On the Change of Refrangibility of Light

Early Journal Content on JSTOR, Free to Anyone in the World

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in
the world by JSTOR.

Known as the Early Journal Content, this set of works include research articles, news, letters, and other
writings published in more than 200 of the oldest leading academic journals. The works date from the
mid-seventeenth to the early twentieth centuries.

We encourage people to read and share the Early Journal Content openly and to tell others that this
resource exists. People may post this content online or redistribute in any way for non-commercial
purposes.

Read more about Early Journal Content at http://about.istor.org/participate-istor/individuals/early-
journal-content .

JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people
discover, use, and build upon a wide range of content through a powerful research and teaching
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please
contact support@jstor.org.

[ 463 ]

XXX. On the Change of RefrangiUlity of Light. By G. G. Stokes, M,A., F.R.S,,
Fellow of Pembroke College, and Lucasian Professor of Mathematics in the
University of Cambridge,

Received May 11,— Read May 27, 1852,

1. IhE following researches originated in a consideration of the very remarkable
phenomenon discovered by Sir John Herschel in a solution of sulphate of quinine,
and described by him in two papers printed in the Philosophical Transactions for 1845*,
entitled ^On a Case of Superficial Colour presented by a Homogeneous Liquid
internally colourless/ and ^ On the Epipolic Dispersion of Light.' The solution of
quinine, though it appears to be perfectly transparent and colourless, like water
when viewed by transmitted light, exhibits nevertheless in certain aspects, and under
certain incidences of the light, a beautiful celestial blue colour. It appears from the
experiments of Sir John Herschel that the blue colour comes only from a stratum of
fluid of small but finite thickness adjacent to the surface by which the light enters.
After passing through this stratum, the incident lights though not sensibly enfeebled
nor coloured, has lost the power of producing the same effect, and therefore may be
considered as in some way or other qualitatively diflTerent from the original light. The
dispersion which takes place near the surface of this liquid is called by Sir John
Herschel epipolic, and he applies the term epipolized to abeam of light which, having
been transmitted through a quiniferous solution, has been thereby rendered incapable
of further undergoing epipolic dispersion. In one experiment, in which sun-light was
used, a feeble blue gleam was observed to extend to nearly half an inch from the
surface. As regards the dispersed light itself, when analysed by a prism it was found
to consist of rays extending over a great range of refrangibility : the less refrangible
extremity of the spectrum was however wanting. On being analysed by a tourma-
line, it showed no signs of polarization. A special experiment showed that the
dispersed light was perhaps incapable, at any rate not peculiarly susceptible, of being
again dispersed.

2. In a paper ^ On the Decomposition and Dispersion of Light within Solid and
Fluid Bodies,' read before the Royal Society of Edinburgh in 1846, and printed in
the 16th volume of their Transactions, as well as in the Philosophical Magazine for
June 1848, Sir David Brewster notices these results of Sir John Herschel s, and
states the conclusions, in some respects different, at which he had arrived by operating
in a different way. The phenomenon of internal dispersion had been discovered by
him some years before, and is briefly noticed in a paper read before the Royal Society

MDCCCLII. 3 o

464 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

of Edinburgh in 1833^. It is described at length, as exhibited in the particular case
of fluor-spar^ in a paper communicated to the British Association at Newcastle in
IS^S-f-. In Sir David Brewster's experiments the sun's light was condensed by a
lens, and so admitted into the solid or fluid to be examined ; which afforded peculiar
facilities for the study of the phenomena. On examining in this way a solution of
sulphate of quinine, it was found that light was dispersed, not merely close to the
surface, but at a long distance within the fluid : and Sir David Brewster was led to
conclude that the dispersion produced by sulphate of quinine was only a particular
case of the general phenomenon of internal dispersion. On analysing the blue beam
by a rhomb of calcareous spar, it was found that a considerable portion of it, con-
sisting chiefly of the less refrangible i*ays, was polarized in the plane of reflexion,
while the more refrangible of its rays, constituting an intensely blue beam, had a dif-
ferent polarization.

3. On repeating some of Sir John Herschel's experiments, I was immediately
satisfied of tbe reality of the phenomenon, notwithstanding its mysterious nature, that
is to say, that an epipolized beam of light is in some way or other qualitatively
different from the light originally incident on the fluid. On making the observation
in the manner of Sir David Brewster, it seemed no less evident that the phenomenon
belonged to the class of internal dispersion:}:. Nevertheless, the singular phenomenon
discovered by Sir John Herschel manifested itself even in this mode of observation^
If indeed the vessel containing the solution were so placed that the image of the sun
in the focus of the lens lay a little way inside the fluid, the phenomenon was masked,
because the increase of intensity due to an increase of concentration in approaching
the focus made up for the decrease of intensity due to passing out of the blue band.
But when the vessel was moved so that the focus of the lens fell either further inside
the fluid or else outside the vessel, the narrow blue band adjacent to the surface was
seen as well as the blue beam which shot far into the fluid. Light which has been
" epipolized " by transmission through a moderate thickness of the solution is indeed
capable of undergoing further dispersion, but not epipolic dispersion, if that term be
restricted to the dispersion by which the narrow blue band is produced. It was no
doubt of great importance to assign to the phenomenon its true place as a member
of the class of phenomena of internal dispersion. Nevertheless the mystery was by
no means cleared up ; rather, we were prepared to expect something of the same sort

* Edinburgh Transactions, vol. xii. p. 542.

t Eighth Report. — Transactions of the Sections, p. 10.

X. By this, I merely mean that, to take a particular example, the exhibition of a blue light by a solution of
sulphate of quinine appeared to be a phenomenon of the same nature as the exhibition of a red light by a solution
of the green colouring matter of leaves, although the latter does not manifest the same singular concentration
as the former in the neighbourhood of the surface by which the light enters ; and the latter had already been
observed by Sir David Brewster, and the phenomenon designated as internal dispersion. I make this remark
because Sir David Brewster has applied this same term to another class of phenomena which are totally
different.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 465

in Other instances of internal dispersion. In fact, the mystery consisted, not in the
narrowness of the stratum from which most of the blue light came, but in the cir-
cumstance that it was possible for light, by passing across such a stratum, to be
deprived of the power of producing the same effect again, without, apparently, being
altered in any other respect.

4. To one who regards light as a subtle and mysterious agent, of which the laws
indeed are in a good measure known to us, but respecting the nature of which we are
utterly ignorant, the phenomenon might seem merely to make another striking
addition to the modes of decomposition with which we were already acquainted.
But in the mind of one who regards the theory of undulations as being for light what'
the theory of universal gravitation is for the motions of the heavenly bodies, it was
calculated to excite a much more lively interest. Whatever difficulty there might be
in explaining how the effect was produced, we ought at least to be able to say what
the effect was that had been produced; wherein, for example, epipolized light differed
from light which had not undergone that modification.

In speculating on the nature of the phenomenon, there is one point which deserves
especial attention. Although the passage through a thickness of fluid amounting to
a small fraction of an inch is sufficient to purge the incident light from those rays
which are capable of producing epipolic dispersion, the dispersed rays themselves
traverse many inches of the fluid with perfect freedom. It appears therefore that the
rays producing dispersion are in some way or other of a different nature from the
dispersed rays produced. Now, according to the undulatory theory, the nature of
light is defined by two things, its period of vibration, and its state of polarization.
To the former corresponds its refrangibility, and, so far as the eye is a judge of colour,
its colour =^. To a change, then, either in the refrangibility or in the state of polari-
zation we are to look for an explanation of the phenomenon.

5. Regarding it at first as an axiom that the dispersed light of any given refrangi-
bility could only have arisen from light of the same refrangibility contained in the
incident beam, I was led to look in the direction of polarization for the required
change in the nature of the light. Since a fluid has no axes, circular polarization is

* It has been maintained by some philosophers of the first eminence that light of definite refrangibility may
still be compound, and though no longer decomposable by prismatic refraction might still be so by other means.
I am not now speaking of compositions and resolutions depending upon polarization. It has even been
suggested by the advocates of the undulatory theory, that possibly a difference of properties in lights of the
same refrangibility might correspond to a difference in the law of vibration, and that lights of given refrangibility
may differ in tint, just as musical notes of given pitch differ in quality. Were it not for the strong conviction
I felt that light of definite refrangibility is in the strict sense of the word homogeneous, I should probably have
been led to look in this direction for an explanation of the remarkable phenomena presented by a solution of
sulphate of quinine. It would lead me too far from the subject of the present paper to explain the grounds of
this conviction. I will only observe that I have not overlooked the remarkable effect of absorbing media in
causing apparent changes of colour in a pure spectrum ; but this I believe to be a subjective phenomenon, de-
pending upon contrast.

3 o 2

466 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT.

the only kind which can here come into play. As some fluids are doubly refi^acting^
transmitting right-handed and left-hemded circularly polarized light with different
velocities, so, it might be, this fluid was doubly absorbing, absorbing say right-handed
circularly polarized light of certain refrangibilities with great energy, and freely
transmitting left-handed. The right-handed light, absorbed, in the sense of with-
drawn from the incident beam, might have been more strictly speaking scattered,
and thereby depolarized. The common light so produced would be equivalent to two
streams, of equal intensity, one of right-handed, and the other of left-handed circu-
larly polarized light. Of these the latter would be freely transmitted, while the
former would be scattered anew, and so on. Yet this hypothesis, sufficiently impro-
bable already, was not enough. New suppositions were still required^ to account for
the circumstance that an epipolized beam, when subjected to prismatic analysis with
a low magnifying power, exhibited no bands of absorption in the region to which, as
regards their refrangibility, the dispersed rays principally belong ; so that altogether
this theory bore not the slightest semblance of truth.

6. I found myself thus fairly driven to suppose that the change of nature consisted
in a change of refrangibility. From the time of Newton it had been believed that
light retains its refrangibility through all the modifications which it may undergo.
Nevertheless it seemed to me less improbable that the refrangibility should have
changed, than that the undulatory theory should have been found at fault. And
when I reflected onthe extreme simplicity of the whole explanation if only this one
supposition be admitted^ I could not help feeling a strong expectation that it would
turn out to be true. In fact, we have only to suppose that the invisible rays beyond
the extreme violet give rise by internal dispersion to others which fall within the limits
of refrangibility between which the retina of the human eye is affected, and the expla--
nation is obvious. The narrowness of the blue band observed by Sir John Herschel
would merely indicate that the fluid, though highly transparent with regard to the
visible rays, was nearly opake with regard to the invisible. According to the law of
continuity^ the passage from almost perfect transparency to a high degree of opacity
would not take place abruptly ; and thus rays of intermediate refrangibilities might
produce the blue gleam noticed by Sir John Herschel, or the blue cylinder, or
rather cone, observed by Sir David Brewster. We should thus, too, have an imme-
diate explanation of a remarkable circumstance connected with the blue band, namely
that it can hardly be seen by strong candle-light, though readily seen by even weak
daylight. For candle-light, as is well known, is deficient in the chemical rays situ-
ated beyond the extreme violet.

7. My first experiments were made with coloured glasses. A test tube was about
half filled with a solution consisting of disulphate of quinine dissolved in 200 times
its weight of water acidulated with sulphuric acid. The tube, having been first
covered with black paper, with the exception of a hole by which the light might
enter, was placed in a vertical position in front of a window^ the hole being turned

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILItY OF LIGHT. 467

towards the light. On looking down from above^in a direction nearly parallel to the
surface of the glass, a blue arc was well seen, extending only a very short distance
into the fluid, and situated immediately behind the hole. As this arc, though
extremely distinct, was not of course what could be called brilliant, I did not at first
venture, for the experiment I had in view, to use any but pale glasses. Having no
direct means of determining which were opake with regard to the invisible rays
situated beyond the extreme violet, I sought among a collection of orange, yellow,
and brown glasses, which, from transmitting mainly the less refrangible rays, seemed
the most likely to absorb the chemical rays. I presently foiind a pale smoke-coloured
glass, which, when placed immediately in front of the hole, prevented the formation of
the blue arc, although when placed immediately in front of the eye it transmitted a
large proportion of the light of which the arc consisted. The colour of the arc was
of course modified, and rendered more nearly white.

On trying other pale glasses, I found one of a puce colour, which, when placed in
front of the hole, allowed the arc to be formed, though it absorbed it when placed in
front of the eye. A yellow, and likewise a yellowish green glass allowed the arc to
be seen in both positions; but its colour was decidedly diflferent according as the
glass was placed in front of the hole or in front of the eye. The breadth, too, of the
arc was differently affected by different coloured glasses placed in front of the hole,
some causing the light to be more, and others less concentrated towards the surface
of the test tube than when the incident light was unimpeded.

8. The sun's light was next reflected horizontally into a darkened room, and allowed
to pass through a hole in a vertical board which was placed in the window. The
hole contained a lens of rather short focus. On placing a test tube containing the
solution, in a vertical position, in front of the lens, at such a distance that the focus
lay some way inside the fluid, the narrow blue band described by Sir John Herschel
and the blue beam mentioned by Sir David Brewster were seen independently of
each other. On trying different coloured glass*es, which were placed, first in front of
the fluid, and then in front of the eye, it was found that the blue beam, as had
previously proved to be the case with the narrow band, was for the most part dif-
ferently affected according as the glass was placed so as to intercept the incident or
the dispersed light. Moreover, the long blue beam and the narrow band did not
behave in the same manner under the action of the same coloured glass.

9. To my own mind these experiments were conclusive as to the fact of a change
of refrangibility. Admitting that the effect of a coloured glass is simply to stop a
certain fraction of the incident light, that fraction being a function of the refrangi-
bility, it is plain that the results can be explained in no other way. It must be
confessed however that these results are merely an extension of that which precisely
constitutes the peculiarity of the phenomenon. For, take the case of the narrow blue
band formed by ordinary daylight. Imagine a glass vessel with parallel sides to be
filled with a portion of the solution, and placed so as to intercept, first the incident.

468 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OP LIGHT-

and then the dispersed light. In the first position the light incident on the fluid
under examination would be ^^epipolized" by transmission through the fluid contained
in the vessel, and therefore the blue band would be cut off, whereas when the vessel
was held in front of the eye the blue band would be freely transmitted. Hence the
effects of the coloured glasses are analogous to, but less striking than, the effect of
a stratum of the solution of sulphate of quinine in the imaginary experiment above
described. There is to be sure one important difference in the two cases, namely^
that in the case of the stratum of fluid the epipolic dispersion which is prevented in
the fluid under examination is produced near the first surface of the stratum, whereas
no such dispersion is produced, or at any rate necessarily produced, in the coloured
glasses. Whatever the reader may think of the results obtained with coloured glasses,
the next experiment it is presumed will be deemed conclusive.

10. The board in the window containing the lens having been replaced by a pair
of boards adapted to form a vertical slit, the sun's light was reflected horizontally
through the slit, and transmitted through three Munich prisms placed one after the
other close to it. A tolerably pure spectrum was thus formed at the distance of some
feet from the slit. A test tube containing the solution was then placed vertically a
little beyond the extreme red of the spectrum, and afterwards gradually moved
horizontally through the colours. Throughout nearly the whole of the visible spec-
trum the light passed through the fluid as it would have done through so much
water; but on arriving nearly at the violet extremity a ghost-like gleam of pale blue
light shot right across the tube. On continuing to move the tube, the blue light at
first increased in intensity and afterwards gradually died away. It did not however
cease to appear until the tube had been moved far beyond the violet extremity of the
spectrum visible on a screen. Before disappearing, the blue light was observed to
be confined to an excessively thin stratum of fluid adjacent to the surface by which
the light entered, whereas when it first appeared, namely when the tube was placed a
little short of the extreme violet, the* blue light had extended completely across iL
It was certainly a curious sight to see the tube instantaneously lighted up when
plunged into the invisible rays : it was literally darkness visible. Altogether the phe-
nomenon had something of an unearthly appearance.

