ff|iiii|iiii|iMi|iiii|iiii|iiii|iiii|im|iiH|
|IIII|IIHI I n 1 1 M 1
1
1 1 1 1

4 I • T •

R9ur*46
2J x44^

duplicated by the readings on the CF- and DF-scales, it

doesn't matter which pair of scales you use for multiply-|

ing. You can use the folded scales exclusively if you wish;
or you can move freely from the ordinary scales to the'
folded scales and back again. ™

There is an advantage to moving back and forth ben

tween the two sets of scales. Suppose C-l is placed oveij

D-314. In that case, half the C-scale moves beyond th<

right end of the sUde rule and the markings from C-31^
into the higher numbers can't be used. Half the CF-scalc
also moves beyond the right end of the slide rule, but her^
it is the numbers smaller than CF-314 that can't be used
(Figure 45).
The situation is, then, that numbers up to 314 ar^

usable on the C-scale and numbers over 314 are usable oi

DF-7U'
Figwre 47

J |iiiii I
I
I
I
'
I
'
I
'

I
'
I
>
I
'
OF
I
'
I
'
I
'
I
'
I
'
I
'
I
'

I
>
I
I
I
I

«'7S9i
i|lll|lllllllllMlllllll|IM|lll|lll|llllM|
I
I I I I I I
I
I I I I I
I
|l ll >|||lll|jJllll>lllllM|l(ll,^l<'/)f||U*lA\',^^

^ *'i. ^ CFI4iJI

t I
lin.liii.llnil.i
4
"'•""* '•'"'''•iTl?v;'rli;;;;,',v.v,l|llll|^^v/|',ul|l;/,'|V;,\lllll|^V;;^^^^^^^
« t • 7 t t 2
m

1
104
1104

l|lVl'^|li^|/|i|iV<'/l|lf<V/|llV|?il|H'/^^^^^^^^^ |m.,....^n„„..,M.,..MH..nn
II

iliiii(iiiiTiiiiiiiiiliiiiiiiiiT.iih
ilWii|/w|iii/t['i'r^'i'l'vtrfiiv,V,N7l|i|i!\v.y///|l|ii\J'/^^

he CF-scale. On one scale or the other, all the numbers

re usable even though half the slide is beyond the right

ind of the slide rule. If, then, in adjusting either C-1 or

3-1 (right) over a D-reading, you are careful to move the
|lide less than half the length of die slide rule you can be
ssured that all possible multiplications can be made
/ithout shifting indices.

Suppose, for instance, you wanted to multiply 2.3 by

1.8. Move C-1 over D-23 and you would find that C-48
v^asbeyond the right-hand end of the D-scale and there-
ore unusable. One way out would be to heave a sigh and
hift indices so that C-1 (right) is over D-23. This, how-
.ver, is not necessary. Leave C-1 over D-23 and simply
lx)k over to the CF-scale. You can CF-48 with no
find
ouble at all. It is directly under a point between DF-

P • t 2
jhiiiliiiiliiiilmiiiinliiiiliiiiliiiiliiiil
^/|fllliV.V.'|'lllll\\\^//l|llllWlll| lllUHlllllIll

C-l (ri9ht)

,4ii.w,|ii','','\'i^iri'(iiiit/fcwi'wii' i|i|i|i|i|i|MMMMi|MMMM'|MMMM'|MMMM'|WW|WW!WW^^
I

0M D-54
110 and DF-111, somewhat closer to DF-110 (Figurcj
46). Call it DF-1104, therefore, and you conclude that

2.3 X 4.8 = 11.04.

In the same way, if you wish to multiply 5.4 by 1.4J|

and move C-l( right) over D-54, you will find that C-14?,

is beyond the left end of the slide rule. Calmly, yoi!

switch to CFa45 and find it just to the left of DF-78fi

( Figure 47 ) . You judge the reading to be DF-783 anc

decide that 5.4 X 1.45 = 7.83.
The usefulness of the folded scales is not confined onl]
to those times when you have misjudged and used th<
wrong end of the C-scale. Suppose ( as frequently hap
pens ) that you must multiply a whole series of number
by a particular factor. For instance, there are just abouj
2.2 pounds in a kilogram and you may be given a numbej
of weights, in kilograms, which you wish to change inti
pounds. Each weight in kilograms must be multipliaj
by 2.2 if that purpose is to be accomplished. Suppose tha
the kilogram weights given you are 12.3, 32.1, 46.7, an<j

74.3.
p

You move C-1 over D-22 and leave it there. Thj

represents a general multiplication of anynumber b
2.2. Now you move the hairline to C-123 and find
I
over D-271; you move the hairline to C-321 and find
over D-706. So far, so good; 12.3 kilograms = 27.
pounds and 32.1 kilograms = 70.6 pounds.
But C-467 and C-473 are beyond the right end of thi

D-scale. No bother! Don't shift indices! Just move ovej

to the CF-scale where you will find CF-467 und(

DF-1027 and CF-743 under DF-1635. Therefore, 46.

kilograms = 102.7 pounds and 74.3 kilograms = 163.

pounds. *

106
 6
Division

Reversing Multiplication

Division is the reverse of multiplication, as subtraction

s the reverse of addition. When I described the addition
ule, I pointed out that we could perform subtraction by
unning the technique for addition in reverse ( see page
1). On the slide rule, we can carry out division by
unning the technique for multiplication in reverse.
In multiplying 2 by 3, we place C-1 over D-2 and look
nder C-3 to find D-6. Therefore 2X3 = 6.
In dividing 6 by 3, we place C-3 over D-6* and look
nder C-1 to find D-2 (see Figure 48). Therefore
^3 = 2. The reason for this is not hard to see. The
osition of D-6 represents a distance of log 6 from the left

nd of the D-scale. The position of C-3 represents a dis-

ince of log 3 from the left end of the C-scale. If we place
^3 over D-6, and follow the C-scale back to C-1, you see
rom Figure 48 that we are subtracting log 3 from the
* In setting a
C-reading over a D-reading, where neither reading
is end of a scale, it is best to use the hairline as a guide.
at the
To place C-3 over D-6 you would first place the hairhne at D-6,
then bring C-3 to the hairline. The hairline is particularly useful
where the readings are not on actual markings. To place C-3 19
over D-413, for instance, without using the hairline, is extremely
difficult, as you can assure yourself if you have a slide rule
handy. With a hairline, it requires a little close attention, but
isnot particularly hard.
107
 1

«PRMnp

] 1|llll|llll|llimMI|IIMIIIII|lllll|ln|UIUIII«|NIMMN|MJI 1 llllttflllllllllflllllllllllllll
i

Uf 4

Uf 4 - Uf 1 - Uf a-

4-i-i-i I

distance of log 6 on the D-scale. What is left of the D

scale beyond C-1 is log 6 - log 3. Since the readin
under C-1 is D-2 (which is a distance of log 2 from th
left end of the D-scale ) we can say that log 6 ,
— log 3 =
log 2. But subtracting logarithms is equivalent to divide
ing antilogarithms; therefore 6 -j- 3 = 2.
Any division can be carried through in the sam|
fashion. If you want to divide 48.5 by 7.4, you plai

C-74 over D-485 and you will find that C-1 is oflF the scale
That needn't bother you, for you need only look undcr
C-1 (right) instead and find 0-655 (see Figure 49). Yo|
decide therefore that 48.5 ^ 7.4 = 6.55. 1

How did you determine the position of the decims

point? In the usual way. Instead of 48.5 -r- 7.4, you mak| j

Figure 49

itSittTStZ
iiti4«tV«lii3v
1
4"" ' ""'
'•v'.^'n|^'v;V|';V/lVlvlv|^'lv;V;/Vlrl;.\|;AlilUl|^^l,V.l^^ll,'.|d
I D I

1
108
 '"'' ^ '" "
<J»v'lwM^Y'lw/(w'lwlwhy'lw/lw'|vA^^^^

Figure 48

ise of the near-equivalent 49 h- 7, to which the answer,

5 you see at once, is 7. This is not the right answer, but
„: is the right order of magnitude.
In dividing you place a C-reading over a D-reiEtding
d then look for either C-1 or C-l( right). No matter
iow you adjust the shde, either C-1 or C-1 (right) will be
n the D-scale so that you can always find a quotient
/ithout reference to the folded scales.
Nevertheless the folded scales do come in handy for the
pecial case of division by ir. For that purpose we can

nake use of the CF-scale, reversing the procedure for

nultiplication of u (see page 113). Since the C-reading
nultiplied by iz equals the CF-reading, the CF-reading
livided by ir equals the C-reading. If you want the

C-74 C-1 |ri9M)

ilil|li|il|Ti|ilililililililililr|il|l
ii|ilii|ilM|iiii|iihl I I 1 1 1 1 1 1 ' ' i'[l'/i'ltli'(i'|'.'l',t.'l'|'i'/lt|V,t,'i| |i|i|i|i|i|i|<|i|)|>|iMll|>rMi|>|i|i|MMM>l<|MM>IMi|>M
I I 1 1 I I

D-4«

74-4Jt
 1

CM
I

ll|l)ll|llll|llll|llll|iill|IMi|iiM|iiiii I
I I I I *
I
I
<
I I M !• I ' I I M *
I
' I * I I I •l>IMM*|M>Mlf*l>l«IMiM|M*M«W

TMIllllllllllllllllillllhlllllllllnillllllllMllinilllllllMllMMrMHlHMill INiHMIIIlllMllMI llMllllllIll'

C-ltl
RgiirvSO A-i-v-

answer to 6 h- tc, you have only to set the hairline on CF-6

and it will simultaneously read C-192 (see Figure 50).
Since r is approximately 3 and % = 2, you have to set
your order of magnitude and can say that 6 h- ^ = 1.92.

Figure 5

«::$Sia:¥:3

i rtiiii|iiii|4iii|iiii|iiii|iiii|iiiiiiiiiiiiiiiiiMi ia'.li\i.\/il/.|AikMMAM|.iM\VnWiv/(iv/tivliv/iiw|S^
1 t > t I

i .

