274 The Hon. J. W.

Strutt on the Light from the Sky,

Station. The fact is, the quarries affording this new and beautiful
form of the mineral are in the township of Windsor, near, how-
ever, the boundary line between this and the township of New-
port ; and the name of the latter got to be assodatcd with the
quarries from its being that of the railway-station in their imme-
diate vicinity, from which the gypsum is despatched to be ship-
ped at Windsor. I have lately observed that this locality occa-
sionally affords, by the side of the separate needles of Ulexite,
the hard form of the same mineral, hardness -~ 3 (Phil. Mag.
January 1868), and that the interior of the hard nodules is some-
times very like chalcedony in appearance. Last summer I found
a good many specimens of Howlite in gypsum from Newport
Station, mostly of the chalky variety; a t~w exhibited in an in-
ferior degree the pearly scales of the Winkworth nodules; so
that the borate is not so rare there as it was thought to be.
I add a summary of the localities of the borates in this county
at present known to me, and the composition and matrix of each
mineral.

Localities of Borates in Hanls County~ Nova Scotia,

Borate and Matrix. Localities.
1. Natroborocaleite(Ulexite, Dana), -] Clifton Quarry,Windsor; Brook-
NaO, 2CaO. 5 BO3, 15 HO, [, ville; Treeothiek'sQuarry,Three
r Mile Plains; Winkwo~h ; New-
in gypsum, at ........................... ~J port Station.
2. Cryptomorphite,
NaO, 3CaO, 9BOa, 12H0, / Clifton Quarry, Windsor.
in glauber-sa]t in gypsum, at .........
3. Silicoboro'calcite(Howlite, Dana), ) Brookville; Winkworth; New-
port Station; Noel (Cheverie
4 CaO, 2 SiO2~5 BO3, 5 HO, ~ and Walton reported, but matrix
in gypsum, at ........................... J unknownto me),
in anhydrite, at ........................ Brookville.
4. Winkworthite,
a. 11 CaO, SiO:, 9 SO3, 3BO8, o0HO,
13. 11CaO, SiO2, 8 SO~, 4 BO3, 20HO, ~Winkworth.
in gypsum, at ........................... J

XXXVI. On the Light from the Sky, its Polarization and Colour.
By the Hon. J. W. STRu~T, Fellow of Trinity College, Cambridge.
['Continued from p. 120.]
N the February Number of the Philosophical Magazine I
I have propounded a theory of the scattering of light by par-
ticles which are small in all their dimensions e~)mpared with the
 its Polarization and Colour. 275
wave-length of light, and have applied the results to explain the
phenomena presented by tile sky. Another theory has been
given by Clausius, who attributes the light of the sky to reflec-
tion from water-bubbles, and has developed his views at length
in a series of papers in Poggendorff's Annalen and Crelle's
Journal*.
Starting fi'om the ordinary laws of reflection and ruff.action,
he has no difficulty in showing that, were the atmosphere charged
with globes of water in sufficient quantity to send us the light
which we actually receive, a star instead of appearing as a point
would be dilated into a disk of considerable magnitude. But the
requirements of the ease are satisfied if we suppose the spheres
hollo~, like bubbles ; for then, on account of the parallelism of
the surfaces, but little effect is produced by refraction on a wave
of light. At the same time, if the film be sufficiently thin, the
light reflected from it will be the blue of the first order, and so
the colour of the sky is apparently accounted for.
Apart from the difficulty of seeing how such bubbles could be
formed, there is a formidable objection to this theory, mentioned
by Briieke (Pogg. Ann. vol. lxxxviii, p. 363)--that the blus~ of
the sky is a much better colour than the blue of the first order.
That it is so appears clearly from the measurements quoted in
the February Number, and from the theoretical composition of
the blue of the first orderf. Nor can we escape from this diffi-
culty by supposing, with Briieke, that the greater part of tha
light from the sky has been reflected more than once.