11. Since the fluid is so intensely opake with regard to rays of extreme refrangi-
bility, it might be expected, that, though it appears transparent and colourless when
examined merely by viewing a white object through it, it would yet exhibit a very
sensible absorbing action with regard to the extreme violet rays when subjected to
prismatic analysis. To try whether such were really the case, I reflected the sun's
light horizontally through a slit, at which was placed a test tube filled with the liquid,
and analysed the line of light by a prism, the eye being defended by a deep blue glass.
I was barely able to make out the fixed line H in Plate XXV., that is, the less refrangible
band of the pair, although in similar circumstances I can generally see about as far
beyond the more refrangible band as it is beyond H. However, to make the result

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 469

more decisive by using a greater thickness^ as well as to render the observation strictly
differential, I placed a tumbler filled with water behind the slit, the blue glass before
it, and then viewed the slit through the prism. 1 saw as far as usual into the violet.
The water was then poured out and replaced by the solution of sulphate of quinine,
which, when viewed by transmitted light, appeared as transparent as the water which
it had replaced. When the tumbler was now placed behind the slit, the blue beam
of dispersed light was observed to extend quite across it, a distance of about three
inches, and. would evidently have gone much further. On viewing the slit through
the prism, the spectrum was found to be cut off about half-way between the fixed
lines G and H. The termination was pretty definite, which indicates that, at least
for that part of the spectrum, the absorbing energy of the fluid rapidly increased with
the refrangibility of the light ; there was, however, an evident diminution of intensity
produced by the fluid, extending from the termination of the spectrum to near G.

12. There could no longer be any doubt, either as to the fact of a change of re-
frangibility, or as to the explanation thereby of the remarkable phenomenon exhi-
bited by sulphate of quinine. Epipolized light is merely light which has been purged
of the invisible, or at most feebly illuminating rays more refrangible than the violet ;
and the term itself, which in fact was only adopted provisionally by Sir John
Herschel, and which has now served its purpose, may henceforth be discarded,
especially as it is calculated to convey a false impression respecting the cause of the
phenomenon. It remained to examine other instances of internal dispersion, of
which, according to Sir David Brewster's observations, the dispersion produced by
sulphate of quinine is only a particular case ; to endeavour to make out the laws
according to which a change of refrangibility takes place; and, if possible, to
account for these laws on mechanical principles.

13. In giving an account of my further experiments, I think it best to describe in
detail the phenomena observed in some of the more remarkable instances of internal
dispersion before attempting to draw any general conclusions. It will save repetition
to explain in the first instance the methods of observation employed, which on the
whole may very fairly be divided into four, though occasionally it was convenient to
employ intermediate methods, or a combination of two of them. Of course I fre-
quently availed myself of Sir David Brewster's method of observation, in which the
effect of the incident light is studied as a whole; but the methods here referred to
relate to an investigation of the separate oflSces of the portions of light of different
degrees of refrangibility which are found in the incident beam. As my researches
proceeded, new methods of observation suggested themselves, but these will be
described in their place.

Methods of Observation employed.
First Method. — The sun's light was reflected horizontally through a small lens,
which was fixed in a hole in a vertical board. The cone of emergent rays was

470 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

allowed to enter the solid or fluid examined. A coloured glass or other absorbing
medium was then placed, first so as to intercept the incident rays, and then between
the substance examined and the eye. For shortness' sake these positions will be
designated as the first and the second. Sometimes a coloured glass was allowed to
remain in front of the hole, and a second glass was added, first in front of the hole
and then in front of the eye.

Second Method.— The sun's light, reflected as before, was transmitted through a
series of three or four Munich prisms placed one immediately after the- other, and
each nearly in the position of minimum deviation. It was then transmitted through
a small lens in a board close to the last prism, and so allowed to enter the body to
be examined, which was generally placed so that the first surface coincided, or nearly
so, with the focus of the lens. The diameter of the lens was much smaller than the
breadth or height of the prisms, so that the lens was completely filled with white
light, the component parts of which however entered in diffferent directions. Re-
garding the image of the sun in the focus of the small lens as a point, we may con-
ceive the light incident on the body under examination as consisting of a series of
cones, corresponding to different refrangibilities, the axes of which lay in a horizontal
plane and intersected in the centre of the lens, the vertices being arranged in a hori-
zontal line near the surface of the body examined.

Third Method. — The sun's light was reflected horizontally through a vertical slit,
and received on the prisms, which were arranged as before,, but placed at the distance
of several feet from the slit. A large lens of rather long focus was placed imme-
diately after the last prism, with its plane perpendicular, or nearly so, to the beam of
light which had passed through the prisms, and with its centre about the middle of
this beam. The body examined was placed at the distance of the image of the slit,
or nearly so.

Fourth Method. — Everything being arranged as in the third method, a board
with a small lens of short focus was placed at the distance of the image of the slit, or
between that and the image of the sun, which was a little nearer to the prisms, inas-
much as the focal length of the large lens commonly employed, though much smaller^
was not incomparably smaller than the distance of the lens from the slit. A second
slit was generally added immediately in front of the small lens. The body examined
was placed at the focus of the small lens. The dispersed light was viewed from above,
and analysed by a prism, being refracted sideways.

The object of these several arrangements will appear in the course of the paper.
The prisms employed consisted, three of them of flint glass and one of crown. The
refracting angles of the former were about 43°, 33°, and 24°, and that of the latter
about 45°. The refracting faces of the smallest of the prisms (the flint of 43°) were
r35 inch high and 1'60 long. The small lens used was one or other of a pair of which
the apertures were 0*34 inch and 0*22 inch, and the focal lengths 0*75 inch and 0*50
inch. The focal length of the large lens generally used was about twelve inches.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 471

Once or twice a lens was tried which had a focal length about three times as greats
but the light proved too faint for most purposes. In the third method it was some-
times convenient to employ a lens of only 6^ inches focal length, but the 124nch lens
was employed in the fourth method, except on a few occasions, when the lens of 36
inches focal length was used. With the 12-inch lens the length of the spectrum from
the fixed line B to H was usually about an inch and a quarter.

It will be convenient for the purposes of this paper to employ certain terms in a
particular sense, but as some of these terms relate to phenomena which have not yet
been described,Jt will be well previously to relate in detail what was observed in one
remarkable instance of internal dispersion.

Solution of Sulphate of Quinine >

14. The effects of some pale coloured glasses in the case of this fluid have already
been mentioned. But there is one glass of which the effect is still more striking. It
is well known that a deep cobalt blue glass is highly transparent with regard to the
chemical rays. Accordingly I found that a blue glass, so deep that only the brighter
objects in a room could be seen through it, produced but very little effect when placed
so as to intercept the light incident on the fluid. When placed immediately in front
of the eye, at first everything disappeared except the light reflected from the con-
vexities of the glass tube ; but when the eye became a little accustomed to the dark-
ness it was possible to make out the existence of the band. The contrast between
the effects of this glass and of the pale brown glass already mentioned was most
striking.

15. When the fluid was examined by the second method, the dispersed light was
found to consist of two beams, separated from each other at their entrance into the
fluid, that is, at the vertical surface of separation of the fluid and the containing
vessel, and afterwards still further separated by divergence. Of course each beam
must have been made up of a series of cones having their axes diverging from the
centre of the lens, and their vertices situated at its focus. The first beam, or that
which was produced by light of less refrangibility, consisted of the brighter colours
of the spectrum in their natural order. It had a discontinuous, sparkling appearance,
and was plainly due merely to motes which were suspended in the fluid. On being
viewed from above through a Nicol's prism, it was found to consist chiefly of light
polarized in the plane of reflexion. Taken as a whole, it served as a fiducial line to
which to refer the position of the second beam, and thereby judge of the refrangibility
of the rays by which it was produced.

This second beam was a good deal the brighter of the two. Its colour was a beau-
tiful sky-blue, which was nearly the same throughout, but just about its first border,
that is, where it arose from the least refrangible of those rays which were capable of
producing it, the colour was less pure. It had a perfectly continuous appearance.
When viewed from above through a doubly refracting achromatic prism of quartz,

MDCCCLII. 3 p

472 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT.

which allowed a direct comparison of the two images, it offered no traces of polar-
ization. It was produced by light polarized in a vertical or horizontal plane as well
as by common light, and in that case, as well as in the former, manifested no traces
of polarization^.

The short distance that the more refrangible rays were able to penetrate into the
fluid might readily be perceived in this experiment, but the second method of obser-
vation was not adapted to bring out this part of the phenomenon,

16. On examining the fluid by the third method, the result was very striking,
although of course only what might have been anticipated. The principal fixed lines
of the violet, and of the chemical parts of the spectrum beyond, were seen with
beautiful distinctness as dark planes interrupting an otherwise perfectly continuous
mass of blue light. To see any particular fixed line with most distinctness, it was of
course necessary to hold the eye in the corresponding plane, when the dark plane
was foreshortened into a dark line. From the red end of the spectrum, as far as the
line G, or thereabouts, the light passed freely through the fluid, or at least was only
reflected here and there from motes held in mechanical suspension. About G the
dispersion just commenced to be sensible, and there were traces of that line seen as a
dark plane interrupting a mass of continuous but excessively faint light. For some
distance further on the dispersed light remained so faint that it might have been
passed over if not specially looked for. It was about half-way between G and H, or
a little before, that it first became so strong as to arrest attention, and a little further
on it became very conspicuous, the tint meanwhile changing to a pale sky-blue. The
light was very copious about the two broad bands of the group H, and for some
distance from H towards G. Some of the fixed lines less refrangible than H were
very plain, and beyond H a good number were visible, which will presently be further
described. The whole system of fixed lines thus visible as interruptions in the di-
spersed light had a resolvable appearance ; but with a very narrow slit and a lens of
long focus at the prisms the light would have been too faint for convenient obser-
vation.

The dispersed light about G, and for some distance further on, was so veiy faint
that I might have overlooked it had it not arrested my attention when observing by
the fourth method ; indeed, I have sometimes specially looked for it in the third
arrangement without having been able to see it. Practically speaking, the dispersion
might be said to commence about half-way between G and H.

* These two results, namely, that the blue beam which constitutes the greater part of the light dispersed
by a solution of sulphate of quinine is unpolarized, or according to his expression possesses a quaquaversus
polarization, and that that still remains the case when the incident light is polarized, have been already
announced by Sir David Beewstek, who appears to have been led to attend to the polarization of the hght
from Sir John Herschel^s observation, that the blue light arising from epipolic dispersion in a solution of sul-
phate of quinine was unpolarized. It seemed important however to repeat the observation on the blue beam
obtained in a state of isolation.

PROFESSOR STOKIS ON THE CHANGE OF REFRANGIBILITf OF LIGHT. 473

17. On refracting the whole system sideways through a prism of moderate angle
held in front of the eye^ the fixed lines became confused, and the finer ones disap-
peared. The edges of the broad bands H were tinged with prismatic colours, like
the edges of two slips of black velvet placed on a sheet of pale blue paper, and
viewed through a prism. This experiment exhibits the compound character of the
dispersed light, notwithstanding the perfect homogeneity of the incident light.

18. The third method of observation is well adapted to bring into view the
variation in the absorbing energy of the medium corresponding to a variation in the
refrangibility of the incident rays. When the eye is placed vertically over the vessel
containing the solution, so that the dark planes corresponding to the fixed lines of
the spectrum are projected into dark lines, of which the length is not exaggerated by
obliquity, the boundary of the dispersed light is projected into a curve, which serves
to represent to the eye the relation between the absorbing power of the medium and
the refrangibility of the incident light. This curve is not exactly that which {:5ir John
HiRSCHBL has treated of in the theory of absorption, and considered as the type of
the absorbing medium to which it is applied, but nevertheless it serves much the same
purpose. It is true, that, independently of any change in the absorbing energy of the
medium, an increasing faintness in the dispersed light would produce to a certain
extent an approximation of the curve to its axis ; but practically, in the case of sul-
phate of quinine, as well as in a great many others, the appearance is such as to leave
no doubt as to the existence of a most intense absorbing energy on the part of the
medium with respect to rays of very high refrangibilities*.

In the case of a solution of sulphate of quinine of the strength of one part of the
disulphate to 200 parts of acidulated water, it has been already stated that a portion
of the rays which are capable of producing dispersed light passed across a thickness
of 3 inches. On forming a pure spectrum, the fixed line H was traced about an inch
into the fiuid. On passing from H towards G, the distance that the incident rays
penetrated into the fluid increased with great rapidity, while on passing in the con-
trary direction it diminished no less rapidly, so that from a point situated at no great
distance beyond H to where the light ceased, the dispersion was confined to the im-
mediate neighbourhood of the surface. When the solution was diluted so as to be
only one-tenth of the former strength, a conspicuous fixed line, or rather band of
sensible breadth, situated in the first group of fixed lines beyond H, was observed to
penetrate about an inch into the fluid. On passing onwards from the band above-
mentioned in the direction of the more refrangible rays, the distance that the incident
rays penetrated into the fluid rapidly decreased, and thus the rapid increase in the
absorbing energy of the fluid was brought into view in a part of the spectrum in

* I should here remark, that, alter the researches described in this paper had far advanced, I met accidentally
witii a passage in the Comptes Rendus, torn. XTii. p. 883, in which M. En. BicaniEiii mentions a solution of
mid sulphate of quinine m a medium emkently remarkable for ite absorbing power with respect to tibe rays
more refrimgible than H .

3p2

474 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT.

which, with the stronger solution^ it could not be so conveniently made out^ inasmuch
as the posterior surface of the space from which the dispersed light came almost con-
founded itself with the anterior surface of the fluid.

The high degree of opacity with regard to rays of great refrangibility which the
addition of so small a proportion of sulphate of quinine is sufficient to produce in
water is certainly very remarkable; nevertheless it is only what I have constantly
observed while following out these researches.

19. In observing by the fourth method^ the part of the spectrum to which the in-
cident light belonged was determined sometimes by the colour, sometimes by means
of the fixed lines of the spectrum. It almost always happened that there were motes
enough suspended in the fluid to cause a portion of the dispersed beam to consist
merely of light which had undergone ordinary reflexion. When the whole dispersed
beam was analysed by a prism, the beam which consisted of light reflected from
motes was separated from the rest ; it was in general easily recognised by its spark-
ling appearance, but at any rate was known by its consisting almost wholly of light
polarized in the plane of incidence^ whereas the truly dispersed light was unpolarized.
It consisted of course of light of definite refrangibility, the same as that of the in-
cident light, and thus served as a fiducial line to which to refer by estimation the
refrangibilities of the component parts of the dispersed light. Of course this part of
the observation was possible only when the incident rays belonged to the visible part
of the spectrum.

On moving the lens horizontally through the colours of the spectrum, in a direction
from the red to the violet, it was found that the dispersion was first perceptible in the
blue. When the dispersed light was separated by a prism from the light reflected
from motes, it was found to consist of an exceedingly small quantity of red; further
on some yellow began to enter into its composition ; further still, perhaps about the
junction of the blue and indigo, the dispersed beam began to grow brighter, and
was found on analysis to contain some green in addition to the former colours. In
the indigo it got still brighter, and when viewed as a whole was somewhat greenish.
Further still it became something of a pale slaty blue, and was found on analysis to
contain some indigo, or at least highly refrangible blue. On proceeding further the
dispersed light became first of a deeper blue and then, a little short of the fixed line
H, whiter. At a considerable distance beyond H the dispersed light was if anything
a shade more nearly white.

By this method of observation the dispersion can be detected earlier in the
spectrum than by the third method, and moreover the change in the colour of the
dispersed light is much more easily perceived ; indeed the most striking part of this
change takes place while the dispersed light is so very faint that it can hardly be seen
in observing by the third method ; moreover, even in the bright part of the dispersed
beam, it is not at all easy by the latter method to make out the change of tint corre-
sponding to a change in the refrangibility of the incident rays, because the tint

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OP LIGHT. 475

changes so gradually and so slightly that the eye glides from one part of the dispersed
beam to another without noticing any change.

20. It has been already mentioned that the blue beam of dispersed light seen in a
solution of sulphate of quinine was produced whether the incident light was polar-
ized in or perpendicularly to the plane of reflexion, or more properly plane of di-
spersion, that is, the plane containing the incident ray and that dispersed ray which
enters the eye. A question naturally presents itself, whether the intensity of the di-
spersed light is strictly the same in the two cases. By combining a lens of rather
short focus and a doubly refracting prism with the four prisms, I satisfied myself that
the difference of intensity, if there were any, was not great, but the experiment pre-
sented some practical difficulties. However, the result of the following experiment
appeared to be as decisive as a negative result could well be.