D-IM

"'' '
•"
;r/:viV;v!T/lvi',W'ivA

110
 CF-2f4

'V k' « li « K V s" r"T

iiiibnihiiriiiuinflMilililililililililimiltlililililililititilililililililili^^^^ ililililililililLililililililiM

c-aof

M i4 -r- r - ijOf

In the same way, you can determine that 25.4 -^ ir =

8.09, and so on.
There is one possibility of confusion in division that

does not arise in multiplication. The order in which you

CM

"''"^' "''""""^""'
i^fiiifitmiii^^

f-4C- IJ4
D-TC

cn CI (rifhl}

v|'I)'jvi?iVi»'iv wh >\mmmnmt\mmm»m*\mmm
 )

multiply does not matter; so that 7.62 X 1.44 will give the
same product as 1.44 X 7.62. Therefore it doesn't matter
whether you put C-l( right) over D-762 and move the
hairline to C-144, or put C-1 (right ) over D-144 and move
the hairline to C-762. In either case, you will end with a
reading of D-110 and you will decide that 7.62 X 1.44 = f

11.0.
I

The order in which you divide, however, does matter,

as it does when you subtract ( see page 34 ) You can see .

that the quotient of 75 ^ 48 is not the same as that of

48 ^ 75. In the former case, you put C-48 over D-75 and
find D-156 under C-1. In the latter case, you put C-75
over D-48 and find D-64 under C-1 (right) (Figure 51).
Since 75 -^ 48 can be replaced by the near-equivalent
75 -^ 50, which equals 1.5, and 48 -^ 75 can similarly be I

replaced by 50 -^- 75 = 0.67, we see what our order of j

magnitude is and where to place the decimal point. We

decide, then, that 75 -^ 48 = 1.56 and 48 ^ 75 = 0.64.
how do you remember which number goes on the
But
C-scale and which on the D-scale? When do you put
C-48 over D-75 and when C-75 over D-48? Actually, the
decision is a simple one.
In any division, as, for example, a ^ h, a is the number
being divided into and is the dividend, while b is the
number doing the dividing and is the divisor. The rule,

then, is that the divisor goes on the C-scale and the

dividend on the D-scale. (This is analogous to the man-
ner in which the subtrahend — analogous to the divisor
— is placed on the slide in the addition rule. See
page 38 .

In the case of 75 -^ 48, 48 is the divisor and so C-48

goes over D-75. In the case of 48 ^ 75, it is 75 that is the

112
 isor, and it is C-75 that goes over D-48. When this

e is followed, the quotient is always found on the

scale underneath whichever C-1 happens to be on
Je.

But what if, by mistake or even by intention, the divisor

)laced on the D-scale. It is not good to place matters

reverse thus, for if you follow one set rule, it will be-

ne automatic and make the use of the slide rule that

ch less of a conscious effort. Still, if you do reverse
rule, all is not lost.

Juppose we take the simple example 42 -^ 7 and try

;Rrst in the ordinary fashion. We place C-7 (the
isor) over D-42 (the dividend) and directly under
l( right), we find D-6. Therefore, 42 -^ 7 = 6.

>Iow we'll try it "upside down." We will place C-42

T D-7 and we must now turn the procedure for finding

quotient upside down as well. Instead of finding

quotient on the D-scale under whichever C-1 is on
le, we find it on the C-scale over whichever D-1 is on
^e. In this case, D-1 (right) is under C-6 and again
-^7 =6 (Figure 52).
Whichever way you carry through your division —
\ h the divisor on the C-scale and the dividend on the
">cale, or the divisor on the D-scale and the dividend
• the C-scale — you will always find the quotient on
same scale as the dividend! You must remember
t.

ciprocals

plvery time you carry through a division on the C-

I D-scales you are actually performing two divisions,

113
one of which is upside-down, so to speak, compared td

the other.
Let us try 23 -^ 5. We will solve it first in the ordinar

way. Since 5 is the divisor, we place C-5 over D-23 an5

under C-l( right) we find D-46. Keeping the order d

magnitude in mind, we see that 23 -^ 5 = 4.6 ( Figui
53).

Take another look at the slide rule in the setting d

Figure 53. That is exactly the setting we would want
we took the example 5 -^ 23 and did it upside-dow>
placing the divisor, 5, on the D-scale diiectly under tl

dividend, 23, on the C-scale. In that case, we will firi

the quotient on the C-scale, where thc^ dividend is, ar

over D-1, we will find a reading just to the left of C-2]i

Figure 52
I
i

.W

ts«t(7t»2 3«
ii2Sii«Vt!i2 If
, . .
,
' '

vvlvv|vvlv|1vvlvl1vli^^Ml/M|il/.v.ll/|^Ml|WlPII'/•\ll'•'l«lpl»^^

¥
1

I TWwfr ...?i I
I/.|iU|.im/i|iW.|jlMWV^

114
we call it C-2175 and adjust the order of magnitude,
decide that 5 -^ 23 = 0.2175.
rhe same setting on the slide rule that gives us the
Dtient of any division, say a-^ b on the D-scale, will

e us the quotient of the division b -^ a on the C-scale.

A^e can write divisions as fractions. We can write
b as a/b and b -^ a sls b/a. We can therefore say
t the same setting of the slide rule that gives us the

ue of the fraction a/b on the D-scale will give us the

" e of the fraction b/a on the C-scale.
rhe fraction b/a is called the reciprocal of a/b and
B versa. It is often important in computations to
ermine reciprocals, and here is one method for doing
t on the slide rule.

CllfiqM)

.t/\'ri'vl'/i'ir,'i','i'i'ii'|i' i|'|i|i|"|MiMI"MM|l|'PM|M"|i|'MI'HMMM|"l'l'|lf WWIpiWlT

42 -=- 7 = 6
D-«

&42 C^

|^fi]vi^ViT.TiiviV|V|V|vIv|\'j?l';|v|vr.V;|livh^^ Wm
i/iiii/<|Viiii/>i<i/iiii/i|>t<'iiiii'<i<i<'iiii4i;iiii
ihl ilihlililililililililililililil.lililililililiMil

42 -r 7 - 6
D.7 D-MrifM)
 1

C-2l7f C4

* < « 7 • t 2
J It
IlilllllllliMtMllllllllllltlltMlHIlllMflnilllMiT 1 1 1 1 1

In I ilv.vj.7/i|iiv.'|jv.'l(iiilpviiii»v|lii|ili^^^^^^^^

0-1 D-23
23 -H I - 4>
Figure 53
Sh-23 -0JI7S

The most important reciprocals are those which hit

volve fractions with a numerator equal to 1. The recip'

rocal of ^/4 is M, but it is customary to drop a 1 tha

appears in the denominator and treat such a number a

consisting of the numerator alone. In other words, Yi f

usually written simply as 4 and we can say that the recip^

rocal of /4 is 4. In the same way, the reciprocal of Vi h i

18, that of V^i) is 29 and so on. On the other hand, th

reciprocal of 17 is Vn, that of 34 is ¥34 and so on.

Reciprocals expressed in fractional form are no prob

lem, but what if you wish them converted to decimal
The reciprocal of 29 is Km>, but what is the decimal fon
of M>o?
Suppose we treat Y29 as a division, as 1 -^ 29. Sin
29 is the divisor, place C-29 over D-1 and under C-

( right) you D-345 (Figure 54), Keeping t

will find

order of magnitude in mind, you decide that Mio

0.0345. you had placed C-29 over D-1 (right), y
If

would have found, under C-1, D-345, and have obtain

the same answer.
Looking at it in general, we can say that if we placj

C-x over D-1, we will find D-l/x under C-1 (right).

116
rtu
^e place C-x over D-l( right), we will find D-l/x under

And, of course, if 1/x is the reciprocal of x, then we

in also say that x is the reciprocal of 1/x. If Ho =
0345, then 1/0.0345 = 29.
roportion

It is possible to combine multiplication and division.

appose you were faced with the problem: 13 X 5.4 -^

83.
^
Place C-1 over D-13 and move the hairline to C-54.
'be hairline now marks oflF a D-reading which is the
"oduct of 13 and 5.4, a product which you needn't look
, but which now becomes the dividend for which 0.83
the divisor. Leaving the hairline where it is, you move
'•83 directly under the hairline and under C-1 (right)
D-846 ( Figure 55 ) .
* Now it is necessary to place the
Li cimal point. Consider that 13 X 5.4 -^ 0.83 might be
:)proximated as 10 X 5 -^-
1, which comes to 50. Ac-
• If the problem had been 13 x 5.4 -?- 0.89, C-1 (right) then the

D-reading under it would have been obscured by the glass edge

of the indicator assembly —
at least on my slide rule. Don't
panic, however. Just move the indicator assembly and look!

117
 I

cw
I

t f 2

>it)lT«t2
lM.llll|il|l.tl»nl|ll|ll|llllllllipili«*»J
illll|il|l.tl»nllll|ll|llllllllipili«*»J l|
I I I • I llll
I I I I I I
I I I
I II II
I
I

iiil|iMi|iili|iiii|iiil|iiir|Uii|iili|ini|iiH|iniiiiii|iiil|iiii|)iii|lni|iili|iih|i«^)
1 t
I
I I f..
lllllll|l|i|lll|illl>|llljlllllllllllll||lllll|ll
I

1 1 1 |llii|iMi|iill|

D.I

I iiiiiiiiiiiiiia»>»iii^nii I i iv > niow>)iiV

i
1
1
mmm »wimiMPtwwpm¥rm^ ......
% 1 i • • 2

Figure 54

cepting that as the order of magnitude, you decide thi

13 X 5.4 ^ 0.83 = 84.6. rf

Some problems involving a combination of multipl

cation and division form a special case.

Consider, for instance, two fractions, a/b and c/c

which are equal in value. We can then say that a/b -

c/d. We can refer to such an equality of fractions as

proportion^ and read it "a is to b as c is to d,"

It may happen that you know the value of three of th

four parts of a proportion and may want to find th

value of the fourth. If one of the numerators is unknowi

the equation becomes a/b = x/d. Solving for x b
ordinary algebraic methods, we find that x = ad/b or i

X d-^ h. If one of the denominators is unknown, w

118
I
M'lM)

««**«««««•«•«•«•«

Kl"l""l"'lM"'l"»| ' I " I ' I ' I

I
I ' I ' M I '
1
1
1
' (ffT^MMMMM' M'IMM'|M"I'I'l'l'l'I'in'|n'I'TTI"l'l'IM"I'l"l"l"MI
I
L 5 1

1/29 = 0.034S
IS

C-N

llllllllllllMlMMtlll(>l>HlMI)<*<<il.iiillllllMlll
L.l(\M,l/.i.|.\I.IV(P|\^V,M'*ivlV/l*|V|v(\'MV

l/M = 0.034S
D-l(ri9M)

ave a/b = c/x which means that x = he/a or b X c -^

. In either case, we can obtain the value of x on the
lide rule in the manner described just above for com-
ining multiplication and division.
I
However, problems involving proportion lend them-
Ives to particularly easy treatment on the slide rule.

Let's substitute numbers for the three known quanti-

ies and take % = 3/x as an example. Consider the frac-

on % first. We know that its decimal equivalent is 0.4

nd if we wish we can check that on the slide rule by
onsidering the fraction to be 2 -^ 5. We place C-5 over
-2, and under C-l( right) we find D-4.