Briicke also brings forward an experiment of great importance
when he shows that mastic precipitated from an alcoholic solu,
tion scatters light of a blue tint. He remarks that it is impos,
sible to suppose that the particles of mastic are in the form of
bubbles.
In his last utterance on this subjeetJ;, Clausius replies to the
objections urged by Brficke and others against his theory, and
shows that, if the illumination of the sky is due to thin plates at
all, those thin plates must be in the form of bubbles. While ad-
* Pogg. Ann. vols. lxxii, lxxvi, lxxxviii. Crellc,vols. xxxiv,xxxvl.
$ I findthat I omitted to explain why it is that the light dispersed from
small particles is of so much richer ahue than that reflectedfrom verythin
films. In the latter case the reflectedwave may be regarded as the sum
of the disturbances originating in the elementary parts of the film, and
these elementary parts may be assimilated to the small particles of the
former supposition. The integration is best effectedby dividing the sur-
face into the zones ofHuyghens; and it is proved in works on physicalop-
tics that the total effect is just half of that due to the first zone. Now the
zones of Ituyghens vary as the wave-length ; and thus it appears that in the
integration the long wavesgain an advantage which diminishesthe original
preponderance of their quicker-timed rivals.
:~ Pogg.Ann. vol. lxxxviii, p. 543.
 276 The Hon. J. W. Strutt on the Light from the Sky,
mitting that if the particles are very small the ordinary laws of
reflection and refraction no longer apply ~', and that therefore
this case is no~ excluded by his argument, he still holds to his
original view as to the nature of the reflecting matter in the sky,
considering that the polarization of the light indicates that it
has undergone regular reflection. His concluding paragraph so
well sums up the case that I cannot do better than quote it.
~cDas Resultat dcr vorstehenden :Betraehtungen kann ich hier-
naeh kurz so zusammenfassen. Soweit man die gewShnlichen
Brechungs- und Reflexionsgcsetze als giiltig anerkennt, glaube
ich auch racine friiheren Schliisse festhalteu zu miissen, n~imlich,
dass in der Atmosphiire Dampfbliischen vorhanden seycn, und
dass sic die Hauptursache der in ihr stattfindenden Lichtreflexion
und ihrer Farben bilden. Nimmt man aber an, die in der
Atmosphiire wirksamen Kiirperchen seyen so klein, dass jene
Gesetze auf sie keine Anwendung mehr finden, dann sind auch
diese Schliisse ungiiltig. Auf diesen Fall ist aber auch die Theorie
der Farben diinner Bliittchen nicht mehr anwendbar, und er
bedarf vielmchr einer neuen Entwickelung, bci welcher noch be-
sonders bcriicksichtigt werden muss, in wiefem diese Annahme
mit der Polarisation des yore Himmel kommenden Lichtes und
mit der angeniihert bekannten GriSsse der in den Wolkea vor-
handenen Wassertheilcheu vereinbar ist."
Clausius does not seem to have followed up the line of research
here indicated. My investigation (written, it so happens, be-
fore seeing Clausius's papers) shows in the clearest manner the
connexion between the smallness of the particles and the polari-
zation of the light scattered from them. Indeed I must remark
that in this respect there is an advantage over the theory of thin
plates, according to which the direction of complete polarization
would be about 76 ° from the sun. It would be a singular coin-
cidence if the action of secondary causes were to augment this
angle to 90°--its observed magnitude. It seems, therefore~ not
too much to say that, if the illumination of the sky were due to
suspended water-bubbles, neither its colour nor its polarization
would agree with what is actually observed.
In his celebrated paper on Fluorescencet, Professor Stokes
makes the following significant remark : ~ " Now this result ap-
pears to me to have no remote bearing on the question of the
direction of the vibrations in polarized light. So long as the