The arrangement being the same as in the third method, but the lens in front of
the prisms having a focal length of only 6*5 inches, the incident light was polarized
in a vertical plane previously to passing through the slit, by transmission through a
pile of plates. The two beams of light were seen as usual in the fluid, namely, the
blue beam due to internal dispersion, and the fainter coloured beam due to motes.
The former of these, which was quite separate from the latter, exhibited the principal
fixed lines belonging to the highly refrangible part of the spectrum. A plate of
selenite was then interposed immediately in front of the vessel, so as to modify the
polarization of the light entering the fluid. This plate was obtained by an irregular
natural cleavage, and was cemented with Canada balsam between two discs of glass.
When examined by polarized light it exhibited a succession of beautiful and varied
tints, according to the various thicknesses of the different parts. Now when the
plate was moved about in front of the vessel, without altering its perpendicularity to
the incident light, different portions of the beam due to motes were observed to dis-
appear and reappear, or at least to become faint and then bright again, so that a
person ignorant of the cause, and not looking at the disc, might have supposed that
the observer had been holding in front of the vessel a piece of dirty glass, having the
dirt laid on in patches ; but in whatever manner the disc was moved in its own plane
without rotation, or turned round an axis perpendicular to its plane, not the slightest
perceptible change was produced in any part of the blue beam.

Explanation of Terms.

21. In all the experiments described in this paper in which a spectrum was formed
for the sake of examining the separate action of portions of light of different refran-
gibilities, the length of the spectrum was horizontal, so that the fixed lines were
vertical. Nevertheless it will be convenient, for the sake of shortness, to use the
prepositions above d^ndi below to ugmiy respectively on the more refrangible sideofand
on the less refrangible side of

The principal fixed lines of the visible spectrum will be denoted by letters in ac~

476 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

cordance with Fraunhofer's admirable map. These lines are now too well known
to need description.

The only map of the fixed lines of the chemical spectrum which I had for a good
while after these researches were commenced is Professor Draper's^ which will be
found in the twenty-second volume of the Philosophical Magazine (1843). Of course
this map cannot be compared for accuracy of detail with Fraunhofer's map of the
visible spectrum, nor does it profess to give more than some of the most conspicuous
lines selected from among a great multitude. The suppression of so many lines^
without any representation by shading of their general effect^ renders it difficult to
identify those which are laid down^ at least if I may judge from my own observations;
besides^ Professor Draper's spectrum was so much purer than the one with which I
found it most convenient to work^ that the two are not comparable with each other.

22. I have made a sketch of the fixed lines from H to the end, which accompanies
this paper. The fixed lines of the visible spectrum are so well known that I thought
it unnecessary to begin before H. A solution of sulphate of quinine is a very good
medium for showing the lines, but a yellow glass, which will be mentioned presently,
is quite as good, or rather better. The map represents the spectrum as seen with the
lens of 12 inches focal length in front of the prisms. The breadth of the slit was not
always quite the same : it may be estimated at about the ^th of an inch. The map
contains 32 fixed lines or bands more refrangible than H, which is the utmost that I
have been able on different occasions to see with this lens, though with a lens of
longer focus and a narrower slit the number of fixed lines which might be counted
was, as might be expected, a good deal larger. As I have not yet identified these
lines, except in certain cases, with those which had previously been represented by
means of photographic impressions, I have thought it advisable not to attempt an
identification, but to attach letters to the more conspicuous lines in my map without
reference to former maps. As the capitals L, M, N, O, P have already been appro-
priated to designate certain fixed lines, I have made use of the small letters /, m, /^,
o, j9, to prevent confusion.

In drawing the map, I have endeavoured to preserve the character of the lines with
respect to blackness or faintness, sharpness or diff'useness. The distances were not
laid down by measurement, except here and there, and they are not, I fear, quite so
accurate as might be desired ; still, I feel assured that no one viewing the actual
object would feel any difficulty in identifying the lines with those in my map, pro-
vided the circumstances under which his spectrum was formed at all approached to
those under which mine was seen when the arrangement as to focal length of the
lens, &c. was that most convenient for general purposes.

The more conspicuous lines in the part of the spectrum represented in the map
may conveniently be arranged in five groups, which I will call the groups H, /, m, n^p.
The group H consists chiefly of the well known paif of bands of which the first
contains Fraunhofer's line H ; the second band I have marked h^ in accordance with

PROFESSOR STOKES ON THE CHANGE OF REPRANGIBILITY OF LIGHT, 477

Professor Draper's map. The most conspicuous object in the next group consists of
a broad dark band^ /. This band is between once and twice as broad as H^ and is
darker in the less refrangible half than in the other. With a lens of 3 feet focal
length and a narrow slit it was resolved into lines, which is probably the reason why
it is altogether omitted in Professor Draper's map, while the first three lines of the
group (if I do not mistake as to the identification) are represented, forming his
group L. Under the circumstances to which the accompanying map corresponds,
the band I appears as a very striking object, perhaps, with the exception of the bands
H, ^, the most conspicuous in the whole spectrum. With a still lower power it ap-
pears as a very black and conspicuous line. A double line beyond I completes the
group /, after which comes another remarkable group m, consisting of five lines or
bands. Of these the first is rather shady, though sharply cut off on its more refran-
gible side, but the others, and especially I think the second and third, are particularly
dark and well-defined. I have marked the middle line m, not because it is more con-
spicuous than its neighbours, but on account of its central situation. After a very
faint group, consisting apparently of four lines, comes another very conspicuous
group % consisting of two pairs of dark bands followed by another pair of bands
which are broad and very dark. The first of these is a good deal broader than the
second, but is not so broad as the band H ; the second is followed by a fine line.
This is as far as it is easy to see ; but when the sunshine is clear, and the arrangements
are made with a little care, a group of six lines is seen much further on. Of these, the
first two are only moderately dark, and the first is rather diffuse ; they stand off a
little from the others, and are a little closer together than the other four. Of the latter,
the first, marked o, is very strong, considering the faintness of the light which it in-
terrupts ; the second and third are faint, and difficult to see ; the fourth, marked j», is
black like the first, and a good deal broader. The line p was situated, by measure^
ment, as far beyond H as H beyond b. Once or twice in the height of summer, and
under the most favourable circumstances, I have observed two broad dusky bands
still further on. The first of these had the appearance of being resolvable into two.
The excessively faint light seen beyond the second seemed to end rather abruptly at
the distance represented by the border of the accompanying plate, as if there were
there the edge of another dark band beyond which nothing could be seen. In order
to see the dusky bands last mentioned, and even to see the group p to most advantage,
it was necessary to allow the central part of the beam incident on the prisms to pass
through them close to their edges, so that evidently a great deal of light was lost by
passing by the prisms altogether. This circumstance, combined with others which
I have observed, convinces me that the great obstacle to seeing the fixed lines in this
part of the spectrum consists in the opacity of glass. Were glass as transparent with
respect to the invisible rays of very high refrangibility as it is with respect to the
rays belonging to the visible spectrum, I know not how much further I might have
been able to see.

478 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

I have endeavoured to identify the fixed lines in my map with the fixed lines repre-
sented in M. Silbermann's map of the chemical spectrum, with a copy of which my
friend Professor Thomson has kindly furnished me. I am still uncertain respecting
the identification. M. Silbermann's map is so very much more detailed than my
own, and must have been made with so much purer a spectrum, that the two systems
of lines are not directly comparable.

23. From the difficulty of identification some persons might be disposed to imagine
that the chemical rays, and those which produced the blue light in a solution of
quinine, were of a different nature, and had each a system of fixed lines of its own.
For my own part, I was too well acquainted with the Protean character of fixed lines
to regard the difficulty of identification as any valid argument in support of such a
view. And that this difficulty arose from nothing more than the different degrees of
purity of the spectra is now put past dispute, for my friend Mr. Kingsley of Sidney
Sussex College, to whom I recently showed some of the experiments mentioned in this
paper, has kindly taken for me some photographs of spectra having nearly the same
degree of extent and purity as those with which I worked, and these show the fixed
lines just as they appeared in a solution of sulphate of quinine and in other media^.

24. The position of a point in the spectrum which does not coincide with one of
the principal fixed lines, will be denoted by referring it to two of those lines, in a
manner which will be most easily explained by an example. Thus |^GH, G^H, GH|^
will be used to denote respectively a point situated at a distance below G equal to
half the interval from G to H, a point midway between G and H, and a point situated
at the same distance above H. In using this notation, the letters denoting fixed
lines will be written in the order of refrangibility, and the fraction expressing the
part of the interval between these lines, which must be conceived to be measured off
in order to reach the point whose position it is required to express, will be written
before, between, or after the letters, according as the measurement is to be taken from
the first line in the negative direction, from the first line in the positive direction, or
from the second line in the positive direction, the positive direction being that of
increasing refrangibility.

25. From the experiments already described, it appears that the beam of dispersed
light which was observed in the experiments of Sir David Brewster consisted of two
very distinct portions, one arising merely from light reflected from motes, and the
other having a far more remarkable origin. It will be convenient to have names for
these two kinds of dispersion, and I shall accordingly call them respectively /a/^e
internal dispersion and true internal dispersion^ or %im^\Y false dispersion mi6^ true
dispersion when the context sufficiently shows that internal dispersion is spoken of.
When dispersion is mentioned without qualification, it is to be understood of true
dispersion. Now that it appears that the mere reflexion of light from solid particles
held in mechanical suspension has nothing to do with that remarkable kind of internal

* See note A at the end.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 479

dispersion which is characterized by the '^ quaquaversus polarization/' the phenomenon
of false dispersion ceases to be of much interest in an optical point of view; while on
the other hand the phenomenon of true dispersion, which had always been very
remarkable, is now calculated to excite a great additional interest. It will be convenient
to mention here the principal characters by which true and false dispersion may be
distinguished, although it will be anticipating in some measure the results of obser-
vations yet to be described.

26. In true dispersion the dispersed light has a perfectly continuous appearance.
In false dispersion, on the other hand, it has generally more or less of a sparkling
appearance, and on close inspection is either wholly resolved into bright specks, or
so far resolved as to leave on the mind the impression that if the resolution be not
complete it is only for want of a sufficient magnifying power.

In true dispersion the dispersed light is perfectly unpolarized. In false dispersion,
on the contrary, at a proper inclination the light is almost perfectly polarized in the
plane of reflexion.

In false dispersion, which is merely a phenomenon of reflexion, the dispersed light
has of course the same refrangibility as the incident light. In true dispersion hetero-
geneous dispersed light arises from a homogeneous beam incident on the body by
which the dispersion is produced.

27. In those bodies, whether solid or liquid, which possess in a high degree the
power of internal dispersion, the colour thence arising may be seen by exposing the
body to ordinary daylight, looking at it in such a direction that the regularly reflected
light does not enter the eye, and excluding transmitted light by placing a piece of
black cloth or velvet behind, or by some similar contrivance. It has been usual to
speak of the colour so exhibited as displayed by reflexion. As however the cause now
appears to be so very different from ordinary reflexion, it seems objectionable to
continue to use that term without qualification, and I shall accordingly speak of the
phenomenon as dispersive reflexion^. Thus dispersive reflexion is nothing more than
internal dispersion considered as viewed in a particular way.

28. The tint exhibited by dispersive reflexion is modified in a peculiar manner by
the absorbing power of the medium. In the first place, the light which enters the
eye in a given direction is made up of portions which have been dispersed by particles
situated at different distances from the surface at which the light emerges. The word
particle is here used as synonymous, not with molecule^ but with differential element.
If we consider any particular particle, the light which it sends into the eye has had
to traverse the medium, first in reaching the particle, and then in proceeding towards
the eye. On account of the change of refrangibility which takes place in dispersion,
the eflFect of the absorption of the medium is different for the two portions of the
whole path within the medium, so that this effect may be regarded as a function of

* I confess I do not like this term. I am almost inclined to coin a word, and call the appearance /wore^ce^ice,
from iluor-spar, as the analogous term opalescence is derived from the name of a mineral.
MDCCCLII. 3 Q

480 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

two independent variables^ namely, the lengths of the path before and after dispersion;
whereas, had the light been merely reflected from coloured particles held in suspen-
sion, the effect of absorption would have been a function of only one independent
variable, namely, the length of the entire path within the medium.

29. When false dispersion abounds in a fluid, it may be detected at once by the
eye, without having recourse to any of the characters ah-eady mentioned whereby it
may be distinguished from true dispersion. When a fluid is free from false dispersion
it appears perfectly clear, when viewed by transmitted light, although it may be
highly coloured, and may even possess to such an extent the property of exhibiting
true internal dispersion as to display, when properly viewed, a copious dispersive
reflexion. On the contrary, when false dispersion abounds, the fluid, if not plainly
muddy, has at least a sort of opalescent appearance when viewed by transmitted light,
which, after a little experience, the eye in most cases readily recognises. In viewing
the phenomenon of dispersive reflexion, as exhibited in a fluid, it might be supposed
that the fluid was water, or else some clear though coloured liquid, holding in suspen-
sion a water colour in a state of extreme subdivision. But on holding the fluid
before the eye, so as to view it by transmitted light, or rather view a bright well-
defined object through it, the illusion is instantly dispelled. The reason of this
difference appears to admit of easy explanation, and will be noticed further on.

30. Light will be spoken of in this paper as active when it is considered in its
capacity of producing other light by internal dispersion, A medium will be said to
be sensitive when it is capable of exhibiting dispersed light under the influence of
light (visible or invisible) incident upon it. In the contrary case it will be called
insensible.

I shall now return to the description of the appearances exhibited by some of the
media most remarkable for their sensibility.

Decoction of the Bark of the Horse- Chestnut (iEsculus hippocastanum).

31. In Sir John Herschel's second paper it is stated that esculine possesses in
perfection the peculiar properties which had been found to belong to quinine. Having
tried without success to procure the former alkaloid, I was content to let this substance
pass, till I found how admirably a mere decoction or infusion of the bark of the tree
answered for all purposes of observation.

This medium is even more sensitive than a solution of sulphate of quinine, and
disperses like it a blue light. The description of the mode of dispersion in the latter
medium will apply in almost all points to the former : the principal difference consists
in the circumstance that in the horse-chestnut solution the dispersion begins earlier
in the spectrum than in the solution of quinine. In a solution of sulphate of quinine
of convenient strength, we have seen that the dispersion came on at about G^H, the
excessively faint dispersion which was exhibited earlier being left out of consider-
ation, whereas in a decoction of the bark of the horse-chestnut, diluted so as to be

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 481

of a convenient strength^ it came on a little before G. This explains the reason of
an observation of Sir David Brewster's^ who has remarked that '^a beam of light
that has passed through the esculine solution disperses blue lights but not copiously,
when transmitted through the quinine solution ; but the beam that has passed through
quinine is copiously dispersed when transmitted through esculine^^."

Green Fluor- Spar from Alston Moor.

32. It is well known that some specimens of fluor-spar exhibit a sort of double
colour. In particular, a variety found at Alston Moor, which is green when seen
by transmitted light, appears when viewed in a certain manner of a beautiful deep
blue. This blue colour seems to have been considered by Sir John Herschel as
merely superficial. It has been shown however by Sir David Brewster to arise
from light dispersed in the interior of the crystal, and to have no particular relation
to the surface.

The crystal with which the following observations were made was of a fine but not
intense green when viewed by transmitted light. On viewing a pure spectrum through
it, there was found to be a dark band of absorption in the red. This band was narrow,
and by no means intense. The crystal exhibited a copious deep blue by dispersive
reflexion.

33. On admitting into the crystal a cone of sunlight formed by a lens of short focus,
and then analysing the dispersed beam, it was found to consist of a very little red
followed by a dark interval, then green, faintly fringed below with less refrangible
colours down perhaps to the orange, then blue, or bluish-green, followed by a great
deal of indigo or violet. Independently of the gap in the red, the spectrum was not
quite continuous, for a band of bluish-green, not very broad, was separated by dusky
bands from the green below and the indigo above. The separate red band and the
two dusky bands were all so faint as to be difficult to see.

The dispersed beam was readily proved to be truly dispersed, for it was unpolarized,
and a pale brown glass cut it off* when placed in the first position, although it trans-
mitted it in a great measure when placed in the second.