But this same- quotient would be given by any other

'action that was equal to 0.4. For instance V\o = 0.4,
119
 C-l

IU|IIIIIIIII|IIII|IIII|IIII ;/.l(,Ml,\\f/,Ml.Ul.'iilIi\liAl|Mlj,«iMiwlMl'ApfMH^^
II
1 «

D-13

Figure 55

^Vzo = 0.4, y2.5 = 0.4, ^15 = 0.4, and so on. If, then, y
tried to divide 4 by 10, or 12 by 30, or 1 by 2.5, or 6 )

15, we would end with the same setting of the sHde

with C-l (right) over D-4. Having achieved this settiiflr

in the case of one such fraction, say %, we ought to fid

on that same setting, all the other divisions I've m

tioned and, indeed, any division at all that yields 0.4
the quotient. Sure enough, if you'll look at Figure
you will find that just as C-5 is over D-2, you have
over D-4, C-3 over D-12, C-25 over D-1, situations whi
are equivalent to %, ^lo, ^%o, and H.b.
To be sure, we can't find %5 on the C- and D-sca
120
 C44

0.(13 X S4)

(Ml Cl(r{9M)

|A^/|h^.'/l|\V/|l|V,V^^'^|l\V/l|\V/fllV^/^il'//l|ly/|l|^ |i|i|i|i|m|i|i|l|i|l|l|l|Mll

0.(13 X S4) DU*

13 X (.4-^0.13 - M^

ecause C-15 is far to the left of the end of the D-scale.

jlowever, if we switch to the CF- and DF-scales (see
Hage 99), we will find CF-15 under DF-6.
In short, once you set up a given fraction on the slide

le, with the denominator on the C-scale directly over

le numerator on the D-scale, then all other fractions of
ual valiie are similarly given ( denominator on C-scale,
erator on D-scale) on other parts of the slide rule.

, on the other hand, you wish to place the fraction with

numerator on the C-scale over the denominator on
e D-scale, then you will find other fractions with
^merator and denominator similarly placed.
121
 '

DF4

OF
|i|i|i|i|i|i|m|i|l|i|i|i|i|i|i|ili|i|i|Mi|<l<l> ,j|l|l|l|l|l|l|l|l|l|l|l|ll|l|l|l|l|l|1|l|ll^V.|^V,|^/,V.|/A\|^^<^^

CF-IS
C-2S C-3 c-s
• T • t 2
llllMllMlilMlllllMlnnllMll I I I I I I ( I I 1 I I I I I I I I I 1 1 I I I llll|llM
It I I
I
I I ll|llll|lMI
lllllllllll ;v.t,',vlv,\1v;|lv.v|v.vl;vilvj/lr.7liv/|.i.l|iJ.|.l.i|l.i.|j.liiJ^ 1 1 1
1
1 1 1 1
1
1 1 1 1
1
1 1 1 li|i
Iilil

Figure 56
D-l DI2 D.2

In a proportion, two fractions are set equal to eacl

other. If one fraction is set upon the C- and D-scales ij

the manner described, the other can also be found.

If we go back to % = 3/x and place C-5 over D-^l
then we should expect to find C-x over D-3. We mov(
the hairline to D-3 and immediately above it is C-7E
Since 2 and 5 in the first fraction are of the same ordel!

of magnitude, 3 and x should also be of the same ord<

of magnitude. We therefore place the decimal point s|

as to say %= 3/7.5.
In the same way, if we are faced with the proportioi
5.12/277 = x/34.3, we place C-277 over D-5.12 and mov|
the hairline to C-343, finding that to be over D-633 (s<

Figure 57). Since 5.12 in the first fraction is two ordei

A
1tStl»7«ll2 c 1 } P ? 1^
MttMMMMHMi '
'

|i'iiiiiii
1
'

ii >
'

M ii '
ii|
'

ii|ifli , , ,
,MM„Mm ,n,M^u4^'HV^'/lllhv.Vt^v/|lll\m^
'1

Figure 57
122
 '*
'*
i?"'iA|yiLlJillHMllljU»lllllllj»<lll)illllllU

-T
MfMiMlAHM"''^''''''''"''''''*'*

C.I(r{«M)

lllllllllll M'|i|'l'|'l'IM"l'|'MllMH|M"l"l»npF mrnm

I
D^

magnitude smaller than 277, we expect x to be two

jders of magnitude smaller than 34.3. We decide, then,
fat 5.12/277 = 0.633/34.3.

verse Scales

he key respect in which slide rule division differs

hm slide rule multiplication is that in multiplication,

ie set C-1 and find the answer under the multiplier,

hile in division, we set the divisor and find the answer
nderC-l.
This is to be expected, since division is the inverse of
ultiplication. When I was explaining the workings of
e addition rule, there was this same inverse eflFect in

«.277 C343

i !i.iA','iVi'iv/iV(i'i'nvhviii
l|l"IW|l4l''l|i'll"ll
I I i)I
(I lllilillllT'lilililililililihTilihlilililihlililil
I
'

i(mI|'IiiMIiiMiyimMiIiVim'i|Immm'iMiiimi|iim#i
l.'.l 'I '
'

D-512 0-«33
i. 1 2/277 - 0.433/34.3
 OF
iiiiiiiiiiiiiiiiiiH^^^^^^^
CF
CIF
'''""'' M""!""!""!""! l'".|....|..M,.,..|„.,-,....,....,„.,,..,.,,.„,.,„ ,,., ,

CI '•'''•'•'•I '••'• 1.1.1,1,1.1.1.1,1,1,1 l.i.r.i.iil.i.i.i.i.TiJit.i.i.l.i.i.i.i.T. I.I 1,1, 1,1,1 ta

C
|iiii|iiii|mi|iiiiyiii|iiii|mi|iiii|mi|iiii|jiii|iiiiW^^^^^

Figure Stf

connection with subtraction as compared to additiof

(see page 33). |
Once you are thoroughly used to working the slid
rule in multiplication and division, you will find .n

difficulty in setting C-1 in multiplication but not in div

sion. Nevertheless, there is some usefulness to preparin

a situation in which you can set C-1 for both division

and multiplication.
To make this possible, we must do for division whj
we did for subtraction in the addition rule. We muM
prepare an inverse scale, one that "runs backward." A| f

inverse logarithmic scale would have its numbers crow(

ing to the left, in mirror-image to a normal scale.
Such an inverse scale is placed in my slide rule imm
diately above the C-scale. It is a "C-inverse" or CI
scale. Above the Cl-scale and immediately under th
CF-scale is another inverted scale, a folded one. It i

Figure 59 cm

CI MiiiidiiiiiiiiiTiiiiiiiiiiiiiiiiiiTiiiiiiiiiiiiiiiiiiiTiMii.iiiiiiiiiiiiTiii,^

1 1 1
1
1 1 1 1
1
1 1 II 1
1 1 1 11 liil|ilil|illl|llll|IMI|llll|llll|IIM|llll|IMI|llll|*>ll|ll'*i""l'l"|l*'t"
1

D-12

12 -=-4-1
 )

I
l'l'l'MP|M'>'IMM'IMi|>M|'IM'IMM'IM>l'M|MM'PI<l'|i|M'|i|i|'l I I
<
I I I
>
I I
'
I
<
t I I I
• |MM|iiii|iMi|iiii|iiti|iiii|iiii|MM|iM

iIiiiiIiimIimiImiiIiiiiIimi 1 1
1
1
1
I iTiM>iiiiiii>iiiiiiiiiiiiiiiiiriMiiiiiiniiiii;iiriiiiiiiiriiiiiiiiiriiiii<iiriiiiiiiiiri i •

e mirror image of the CF-scale, in fact, and is, there-

re the CIF-scale ( Figure 58.
To distinguish these inverse scales from ordinary ones,
i! d to make sure that the casual slide rule user notices

jimething is different, the CI- and CIF-scales are usually

] in ted in red, whereas all the other scales I have dis-

cussed are printed in black.

The CF-scale can be used in division, just as an in-

^^rted scale was used in subtraction on the addition

lie (see page 37).
Consider the problem 12 -^ 4. If this is carried
though in the ordinary manner, using the C- and D-
sile, C-4 is brought over D-12 and under C-l( right)
i Found D-3, showing that 12 ^ 4 = 3.
jSuppose, however, we use the Cl-scale along with the
I'scale. Again, we will deal with the problem 12 -^ 4.

lis time we place the CI-1 over D-12 and move the

I.
It

flllllMllllMlllllllMIMIIllllMlMllllMtlllllTlMllllKllllllMlllM t ill I I I I I I M I I I t 1 1 I I I I I I I I iTlillliMlftlltllMirMllMMIII t till I ill t M ll M fl

I 1 1 1 1 < M 1 1 M I I I I I I I fi i • i I I t i I li i i I I I I I r I I > r > ( i • • i I

^lli|liHtiiii|Mii|ti)i|iiii|HHHiii|ili(|iiMi > I
I
I •! M'|MMMMMMMMMi|MMMMMMMUM"|MMMMMMMMM'|MMMMMMMUMflMMM'MI'l'IM»|»IMMMM

125
 )

D-l

...MIMM |I|IMMM> I<| |M<MV/|1|lite^

• • • I • • • •
I '
'
' * i*i*'|"*i|*i>*|Hii|Mii|Mii|itii|iiii|iiii|MiiMHi|nii| tilt (III ill iM|iii 1
1 1 1 1
1
1
1
M 1
1
1 M 1 1 1 II I
t|*|i|M>|*l*tM'M 'I'l'MI'I'IMMMM' t M '1*1*1 ' ' ' ' ' '
'
'
'
I
' I ' t ' I * I ' I * I ' I ' I ' I * I '
I ' I 'I if

IiIiIkiIiIiIiIiIiIiI il ill 1 1 li 1 1 1 1 1 1 1 1 1 1 ( I I ill I I I I I I I I I I I I I I I I I I TmiiIiimImiiImmIi^mIiimIiiiiIimiIiimIiiiiTmi \^I^t liiiilmilii nli i I i i i i I i i i i I i i i iThiiIiihmimIhhIiJ

^"^"^ lllll|IIMniM|MM|IIM|IIM|IIM|IIM|ini|IIIIMMI|MIIIIIII|lllllllll|IMI|llll|IIM MM|1IIII I I I I | I M I | MM M | 1

1
1 Ulll

R9ur«60

hairline to CI-4. Under it is D-3, the quotient (Figuri

59 ) . Suppose we had tried to solve 45 -j- 9 in this way|

We place CI-1( right) over D-45, but find we cann

move the hairline to CI-9, for that is far off the D-scalei

We switch to the CIF-scale, therefore. Placing the hair

line on CIF-9, we find it under DF-5, so that 45 ^9=
( Figure 60 ) . ( Had we placed CI-1 over D-45, we woul
not have had to switch to the folded scales.
By using the inverse scales, in other words, we evolv
a routine in which we set the end of a scale (C-1 i
multiplication and CI-1 in division) in carrying througl|i

both multiplication and division.