* In many departments of science a temleney may be observed to extend

the field of familiar laws beyond their proper limits. Thus the properties
of gross matter are often assumedto hold equallygood for molecules. An
example more analogousto that which suggeststhis remark is to be found
in the commonexplanation of the mode of action of the speaking-trumpet.
t Phil. Trans. 1852, p. 526.
 its Polarization and Colour. 277
suspended particles are large compared with the waves of light,
reflection takes place as it would from a portion of the sur-
face of a large solid immersed in the fluid, and no conclusion
can be drawn either way. But if the diameter of the particles
be small compared with the length of a wave of light, it seems
plain that the vibrations in a reflected ray cannot be perpendi-
cular to the vibrations in the incident ray." This is the only
passage that I have met with in which the theory of the reflec-
tion of light from very small particles is touched upon.
If it be assumed, as in the theories of Green and Cauehy of
reflection at plane surfaces, that the effect of dense matter is
merely to load the ether, it follows rigorously that the direction
of vibration cannot be turned through a right angle when light
is scattered from small particles. But all we know in the first
instance is that the velocity of propagation of luminous waves is
less in ordinary transparent matter than in vacuum ; and this
may be accounted for as well by a diminished rigidity as by an
increased density. In the first case a scattered ray might be
composed of vibrations perpendicular to those of the incident
beam ; so that the matter is not quite so clear as it would seem
from the argument of Professor Stokes. I believe, however,
that good reasons may be given for rejecting the view that the
difference between media of varying refrangibility is one of rigi-
dity. The point is an important one, and I propose to recur to
it later.
The experiments of Professor Tyndall* with precipitated
clouds exhibit more clearly than had been done by Brficke the
relation between the size of the particles and the nature of the
dispersed light. The observation that the polarization is complete
perpendicular to the track of the incident light is in itself sufficient
to disprove the theory of bubbles. As the particles increase in
magnitude, the azure and polarization are gradually lost. During
the transition a different and more complicated set of phenomena
present themselves, which will furnish a test for the theory when
it is extended so as to include the consideration of particles which
are no longer very small in comparison with the waves of light.
All who have written on this subject seem to have taken for
.granted that the foreign matter in the atmosphere is water or
lee. Even Tyndall, who expressly says that any particles, if
small enough, will do, still believes in the presence of water-par-
ticles. But this view is encumbered with considerable difficulty ;
for even if, in virtue of its transparency to radiant heat, the air
in the higher regions of our atmosphere is at a very low tempe-
rature, it would still be capable of absorbing the very small
quantity of water which is sufficient to explain the blue of the
* Phil. Mag. vol. xxxvii, p. 385. Phil. Trans. 1870.
Phil. Mag. S. 4. Vol. 41. No. 273. April 1871. U
278 On the Light from the Sky, its Polarization and Colour.
sky. At any rate it is difficult to imagine particles of water
smaller than the wave-length endowed with any stability. These
difficulties might perhaps be got over if there were any strong
argument in favour of the water-particles ; but of the existence
of such I am not aware. Every one knows that a blue haze
evidently akin to the azure of the sky obliterates the details and
modifies the eolour of a distant mountain ; and this, when it
occurs on a hot day, cannot possibly be attributed to aqueous
particles. On the face of it, there is no reason for supposing
that near the earth's surface the foreign matter is of one kind
and at a great altitude another. If it were at all probable that
the particles are all of one kind, it seems to me that a strong
ease might be made out for common salt. Be this as it may,
the optical phenomena can give us no clue.
The apparatus by means of which the comparison was made
between sky light and that of the sun diffused through white
paper, was originally arranged for measurements of the absolute
absorption of eoloured fluids for the various rays of the spectrum,
and had been applied rather extensively in experiments having
that object. In the shutter of a darkened room were placed two
slits in the same vertical line, each about three inches long, and
a foot apart. At the other end of the room was an arrangement
of prisms and lenses for producing a pure spectrum on a screen
in the ordinary way. At first only one prism was used ; but I
soon introduced another; and the number might probably be
further increased with advantage. It is even more important to
have a great dispersion in these experiments than in the ordinary
spectroscope. Two spectra would thus be thrown on the screen
one over the other, but by means of a very obtuse-angled prism
situated in front of the dispersion-prisms they are brought
together so as exactly to overlap. The double spectrum thus
formed passes through a horizontal slit in the screen placed so
as to receive it. Close behind is an opaque card carrying a
small vertical-slit, which can be slid along so as to allow any
desired part of the spectrum to pass through. At the beginning
and end of a set of experiments the card is removed, and the
principal fixed lines are observed through an eyepiece and re-
ferred to a scale situated just over the hol;izontal aperture.
When the experimenter looks through the eye.slit iu the di-
rection of the lens, he sees the two parts of the obtuse prism
illuminated with light, in each case homogeneous and, if the
adjustments are properly made, belonging to the same part of the
spectrum. By varying the breadths of the original slits, the two
parts of the field may he made equally bright; and when the
match is attained, the breadths are inversely proportional to the
richness of the lights behind them in the homogeneous ray under
 On Attraction caused by Vibrations of the Air. 279
consideration. But if the object be to make a complete com-
parison between two lights, it is often more convenient to leave
the widths of the slit~ arbitrary, and then, by sliding the card, to
seek that part of the spectrum which allows a match. It was in
this way that the observations on the light of the sky were made.
To give an idea of the degree of accuracy to which the compa-
risons may be made, I may mention that in my experiments on
absorption, the means of six observations were usually correct
to about one in 50 or 60. In the less-luminous parts of the
spectrum the error might be somewhat greater.
The difficulty, however, of getting a satisfactory result with
the blue of the sky does not lie in the inaccuracy of the measure-
ments, but in the arbitrary character of the light with which it
is compared. In order to test the theory in a strict manner, the
second light ought to be similar in composition to that which
lights up the sky. Now the sky is lit not only by the direct
rays of the sun, but also by itself and by the bright surface of
the earth. It is evident, therefore, that the requirements of the
case are very imperfectly met by taking as the second light that
of the sun as received by us, even if the translucent material
through which we diffuse it effects no change in the quality. A
nearer approximation to what we want would probably be found
in the diffused light of a thoroughly cloudy day. But here we
meet with an experimental difficulty; for the method described
is only available to compare two lights both given at once. A
suitable artificial light might no doubt be used as a middle term
to be afterwards eliminated ; but a candle or a lamp would hardly
be available, on account of the yellowness of their light. On the
other hand, the bluer radiation from burning magnesium would
probably be inconvenient, and difficult to keep constant in quality
from day to day. I am, however, in hopes that, by a method
founded on a different principle, I may be able to compare the
blue (ff one day's clear sky with the white light from the clouds
on another.
Terling Place, Witham,
March 6, 1871.

XXXVII. On Attraction caused by Vibrations of the Air.

By Professor CHALLIS,M.A., F.R.S., F.R.A.S.*

A PTER concluding from reasoning founded on hydrodyna-

mieal principles, in my work ' On the Principles of Ma-
thematics and Physics,' that the vibrations of an elastic fluid,
such as the air, are capable of causing a permanent motion of
• Communicatedby the Author.
U2