34. When the crystal was examined by the third method, the general result closely
resembled that produced by sulphate of quinine. The dispersion commenced about
half-way between G and H, and continued from thence onwards far beyond H. It
was strongest about H. The fixed lines were seen with beautiful distinctness as dark
planes in the crystal. The groups H, /, m were quite evident, and n might be seen
without difficulty. I have even seen some of the fixed lines of the group p. The
tint of the dispersed light appeared as nearly as possible uniform throughout. The
distance to which this light could be traced from the surface, did not at all diminish
so rapidly in this crystal, with an increase in the refrangibility of the incident light,

* Philosophical Magazine, vol. xxxii. (June 1848), p. 406.
3 Q 2

482 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

as it had done in the case of a solution of sulphate of quinine. Indeed, it was difficult
to say how far the decrease in the depth to which the incident rays could be traced,
by means of the dispersed light which they produced, was due merely to the increasing
faintness of the light, and how far it indicated a real increase in the absorbing energy
of the crystal ; whereas in the case of sulphate of quinine the appearance presented
unequivocally indicated a very rapid increase of absorbing power.

35. On examining the crystal by the second method, the general appearance was
the same as in the case of sulphate of quinine, but the beam of falsely dispersed light
was absent. In addition to the copious beam of deep blue light dispersed by the
most refrangible rays, there was however a faint beam of red or reddish light dispersed
by rays of low refrangibility. This beam was too faint to be seen by the third method
of examination. It will be remembered that the prismatic analysis of the transmitted
light gave a band of absorption in the red. Another crystal of a pale colour, which
did not give a similar band of absorption in the red, exhibited nothing but the blue
beam of dispersed light when examined by the second method.

36. On examining the crystal by the fourth method, the extreme red proved inac-
tive. The activity commenced about the most refrangible limit of the red trans-
mitted by a deep blue glass, when the dispersed light was red, but extremely faint.
On moving the lens onwards through the spectrum, the dispersed light rapidly
became brighter, and then died away. When at its brightest, although even then it
was almost too faint for prismatic examination, it appeared to consist of not quite
homogeneous light a little lower in refrangibility than the active light. For a con-
siderable distance further on there was no sensible dispersion produced. The di-
spersed light became again perceptible when the active light belonged to the greenish
yellow, or not till the blue, according to the intensity of the incident light. As the
lens moved on the dispersed light remained faint for a considerable time. It was
first reddish and then brownish, with a refrangibility answering to its colour. When
the active light was at G^ H, or thereabouts, the dispersed light rapidly grew much
brighter, and became of a fine blue. On analysis it was found to consist of rays the
refrangibility of which ranged within wide limits. The red rays were, however,
almost wholly wanting, while the rays belonging to the more refrangible part of the
spectrum resulting from the analysis of the dispersed beam were particularly copious.
The most refrangible limit of the dispersed light did not quite reach in refrangibility
the active light. The dispersed light was most copious when the active light belonged
to the neighbourhood of H. As the lens moved on the dispersed light grew less
bright, and gradually died away.

Solution of Guaiacum in Alcohol.

37. This is one of the media mentioned by Sir David Brewster, who remarks that
it ^^ disperses, by the stratum chiefly near its surface, a beautiful violet light."

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILFJ Y OF LIGHT. 483

When this fluid is examined by the third or fourth method, it is found to exhibit a
copious internal dispersion, which begins to be conspicuous much lower down in the
spectrum than in the cases already described. In observing by the third method,
the true dispersion appeared to commence about the end of the green, the dispersed
light being reddish-brown. By the fourth method the dispersion could be traced as
low down as D| 6, the dispersed light being reddish. As the lens moved onwards,
in a direction from the red to the violet, the more refrangible colours entered in suc-
cession into the dispersed beam, and it became successively brownish, yellowish,
greenish, and bluish. In whatever part of the spectrum the lens might be, it was
found that the most refrangible part of the dispersed beam was of lower refrangibility
than the active light. This could be easily determined by means of the beam of
falsely dispersed light, which was always visible so long as the active light belonged
to the visible part of the spectrum.

38. With the third arrangement the fixed lines were seen as before by means of
the dispersed light, but in this fluid they could be seen much lower down in the
spectrum than in the solution of sulphate of quinine. The group H was seen on a
greenish ground. About the group I the ground was still greenish, but the dispersed
light was not very copious. The beautiful violet light mentioned by Sir David
Brewster is produced only by rays of extremely high refrangibility, and is found to
extend from the beginning of the group m to the end of the group n, and even further.
This part of the dispersion is best seen with a rather dilute solution.

39. In a solution of guaiacum, just as in the solution of sulphate of quinine, the
absorbing power of the medium increases very rapidly with the refrangibility of the
light. This is shown by the rapid decrease in the distance from the surface to which
the dispersed light can be traced. The reason why the violet dispersed light is con-
fined to a very thin stratum adjacent to the surface by which the light enters, is simply
that the medium is so nearly opake with regard to the invisible rays beyond the ex-
treme violet that all such rays are absorbed by the time the light has passed through
a very thin stratum of the fluid.

40. If the solution be strong the colour is of considerable depth. In all such cases
it is necessary to take the precaution, mentioned by Sir David Brewster, of trans-
mitting the incident beam as near as possible to the upper surface, so as just to graze
it. The absorption of the medium would otherwise modify the tint of the dispersed

beam.

41. The solutions of quinine and guaiacum present a striking contrast with respect
to the change of tint of the dispersed beam. In the former solution the change is
but slight, if we except that part of the dispersion which is very faint; whereas in
the latter, the prismatic colour which makes the nearest match to the composite tint
of the dispersed beam runs through nearly the entire spectrum, as the refrangibility
of the active light changes from that of the green rays to that of invisible rays situ-
ated far beyond the extreme violet.

484 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT,

Tincture of Turmeric.

42. This fluid is very sensitive, and exhibits a pretty copious dispersive reflexion
of a greenish light. In its mode of internal dispersion it strongly resembles a solution
of guaiacum, but the final tint of the dispersed light does not correspond to so high a
mean refrangibility. When the fluid was examined by the third method, the true
dispersion appeared to commence about b. The absorbing power was so great for
the rays of high refrangibility, that from a little above F (in the case of tincture not
diluted with alcohol) to the end the dispersed light seemed to be confined to the mere
surface. By the fourth method the dispersion was as usual traced a little lower down
in the spectrum. When the dispersed beam was first perceived it was nearly homo-
geneous, and its refrangibility was only a very little less than that of the active light.
As the refrangibility of the active light increased, new colours, in the order of their
refrangibility, entered into the dispersed beam, which became more and more com-
posite, while at the same time its upper limit became distinctly separated from the
beam of falsely dispersed light, which, when the whole dispersed beam was analysed
by a prism, was always found in advance of the other. The tint of the dispersed
beam passed from orange through yellow to yellowish green, which was its final tint.
Tincture of turmeric is well adapted for exhibiting the fixed lines in the invisible part
of the spectrum, though perhaps not quite so well as a solution of sulphate of quinine.

Alcoholic Extract from the Seeds of the Datura Stramonium.

43. This fluid, which I was led to try in consequence of Sir David Brewster's
paper, proved to be remarkably sensitive, and exhibited a copious dispersive reflexion
of a pale but lively green. The general phenomena are so nearly the same as in a
solution of sulphate of quinine that there is no need of a separate description. The
principal difference consists in the tint, which is green instead of blue. In the pre-
sent case, however, the fluid, in addition to its dispersion of green, dispersed a red
beam under the influence of certain red rays. As the lens employed in the fourth
method of examination was moved from the extreme red onwards, the light was at
first inactive, but when the lens reached a certain point of the spectrum, a red beam
of truly dispersed light suddenly appeared, which disappeared with almost equal sud-
denness as the lens moved on. In this mode of observation the refrangibility of the
dispersed could hardly be distinguished from that of the active light ; but on com-
bining the first and third methods, by removing the lens, placing the vessel truly in
focus, and holding a blue glass alternately in front of the vessel and in front of the
eye, I satisfied myself that the truly dispersed beam, taken as a whole, was of lower
refrangibility than the light by which it was produced. The utility of the blue glass
depended upon the circumstance that the upper extremity of the extreme red which
it transmitted nearly coincided with the point of the spectrum at which the red beam
occurred. This red beam was doubtless due to the presence of a small quantity of

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT. 485

chlorophyll, or one of its modifications. The light transmitted by the fluid exhibited
on prismatic analysis the absorption band in the red which is so characteristic of that
substance.

The colour of the solution was a pale brownish yellow; it would no doubt have
been still paler, and perhaps nearly colourless, had the sensitive principle to which
the green dispersion was due been present in equal quantity but in a state of purity.
As it was, the fluid was pale enough to exhibit well, when poured into a test tube and
held in front of a window, a narrow arc on the side of the incident light, like sulphate
of quinine, only in this case the arc was green instead of blue.

Frequency of the occurrence of true internal dispersion having the same general
character as that which takes place in the cases above described.

44. If we except the red dispersed beam produced by red rays in the crystal of
fluor-spar and in the stramonium extract, a strong similarity may be observed in the
mode of internal dispersion which takes place in the cases hitherto described. As
the refrangibility of the incident light continually increases, the rays are at first in-
active. At a certain point of the spectrum, varying according to circumstances, the
true dispersion begins to be sensible, but is faint at first. After remaining faint for
some distance it presently becomes more copious. It remains very conspicuous
through the whole of the violet and beyond, and then gradually dies away. It con-
sists at first of light of comparatively low refrangibility, and then new colours in the
order of their refrangibility enter into it. Frequently the greater part of the change
of prismatic composition takes place while the dispersed light is very faint, so that
practically speaking we may almost say that the tint is uniform. Sometimes, when
the dispersion just commences, the dispersed light is nearly homogeneous, and has a
refrangibility so nearly equal to that of the active light that the beams due to true
and false dispersion can hardly be separated.

45. Now this, so far as I have observed, is much the commonest kind of true in-
ternal dispersion, although sometimes the phenomenon presents very striking singu-
larities. In the paper in which Sir David Brewster first announced the discovery
of internal dispersion, he remarks " that it is a phenomenon which occurs almost
always in vegetable solutions, and almost never in chemical ones or in coloured
glasses=^." For my own part, I have rarely met with a vegetable solution which did
not exhibit more or less the phenomenon of frz^e internal dispersion. Its existence
may in general be easily detected in the following manner. The sun's light being
reflected horizontally through a lens, a deep blue glass is left in such a position as to
intercept the light incident on the vessel containing the fluid, which is placed at the
focus of the lens. A pale brown glass of the proper kind is then placed so as to
intercept, first the incident, and then the dispersed light. A vessel with flat sides filled
with a solution of sulphate of quinine would be better, and then the placing of the

* Edinburgh Transactions, vol. xii. p. 542.

486 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

medium in the second position might be dispensed with, the medium being sensibly
transparent. Sometimes it is useful to have recourse to analysis through a doubly
refracting prism, or a rhomb of calcareous spar. In this way true internal dispersion
may often be detected in a fluid which is actually muddy, in which case, were the
effect of the incident light observed as a whole, the true would be masked by the
enormous quantity of false dispersion which such a medium would offer.

46. The fluids obtained by treating the leaves and other parts of plants with
alcohol or hot water are almost always sensitive, so far as I have observed. The
solutions in water presently ferment, and are frequently highly sensitive in the early
stages of fermentation ; they are usually more or less sensitive in all stages. DiflFerent
kinds of fungus furnish very sensitive solutions. When aqueous solutions become
muddy by decomposition, other clear and often highly sensitive liquids may be
obtained from them by various chemical processes. Port and sherry are decidedly
sensitive. In such cases the fluid is a mixture of several substances, of which
some may be sensitive and others insensible. When vegetable substances are
isolated they are frequently insensible, or else so very slightly sensitive when examined
under great concentration of the highly refrangible rays, that it is quite impossible
to say whether the sensibility thus exhibited may not be due to some impurity: thus,
several solutions containing sugar, salicine, morphine, or strychnine were found to
be insensible. A solution of veratrine in alcohol proved to be sensitive in a pretty
high degree, dispersing internally a bluish light. Sir David Brewster has
remarked that a solution of sulphate of strychnine in alcohol dispersed light after it
had stood for some days. This observation I have verified with reference to true
dispersion, which the solution exhibits, though not very copiously, after it has been
made some time. There can be little doubt that the sensitive principle in this case
is not strychnine, but some product of its decomposition. I now come to some
instances of internal dispersion which are far more striking.

Solution of Leaf Green in Alcohol.

47. It was in this very remarkable fluid that the phenomenon of internal dispersion
was first discovered by Sir David Brewster, while engaged in researches relating to
absorption. The character of the internal dispersion of a solution of leaf-green is no
less remarkable than the character of its absorption. On account of the close con-
nexion which seems to exist between the two phenomena, it will be requisite first to
say a few words about the latter.

When green leaves are treated with alcohol, a fluid is obtained which is of a
beautiful emerald-green in moderate thicknesses, but red in great thicknesses, and
which has a very remarkable effect on the spectrum. A good number of the following
observations on the internal dispersion of leaf-green were made with a solution
btained from the leaves of the common nettle, by first boiling them in water and
then treating them with cold alcohol, the leaves having previously been partially

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 487

dried by pressing them between sheets of blotting paper. Nettle was chosen partly
because it stands boiling without losing its green colour, and partly for other reasons.
My object in boiling the leaves was to obtain the green colouring matter more nearly
in a state of isolation, but it seems to have the additional advantage of giving a
solution less liable to decomposition. Indeed, this fluid seemed disposed to remain
permanently unchanged when kept in the dark ; but a small portion of it which was
exposed to strong light had its colour rapidly discharged.

48. When fresh leaves are left in contact with alcohol in the dark, or in only weak
light, the colour of the fluid changes by degrees, and it seems to approximate
(making allowance for impurities) to a type which is nearly represented by the fluid
obtained in this manner from laurel leaves, or that obtained by treating with alcohol
tea leaves from which a good deal of brown colouring matter has first been extracted
by water. This type was rather ideal than actual, being derived from a comparison
of different cases, until it seemed to be realized in the case of a fluid obtained by re-
dissolving in alcohol a crust which had formed itself at the bottom of a test tube
containing leaf-green. The principle to which the peculiar absorption and internal
dispersion of such a fluid seems due may be called modified leaf-green. The fluid
itself is not green but olive-coloured, becoming red at great thicknesses.

49. When solutions of leaf-green, and of its various modifications, are examined in
different thicknesses by the light of a candle, there are five bands of absorption which
may be observed in the spectrum. These will be called, in the order of their refran-
gibility, Nos. 1, 2, 3, 4 and 5, the bright bands below the respective dark bands
being also numbered in the same manner. Of the dark bands, Nos. 1, 2, 3 and 5,
are the first four in Sir David Brewster's plate^. No. 4 is mentioned in the memoir,
but not represented in the plate, which corresponds to a thickness not sufficient to
bring out this band. The last band in the plate could not be seen without strong
light. The dark bands Nos. 1 and 2 are situated in the red, No. 3 about the yellow
or greenish yellow, No. 4 in the green, and No 5 early in the blue. Of these, No. 1
is in small thicknesses by far the most intense, and it may be readily seen even in a
very dilute solution ; it might apparently be used as a chemical test of chlorophyll, or
one of its modifications. The test would be of very easy application, since it would be
sufficient to hold a test tube with the liquid at arm's length before a candle at a little
distance, and view the linear image of the flame through a prism applied to the eye.

50. Fresh and modified leaf-green differ much in the order in which the bright
bands are absorbed, and in the degree to which the dark bands are developed before
they cease to be visible by the absorption of the part of the spectrum in which they
are situated. In the green fluid, the dark band No. 5 is not usually seen, because
the spectrum is there cut off, unless a very small thickness be used. With a mode-
rate thickness, Nos. 2 and 3, especially the former, are well seen, and No. 1 is very
intense. As the absorption goes on, the bright bands Nos. 2 and 3 are absorbed,

* Edinburgh Transactions, vol. xii.
MDCCCLII. 3 R

488 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT.

and there is left the red band No, 1, and a double green band, consisting of the
bright bands Nos. 4 and 5, separated by the dark band No. 4, which by this time
has come out. In modified leaf-green, the dark bands Nos. 4 and 5 are much more
conspicuous than in the green fluid, but No. 3 is wanting, or all but wanting. With
a thickness by which the absorption is well dev:eloped, the conspicuous bright bands
are in this case Nos. 1 and 3, and next to them No. 2, whereas in the green fluid
Nos. 2 and 3 were quickly absorbed, or at least the whole of No. 2, and the greater
part of No. 3.