The inverse scales can be used for another purpo

CI iiiiiiiiiiliiiiiiiiTiiiiiiiiiliiiiiiiiiTiihiiiiliiiiiiiiiTiiiiiiiiiliiiiiiiiiTiiiiihiiliiiiiiiiiTiiiiiii

c lllMlllMlllllhlMnHlhlllllMlllllllnnlMlllHIllllMlMltlMlllmilllllilnilMIIMMllHuTlllllM

—^— leq I
Figure 61

126
 > > 4 I • 7 fl t
''"'''"
I A\itll /M\U lH/ /MU lll//MlV''''''

mll|llll|IIH|IMI|HII|MM
gj^ | I ^-(,^1

I
1 1 1 1 1 M 1 1 1 1 1 1 1 1 fl M I I t I I I

|l|llll|llll|)IH|IUI| I
I
• I
I
I
I
I
I i|i|i|i|in|i|iii|i|t|i|i|i|iMliMI'l'|'l'mTf
t

oo. To see what that is, let us consider some reading

>n the C-scale, which we can call x, and the reading
nmediately above it on the Cl-scale, which we will

ally.

The distance of x from C-1 on the left represents the

tgarithm of x. The distance of y from CI-1( right) —
bmember the Cl-scale is a "backward*' one — represents
le logarithm of y. Therefore, the entire stretch from
-1 to CI-1( right) is equal to log x + log y (see Figure

But the distance represented by the entire stretch

3m C-1 to CI-1( right) is equal to log 10. We can
erefore say that log x + log y = log 10. If we switch

I il '09 Y

('•MtiMiTMi iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii nil II iTiiHiniihiiiiiiiiniiiiiiiiriiiiiiiiiiiiiiiiiiiMiiiiiiiiiiMiiiiiiriMiiiiMriiiiiiniriiiiii ii

Mill
iiTl lllllllllliil
il iiliiiiliiiiliiiiliiiihiHli
til I 1 1 1 1 1 1 1 1 1 1 1 II lihTililililililHililiTilililililililililifihlilililililililiTililililililihlihTilililil^^^

Vf « + lofl If
 T

cuts

CI

c
1, 1,1. 1. 1, 1. Ill, I.

1
I I 1 I I ( I I I I
M.I. I, lilil

1
I I I I I I I I M
2
I I
1,1.1.1. I.I,

I I t I I I t •
3«S<7I«2
I I
l,l,l.lil,l,l,l,l,l,lilil.liM,Tililililitili

I I I I I I M t I I I I < M
I ill III,

illllltllMllltMllllllll lIllMrilMllllllllMllltlTl I
1

I I
1

I
1

I
1

I
1

I
1

I
1

I I
1

I
1

I
1

I
1 1

Mlllll
1 1 1 1

lllllllll

C-2M
Figure 62

from logarithms to antilogarithms we say that xX y

10, or X = lO/y.
We can, of course, adjust the decimal point to sui
ourselves. For instance, we can move the decimal poin
of y in such a way as to divide its value by ten. In tha
case, in order to keep the value of 10/y unchanged, w
must also divide the numerator of the fraction by ter

We end by saying that x = 1/y. ( For that matter we ca|

also, with full justification, say that x = 100/y or x

0.001/t/, but it is most convenient to say that x = 1/t/.

If X = l/y, then x is the reciprocal of y and vice verssj]

It follows then that when the haiiline crosses the slidfi

it marks off simultaneous readings on the C-scale an

Cl-scale that are reciprocals of each other. You can se

for instance that CI-2 is directly over C-5. This tells ii

( if we make sure to keep the decimal point in the propc

position) that the reciprocal of 2 is 0.5 and vice versi

since 2 ~ 1 0.5 and 0.5 = /!'. The same reading tells ii

that the reciprocal of 20 is 0.05, that the reciprocal (j

5 is 0.2, that the reciprocal of 0.005 is 200, and so oi

since y^o = 0.05, % = 0.2, and 1/0.005 = 200.

For none of these reciprocals is the slide rule reallj-
128
 1 )

CI-206

llllill il

I.

Hhiiiliiiil
I

1
t I

1
I

1
I I

1
1

1
I I

1
I

1
I

1
I

1
I >

1
I

1
>

1
>

1
I

1
I I

1
I ,,i
5
'„ ',.i,...i.

6
* I

7,8,9,1
.M,i,,..i,.Mr. i.i.i 1 1

lihlililililililililihlililililililililililililililililililihlilililiiililililililililililililililil

C-415 1/485 = 0.00206

1/206 = 0.00485

eeded; we can work it out in our heads without trouble,

"he principle is established, however.
Suppose, now, you wanted the reciprocal of 485. First
?t's place the decimal point. Fractionally, the reciprocal
if 485 is V-is'. which is sHghtly larger than lUo, Since

jf»oo = 0.002, we expect the reciprocal of 485 to be

ilightly larger than 0.002.
Now place the hairline on C-485 and you will find it

) be marking CI-206 at the same time (Figure 62).

'he reciprocal of 485 is therefore 0.00206. (If you had
laced the hairline on CI-485, you would have found it

> be marking off C-206. The same is true of the CF-

ale and the CIF-scale taken in combination. A reading
F CF-485 corresponds to one of CIF-206, and a reading
F CF-206 corresponds to one of CIF-485.
I
Reciprocals can also be found on the C- and D-scales

Y the method I described earlier ( see page 130 ) . The

e of the inverse scales, however, is the simpler for such
urpose.
I will describe one other adventure in reciprocals. As
u know, for any C-reading, say C-x, the corresponding
P-reading is r.x (see page 101). In other words, cor-
129
 .

responding to C-2 is CF-628, so that 2t: = 6.28.

Since the GIF-readings are the reciprocals of the cor-
responding CF-readings, for the setting C-x the corre-

sponding reading on the GIF-scale is equivalent to —

Let's take a specific example. Gorresponding to G-2,

we have GIF- 1591 (Figure 63). This means that ^=\

0.1591. Placing the decimal point is not difficult, if wd
remember that r. is close to 3. In that case — is approxin

mately % or 0.167. That sets the order of magnitude at

once.

Similarly, if you want the value of ^-^- , place th

o.Zt:

hairline of G-82 and find the simultaneous reading GIF]

388. Since 8.27: is close to 25, we can say that ^-r-

close to !i.-. or 0.04. That fixes the order of magnitude

1
and we conclude that 0.0388.
0.2i7Z

vei
 7

Powers

The Other Side

So FAR, I have described seven scales, one under the

bther, on the sHde rule. On the upper part of the body
is the DF-scale. On the slide are the CF-scale, GIF-
jcale, Cl-scale, and C-scale. On the lower part of the
i
oody are the D-scale and the L-scale.
, There is room for more and, indeed, one slide rule I

jse has several scales in addition to those I have listed.

i
These additional scales, however, involve operations
vhich will not be taken up in this book, and therefore,
hese scales will not be discussed.
Even so there remains the other side of the slide rule
vhich is available for other scales, some of which can
)e used for operations we will consider in this book.

up the other side of the slide rule, then. We

Let's take
:an begin by pointing out that there are two ways of
eversing the slide rule. It can be rotated horizontally
n a long swing, or vertically in a short flip.

Obviously, the short flip is easier and it is the method

rotation used. The proof of that is, that if you rotate
I slide rule horizontally, you will find that the markings
in the reverse side are upside-down. If you flip it

vertically, the markings on the reverse side are right-

ide up. You are clearly expected to perform the latter
aovement.

.3,
I
 In flipping the slide rule, however, we reverse the
position of the top and bottom of the body. What was
the upper body in front becomes the lower body in back
and vice You can check this by placing one
versa.

finger on the upper body and keeping it there while you

flip.

The scales on the other side of my slide rule are, with

one exception, different from those on the front. That
one exception is the D-scale, which is present on the
reverse side in the same D-position (see page ^67) as

in front.

To distinguish this D-scale from the one we have been

using previously, we will call it the D (back) -scale.
Notice that the D( back) -scale is not on the same part
of .the body as the D-scale. They both seem to be on
the lower body if we look at each separately, but in
order to look first at one, then at the other, we must flip

the slide rule. If you place your thumb on the D-scale

and your forefinger on the D( back) -scale, you will see

that the two are on different parts of the body.

If the slide rule is properly adjusted; if one half of the

body is fastened directly over the other and if the two
glass windows of the indicator assembly are properly

CIF-IS9I
k
•.a

ii|iiii|iiii|i III |(iii|Milltlll|llfl|llll|)ill|i>ii|iiitj ll|ltll|ll)l|MM II 11 I I I l| I

nil) 1 1 i|i 1 1
1
1 1 1 1 I I 1 1 I
I
I I I I
I
I
I
II I I
I
I I I I I I I I
I I
i • • <
I
I t I I
j

CIF n 3
I
'
2 9 • 7 ^
5
I

« 3 2

C 1
I I t I I 1 I I I
1
t I ( I I I I I M ll
2
1 I I I II ••
345S7t9:
lllllllllllllillllMlMlllHMllHllMlliMllllllilllHlllllltllllMII I I M I I II llll lllllllllllllllllloilhlil

C2
Figure 63
l/2ir 0.1591

132
 M .

positioned, so that the hairhne of one is directly opposite

the hairhne of the other, then the two D-scales will be

coordinated. If the hairline gives a reading of D-x, it

vvill be found to give a reading of D ( back ) -x, when the

jhde rule is flipped.

The D( back) -scale acts as a connecting link between

e two sides of the slide rule. Most often, the answer
btained in an ordinary multiplication or a division will
oe found on the D-scale. If there is then anything you
vish to do to that answer which will involve scales

bund only on the other side, you need only flip your
lide rule. There is the answer found on the D-scale,
)roperly marked off on the D( back) -scale, without any
lecessity on your part to make a new setting. And you
nay then continue.
Under the D ( back ) -scale, for instance (on my slide

ule), is a Dl-scale, which is the red-marked inverse of

?he D-scale, just as the Cl-scale is the red-marked inverse
i the C-scale ( see page 144 )
The D ( back ) - and Dl-scales can be used, together, to
nnd reciprocals, just as the C- and Cl-scales or the CF-
nd CIF-scales can be so used {see page 128)-
There are two advantages, though, to using the D-

CIF-388

I'p ;.IM|.I11|MII|,1M|..I
i'i'i'i'iM'i'i'i'i'i'i'ii|'i'iii'i'i'i'i'i'i' I' I '
r I ' I ' I '
1
' I ' I
• I ' I ' I ' ' I ' I ' I ' I ' r"'i "'I'
8 '
7 6 '
5
I

'
* ^ n

f ,
5 ,6,7,8 ihiiliiiiiiiiiLiiiiiiiiliiiiiiiiil

C 82

1/8.2* 0.0388
and Dl-scales. In the first place, if you are working on
the other side of the sUde rule, you can find your recip-i

rocals right there.