51. It seems worthy of remark, that, especially in the case of the green fluid, the
absorbing power alters with the refrangibility of the light at a very different rate on
the two sides of the intense dark band No. 1. This might be inferred from the order
in which the bright bands disappear ; but it was rendered visible to the eye by the
following easy experiment. A narrow test tube was partly filled with a solution of
leaf-green, and then a few drops of alcohol were added, which remained at the top,
and there diluted the solution. The tube was then held before a candle, and the
linear image of the flame was viewed through a prism. In the under part the dark
band No. 1 was broad, the bright band No. 2 being narrow, and almost obliterated,
but in the upper part the dark band No. 1 was very narrow. Now on tracing
upwards the sides of this dark band, it was found that the less refrangible side was
almost straight, and the diminution in the breadth of the band was produced by the
encroachment of the bright band No. 2. Speaking approximately, we may say that
in proceeding from the extreme red onwards, at a certain point of the spectrum the
fluid passes abruptly from transparent to opake, and then gradually becomes almosD
transparent again.

52. It may here be remarked, that although the absorption produced by leaf-green
is best studied in a solution, its leading characters may be observed very well by
merely placing a green leaf behind a slit, as near as possible to the flame of a candle,
and then viewing the slit through a prism.

53. After this digression relating to the absorption of leaf-green, it is time to
come to its internal dispersion. And first, when a cone of white light coming from
the sun is admitted horizontally into the fluid, as close as possible to its upper
surface, and the beautiful red beam of dispersed light is analysed by a prism, the
spectrum is found to consist of a bright red band of a certain breadth, followed by
a dark interval, and then a much broader green band not near so brilliant. There
is usually but little false dispersion^ and what there is may be almost entirely got
rid of by analysing the beam by a Nicol's prism, so as to view it by light polarized
in a plane perpendicular to the plane of dispersion. Now on raising the vessel with-
out removing the prism from the eye, it was found that a dark band, which was in
fact the absorption band No. 1, appeared almost exactly in the middle of the bright
red band. On continuing to raise the vessel, so as to make the dispersed rays pass
through a still greater thickness of the medium before reaching the eye, the dark

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBlLlTY OF LIGHT^ 489

band increased in width, and when the red beam was almost absorbed, the part that
was left consisted of two cones of red, one at each side of the dark band, which by
this time had become broad. The whole appearance seemed to indicate that the
bright red beam of dispersed light had a very intimate connexion with the intense
absorption band No. 1.

54. Among coloured glasses, there is one combination which produces a very
striking effect. When a deep blue glass is placed in the first position, the dispersed
light, if the solution be at all strong, is confined to a very thin stratum adjacent to
the surface, and is best seen by placing the vessel so that the surface of the fluid at
which the light enters is situated at a little distance on either side of the focus of
the lens, when there is seen a bright circle of a most beautiful crimson colour. It
might be supposed that the red of which this circle mainly consists was nothing
but the extreme red transmitted by the blue glass. But it is readily shown that
such is not the case. For in the first place, the fluid transmits pretty freely the red
transmitted by the blue glass, whereas the red light found in this circle is almost
confined to the surface of the fluid. Again, it was found that a pale brown glass,
which transmitted freely the extreme red, almost entirely cut off the bright circle,
when placed in the first position without removing the blue glass, although it freely
transmitted it when placed in the second position. It appears, therefore, that the
bright circle is due, not to the red, but to the highly refrangible rays transmitted by
the blue glass.

55. When a solution of leaf-green was examined by the third method, the appear-
ance as seen from the outside was very singular. The fixed lines in all the more
refrangible part of the spectrum were seen as interruptions in a bright red ground
verging to crimson. The beauty and purity of the tint, and the strange contrast
which it presented to the colours belonging to that part of the spectrum, were very
striking. About H the tint began to verge towards brown, and the fixed lines
beyond H were seen on a brownish red ground. That the ground on which the
fixed lines of somewhat less refrangibility were seen was rather crimson than red,
arose, no doubt, from the mixture of a little blue or violet light due to false disper-
sion, and to the scattering which took place at the surface of the glass.

56. On looking down from above, the places of the more conspicuous bands of
absorption were indicated by dark teeth, with their points turned towards the inci-
dent light, interrupting the dispersed light. It is to be understood that the light
was transmitted as close as possible to the upper surface, so that the absorption by
which these teeth were formed took place before dispersion. In this way the places
of the absorption bands Nos. 1, 2 and 4, were perfectly evident. No. 3, it will be
remembered, was by no means conspicuous. When the solution is of convenient
strength, the absorption is so rapid beyond the bright band No. 5, that the disper-
sion is confined to a thin stratum close to the surface by which the light enters, and
therefore no dark tooth would be seen corresponding to the dark band No. 5.

3 r2

490 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBIUTY OF LIGHT.

57. On following the active light through the spectrum/ in the direction of
increasing refrangibility, the dispersion was found to commence with a bright, but
narrow tail of pure red light, which shot right across the vessel. The light by
which this tail was produced belonged to the more refrangible part of the extreme
red band which is transmitted by a moderate thickness of the fluid. The activity
of the incident light commenced almost abruptly: the same, it will be remembered,
was the case with the absorbing power of the medium. After the tail of red light
came the intense absorption band No. 1, where the dispersed light was confined to the
immediate neighbourhood of the surface by which the active light entered. At this
place a very bright band of dispersed light was visible on looking at the vessel from
the outside. In this part of the spectrum the active and the dispersed light were both
red ; but that dispersion was accompanied by a change of refrangibility, was shown
by the effect of absorbing media. Thus the long red tail and the bright band adja-
cent to the surface were differently affected by a blue glass, according as it was held
in the first or the second position ; and the bright band, though much enfeebled, was
still plainly visible through a considerable thickness of the fluid, although a stratum
having a thickness of only a very small fraction of an inch was sufficient to absorb
the rays by which the band was produced. Although the dispersion continued
throughout the whole of the visible spectrum and beyond, it was comparatively feeble
in the brightest part of the spectrum. It became pretty copious again in the neigh-
bourhood of the dark band No. 4, and remained copious throughout the blue and
violet. In the green, the dispersed light was red, slightly verging towards orange,
and in the blue and violet it was red verging a little towards brown.

58. It may seem superfluous, after what precedes, to bring forward any further
proof of the reality of a change of refrangibility. Nevertheless the following experi-
ment, which was in fact performed at an early stage of these researches, may not be
deemed wholly unworthy of notice, as not involving the use either of absorbing
media or of false dispersion.

A small narrow triangle of white paper was stuck on to the outside of the vessel
containing the leaf-green, in such a manner that its axis was vertical, and its vertex,
which was uppermost, was situated at the height of the middle of the spectrum. A
narrow vertical slit was then placed at the distance of the image of the first slit,
where the fixed lines were formed, and moved sideways till the light immediately
beside the fixed line G passed through it. The vessel was then placed a few inches
behind the slit, and moved sideways till the riband-shaped beam of homogeneous
light, which passed through the second slit, was incident on the vertex of the
triangle. On looking at the vessel from the front, as nearly as was convenient
in the direction of the incident light, there appeared a bright vertical bar cor-
responding to a section of the incident beam. This bar was of two colours,
namely, red in the upper half, where the light fell on the fluid, and indigo in the
under half, where it fell on the paper. On refracting the whole system sideways,

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OP LIGHT. 491

through a prism of moderate angle applied to the eye, the objects appeared in the
following order as regards refrangibility. First came the upper half of the bright
bar, which was only a very little widened by refraction^ so that it consisted of red
light which was approximately homogeneous. Next came the triangle, with its
vertex a little rounded, and its edges tinged with prismatic colours. The vertex,
which had formerly coincided with the bright bar, now lay a little to one side of its
upper half. The triangle was of course seen by means of the diffused light of the
room, which was not perfectly dark, and therefore its refrangibility must have cor-
responded to the brightest part of the spectrum, or nearly so. Lastly came the
under half of the bright bar, which was much more refracted than the triangle, so
as to be shifted almost completely off it. The paper triangle was far too close to
the first surface of the fluid to allow of attributing the dislocation of the bright bar
to any error depending upon parallax ; but to prevent all possible doubts on this
score, I took care to refract the system both right and left, and the result was the
same in the two cases. The conclusion is therefore inevitable, that the indigo light
which had changed its colour by dispersion from leaf-green had changed its refran-
gibility at the same time.

59. In viewing a solution of leaf-green in a pure spectrum, I noticed a phenomenon
which further indicates the close connexion which seems to exist between the
absorption and internal dispersion of this fluid. On holding the eye vertically over
the fluid, and looking down at the dispersed light through a red glass, I observed
five minima of illumination, having for the most part the shape of teeth with their
bases situated at the surface by which the light entered, and their points turned
inwards. These minima occupied positions intermediate between the bands of
absorption, so far at least as the positions of the latter were indicated by dark teeth
pointing in the contrary direction. The first minimum was situated a little beyond
the intense absorption band No. 1, and corresponded in position to the bright band
No. 2. The second was situated a little further on. The maximum intervening
between this and the third was but slight, so that the second and third together
formed pretty nearly one broad minimum. The third and fourth were situated one
at each side of the dark band No. 4, so as to correspond in position to the bright
bands Nos. 4 and 5. The fifth was situated a little way beyond the bright band
No. 5. This last minimum was not tooth-shaped, inasmuch as it occurred at a part
of the spectrum where the dispersed light was almost confined to the surface of the
fluid. These minima are best seen when the solution is rather weak. They may be
perceived without using a red glass, though not so easily as with its assistance.
With a stronger solution it was observed that the first minimum ran obliquely into
the dark tooth corresponding to the absorption band No. 1.

60. The reason of the occurrence of these minima appears to be simply this, that
the more copiously dispersed light is produced, the more rapidly the incident light
is used up in producing it, so that minima of activity correspond to points of the

492 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

spectrum at which the incident light penetrates to comparatively great distances
into the fluid before it is absorbed. The oblique position observed in the first mini-
mum is readily explained by considering that the illumination at any point of the
field of view depends conjointly upon the activity of the incident light, which is
a function of its refrangibility, and upon the fraction of the incident light left
unabsorbed, which last is a function both of the refrangibility and of the distance
from the first surface.

61. It seems worthy of remark^ that while the quantity of dispersed light is liable
to fluctuations having an evident relation to the bands of absorption which occur
throughout the spectrum, the quality of the light dispersed, as regards its refrangi-
bility, appears rather to have reference to the intense absorption band No. 1.

Extract from blue leaves of the Mercurialis perennis.

62. The juice of this plant has the property of turning blue by exposure to the air.
Some leaves and stalks which had turned blue were treated with alcohol, and a green
fluid was thus obtained much resembling in colour the ordinary solutions of leaf-
green, but I think of a rather bluer green than usual. In its mode of absorption, too^
it much resembled ordinary solutions of leaf-green, to which substance no doubt the
greater part of its colour was due. Its internal dispersion however was very peculiar,
for it dispersed a copious orange in place of a blood red like the extracts from fresh
green leaves in general, those of the Mercurialis perennis included. On analysis the
dispersed beam was found to consist chiefly of a red band, similar to that which
occurs in solutions of leaf-green, and of a yellow or orange and yellow band, a good
deal brighter than the former, from which it was separated by an intervening dark
band. When the fluid was examined by the second method, it was found that
the yellow dispersion was produced principally by the brightest part of the spectrum.
After a considerable time the fluid lost its fine green colour, as is very often the case
with solutions of leaf-green, and became yellowish brown, but the red and yellow
dispersions still continued.

When the fluid was examined by the fourth method, it was found that the red
rays dispersed a red, just as in a solution of leaf-green. The additional dispersion
which was so conspicuous in this fluid began almost abruptly about the fixed
line D. When it was first observed, the i-efrangibility of the orange dispersed light
could hardly, if at all, be separated from that of the active light. As the lens moved
on, the orange beam rapidly grew brighter, and yellow entered into it ; and now it
was easy to see that the beam of falsely dispersed light lay at its more refrangible
limit. The orange and yellow dispersed beam was brightest at about DfE; but
though it decreased in intensity it could be traced far beyond that point, in fact,
throughout the spectrum.

63. I have generally found that when a copious dispersion commences almost
abruptly at a certain point of the spectrum, it is followed by a band of absorption in

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 493

the transmitted light. This law did not seem applicable to the orange dispersion
exhibited by the solution just mentioned ; but then it is to be remembered that the
solution contained a quantity of chlorophyll, which produces absorption bands with
such energy that it would naturally mask the bands which might be due to an-
other colouring principle with which it was mixed. To try whether the law would
be obeyed if the chlorophyll were got rid of, I boiled in water some portions of the
root and young shoots which had turned blue, chlorophyll being insoluble in water.
The solution thus obtained was red, in small thicknesses pink, and dispersed copiously
a yellow or rather orange light. On subjecting the jfluid to prismatic analysis, a band
of absorption was seen at the place expected. Since aqueous solutions of this nature
are liable to decomposition, frequently decomposing before sunlight can be obtained
by which to examine them, the red solution was concentrated by evaporation and
purified by alcohol, in which the orange-dispersing principle is soluble, as had already
appeared from the properties of the alcoholic solution. The alcoholic solution thus
obtained remained unchanged, at least for a long time, and had the further advantage
over the aqueous solution of presenting the sensitive principle more nearly in a state
of isolation, though it was still contaminated by some principle which dispersed a
whitish light under the influence of rays of high refrangibility.

64. The blue colouring matter may be readily extracted by cold water, but is de-
composed by boiling. The blue solution dispersed an orange light like the other, but
the dispersed light could not be nearly so well seen, just as would be the case were
the red orange-dispersing fluid mixed with an insensible blue fluid of a much deeper
colour, so that the mixture of the two would be blue. And in fact when the blue
fluid was changed to red by boiling the colour became far less intense.

Archil and Litmus.

65. It is stated by Sir David Brewster that a very remarkable example of internal
dispersion, which had been pointed out to him by Mr. Sohunk, is exhibited in an
alkaline or in an alcoholic solution of a resinous powder produced from orcine by
contact with the oxygen of the air. Not being able readily to procure a specimen of
orcine, I tried archil, and obtained from it and litmus some very remarkable solutions.

In the fluid state in which archil is sold, the colour is much too deep for convenient
optical examination. When a small quantity of archil is diluted with a great deal of
water, the diluted fluid is very sensitive. It is red by transmission, or in small thick-
nesses purple, but exhibits by dispersive reflexion a pretty copious but rather sombre
green.

66. When the fluid was examined by different methods, it was found to disperse a
little red, some orange, and a great deal of green. The red dispersion was so slight,
that in observing by the third method it appeared doubtful whether there was any
except false dispersion. It commenced in the red, when the active and dispersed
lights had the same refrangibility, or nearly so. The orange dispersion commenced

494 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

about the fixed line D^ the dispersed light being at first nearly homogeneous^ and of
the same refrangibility as the active light. On proceeding onwards in the spectrum^
in observing by the fourth method, the orange beam became brighter, and yellow
entered into it, but no colour beyond that, so that the orange and yellow beam was
left behind by the beam of falsely dispersed light, from which it was separated by a
perfectly dark interval. The green dispersion began about 6, or a little beyond,
coming on almost abruptly. The manner of its commencement was best observed
by the fourth method, by holding a prism to the eye while the lens was moved
through the spectrum. In this way it was found that on arriving at the point of the
spectrum above mentioned, a gleam of green light shot across the dark space which
before separated the beam of falsely dispersed light from the orange beam of truly
dispersed light. As the lens moved on, the green dispersed light grew brighter, but
its more refrangible limit did not seem to pass, or at least much to pass, the refran-
gibility it had at first ; so that the green beam of truly dispersed light was almost
immediately left behind by the beam of falsely dispersed light. The former, on being
left behind, soon died away.