The less obvious one (for my slide rule) is that the"

Dl-scale is the only inverse scale on the body and not

on the shde. This means that reciprocals can be found

C-lt

front

D-(62 X It)

D(b«ck).(62 X II)

r«v«fM I

ii II iiii ii| iiiiiiii|iiiiiiiii|iiii|iiii|iiii|iiiiiiiiiiiiii|iiiiiiiiiMiii|i>ii|iiiiiii»|<><<M*'*i '

ii>i'i"iii*iMii*i*"y

m lililililililililHi i,i,i,i,i,i,i,i,i.i,liiiiii.iiiii<iiiiiiliiiiiiiiiTiiiiiMiliiiiii<i>Tiiiiiiiiiliiii< I it J

1/4.2 X 1.1 = 0.0196

DI-196
Figur* 64

A 2 311 , 4 , S « 71
lllllllllllll iiNi|i|i|i|i|iH

1
I I I I
I
I I I I

ias4f«Vt«2
I
I II l| I I ll|llll|IIM|llll|llll|llll|llll|llll|llll|llll|lllllllll|llll|llll|llll|llll|lllll I I I I
I
I I I I
I
I II l|MII|lllllllll|llt

Figure 65
 regardless of the position of the sHde.
Suppose, for instance, you wanted the solution to a

problem like
g^^ ^g
. To obtain the product of 6.2 X
1.8 is simple, but once that is obtained what you want is

the reciprocal.

C.| (right)

rm imkmi^^m^k^^^^^^^^^ i|i|i|(|i|i|Himi|i|Mi|MMM>tMMMMM)fMMMMimffll|Wffr

D-62

mmm

it|ii ii|iiii|iiii|iiii|iiii|Mii|iiMi I
I
>
I
•
M I |MMMMMMMMMi|MiMIM>IMilMMi|<|i|MM*MIMMi|*|MMi|>IMMH<l<l>|Mi|i|'IM|l
I
Hi
II I
2 • • 7 • I 4 a 2 1 1
>: iiiiliiiiliiiiliiiiliiiiliiiiliiii liii I li 1 1 1 1 1 1 1 I IiiiiIiiiiIiiiiIiiiiIiiiiIiiiiImiiIiimIiimIiiiiIiiiiIiiiiIiiiiIiiiiIiiiiIiiiiIiiiiIiiiiIi I I I Ii I I I I

*H**i**^^

IHrM|iiii|iiii| I
I
I
I
I
lM>|MMnMMMMMn>|nMMM'IMMMM>|MMMMMMnMM*|MMMMi|MMn>n|<l'MMIM)|>l'IMi|M>l>IMi|

135
 doubU
L
0.1 0.2 0.3 0.4 OJ 0.6 0.7 0.1 C
KlIllHnltHltlMihMlllMlllMthlllllHlllllltlMllinilllllhllllltltlMlllllllUlttlMMllllllHnhlltUMllMMllllllllllllHlllllllHMllHllllllllMlhlllhlillllllllllll^
^

A 1 2 3 4 S 6 7 I

'iiii|iiii|iiii|iiii|iiii|iiii|iiii|iiii|iiiiMi|i|i|i|i|i|H^
B
2 3 4 S 6 7

Figure 66

Let us begin. We place C-l( right) over D-62, move

the hairhne to C-18 and find that to be over D-1116.
What we need, now, is the reciprocal of 1116. This
means, if we are to use the C- and CI- scales, or the CF-
and CIF-scales, that we must move the hairline to 1116
on one of those four scales.
Yet that is not necessary. If we flip the slide rule, the
D-reading is duplicated in the D ( back ) -reading and the
Dl-reading is the reciprocal of that. We need move
neither slide nor hairline.

In short, let's try to solve the problem f>ns^^o again.

Place C-1 (right) over D-62, move the hairline to C-18,|

flip the slide rule and note that the hairline marks DI-
896 ( Figure 64 ) . Next we will check the decimal point.
1 , . . , 1
Since can be approximated as the
6.2 X 1.8 6X2'
1
answer must be about Vi 2 or 0.083. Hence,
6.2 X 1.8
0.0896.

The A' and B-scales

On the reverse of the slide rule in the A- and B-posi-

tions (see page 66) are two scales -actually named foi

those positions, the A-scale and the B-scale (Figure 65).

The A- and B-scales are a new variety. Except for the
136
 .

'•2 ti 1.4 1.5 1.6 1.7 1.8 1.9 2.0

lllllllililllillM>i|il<lliilililiitiM'|i>r>lrii.h<-i (mI..ii<>.mIm.iIi>mIiiiiI Ii< ilt...l- mI..i.Ii.i>1>'..Ii»>I ,..' ,.. ''
ihIiiiiIi>I'I>miIi>mIii(|Iih>I. ..ImiiIimiI

20 30 40 SO 60 70 80 90 100
M | H i i |iiii|iiii iiii|iili|i|i|i|i|i|i|i|iM
|

20 30 40 SO 60 70 80 90 100

L-scale, all the scales I have discussed so far have really

been members of a single family represented most

:leaily by the C-, D-, and D( back) -scales. The CF-

md DF^-scales differ from these only in that they start

[it the middle; the CI- and Dl-scales only in that they
un backward; the CIF-scale only in that it starts at

the middle and runs backward also. All these, however,

lave the same system of primaries, secondaries, and
ertiaries.

Not so the A- and B-scales. They begin at the left at

and pass through the units, reaching 1 again in the

aiddle. There is then room for an exact repetition,
leaching 1 still again at the right end. It is as though
he C- and D-scales have been compressed to half their
3ngth so that two of them will fit into the usual slide
iile length.
We can refer to the two identical halves of these
ales, as the A (left) -scale and A{right) -scale, and, of
urse, the B (left) -scale and B (right) -scale.
Let's stop to consider what this means. You may
member that I described the construction of the C-
d D-scales by lining up their markings as antiloga-
ms matching the logarithmic readings of the L-scale
ee page 67 )

The L-scale runs from 0.0 to 1.0, but what if we were

I
construct such a scale with the units spaced more
I 137
closely together, so that there would be room for j

stretch from 0.0 to 2.0. If antilogarithms were lined uj

with such a "double L-scale," the results, as shown ir

Figure 66, would be a logarithmic scale running fron

1 to 100.

Since the slide rule takes no account of the decima

point, the primaries of the A- and B-scales are no
marked 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50
60, 70, 80, 90, 100, but 1, 2, 3, 4, 5, 6, 7, 8, 9
1, 2, 3, 4, 5, 6,, 7, 8, 9, 1. There is thus an A-1, an A-1
(middle), and an A-1 (right.)
On the whole it is best to let the A ( left ) -scale repre

sent the stretch of the numbers from 1 to 10 and the

A ( right ) -scale from 10 to 100, varying the decima

point two places at a time to get other stretches. Thus
the A (left) -scale includes not only, let us say, the num
ber 5, but also 500; 50,000; 5,000,000; and so on. Work
ing in the other direction, it also stands for 0.05, 0.0005

0.000005, and so on. On the other hand, the A (right)

scale, includes 50; 5,000; 500,000; 50,000,000; and s.

on; as well as 0.5, 0.005, 0.00005, and so on.

This is not really difficult to remember. You wi
notice that the A ( left ) -scale includes the numbers wit!
an even number of zeros neighboring the decimal poi
when written in the form given above. Thus, 500 h
two zeros neighboring the decimal point, as has 0.0<

and both 50,000 and 0.0005 have four zeros neighborin

the decimal point. (Naturally, numbers less than o
have to be written with the preliminary zero, as 0.

and not as .05 if this rule is The number 5 h

to hold. )

no zeros and if this is considered an even number

zeros, this, too, fits the rule.
On the other hand, the A ( right ) -scale includes t

138
 numbers with an odd number of zeros neighboring the
decimal point. Thus, 50 and 0.5, both have one zero
neighboring the decimal point, 5000 and 0.005 both
have three, and so on. All this, of course, holds for the
B-scale as well.
The A (left) -scale crowds the full stretch of values
found in the C- and D-scales into half the usual length.
This means that the A (left) -scale cannot be subdivided
as finely as the C- or D-scale (or any of the others of
that family). There isn't the room.
The space between the numbered primaries of the
A (left) -scale are divided into secondaries representing
enths, which are spaced more closely, of course, than
ire the secondaries in the C- and D-scales. Even be-
ween the primaries 1 and 2, where there is the most
oom, the secondaries are crowded too closely together
o allow a convenient marking off by number as is done
n the C- and D-scales and the rest of that family. None
)f the secondaries of the A- and B-scales are numbered.
The space between the secondaries lying in the range
om 1 to 2 on the A ( left )
-scale are split into five sub-
*
divisions marked off by tertiaries representing 0.02 each.
Tom 2 to 5, the space between the secondaries are
livided in half, with a single tertiary representing 0.05.
Jetween 5 and 1, there are no tertiaries at all. The
arkings on the A ( right ) -scale and of course on the
(left)- and B( right) -scales are identical with those
the A ( left ) -scale.

If you make use of the A- and B-scales, it is important

keep a wary eye on their markings and make sure
u do not carry over the habits developed on the C-
d D-scales.
You can multiply and divide on the A- and B-scales
139
 A(l«ff)-6

A 1 2 3lt 4 S 7
''''''''''''''"''''''''''"'''"'''' '''''''''''^ i.l.i'.i.i.i.l.i.i.i.i.l 1,1,1 1

\v,'l|V,'iinWn
B

precisely as you can on the C- and D-scales. There are

both advantages and disadvantages to this. I will

mention the advantages first.

i

Since there are two scales, a left and a right, included

in the A- and B-scales, there is no likelihood of the slide;

moving off the scale and you do not have to interchange]

indices.

Suppose, for instance, you multiply 6 by 2 on the C-

and D-scales and carelessly begin by putting C-1 over
D-6. You must next move the hairline to C-2, which,
however, is well beyond the right edge of the slide rule.
Nor can you switch to the CF- and DF-scales where, in
this case, CF-2 is even farther beyond the right edge

There is no recourse but to switch indices.

If you did the same thing on the A- and B-scales, you

would have no difficultv- It is the B-scales that is oi

the slide, so it is that which you move. Place B(left)-]

under A ( left ) -6 and now move the indicator to B ( left ) -2

To be sure, B(left)-2 is beyond the right end of the

A ( left )
-scale, but here we have an A (right) -scale tc

take up the slack (Figure 67). B(left)-2 is under A

(right) -12 so 6 X 2 = 12. It is for this reason that then
is no need for a folded A- or B-scale. j

The A- and B-scales can be used for division, too. in •

|T'iv||.^T-1vl'Y.fi^|.'l/|ln^kl,||l/.|Vll|llll'.^'/|lV^^ iiiiiiiiiiii <
I
'
I
'
I
'
I
'
I
'
I
'
I

liTianner analogous to the C- and D-scales. Here, too,

the divisor is best placed on the sUde; that is, on the
B-scale.