67. We might suppose either that the red, orange and green dispersions are due to
the same sensitive principle, or that they are produced by three distinct sensitive prin-
ciples mixed together in the solution. The latter would appear the more probable
supposition, to judge by the apparent want of connexion between the three disper-
sions. This view is strongly confirmed by the following results. Some ether was
poured on archil in the fluid state, and after being gently moved about and allowed
to stand, a little was withdrawn without agitation. A purplish rose-coloured fluid
was thus obtained, which was highly sensitive, exhibiting the orange and green di-
spersions but not the red. The orange dispersion was far more copious, in proportion
to the whole quantity of dispersed light, than had been the case with archil diluted
with water.

Some archil was violently agitated with ether, and after subsidence the ether was
withdrawn. This ethereal solution was much deeper in colour than the former, and
exhibited the red dispersion in addition to the orange and green. On adding a small
quantity of water, and agitating, a separation, or at least partial separation, of the
sensitive principles took place ; for the upper fluid exhibited the orange dispersion
abundantly, but none of the red, and little or none of the green, while the under fluid
exhibited the green and red dispersions with little, if any, of the orange. The upper
fluid exhibited a pretty copious dispersive reflexion of reddish orange, and the under
fluid a remarkably copious reflexion of a fine green. A similar separation, more or
less perfect, took place in other cases, the dispersion of orange bearing to that of
green a greater ratio in the ether than in the water. Some of the green-dispersing
fluids thus obtained were most remarkable on account of the extraordinary copious-
ness of the reflected green, and the strange contrast which it presented to the trans-
mitted tint, which was a purplish red.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILTTF OF LIGHT. 495

The red dispersion in the second ethereal solution, though decided^ was by no
means copious. In the case of archil merely diluted with water, it had been so
slight that its existence might have been considered doubtful. It might be supposed
that the first solution was not sufficiently concentrated to exhibit the red dispersion,
in which case the red and green dispersions might have been due to the same sensitive
principle. But an ethereal extract from dried archil, which was plainly concentrated
enough, did not exhibit the red dispersion, although it did exhibit the orange and
green dispersions. None of the sensitive principles appear to constitute the chief
part of the colouring matter of this dye-stuff.

68. When some of these ethereal solutions were examined by the third method,
with a lens of shorter focus than usual, the appearance was very singular. At the
less refrangible end of the spectrum the incident light was quite inactive ; and then,
on reaching a certain point, a copious dispersion of orange commenced abruptly.
This continued with no particular change for some distance further on, when it
passed abruptly into green. The fourth method showed however that the former
dispersion continued, and was only masked, in the third method of observation, by a
new and more powerful dispersion of green which then commenced. And in fact
when the green-dispersing principle was separated, or partially separated, by water,
the orange dispersion was seen to continue where before it appeared to have been
exchanged for green.

69. I ought here to mention that a similar separation did not take place on the
addition of water only to an ethereal extract from archil previously dried. The con-
dition which determined the separation in the first case appeared to be the presence
of a small quantity of ammonia, which would evaporate on drying the archil. And
in fact when a small quantity of ammonia was added to the extract from dried archil,
a partial separation was effected. I do not here enter into the question whether one
of the sensitive principles may be obtained from the other, whether, for example, a
chemical combination of the orange-dispersing principle with ammonia might disperse
a green, or a green with a little orange. A solution containing a mixture of the same
substance in two different states of chemical combination, both compounds being
sensitive, is not the less justly regarded as containing two distinct sensitive prin-
ciples.

70. The preceding results are mentioned, not for their own sake, but merely for
the sake of the method of examination employed. The results indeed are so imper-
fect as to be worthless on tlieir own account. A complete optico-chemical examina-
tion of archil and litmus would itself alone furnish a subject for research of no small
extent; but it belongs rather to chemistry than to general physics. It is quite
possible that internal dispersion may turn out of importance as a chemical test. The
dispersing such a tint, and the having the dispersed light produced by light of such
a refrangibility, form together a double character of so peculiar a nature that it
enables us, so to speak, /o see a sensitive principle in a solution containing many sub-

MDCCCLII. 3 s

496 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

Stances, some of them, perhaps, coloured, so that the colour of the solution may be
very different from what it would be if the sensitive principle were present alone.

71. The law mentioned at the beginning of art. 63 did not seem very applicable
to archil when the fluid was merely diluted with water. But when the orange-di-
persing and green-dispersing principles were obtained, as it would appear, more

nearly in a state of isolation, by means of ether and water, the law was found to be
obeyed. Thus, when the ethereal solution which exhibited the orange dispersion
and little else was examined by the third method, the disperson was found to com-
mence with a tail of light followed by a dark tooth, indicating the position of a band
of absorption. When the light transmitted by a certain thickness of this fluid was
subjected to prismatic examination, it was found to consist of red followed by some
orange, when the spectrum was cut off with unusual abruptness. After a broad dark
interval came the most refrangible colours faintly appearing. Those solutions which
exhibited a copious dispersion of green gave, in addition to a band obliterating the
yellow, a very distinct band separating the green from the blue. A similar band,
but by no means distinct, might be seen in archil merely diluted ; and it is particularly
to be observed that this band, which occurred a little above the point of the spectrum
where the green dispersion commenced, became more conspicuous when the green-
dispersing principle was present more nearly in a state of isolation.

72. Two portions of litmus were treated, one with ether and the other with alcohol,
which were allowed to remain in contact with the solid. Both extracts, but espe-
cially the latter, were highly sensitive, exhibiting dispersions of orange and green
similar to archil, and due apparently to the same sensitive principles. The ethereal
extract dispersed' chiefly orange, while the alcoholic extract dispersed orange and
green in nearly equal quantities. The latter extract exhibited a remarkably copious
dispersive reflexion of a colour nearly that of mud, and was altogether one of the
strangest looking fluids that I have met with. On viewing it in such a manner that no
transmitted light entered the eye, one might almost have supposed that it was muddy
water taken from a pool on a road. But when the bottle containing it was held be-
tween the eye and a window the fluid was found to be perfectly clear, and of a beautiful
purple colour.

Canary Glass.

73. Among media which possess the property of internal dispersion in a high
degree. Sir David Brewster mentions a yellow Bohemian glass, which dispersed a
brilliant green light. This led me to seek for such a glass, and it proved to be pretty
common in ornamental bottles and other articles. The colour of the glass by trans-
mitted light is a pale yellow. Its ornamental character depends in a great measure
upon the internal dispersion, which occasions a beautiful and unusual appearance in
the articles made of it. The commercial name of the glass is canary glass. The
following observations were made with a small bottle of English manufacture.

74. When the sun's light was admitted without decomposition the dispersed beam

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 497

was yellowish green. The dispersion was so copious that when a large lens was used
the dispersed beam approached to dazzling. The prismatic composition of this beam
was extremely remarkable. The beam was found on analysis to consist of five bright
bands, which were equal in breadth and Equidistant, or at least very nearly so, and
were separated by narrow dark bands. The first bright band was red, the second
reddish orange, the third yellowish green, the fourth and fifth green. I have very
frequently observed dark bands, or at least minima, in the spectrum resulting from
the prismatic analysis of dispersed beams, but I have not met with any example so
remarkable as this, except in a class of compounds which the properties of canary
glass led me to examine.

75. On analysing a beam of sun-light transmitted through a certain thickness of
the glass, there was found to be a dusky absorption band a little below F, another
less distinct at F| G, and the spectrum was cut off a little below G.

76. When the glass was examined by the third method, the dispersion was found
to commence abruptly about the fixed line h. It remained remarkably copious
throughout the whole of the visible spectrum and far beyond, with the exception of
a band beginning a little above F, and having its centre at about F|^G, where there
was a remarkable minimum of activity. This band, it will be observed, was situated
between the bands of absorption already mentioned. The tint of the dispersed light
appeared to be uniform throughout, except perhaps where the dispersion was just
commencing. This was the best medium I have met with for showing the fixed lines
of extreme refrangibility, though some others were nearly as good.

77- On examining the glass by the fourth method, it was found that the dispersion
commenced nearly where the dispersed light ended, that is, the lowest refrangibility
of the rays capable of being dispersed was nearly the same as the highest refrangibility
of the rays constituting the dispersed beam exhibited by white light as a whole. The
dispersion appeared indeed to commence a little earlier, at about the refrangibility of
the fourth dark band in the spectrum of the entire dispersed beam. When the small
prism was held to the eye with one hand, while the small lens in the board was gra-
dually moved with the other, in a direction from the red to the violet, through the
part of the spectrum where the dispersion commenced, it was found that the region
of the first four bands was lighted up almost simultaneously, the whole field of view
having been previously dark. When the lens was moved a very little further on the
dispersed beam with its five bands was formed complete. Indeed the whole five ap-
peared almost simultaneously. Speaking approximately, and in fact with almost
perfect accuracy, we may say that if white light be conceived to be decomposed into
two portions, the first containing rays of all refrangibilities up to that of the fixed
line &, or thereabouts, and the second containing rays of all greater refrangibilities,
the dispersed light produced by white light as a whole belongs exclusively to the first
portion ; and yet, were the bottle illuminated by the first portion alone, no dispersion
whatsoever would be produced, whereas were it illuminated by the second portion

3 s 2

498 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

alone^ which contains not a ray having the same refrangibility as any one of the
dispersed rays, the dispersion would be exhibited in full perfection.

Common Colourless Glasses,

78. Sir David Brewster states that he has met with many specimens, both of
colourless plate and colourless flint glasses, which disperse a beautiful green light.
All the colourless glasses which I have examined dispersed light internally to a
greater or less extent, with the exception of some few specimens belonging to
Dr. Faraday's experiments. A beautiful green seems to be the commonest tint of
the dispersed beam, and this I have found in wine glasses, decanters, apothecaries'
bottles, pieces of unannealed glass, &c. ; also in many specimens of plate and
crown glass. The green was generally of a finer tint than that dispersed by the
canary glass, but was not near so copious. On analysis it was found to consist
usually of red and green separated by a dark band, or rather a minimum of bright-
ness. Those specimens which were examined by the third and fourth methods were
found to exhibit a little false dispersion, produced chiefly in the brightest part of the
spectrum, but the greater part was true dispersion. This dispersion was produced
chiefly by a rather narrow band, comprising the fixed line G, where there appeared
to be a remarkable maximum of sensibility. The line G lay a little above the lower
limit of the band. Below the band dispersion also took place, though not near so
copiously, and there appeared to be another maximum of sensibility some way further
down in the spectrum; but above the band dispersion almost entirely ceased of a
sudden; a very unusual circumstance when the active and the dispersed light are
well separated in refrangibility. The position of the band in the spectrum, and the
distribution of the illumination in it, which are very peculiar, were the same in all the
specimens which were sufficiently sensitive to admit of being examined by the third
method, but the tint of the dispersed light was not quite the same.

79. Orange-coloured glasses are frequently met with which reflect from one side,
or rather scatter in all directions, a copious light of a Wuish-green colour, quite dif-
ferent from the transmitted tint. In such cases the body of the glass is colourless,
and the colouring matter is contained in a very thin layer on one face of the plate.
The bluish green tint is seen when the colourless face is next the eye. As this phe-
nomenon was supposed by Sir John Herschel to offer some analogy with the re-
flected tints of fluor-spar and a solution of sulphate of quinine, I was the more de-
sirous of determining the nature of the dispersion. It proved on examination to be
nothing but false dispersion, so that the appearance might be conceived to be pro-
duced by an excessively fine bluish-green powder contained in a clear orange stratum,
or in the colourless part of the glass immediately contiguous to the coloured stratum.
The phenomenon has therefore no relation to the tints of fluor-spar or sulphate of
quinine. It is true that the very same glass which displayed a superficial reflexion
of bluish green, when examined by condensed sun-light exhibited also, in its colour-

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT. 499

less part, a little true dispersion, just as another colourless glass would do. But this

has plainly nothing to do with the peculiar reflexion which attracts notice in such a

glass.

Observations on the preceding results.

80. There is one law relating to internal dispersion which appears to be universal,
namely, that when the refrangibility of light is changed by dispersion it is always
lowered. I have examined a great many media besides those which have been men-
tioned, and I have not met with a single exception to this rule. Once or twice, in
observing by the fourth method, there appeared at first sight to be some dispersed
light produced when the small lens was placed beyond the extreme red. But on
further examination* I satisfied myself that this was due merely to the light scattered
at the surfaces of the large prisms and lens, which thus acted the part of a self-lumi-
nous body, emitting a light of sufficient intensity to affect a very sensitive medium.

81. Consider light of given refrangibility incident on a given medium. Let some
numerical quantity be taken for a measure of the refrangibility, suppose the refractive
index in some standard substance. Let the refrangibilities of the incident and
dispersed light be laid down along a straight line AX (fig. 2) taken for the axis of
abscissae ; let AM represent the refrangibility of the incident light, and draw a curve
of which the ordinates shall represent the intensities of the component parts of the
truly dispersed beam. According to the law above stated, no part of the curve is
ever found to the right of the point M ; but in other respects its form admits of great
latitude. Sometimes the curve progresses with tolerable uniformity, sometimes it
presents several maxima and minima, or even appears to consist of distinct portions.
Sometimes it is well separated from M, as in fig. 2 ; sometimes it approaches so near
to M that the most refrangible portion of the truly dispersed beam is confounded
with the beam due to false dispersion.

82. Let/(^) be the ordinate of the curve corresponding to the abscissae, a the
abscissa of the point M. Since/ (.t?) is equal to zero when w exceeds a, the curve
must reach the axis at the point M at latest, unless we suppose the function capable
of altering abruptly, as is represented in fig. 3. I do not think that such an abrupt
alteration, properly understood, is necessarily in contradiction with the law of con-
tinuity. For the sake of illustration, let us consider the phenomenon of total internal
reflexion. Let P be a point in air situated at the distance z from an infinite plane
separating air from glass. Conceive light having an intensity equal to unity, and
coming from an infinitely distant point, to be incident internally on this plane at an
angle y+d^ where y is the angle of total internal reflexion. The intensity at P is
commonly, and for most purposes correctly, considered as altering abruptly with ^,
having, so long as 6 is negative, a finite value which does not vanish with d, but
being equal to zero when d is positive. The mode in which the law of continuity is
in this case obeyed is worthy of notice. In the analytical expression for the vibra-
tion, when 6 passes from negative to positive, the coordinate z passes from under a

500 PROFESSOR STOKES ON THE CHANGE OF REFRANGlBILITY OF LIGHT.

circular function into an exponential with a negative index, containing in its deno-
minator X, the length of a wave of light. As ^ increases through zero, the expression
for the vibration alters continuously ; but if z be large compared with X it decreases
with extreme rapidity when d becomes positive. On account of the excessive small-
ness of X, it is sujQScient for most purposes to consider the intensity as a function of
6 which vanishes abruptly; and indeed it would be hardly correct to consider it
otherwise. For the use of the term intensity implies that we are considering light as
usual, whereas those phenomena which require us to take into account the disturb-
ance in the second medium which exists when the angle of incidence exceeds that
of total internal reflexion, lead us to consider the nature as well as the magnitude of
that disturbance, which no longer consists of a series of plane waves constituting
light as usual. It is in some similar sense that I mean to say that we may suppose
the function/* (^), which expresses the intensity of the truly dispersed light, to alter
abruptly, without thereby implying any violation in the law of continuity. In
observing by the fourth method, the portion of the spectrum operated on, though it
may be small, is necessarily finite, and in some cases no separation could be made
out between the beams of truly and falsely dispersed light. Hence I cannot under-
take to say from observation, whether the variation off {oo) be always continuous,
though sometimes very rapid, or be in some cases actually abrupt. I think, however,
that observation rather favours the former supposition, a supposition which, inde-
pendently of observation, seems by far the more likely.

^ 83. Although the law mentioned in Art. 80 is the only one which I have been
able to discover, relating to the connexion between the intensity and the refrangi-
bility of the component parts of the dispersed beam, which appears to be always
obeyed, and which admits of mathematical expression, there are some other circum-
stances usually attending the phenomenon which deserve notice.

When dispersion commences almost abruptly on arriving at a certain point of the
spectrum, the dispersed beam is very frequently almost homogeneous at first, and of
the same refrangibility as the active light. If the dispersed beam, when first per-
ceived, be decidedly heterogeneous, its refrangibility extends almost, if not quite, to
that of the active light, so that it is difficult, if not impossible, to separate the beams
of truly and falsely dispersed light. On the other hand, when dispersion comes on
gradually, it is generally found that the refrangibility of even the most refrangible
part of the dispersed beam does not come up to that of the active light.