Consider the problem 5.4 -^ 2.2. The divisor is 2.2

>o you place it on the B-scale; B( left) -22 is put directly
under A (left) -54 with the help of the hairline, and over
3(left)-l is A(left).245 (Figure 68). Adjusting the
lecimal point, you decide that 5.4 -^ 2.2 = 2.45.
The A- and B-scales can also be used to find recipro-
cals in just the same w^ay that the C- and D-scales can
)e used (see page 113).

squares

The A- and B-scales can do everything, it seems, that

he C- and D-scales can do, without the danger of having
switch indices in mid-problem. Nevertheless, there
5 a serious disadvantage to their use.
Since the markings on the A- and B-scales are only
alf as far apart as the equivalent markings on the C-
nd D-scales, the A- and B-scales cannot be read as
osely. This means that any answers that you get on
lie A- and B-scales are not as accurate as those you
et on the C- and D-scales.
141
 !!

A(Uft)-24S A(l«ft).54

^^^^H I I I I It t I lIllllhllllllllllMllMlllllllllttll I I I I I I I I lIllllllll.llllllllMlllllllllliltll ||||||ll>ll>> nil

S
I
7,1
11 1
1 ( 1
1 1 t|i iii|iiii|iiti|iMi)9ii|nn|Min
1 1 < <
v.''r.v)ijv,'iiv,'iir.'|i|i
i

-
S.4 2.2 2.4S
•(l«ft)l (Uff|-22

This question of accuracy is so important that ordinary

muItipUcation and division should always be done on
the C- and D-scales and never on the A- and B-scales.
Increasing the accuracy is so desirable that one should
willingly pay the price of an occasional shift of indices,

or of having to transfer attention to the folded scales.

Yet I am not suggesting that the A- and B-scales be j

completely ignored, that they never be referred to. The

A-scale, at least, has a particular convenience that is

quite suflBcient to justify its existence.

To see what that is, let me remind you again that the
C- and D- scales are built up on the L-scale, which runs
the length of the slide rule with a range from 0.0 to 1.0, |

while the A- and B-scales are built up on a similar

"double scale" with a range from 0.0 to 2.0 (see page
137).

Ie9 f or 2 I09 i

liiiihiiiliiiiliinliinliuiliiHliiMlHuhJililihlilililihlitiltlihlilitililililililihhb^

iiiii|niinin|Miiitiii|iiniiiii|iMiiiiM|iHiiiiH|)Hinm|uiNNH|MN|N^^

lof «•
 mm

t 7 • t 1

|iiViiiv.'iiv,vii',jiiv#j(.\,i.iLJ,ailii/|/i/i//.'/^^^^^^^^^^^^ ,., ,«,1 1 1 iiiiii|iiiiiii

Figure 68

Since the double scale covers twice the range in the

1 same distance, it moves, so to speak, twice as fast. At a
given point where the L-reading is, let us say a, the
"double L"-reading would be 2a.
Let us now suppose that the hairline is over a partic-
ular D ( back ) -reading and is simultaneously over a
particular A-reading; we can call them D(back)-x and
A-y. The distance of D(back)-x from D(back)-1 rep-
resents the logarithm of x and the distance of A-i/ from
A(left)-1 represents the logarithm of y on the double
I'Scale. It is the same distance in both cases, but the
double scale logarithm is twice the size of the ordinary
logarithm at the same point, so that we can say log
y = 2logx, or log y = logx + logx (Figure 69).
If we shift to antilogarithms, remembering to convert
an addition into a multiplication, we have y = xXx,or
A-y
"1 Rgur« 69

ilHH MIMllllrflllllllllll liliiiiliiiilMiHiiiliiiliiHlNMl 1 1 1 lililililihhliTililililililililililililililiiiMMiriiiiliiiiTiiiiliiiiliiiiliiiiTiiiiliiiiTiiul«<J

'1*1*1 iinMi|iiii|iin|iiii|iiM HI l|l |l|lll|i|l|i|i|i|i|i|i|i|i|i|i|i|i|i|i|i|i|Hi|i|i|i|i|i|i|i|i|i|i|i|i|i|i|i|HMMpi|>|i|'|<«'l<n'|

I
I 1 1 1 1 1 1 1 1 1

-^-

D(bMk)l
143
 1

Figure 70

IimiImiiIimiIiiiiIiimIiiiiIiiiiIiiiiIiiiiIiiiiI I I I I I lilililililllllltlllllllllllhllllllllllllllllllllllhliiiilii
I

1
• I t I
I
I I I I
I

1
I t I )
I
I I t I

23&SC9S92
I
I I I I
I
M I I Itl lt|llll|nM|lin|llll|IMIIIIII|IIIIHIIl|llMllli||HM|IHI|IMI| llll|llll|MI|||||||

to use the usual mathematical teraiinology, y = x^.

What we have now shown is that when the' hairline
marks off simultaneous readings on the D( back) -scale
and the A-scale, the A-reading is the square of the
D ( back ) -reading.
To determine a square, then, we need only adjust
the hairline without moving the slide. To determine the
square of 5.1, place the hairline over D( back) -51 and
note that it also gives a reading of A (right) -26. We
can establish the order of magnitude by remembering
that 5.1 is not much different from 5 and that the
square of 5 is 25. Therefore 5.1- = 26.
In the D( back) -384 corresponds to
same way,
A (right) -147 (Figure 70). To place the decimal point,
we substitute 4 for 3.84. Since we know that 4^ = 16,
that 40^ = 1600 and that 0.4^ = 0.16, we can tell that
3.84- = 14.7, that 38.4- = 1470, and that 0.384^ = 0.147.
It is helpful, in connection with finding squares and
placing the decimal point, to remember the odd-even 1

rule concerning the right and left halves of the A-scale.

The A (right) -scale involves numbers with an odd num- i^,

ber of zeros neighboring the decimal point. For that ji

reason, you may expect its readings to be of the order

144
 A(ri«M)l47 A(ri9kt)-26

3 ,4,5 6 7 • 9 1

|< |< I '('I 'Ml 'I' I

|'|M'l'l'l'|'l'l'l'|i|'l'l'l'l'|'l'l'IM'|'|i|i|l|'|'l'IM'l'p'l'l'l'l

(S-l)^ = 26
D(b«a)-3I4 D(b«a)-si
|3.i4)- = 14.7

){ magnitude, let us say, of 10 or 1000, but never 100.

^rherefore since the squares of 384 and of 51 fall in the
||( right) range, their squares might be 14.7 or 1470 in

he first case or 26 or 2600 in the second. Never, no

latter how you place the decimal point in 384 or 51,
an the square turn out to be 1.47 or 147 or 2.6 or 260.
Another point to remember is that the D ( back ) -scale
> not an absolute necessity for determining squares on
he A-scale. The D-scale itself can also be used. Once
D-reading is set, one need only flip the slide rule to
he other side to find the squares marked off as a
imultaneous A-reading.
To be sure, squares can be determined without the
-scale too. If 3.84- is considered to be 3.84 X 3.84,

lat multiplication can be carried through on the C-

nd D-scales in the ordinary way. If this is done, the
iiswer is obtained with greater accuracy, because the
nal reading is made on the D-scale and not on the
ss finely divided A-scale.
If, indeed, we carry through 3.84 X 3.84 on the C-
nd D-scales by putting C-l( right) over D-384, then
oving the hairline to C-384, we will find, under it,

*'-1475. We will conclude that 3.84- = 14.75. This is

145
 "

closer to the truth than the 3.84^ = 14.7 determined bi

use of the D- and A-scales; for the true answer, worked
out in full, is 14.7456. ^i

Nevertheless, working out the square by use of thJ

A-scale involves no motion of the slide, and for quid
work that maneuver comes in handy.
By substituting the Dl-scale for the D ( back ) -scale

we can get the square of a reciprocal: say

( j^j . W(
can set the hairline on DI-48 and obtain a simultaneouj
reading of D( back) -208, which gives us the reciproca

—Q . However, we needn't stop to work out the decima

4.0
place or even look at the digit combination. We pass

directly on to the A-scale, which will give us the square

of the D ( back -reading and hence the square of

) the

reciprocal of the Dl-reading.

|
We find the A-reading to be A (left) -433 (Figure 71)
and now we are ready for the decimal point. In place of
4.8, let us take 5. We therefore have % )- ( which equals
(0.2)^ or 0.04, an order of magnitude with two zeros

neighboring the decimal point (checking the fact that

we ended with an A (left) -reading, which requires an

/ 1 H
even number of zeros).
'
We conclude then that (-r-z
\4.8 h
\

A(Uft)-433

1 ,
2 3n 4
''"
I I I I I I I I III llllllllllMllllHllllllllllilllllllllll I I I I I I I I I I I I I I I I I I lllllllllllllllllllllllllll lllilililihliililiilillliiliiiili

1
I I I I
I

.
I I I I
I

1
I

9
I II
I

.
2

*
3(SCyt*2
llll|llll|llll|llll|llll|llll|llll|llll|llll|IMI|lllllill

7 •
|llll|llll|llll|llll|lllll

«
I I I |i III I
iiii|iiii|iiii|i>ii|iili
|iiii|iiii|iii

I
Dl lilihiililililililililililililililililililllilililililillhlilililililililililililil lilililililili I ill ii I 1 1 1 1 1 1 I I I I I I iTiiiiliiiilm

(1/4.11)- - 0.0433
Figure 71 DM81
w = 0.0433.
The fact that the shde rule gives us a method for

quickly determining squares means that we can carry

through a multipHcation or a division and find the
square of the result at once.
If there were a C ( back ) -scale adjoining the D(back)-
scale on the reverse side of the slide rule (as there is, in

some slide rule designs), we could run the multiplica-

tion or division on those scales, get the product or
quotient on the D( back) -scale and take the simultaneous
A-reading.
On my slide rule, however, there is no C( back) -scale.
The reverse of the slide, where such a scale should be,

iis given over, instead, to several scales involving trigo-

nometric functions which I will not discuss in this book.