Thus in the cases of the red dispersion exhibited by a solution of leaf-green, and
of the orange dispersions exhibited by solutions obtained from archil and from the
Mercurialis perennis^ the dispersed light was at first nearly homogeneous, and of the
same refrangibility as the active light. In the case of the green dispersions shown
by a solution obtained from archil, and by canary glass, the dispersed light was
heterogeneous from the first; but still, when it first commenced, a portion of it had
nearly the same refrangibility as the active light. In a solution of sulphate of quinine

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 501

the dispersion came on gradually, being perceptible when the active light belonged
to the middle of the spectrum ; and in this case the dispersed light consisted of
colours of low refrangibility. The bright part of the dispersion however came on
pretty rapidly, when the active light approached the extreme limit of the visible
spectrum, and accordingly the dispersed beam consisted in that case chiefly of light
of high refrangibility.

84. The mode of absorption of any medium may very conveniently be represented
by a curve, as has been done by Sir John Herschel. To represent geometrically
in a similar manner the mode of internal dispersion, would require a curved surface.
Let the refrangibility of light be measured as before, and suppose for simplicity's
sake the intensity of the incident light to be independent of the refrangibility, so
that dy may be taken to represent the quantity of incident light of which the refran-
gibility lies between y and y-^-dy. Considering the effect of this portion of the inci-
dent light by itself, let x be the refrangibility of any portion of the dispersed light,
and %dxdy the quantity of dispersed light of which the refrangibility lies between x
and x+dx. Then the curved surface, of which the coordinates are <r, y, z^ will
represent the nature of the internal dispersion of the medium. We must suppose
the intensity of the incident light referred to some standard independent of the eye,
since the illuminating power of the rays beyond the violet, and even of the extreme
violet, is utterly disproportionate to the effect which in these phenomena they
produce.

From the nature of the case, the ordinate z of the surface can never be negative.
The law mentioned in Art. 80 may be expressed by saying, that if we draw through
the axis of z a plane bisecting the angle between the axes of .r and i/, at all points
on the side of this plane towards x positive, the curved surface confounds itself with
the plane of xy.

85. Let us consider the form of this surface in two or three instances of internal
dispersion. For facility of explanation, suppose the plane of xy horizontal, let x be
measured to the right, y forwards, and z upwards. Let a line drawn in the plane of
xy through the origin, and bisecting the angle between the axes of a? andy, be called
for shortness the line L. In all cases the surface rises above the plane of xy only to
the left of the line L.

In the case of a solution of leaf-green, the surface consists as it were of two
mountain ranges running in a direction parallel to the axis of «/, or nearly so. The
first range, if prolonged, would meet the axis of <r at a point corresponding to the
place of the dark band No. 1 in the red, or nearly so. The second would meet it
somewhere in the place corresponding to the green. The green range is much
broader than the red, but very much lower, and is comparatively insignificant. The
ridge of the red range is by no means uniform, but presents a succession of maxima
and minima. The range commences at the end nearest to the axis of x with a very
high peak, by far the highest in the whole surface. In following the ridge forwards,

502 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

five minima or passes may be observed, with hills intervening. The ordinates y of
the* first four of these minima correspond to the refrangibilities of the bright bands
Nos. 2, 3, 4 and 5, The last minimum lies a little further on. Whether similar
minima exist in the green range is not decided by observation, on account of the
faintness of the green dispersed light.

In the case of canary glass, the surface consists of five portions like mountain
ranges running parallel to the axis of «/, and having abscissae belonging to the red,
reddish orange, yellowish green, green, and more refrangible green, respectively. These
ranges do not all start from the immediate neighbourhood of the line L, but on the
side towards the axis of x end almost in cliffs, at points at which the ordinate y is
nearly equal to the abscissa of the fifth range, perhaps a little less. Thus the first
three ranges are well separated from the line L. The ranges are intersected by a
sort of valley running parallel to the axis of a?, and having for its ordinate y the
refrangibility of F^G. With the exception of the minima which occur where the
ranges are intersected by this valley, the ridges run on very uniformly, and it is only
very gradually that the ranges die away.

The form of the surface which expresses the internal dispersion of a solution of
sulphate of quinine, may be gathered from the description of that medium. In this
case the surface resembles a rising country, not intersected by any remarkable
mountain ranges or valleys.

Fig. 4 is a rude representation of the internal dispersion in a solution of leaf-
green. The curves represented in the figure must be supposed to be turned through
90"^ about the lines on which they stand, and will then represent sections of the
surface already described, made by vertical planes parallel to the axis of x, OL is
the straight line bisecting the angle xOy. The figure is merely intended to assist the
reader in forming a clear conception of the general nature of the phenomena, and
must not be trusted for details. No attempt is made to represent the several maxima
and minima in the intensity of the red beam of dispersed light. In any such
figure, if we suppose homogeneous light to be incident on the medium, and wish to
lay down the place of the falsely dispersed beam, we have only to draw a straight
line parallel to the axis of x^ through the point in the axis oiy which corresponds
to the refrangibility of the incident light, and find where this line cuts the straight
line OL which bisects the angle wOy,

On the cause of the clearness of fluids^ notwithstanding a copious internal dispersion

which they may exhibit,

86. It has been already remarked, that though water holding a water colour in
suspension makes an admirable imitation of a highly sensitive fluid, when the latter
is viewed by dispersive reflexion alone, the two fluids have a totally different appear-
ance when viewed by transmitted light. The cause of this difference appears to be
plain enough. The light due to internal dispersion emanates from each portion of

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 503

the fluid which is under the influence of the active light, and emanates apparently
in all directions alike. I have not attempted to determine experimentally vrhether
the intensity is strictly the samp in all directions. The experiment would be very
difficult, especially for directions nearly coinciding with that of the active light,
because in that case the light which was really due to internal dispersion would be
mixed up with the glare which is always found in the neighbourhood of light of
dazzling brightness. However, I have seen nothing which led me to suppose that
the intensity was different in different directions. We may express the results of
observation extremely well, by saying that the fluid or solid medium is self-luminous
so long as it is under the influence of the active light.

Accordingly, when a bright object, such as the sky, or the flame of a candle, is
viewed through a highly sensitive fluid, the regularly transmitted light is accompa-
nied by some side light due to internal dispersion. The latter, however, emanating
in all directions alike from the influenced particles, is too faint, when contrasted
with the regularly transmitted light, to make any sensible impression on the eye.
But when a fluid, itself insensible, holds in suspension a great number of solid
particles of finite size, the light reflected from such particles is reinforced, in direc-
tions nearly coinciding with that of the incident light, by a great quantity of diff*racted
light, so that a bright object viewed through such a fluid is surrounded by a sort of
nebulous haze, giving the fluid a milky appearance.

Washed Papers,

87. In a paper '^On the Action of the Rays of the Solar Spectrum on Vegetable
Colours," Sir John Herschel mentions a peculiarity whicli he had observed in
paper washed with tincture of turmeric, which consists in its being illuminated,
when a pure spectrum is thrown on it, to a much greater distance at the violet end
than is the case with mere white paper ^. This phenomenon was attributed by Sir
John to a peculiarity in its reflecting power, and was considered as a proof of the
visibility of the ultra-violet rays. The colour of the prolongation of the spectrum
was yellowish green. Sir John appears to have been in doubt whether the greenish
yellow colour was to be attributed to the mixture of the true colour of the ultra-
violet rays with the yellow of the paper due to diffused light, or to the real colour of
the ultra-violet rays themselves, which on that supposition would have been incor-
rectly termed '^ lavender."

88. The fact of the change of refrangibility of light having been established, there
could be little doubt that the true cause of the extraordinary prolongation of the
spectrum on paper washed with tincture of turmeric, was very different from what
Sir John Herschel had supposed, and that it was due to a change of refrangibility
in the incident light, which was produced by the medium in a solid state. Tincture
of turmeric has already been mentioned as a medium which possesses in a high

* Philosophical Transactions for 1842, p. 194.
MDCCCLXI. 3 T

504 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT.

degree the property of internal dispersion. It was the observation of Sir John
Herschel's already mentioned, which led me to try this medium. But it is by no
means essential that a sensitive substance should be in solution, or in the state of a
transparent solid, in order that the change of refrangibility which it produces should
admit of being established by direct experiment, although of course the mode of
observation must be changed.

89. A piece of paper was prepared by pouring some tincture of turmeric on it, and
allowing it to dry. In this way the part which was deeply coloured by turmeric
was in juxtaposition with the part which remained white, which was convenient in
contrasting the effects of the two portions. The sun's light being reflected hori-
zontally into a darkened room through a vertical slit, the paper was placed in a
pure spectrum formed in the usual manner. On the coloured part the fixed lines were
seen with the utmost facility far beyond the line H, on a yellowish ground. The
colours too of all the more highly refrangible part of the spectrum were totally
changed. From the red end, as far as the line F, or thereabouts, there was no mate-
rial change of colour ; but a little further on a very perceptible reddish tinge came
on, which was quite decided at FJG, where it was mixed with the proper colour of
that part of the spectrum. About Gf H the colour became yellowish. The reality
of a change of refrangibility was easily proved by refracting the spectrum on the
screen by a prism applied to the eye. When the refraction took place in a plane
parallel to the fixed lines, they were seen distinctly throughout the spectrum ; but
when it took place in a plane perpendicular to the former, the fixed lines in the less
refrangible part of the spectrum, and as far as F, were distinctly seen ; but in the rest
of the spectrum they were more or less confused, or even wholly obliterated, accord-
ing to their original strength, the refracting angle and dispersive power of the prism^
and its distance from the paper. With a prism of small angle the edges of the broad
bands H were seen tinged with prismatic colours.

90. The change of refrangibility was further shown by the following observation.
The paper was placed in the pure spectrum in such a manner that the line of junc-
tion of the coloured and uncoloured parts ran lengthways through the spectrum, so
that the same fixed line was seen partly on the coloured and partly on the uncoloured
portion. On viewing the whole through a prism of moderate angle applied to the eye^
and so held as to refract the system in a direction perpendicular to the fixed lines,
the line F was seen uninterrupted, but G was dislocated, the portion formed on the
yellow part of the paper being a good deal less refracted than that formed on the
white. The latter was indeed faintly prolonged into the yellow part of the paper, so
that on this part G was seen double; but the image which was by far the more in-
tense of the two was less refracted than that formed on the white paper. The whole
appearance was such as to create a strong suspicion of some illusion, as if some other
gruop of fixed lines formed on the yellow part of the paper had been mistaken for G,
though certainly no reason appears why such a group should not have had its coun-

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 505

terpart on the white part. However, to remove all doubts^ I refracted the Ksystem in
the direction of the fixed lines, and then turned the prism round the axis of the eye
through 90^5 when the plane of refraction was situated as before. At first the two
portions of the line G were of course seen in the same straight line ; and the perfect
continuity with which, as the prism turned round, the appearance changed into
what had been first seen, left not the shadow of a doubt as to tlie reality of the dis-
location.

91. The cause of the whole appearance is plain enough. The light coming from
the illuminated part of the yellow paper consisted, in the neighbourhood of G, of two
portions ; the first, indigo light, which had been scattered in the ordinary way ; the
second and larger portion, heterogeneous light having a mean refrangibility a good
deal less than that of G, which had arisen from homogeneous light of higher refrangi-
bility. The absence of the first occasioned the faint prolongation of the more refracted
part of the line G ; the absence of the second gave rise to the less refracted part.

92. The broad bands H were seen faintly but quite distinctly on the white paper.
On refracting them sideways by a prism of moderate angle held to the eye, they
became confused, and tinged with prismatic colours. The confused images of these
bands, seen in the white and coloured parts, were nearly continuous. It thus appears
that the visibility of the bands H on the white paper was due to a change of refrangi-
bility which that substance had produced in violet light of extreme refrangibility.

93. Effects similar to those produced by paper coloured by tincture of turmeric
are also produced by turmeric powder, or even by the root merely broken across.
Notwithstanding the roughness of the latter, the bands H and fixed lines far beyond
are seen with the utmost facility.

94. These phenomena are much better observed by covering the slit with a deep
blue glass, which absorbs all the bright part of the spectrum, while it freely trans-
mits the violet and invisible rays, which are mainly efficient in this class of pheno-
mena. In this way fixed lines may be seen on common white paper far beyond H.
These lines may be seen without the use of the blue glass, by allowing the bright
colours to pass by the edge of the paper, and receiving on it only the extreme violet
and invisible rays.

95. Paper coloured by turmeric having exhibited so well the sensibility of that
substance, I was induced to try various other washed papers, in fact, papers washed
with most of the fluids with which I had made experiments. I found almost always
that sensitive solutions gave rise to sensitive papers, exhibiting a change of refrangi-
bility of the same character as that shown by the solution. Besides the turmeric
paper, the two most remarkable were paper washed with a pretty strong solution of
sulphate of quinine, and paper washed with the extract from the seeds of the Datura
stramonium. I should here observe, that it was not till long after the time when these
experiments were made that I was acquainted with the high sensibility of a decoc-
tion of the bark of the horse-chestnut. The former of the papers just mentioned ex-

3 T 2

506 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OP LIGHT.

hibited the fixed lines of the invisible rays on a blue^ and the latter on a green
ground. The dispersion produced by the quinine paper was not exhibited so early in the
spectrum as in the case of turmeric, nor was it so copious in the extreme violet rays,
and for some distance further on, but the quinine paper seemed superior to the other
for showing the fixed lines of extreme refrangibility. With the turmeric paper the
group n M^as plain enough, but with the quinine paper I have seen some fixed lines of
the group f. The stramonium paper was, on the whole, I think superior to the qui-
nine paper in point of the copiousness of the dispersed light, but seemed hardly equal
to it for showing the fixed lines of extreme refrangibility. However, it is likely that
paper washed with a solution of the sensitive principle in a state of purity would have
been quite equal to the quinine paper in this respect.

96. A washed paper is a little more convenient for use than a solution, but, as
might be expected, it does not show the fixed lines with quite as much delicacy, nor
is it quite so good for tracing the spectrum to the utmost limits to which it can be
traced with the substance employed.

97. The sensibility of fresh leaf-green could not be made out on a washed paper
by this mode of observation, but the sensibility of the substance extracted by alcohol
from black tea, from which the brown colouring matter had been removed by hot
water, was plainly exhibited by the redness which it produced in the highly refran-
gible part of the spectrum.

98. Paper washed with a solution of guaiacum seemed an exception to the general
rule ; but this is not to be wondered at, since a paper prepared in this manner is
turned green when exposed to the light, and it is diflicult to prevent some degree of
discoloration. That the fluid state is not essential to the exhibition of the sensibility
of this substance, was however plainly shown by the high degree of sensibility of the
solid resin from which the solution was made. In this case the bands H were seen
on a greenish ground. The dispersion of a fine blue light under the influence of rays
of still higher refrangibility was hardly, or not at all, exhibited by the solid resin.

99. Shell-lac, common resin, glue, are all highly sensitive. The ground on which
the fixed lines in the neighbourhood of H are seen is brown in the case of shell-lac,
and greenish in the case of resin and glue. The sensibility of glue is evidently not
due to gelatine, for isinglass is almost, if not quite, insensible. These are merely a
few instances of sensibility : I shall defer further mention of the subject till I have
described a better mode of observation. I will merely observe for the present, that
several washed papers proved not greatly inferior to turmeric paper for showing the
fixed lines about and beyond H.

Effect of refracting a Narrow Spectrum in a Vertical Plane,

100. In the arrangement last described, when a short slit is used, the spectrum
received on the washed paper or other substance is of course narrow, so that the
fixed lines formed on the paper are but short, and may roughly be regarded as mere

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 507

points. If, now, the whole be viewed through a prism, so as to be refracted in a verti-
cal plane, the effect is very striking. For facility of explanation suppose the red to
be to the left, and the rays to be refracted upwards, so that to the observer the image
is thrown downwards. The original spectrum on the screen is decomposed by the
prism held to the eye into two spectra, which diverge from each other. The first oi
these runs obliquely downwards from left to right, and contains the natural colours
of the spectrum from red to violet. It consists of light which has been scattered in
the ordinary way by the substance on which the primary spectrum is received, and
the cause of its obliquity is evident. The second spectrum is horizontal, that is to
say, it approximates to the form of a long rectangle having its longer sides horizontal.
Of course it would be theoretically possible to render the vertical sides the longer,
but when the whole arrangement of the apparatus is such as to be convenient for ob-
servation, the horizontal sides are much longer than the others. In this second spec-
trum the colours run horizontally^ that is to say, the lines of equal colour are hori-
zontal. The interruptions of the primary spectrum corresponding to fixed lines,
almost reduced to points, are now elongated, so that in this strangely formed spec-
trum the principal fixed lines of the solar spectrum are seen running across the
colours.