I
Fortunately, this introduces no great difficulty. We
ican use the C- and D-scales on the face of the slide rule

instead.
^ Suppose, for instance, we want to solve the problem
,(9.44 X 2.75)'-. Beginning at the front of the slide rule,

jtve place C-l( right) over D-944, move the hairline to

i;-275, and find a D-reading under it that represents the

product of 9.44 X 2.75. However, we are not interested

in the product but in its square, so we needn't even

H iIiiiiImmIiimIiiiiIiiiiIiiiiI
2
i I i I i I i I i I i I i I i I i I i
3,4,S S7S91
Ii IiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiI i i i i I i I i i I iiMliiiiliiiiltiMliiiiliiiiliiiiliiiil

"I i>i>Miiii|iim I
I
I
I
I
I
i|i |i|i|i|MMMMMM*|MMMMMMM<MI<|>|i|MMT|i|MMi|MM<MI<lMMM>l>|iMI'IM>|>IM>IM>|M>IM<n

IJ lllMlllllllHlllllllllllllllllllllllllllllllllllirilllllllllllllllllirilllllllirilMll
I I I I I I I
I'll ImiiImIiIiIIiIiIIiIi I I I ll I I I I I I
I I I I I I

147
 '

k-lelt b^ k-middle

|iMi|Mii|iiii|iiti|iiiiniM|iiii|iiii|itii(iiit| •
I
•
I ' r I
'I •|>IMH'lMM'IH'tMMMM'|1M'l1TI'''''^''''''''M^*^'''*'H'''''*'''l"''l"'V*'''l''*V'''l''''l'''M"*M''*M'*''l**"l**'*)*<'T"1

1
1 1 1
1
1 1 1 1
1

123fcS67«t2
1 1 1 1
1
1 1 iii|iM i|ini|iiii|Mii{iiii|nM|iiii|iiM|iiii|iiii|iiii|iiii{iiii|iiii|iiii|Mii{ 1 1 1
1
1 1 1 1
1
1 III i|iiii|iiii|iiii|iiM|iiii|H

Figure 72

bother looking at the D-reading. Instead, we flip the

shde rule in search of the A-reading and find it to be
A (left) -673.
We get the order of magnitude by substituting 10 for
9.44 and 3 for 2.75. The problem becomes (10 X 3)^
= 30^ = 900, which checks with the fact that an A (left )-
reading requires an even number of zeros in the
neighborhood of the decimal point. We conclude, then,
that. (9.44X2.75)2 = 673.
You can use the same principle to solve (6.3 -^ 1.18) 2,
provided you remember to manipulate the slide rule in

such a fashion as to let the quotient of 6.3 -^ 1.18 fall

upon the D-scale. Only if it falls upon the D-scale can

the square of the quotient simultaneously be found on
the A-scale. However, if we follow the rule of placing
the divisor on the C-scale ( see page 128 ) , this point will
take care of itself. I

Since 1.18 is the divisor, it goes on the C-scale. Place

C-118 over D-63. Move the hairhne to C-1 and under it

is D-535, which represents the quotient. Without

bothering with the D-reading at all, however, you flip '

the slide rule and find A ( right ) -285. To place the

decimal point you replace 6.3 by 6 and 1.18 by 1. The
problem becomes (6-4-1)2 = 6^ = 36. This order of
magnitude is in the 10*s, which checks with the fact that
148
 k-ri9M

||lll|limiH.|MMpitll—llip—(MllJHll|IHH«««riipi»i«P«^^

[.JII|IIII|IMII I
I
I
I
I
I • I
I
I
• I < I • I
) IM'l'l'l'l'l'l'l'1'l'l'l'l'|i|'l'l'l'l'|'l'l'IMili|'l'l'l'l't'l'l'l'MI'l'l'l'|<l>1 1'l'l'l'l'l

the final reading was found on the A (right) -scale,

j-equiring an odd number of zeros in the neighborhood
)f the decimal point. Therefore (6.3 ^ 1.18 )2 = 28.5.

V ubes

An exponential number like x^ is said to be a power of

The higher the value of the exponent, the higher the
:.

lower. The average slide rule can reach a higher power

han the square; it can, in fact, handle the cube (x") just
\s easily.

1 The square involved a double L-scale so it is not

unexpected that the cube will involve a "triple L-scale."
1
ince the ordinary L-scale covers the range of logarithms
'rom 0.0 to 1.0, the triple L-scale will cover the range
ifom 0.0 to 3.0. In terms of antilogarithms this is the
singe from 1 to 1000.

The antilogarithm scale with markings from 1 to 1000

i
present on the front of my slide rule, all the way at

iie top, and is the K-scale (Figure 72). (In another

ide rule I have, it is on the back, all the way on the
ottom.) If you follow its markings you will see that its

rimaries represent 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40,

0, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800,
JO, 1000. Since the decimal point is arranged to suit
149
ourselves, it is only necessary to write the digits: 1, 2, 3, 4,

5, 6, 7, 8, 9, 1, 2, 3, 4, 5, 6, 7, 8, 9, 1, 2, 3, 4, 5, 6, 7, 8, 9,

1.

Just as the A- and B-scales are divided into two iden-

tical halves, the K-scale is divided into identical thirds, i

so that there is a K{left) -scale, a K{middle) -scale, and a

K{right) -scale.
The decimal point jumps three places at a time in.
each of the three sections, thus:
i

K( left). scale 0.002 2 2,000 2,000,000

K( middle) -scale 0.02 20 20,000 20,000,000
K( right) -scale 0.2 200 200,000 200,000,000 etc.

The individual thirds of the K-scale are even more

crowded than the halves of the A- and B-scales. The
spaces between 1 and 3 are marked off by secondaries
with values of 0.1 and tertiaries with values of 0.05. The
spaces between 3 and 6 are marked off by secondaries
only, with values of 0.1 each; and the spaces between 6|

and 1 by secondaries with values of 0.2.

The K-scale exists by itself; there is no similar scale

that can be moved along it as the B-scale can move along

the A-scale, or the C-scale along the D-scale. For that-

reason there is no way of using the K-scale directly for;

ordinary multiplications and divisions, and that is just

as well, for the K-scale would be extremely inaccurate

for that purpose.

The important use for the K-scale is in connection

with cubes. In fact, the "K" of the K-scale stands for

Kubus, the German word for cube.
The manner in which it was shown that the readings

150
^ on the A-scale represent the squares of the correspond-
ing readings of the D( back) -scale (see page 144) can
be used (with appropriate modification for a triple L-
I
scale, rather than a double) to show that K-readings
represent the cubes of corresponding D( back) -readings.
For instance D-2 corresponds to K(left)-8; D-43
;
corresponds to K( middle) -8 and D-928 corresponds to
'K( right) -8. Remembering the discussion on decimal

I
points above, we see that 2*^ = 8, 4.3'^ = 80, and 9.28-^ =
'800.

The method for solving problems such as ( 64 X 0.4 )

^

and (3.11 -^ 2.6)^ is just the same as for solving anal-

ogous problems involving squares ( see page 144 ) . We

iperely look for the K-reading, rather than the A-reading.
 I

8
Roots

Square Roots

When we dealt with addition, we found we could

also deal with its reverse — subtraction. Again, in deal- fl

ing with multiplication, we dealt with its reverse —

division. Now, in connection with powers, we find that {

the reverse operation can also be dealt with, and it in-

volves roots.
The square root of a is that number which, when
multiplied by itself, yields a as a product. In other
words if b X b =a (an equation which can also be
written b' = a), then b is the square root of a, and this
can be written b = Va. The square and the square root
are thus opposite sides of a coin, so to speak. If a is the
square of fo, then b is the square root of a.
^

We can make this plainer with numbers, perhaps.

Since 5 X 5 = 25, 25 is the square of 5, and 5 is the square
root of 25. Since 6 X 6 = 36, then 36 is the square of 6,

and 6 is the square root of 36.

To find the square of a number, using the C- and D-
scale, as I pointed out earlier (page 144), is simple
enough. To determine 7.8', you need only solve for
7.8 X 7.8 in the usual way and find the answer to be 60.8.
To find the square root of a number using the C- andj
D-scale is quite another thing. Suppose you want the
solution to \/7.8. That means you want a number which,!
when multiplied by itself, will yield 7.8 as a -product.

You find that product on the D-scale and mark D-78'

152
I

with the hairUne.

Next you must adjust the sUde in such a way that the
C-reading at the hairUne over D-78 must be the same as
the D-reading under C-1. By careful manipulation of
the shde you will find that when C-279 is over D-78, C-1
is over D-279. That means that 2.79 X 2.79 = 7.8 and,
therefore, that VT8 = 2.79.
Adjusting the slide to find a square root is not an-
impossible procedure, but it is a tedious one, and must be
carried through carefully, with frequent glances back
and forth between C-1 and the hairline. It is not at all

the usual flick-flick of the slide rule as in carrying through

so many other computations.

Consequently, if another technique will suffice to
pfind the square root, it should be used.
The answer lies with the A-scale and with the knowl-
edge that finding a root is the reverse of the operation
[of finding a power. Since the A-reading is the square of
the simultaneous D( back) -reading, the D( back) -read-
ing is the square root of the simultaneous A-reading.
To find the square root of 7.8, then, it is only necessary
'to place the indicator on A (left) -78. The simultaneous
ireading of D ( back ) -279 ( Figure 74 ) is sufficient to tell

us that \/7;8 = 2.79, with that single setting of the

indicator. Not only is it not necessary to adjust the
slide carefully back and forth to find the square root, it

isn't necessary to move the slide at all.

By combining the A- and Dl-scales, we can obtain

he reciprocal of the square root. If we wanted —
. , we
vould set the hairline at A (
left ) -78 and note the simul-
aneous reading of DI-358 (Figure 73). We therefore

conclude that -7= = 0.358. ( To locate the decimal

V7.8
153
 I I 1 )

A(Uft|:

I
I imIiiiiIiiiiIiiiiIiiiiIiiiiIiii>IimiIiiiiIiiiiI
,
2
i I i I i I i I i
,31145
I i I i I i I i IiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiI i i i iI i
6.
iiiliiiiliiiiliiiili
7

li

123(SC7««2
D-27|
I I I I I
I
M I I

I
I I I I
I
M I t

I
I I I I
I
M 1 1
j
1 1 n|l M|ilil|nil|iMi|nri|iMi|int|ini|lMi|iili|llil|liii|iiii|
I
I I I I
I
I I t I
I
I I < I
I
I i i
|
i i i 1 1 1 1 1
1
1 1 1 1
1| iiii

Dl I h\ I J. I.M.I, Ill, 1. 1, 1, l.l.l.lihhhLl.l.MJ, hlililllilil ill I < I I I I I || I I I I I I I I . I . I > Inlil I

/7J = 2.79

l//7.8 = 0.358
'*'•*

point, we need only replace 7.8 by the fairly close ap-

i

proximation 9. We then see that —= = Vz = 0.333.

Notice that we determined the square root of 7.8 to
be 2.79 by setting the hairline at A (left) -78. What if we
had set the hairline at A (right) -78? If we do so we
would find a simultaneous reading of D-878. The dif-

ference is a matter of decimal point. We are now deal-

ing not with 7.8 but with 78, and we can say that
V7:8 = 8.78.
In finding a square root, we must first determine
whether to use the A (left) -scale or the A (right) -scale.
To do we deal only with
that the first digit of the number
whose square root we want, converting the other digits
into zero. We then count the number of zeros in the
neighborhood of the decimal point, and for even num-
bers use the A (left) -scale, and for odd numbers the

A (right) -scale, in accordance with the rule given on

page 161.
For the square root of 723, we use the number 700 as

a guide. Since there are an even number of zeros, we

use the A (left) -scale. For the square root of 0.562, we
use 0.5 as the guide. Now we have an odd number of

zeros and use the A (right) -scale.

If your memory fails you, you can always use the
154 I
 1 ,

uilillilllllllllllllllllllll
2
I I I I I I I I
3,«,5
illlilililililililihlihlilililililililililililililonli
6 78911}]
i>lMMhinliinli.iiliHil<,iil..i<liiiil m> J....I ...l

>iiiuiiMii 1
1
1
• I I 1 r I I 1
1
I I
1

1'l
' I'l 'I'Mi'i ('iMi|'|i|'i'ii|i|i|M'i'i'iTi'i'i'iT''''''''l'''''''''l'''''''''l "" I' "<i"'

2tt7eS«3
1 ' ' 1 ' ' 1 ' ' '

9 1 1

2 1 1
,Mllill<l<l<il'< i i li I < I I M I I lrilllllllllllll>l.llil.ll.M.l.llllMI.|il.llill.l.llillMlllili>i ,1 n.M i,,..i,.,.i

Figure 73

method of approximation. Suppose that you want the

square root of 67. If you use the A (left) -67 setting,
you obtain a reading of D( back) -259. On the other
hand, if you use the A (right) -67 setting, you obtain a
reading of D( back) -819. You need not hesitate between
I the two. Since 67 is not far removed from 64, and you
know that the square root of 64 is 8 ( since 8 X 8 = 64 )

I
you have no need to pause in saying that V67 = 8.19.
j(On the other hand, \/6.7 = 2.59.)
Of comse, if you
ihad treated 67 as 60 and noted the odd number of
zeros — just one, that is — in the neighborhood of the
decimal point, you would have used the A ( right ) -scale

at once.

Cube Roots
Just as we could reverse the procedure of finding
squares in order to make it possible for us to find square
roots, so we can reverse the procedure of finding cubes
so as to make it possible for us to find cube roots.
The cube root of b is the number which, when multi-
plied by itself twice, gives fo as a product. If a X a X
a = b (which can also be written a^ = b), then a is the
cube root of b, or, as it is usually written, a = n/£. Cubes
and cube roots are inverses of each other. If b is the
cube of fl, then a is the cube root of b.
We already know that if we simultaneously take a
155
K-reading and a D-reading, the K-reading is the cube of
the D-reading ( see page 176 ) . It follows, therefore, that

the D-reading is the cube root of the K-reading.

Suppose we wanted ^^^T2, Place the hairline at

K( left) -72 and the simultaneous reading of D-193 tells

us that >3^ = 1.93 (Figure 74). The decimal point is

located by the fact that if we replace 7.2 by 8, we know
at once that the ^8 = 2, since 2^ = 8.
By combining the K- and Dl-scales, we can obtain
the reciprocal of the cube root. Since K ( left ) -72 gives a

simultaneous reading of DI-518, we know that ,,

= P
>;J^7.2

0.518. Again, we locate the decimal point by replacing

7.2 with 8. Since — = H = 0.5, == we have our answer.

^8
In finding cube roots, however, we have three K-
scales to choose from for our setting. In finding the cube
root of 7.2, for instance, do we use K( left) -72, K(mid-
Figure 74 K-72

D(tMa)-l93

fl

1
I I I I
I
I I I I
I

1
I I I I

t34«C7«»
|MII|llll|IHI{llll|llll{llll|MII|llll|llll|llll|llll{llM|M(l|llll|llll{ll iiM| 1 1 1
1
1 1 1
1
1| 1 1 1 i|ii ii|i iii|iiii|iiii|iiiitin*4*

m lilililililililililXlilililililllililXlililililililililililililrhlidl.lil.l.lilihliL hill I ill 1 1 I I I I I I I I I I I I I I I I ImmIiiiiIiimIiiiiIii pml

irn^i.n i/V^-ojii Dl-ill

 K( right) -72? The answer hes in the table on
die) -72, or
page 175. For numbers from 1 to 10, we use the K(left)-
scale, between 10 and 100 the K ( middle ) -scale, be-

tween 100 and 1000 the K( right) -scale. For numbers

from 1000 to 10,000, we are back to the K( left) -scale,
and so we progress through the three scales in order

indefinitely.

The same holds true for numbers less than 1. From

1000 down to 100 we use the K( right) -scale, as I have
already said, from 100 down to 10 the K( middle) -scale,
and from 10 down to 1 the K( left) -scale. Progressing
I
then in this direction, we use the K( right) -scale for

j
numbers from 1 down to 0.1, the K ( middle ) -scale for
0.1 down to 0.01 and so on.
• • « « •

In this book I have dealt with twelve scales on the

idide rule. With those twelve scales, we can multiply,

divide, take proportions and reciprocals, obtain squares.

lll|IIHIIIII|HIIUUI)IIIJ«|MMIIII|llll|IIM|HII|HH|Mi|Mr^MI|l|l|l|l|l|l|l|l|m|l|^^^

:
"t|""MHM I
I
I |i MM|MMMM<jMMMM>|MMMM>IMHMMi|MMMMMMMMMi|MMMM'IMMMM<|MMMMMHMMiSi|MM^^

N*l j^ IXlllinllM it M I I I I M I I I I I I I I I I I tMlllMMllMtlMlllllMllMlllllllllMnMlhllllHMlllliniMlt llllMMldlllMlllllltlillllillil

4 ..

157
square roots, cubes, cube roots, logarithms, and anti-

logarithms.
Such calculations (together with additions and sub-
tractions for which we don't need slide rules) probably

take care of virtually all the calculations that we en-

counter in ordinary life.

There are, to be sure, other types of computations

that can be performed on slide rules. Those scales on

my slide rule which I haven't discussed in this book

make it possible to deal with fractional exponents, with
natural logarithms, with trigonometric functions, and
so on.
Then, too, special scales not found on ordinary slide

rules are adapted to the particular needs of those- who

must do calculations in special fields, such as those of
electrical engineering or of business. By bending the
slide rule into circles or spirals, a greater length of scale

(and hence greater accuracy) can be squeezed into a

particular space.
What we have described in this book, therefore, by no
means exhausts the usefulness of the slide rule.
But we have made a good beginning. Armed with the

slide rule, there will be few ordinary problems that need

hold any terrors for you. A push of the slide this way
and that, an adjustment of the hairline, and the answer
is yours!
No wonder that to many people the feel of the slide
rule in the hands spells security.

As I said in the introduction — a man who must carry

out numerical computations constafntly would be as lost

without his slide rule as a doctor without his stethoscope

or a painter without his brush.

158
 1 1

INDEX Decimal ( s ) addition of,

,

25-29
Decimal point, 52-57
slide rule and, 81,88-92
A-scale, 136-137
Direct scale, 35
markings of, 137-139
Division, exponential numbers
multiplication and,
and, 41-42, 44
140-141
and, 44, 41-42
squares and, 144-145, 170
inverse scales and, 126-128
Addition, 10-11
logarithms and, 51-52
addition rule and, 19-31
pi (tt) and, 109-110
decimals and, 25-29
slide rule and, 129,
fractions and, 23-25
107-113
lengths and, 11-12
Addition rule, 17-18, 19ff.
construction of, 17-18
subtractions on, 32-38 Estimations, 30-31, 90-92
Analog computer, 12 Exponent, 40
Answers, estimations of, Exponential numbers, 40
30-3
Antilogarithms, 42-47
slide rule and, 64-69 Folded scales, 92-99, 1 1

Foot-rule, 9-10
Fractions, 23-25
I

B-scale, 137-137
Body, 19
i Hairline, 65

C-scale, 83-84
CF-scale, 99 Index, 84
Characteristic, 53-54 Indicator assembly, 64-65
Cl-scale, 124-125 Inverse scale, 35, 123-130
CIF-scale, 125
Computer, analog, 12
Cube(s), 149-151 K-scale, 149
Cube roots, 155-157, 184

L-scale, 61-63
D-scale, 67 Left index, 84
markings on, 68-78 Logarithm (s), 42, 46, 49
D(back) -scale, 132 decimal, 49-50
DF-scale, 99 decimal point and, 52-57
Dl-scale, 133-134 fractional, 48-50
 1 1

negative,55-57 CIF, 125

58-64
slide rule and, D, 67
Logarithm tables, 50-51 D(back), 132
Logarithmic scale, 71-72 DF, 99
DI, 133
folded, 92-99
inverse, 35, 123-130
Magnitude, order of, 89-90
K, 149
Mantissa, 53
L, 61-63
Multiplication, exponential
logarithmic, 71-72
numbers and, 40-41
markings of, 60-63, 68-78
folded scales and, 102-106
position of, 61-62, 66-67
logarithms and, 51
Secondaries, 60
pi (tt) and, 100-101
Slide, 19
slide rule and, 82-87
Slide rule, 7-8, 14-16 I

accuracy of, 78-79

antilogarithms and, 64-68
Order of magnitude, 89-90 79-80
back of, 131 ff.
cube(s) and, 149-151
Pi (tt), 99-102 cube roots and, 155-156
division by, 1 1 1-1 12 -|;/decimal points and, 8
multiplication by, 100-102 division and, 107-113
reciprocals and, 129-130.> logarithms and, 58-63,
Primaries, 60 79-80
Proportions, 117-123 multiplication and, 82-87 |i

proportions and, 117-122

reciprocals and, 113-117,
128-129
Readings, estimated, 29-3
square(s) and, 144-145
Reciprocals, 1 13-1 17
square roots and, 152-155
inverse scales and, 128-129
Square(s), 141-148
squares of, 146
Square roots, 152-155
Right index, 84
Straightedge, 9
Ruler, 9
Subtraction, 11
addition by, 10-11
addition rule and, 32-38

Scale(s), 35 10-multiples, 46-47

A, 136-137 Tertiaries, 60
B, 136-137 2-multiples, 39-44
C, 83
CF, 99
CI, 124-125 Univac, 12

&/(D6?Dt
V

C/5

CI
M
O
CD

• «

U. OS
O yj
Cm

a.
o <t

pf-ip pOKilDi i-rc-n Hill ir^i-.

NTER

026 2340 2
 (

A COMPUTER
IN YOUR POCKET...
In this bookj Professor With his usual
iVj noted seienti .enthusiasm and talent
kBf, and aetfior^ for making things
Iiitfodycas the reader bgth entertaining and
to the deliglits of j!sij|t^'erstandable,
the slide ryle. *'A Professor Asimov
5?
3 '^ ! says, explains the principles
«s.
[' 'f^ B, :
ni,L., i\-.-R-
^^*r, p\--'^
t*,*:^ ^»tta-
r-^— e
,(>Jt:
%^-
* »
5
Y.
# «
1}-
v-
^c(d V;l^
of the slide rule.
illlpf©SSi¥©3S Once you thoroughly
electronic computer, be understand these
it has many adwaetages, principles — once you
it small enoogh
is know what ybu are
to pyt in f oyr pocket, doing and why— it will be
it need not eost simple to use the slide
^re than a coyple rule on problems that
Jollafs, it can't arise from day to day.
go Oil! of order, (

A delightful book for

andj i»est of a!ij
learning."
it can solve almost arw -HORN BOOK
(o* S « '
*»f- fie t '5w«- 2 illcl

you meat up will

under ordlna

to aii
is simpie to operate.
If yoo know grade*
school aritlinietio,
you can ysa a slide
rule, even thoygh
you may not qyite see
how it works!*'

FAWCETT WORLD LIBRARY