101. It will be convenient to have a name for the second of the two spectra above
mentioned. As the term secondary spectrum is already appropriated to something
altogether different, I shall call it the derived spectrum. The first of the diverging
spectra may be called the primitive spectrum^ while the original spectrum, considered
as not yet decomposed by the prism held to the eye, may be called, for distinction,
as in fact it has been already called, primary.

102. In accordance with the law enunciated in Art. 80, it is found that the derived
spectrum appears always on one and the same side of the primitive, being less refracted.

103. The brilliancy of the derived spectrum, its extent, both vertically and hori-
zontally, the colours of which it mainly consists, the distribution of its illumination
in a horizontal direction, all depend upon the nature of the substance upon which
the primary spectrum is received. As a general rule, it may be stated that it starts
from the neighbourhood of the brightest part of the primitive spectrum, and extends
from thence onwards to a good distance beyond the extreme violet ; and that with a
given substance its colour is pretty uniform, that is, does not much change in passing
from one vertical section to another. Sometimes the derived spectrum remains very
bright up to its junction with the primitive, or at least till it gets so near that the
superior brilliancy of the primitive spectrum prevents all observation on the derived ;
sometimes it remains dull to a considerable distance from the primitive spectrum,
and then, opposite a highly refrangible part of the primitive spectrum, a strong illu-
mination comes on in the derived, lasts for some distance, and afterwards gradually
dies away. Many of the results mentioned in this paragraph are better observed by a
sdmewhat different method, which will shortly be described.

508 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

104. It has been already stated that the bands H were distinctly seen on common
white paper, the substance usually employed as a screen in experiments on the
spectrum, but that this was due to a change of refrangibility produced in the extreme
violet rays. These same bands have been seen on paper in the experiments of
others, though of course their visibility was not attributed to its true cause. By
the method of observation described in Art. 100, or still better, by a method not yet
explained, it may he seen that the change of refrangibility produced by white paper
is by no means confined to the extreme violet rays, and those still more refrangible,
but extends from about the middle of the spectrum to a good distance beyond the
extreme violet. The distance to which the illumination can be traced by means of
light merely scattered in the ordinary way, may be seen by examining the primitive
spectrum. In the primitive spectrum formed on white paper and other white sub-
stances, I have not been able to trace the illumination beyond the edge of the broad
band H, which accords very well with the illuminating power of the extreme violet
when received directly into the eye.

Illuminating Power of the Rays of high Refrangibility.

105. The prolongation of the spectrum seen on turmeric paper was brought
forward by Sir John Herschel as a proof of the visibility of the ultra-violet rays, or
rather as a confirmation of other experiments which had led him to the same con-
clusion. Of course, the experiment with turmeric must now be regarded as having
no bearing on the question ; but from the way in which Sir John speaks of it, it
would appear that he thought the other experiments not so conclusive as to be inde-
pendent of the confirmation which they received from this. The experiment with
the distorted spectrum, indeed, must now be put out of account, because in this
experiment, as I have been informed by Sir John Herschel, the light was only
thrown on a screen. Accordingly, the question of the visibility of these rays may be
regarded as open to further investigation.

While engaged in some of the experiments described in Art. 89, I had occasion to
form a pure spectrum in air in a well-darkened room, the slit itself by which the
sun's rays entered being covered by a deep blue glass, so that no great quantity of
light entered even at this quarter. Now, if ever, it would appear that the ultra-
violet rays ought to be seen by receiving them directly into the eye ; for the blue glass
was so transparent with regard to these rays that the fixed lines far beyond H were
seen with facility, even on substances, such as white paper, which stand low in the
scale of sensibility ; and the length of the spectrum from B to H was about an inch
and a quarter, so that when the extreme violet rays entered the pupil, supposed to
be held near the pure spectrum, ngt only the extreme red rays transmitted by the
blue glass, but even the brighter part of the transmitted blue and violet rays fell
altogether outside it. However, on holding the eye a few inches in front of the pure
spectrum, so as to see the fixed lines distinctly, the bands H were indeed seen with

PROFISSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 509

great facility ; but I was not able to make out fixed lines beyond the end of the group
/, that is^ about the end of Fraunhofer's map. However, the eyes of different
individuals may differ much in their power of being affected by the highly refrangible
rays, [t must be confessed, that on looking in the direction of the prisms, a good
deal of blue light was seen, consisting of light which had been scattered at the
surfaces of the prisms and lens. This light, though far from dazzling, was sufficient
to prevent the eye from seeing excessively faint objects, even though they might be
well defined. For want of a heliostat, I did not attempt an experiment I was medi-
tating for securing a more perfect isolation of the ultra-violet rays^.

However, it seems to me to be a point of small importance, so far as regards its
bearing on other physical questions, whether the illuminating power of these rays is
absolutely null or only excessively feeble. It is quite certain, that if not absolutely null,
their illuminating power is at least utterly disproportionate to the effect which they pro-
duce in the phenomena to which the present paper relates, and indeed that is true
even of the violet rays. By illuminating power ^ I mean of course, power of producing
the sensation of light when received directly into the eye ; for by giving rise to light
of lower refrangibility, they are able to illuminate strongly an object on which they
tall

Mode of Observation specially applicable to Opake Bodies.

106. In some of the experiments already described, the change of refrangibility
was exhibited, which was produced by washed papers and solid bodies. There exists,
however, a mode of observation far preferable to those which have already been
explained as applicable to such cases, and which may even in some instances be
employed with advantage in the examination of transparent bodies. In the experi«
ment described in Art. 100, the primitive spectrum is pure, but the derived spectrum
impure, on account of the finite length of the slit. Were the slit reduced to a point,
it is true that the derived spectrum would become pure like the primitive, but then
the quantity of light would be so small that the primary spectrum would hardly
bear prismatic analysis. It is well, once for all, to examine a few sensitive opake
substances in a very pure spectrum, because then the exhibition of fixed lines running
across the colours in the derived spectrum removes even the shadow of a doubt as to
the reality of the change of refrangibility of the incident light. Besides this, the
only theoretical advantage in having the primitive spectrum very pure is, that it
might be expected to enable us to detect any very rapid fluctuations in the colour or
intensity of the dispersed light. Of course, I am now speaking only with reference
to experiments in which the observer is employing the spectrum to examine some
substance, not employing the substance to examine the spectrum. But practically,
I have not found any advantage on this account; for abrupt, or almost abrupt
changes in the colour or intensity of the dispersed light hardly ever, if ever, occur,

* See note B.

510 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT.

except when the active and the dispersed light have very nearly the same refrangi-
bility. But such changes could not be observed even with a pure primitive spectrum^
because in the place where they occur the primitive and derived spectra overlap ;
and independently of this, the brilliancy of the primitive spectrum would prevent all
exact observation of the derived. It is true, that in the case of chlorophyll, or some
of its modifications, changes of intensity having apparently somewhat the same
nature were observed when the active and the dispersed light were widely separated
in refrangibility. But the sensibility of this substance is difficult, if not impossible, to
observe in the case of a washed paper or a green leaf, except by one of the methods
not yet described, so that it is not to be expected that such fluctuations could be
made out. Besides, it is to be remembered that the fluctuations observed in the case
of solutions of chlorophyll, were fluctuations in the rate at which dispersed light was
produced, not fluctuations in the sum total of the dispersed light produced by the
time the active light was exhausted. Fluctuations of the former kind by no means
imply fluctuations of the latter; and indeed, the circumstance, that maxima of
activity in the solution correspond to minima of transparency, would seem to show
that the total quantity of light dispersed, considered as a function of the refrangibility
of the active light, is not subject to these fluctuations, or at least not to anything
like the same extent. Now the total quantity of red light dispersed by a green leaf^
or by a paper washed with a solution of chlorophyll, must depend upon the sensibility
of this substance and upon its transparency conjointly, and therefore it is likely
enough that such maxima and minima would not be observed, even were the
dispersed light much stronger than it is.

107. Suppose now the slit by which the light enters to be placed in a horizontal
instead of a vertical position, so as to lie in the plane of refraction. Corresponding
to light of any given refrangibility, the image of the slit formed after refraction
through the prisms and lens will now be a narrow parallelogram, which may be
regarded as a horizontal line. The series of these lines, succeeding one another in a
horizontal direction, and consequently overlapping, forms the spectrum incident on
the body examined. This spectrum is now no longer pure, but only approximately
so, a point, however, which, as we have seen, is not of much consequence. But by
this trifling sacrifice two very great advantages are gained. The first is increase of
illumination. When the slit is vertical, the spectrum received on the body occupies
a rectangle having for breadth the length of the image of the slit ; but when it is
horizontal, the same, or very nearly the same quantity of light is concentrated into a
rectangle having the same length as before (the length of the image of the slit being
disregarded compared with that of the spectrum), but having for its breadth only
the length of the image of a line drawn across the slit. Hence the intensity of the
incident light is increased in the ratio of the breadth to the length of the slit. The
second advantage is purity in the derived spectrum, a point of much consequence,
because sometimes the composition of this spectrum presents very remarkable

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 511

peculiarities. If the slit be not too long^ the spectrum formed in air is still sufficiently
pure to allow us to make out in a general way what are the refrangibilities of those
portions of the incident light which are most efficient in producing dispersed light ;
and this is nearly all that can be done even when the spectrum is very pure.

108. The method of observation which has just been described is that which latterly
I have almost exclusively employed in examining opake substances. As it will be
convenient to have a name for it^ I shall speak of examining a substance in a linear
spectrum. In examining substances which are only slightly sensitive, it is often
highly advantageous to cover the slit with a blue glass.

109. Fig. 5 is intended to represent the usual appearance of the primary linear
spectrum^ and of the primitive and derived spectra. XY is the primary spectrum, as
seen by the naked eye, RV, ST are the primitive and derived spectra into which it
is separated by the prism held to the eye. The direction of the shading in RV is
intended to represent the composition of this spectrum, which may be regarded as
consisting of an infinite number of images of the slit arranged obliquely in the order
of their refrangibility. The direction of the shading in ST is that of the lines of the
same colour and same refrangibility. Of course the figure does not represent the
amount of vertical displacement of the primary spectrum when viewed through the
prism held to the eye.

110. There is another mode of observation which I have occasionally found con-
venient when the object was to determine whether a substance exhibited so much as
a low degree of sensibility. In this method the sun's light was reflected horizontally
through a large lens, and then transmitted through a small lens placed in the con-
densed beam. The small lens was covered by a small vessel with parallel sides of glass,
containing a blue ammoniacal solution of copper, or else by a deep blue glass com-
bined with a weak solution of nitrate or sulphate of copper. The object of the latter
solution was to absorb the extreme red which is transmitted by a blue glass. The light
coming through the lens was then analysed by a prism, being received directly into the
eye, or else allowed to fall on a white object which had been previously ascertained not
to change the refrangibility of the light incident upon it. I found clean white earthen-
ware to serve very well for such an object, but each observer ought to test for him-
self the substance he employs. When a test object, such as white earthenware, is
used, it is placed at the focus of the lens, and the spot of blue light formed upon it is
analysed by a prism to see if the absorption is sufficient. When the visible rays are
considered to have been sufficiently absorbed, the object to be observed is placed at
the focus of the lens, and the spot of light formed upon it is viewed through a prism.
The spectrum then seen is compared with that given by the test object. This
method of observation is rather easier than that of a linear spectrum, and is at least as
delicate if the object be merely to determine whether a substance is sensitive or not,
but on the whole it is not near so useful. It may sometimes be used with advantage
in the case of translucent bodies.

MDCCCLTI. 3 V

512 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

111. An extremely pale solution of nitrate or sulphate of copper is sufficient to
absorb the extreme red transmitted by a deep blue glass. This is not the case with
the ammoniacal solution^ which does not absorb the extreme red till it is of a pretty deep
blue. Its absorbing power is greatest, not at the extreme red, but about the orange,
as may be seen by using candle-light, which is richer in red rays than daylight.

112. Another method of observation which is sometimes useful, consists in employ-
ing a large lens and absorbing medium, as described in Art. 110, but leaving out the
additional small lens. The substance to be examined is placed in the condensed
beam, and viewed through an absorbing medium which is approximately comple-
mentary to the former. This method is chiefly useful in examining a confused mass
of various substances. The most minute fragments of sensitive substances show
themselves in this manner.

Results obtained with a Linear Spectrum,

113. When this method is applied to the examination of common objects, it is
found that the property of producing a change of refrangibility in the incident light
is extremely common. Thus, wood of various kinds, cork, horn, bone, ivory, white
shells, leather, quills, white feathers, white bristles, the skin of the hand, the nails,
are all more or less sensitive. To make a list of sensitive substances would be end-
less work ; for it is very rare to meet with a white or light-coloured organic sub-
stance which is not more or less sensitive. I am not now speaking of organic sub-
stances obtained in a state of chemical isolation, of which some are sensitive and
others insensible. That substances of a dark colour should frequently prove insen-
sible is only what might have been expected, because the dispersed light is not reflected
from the surface, but emanates from all points of a stratum of finite thickness \ and in
order that dispersed light should be forthcoming, it is necessary that the active light
entering, and the dispersed light of a different refrangibility returning, should both
escape absorption on the part of the colouring matter. Such substances usually con-
sist of a mixture of various chemical ingredients, of which one or more may very
likely be sensitive, in which case the substance may be compared to a solution of sul-
phate of quinine mixed with ink. Frequently however the colouring matter is itself
sensitive.

114. Among sensitive substances I have mentioned the skin of the hand, which
stands rather low in the scale. 1 have found the back of the hand a convenient test
object. When the sunlight is not strong enough to show with ease the derived spec-
trum in the case of the hand, there is little use in attempting to observe.

115. It is needless to say that papers washed with tincture of turmeric, or with a
solution of sulphate of quinine, display their sensibility in a remarkable manner when
examined in a linear spectrum. The sensibility of turmeric paper is rather impaired
by exposing the paper to the light, but on the other hand is materially increased by
washing it with a solution of tartaric acid.

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT. 513

116. Paper washed with an ethereal solution from dried archil exhibited very well
the sensibility of that substance. The derived spectrum consisted chiefly of two
distinct portions, one containing orange and a little red, the other consisting chiefly
of green, just as in the beam of dispersed light, produced by white light taken as a
whole, which the solution itself exhibited. Indeed, I have found that the prismatic
composition of dispersed light could be determined even more conveniently by means
of a linear spectrum than by means of the beam dispersed by a solution.

117. The inside of the capsules of the Datura stramonium is nearly white, and
apparently uniform. But when the capsules are examined in a linear spectrum,
certain patches shine out like bright clouds in the invisible rays. The whole of the
inside is sensitive, as such substances almost always are, but these patches, which
are probably spots against which the seeds have pressed, are remarkably so. The
capsules were examined after they had begun to burst.

118. By means of a linear spectrum the sensibility of chlorophyll may be detected
in a green leaf. It is exhibited by the appearance in the derived spectrum of a
narrow pure red band of remarkably low refrangibility. The refrangibility is so low
that I have always found this band separated from the derived spectrum due to
other sensitive substances with which chlorophyll or one of its modifications might
have been mixed.

119. The petals of flowers, so far as I have examined, are as a class rather
remarkable for their insensibility, some appearing quite insensible, and others only
slightly sensitive. The bright yellow chaffy involucre of a species of everlasting,
proved, however, highly sensitive, and its sensibility was also displayed in an alcoholic
solution. This medium was sensitive enough to exhibit a pretty copious dispersive
reflexion of a pale greenish yellow light. Its sensibility was more confined than
usual to the rays of very high refrangibility.

120. Among petals, the most remarkable which I have observed are those of the
purple groundsel (ASewedo elegans). These petals disperse a red light, more copious
than is usual among petals. If a petal be placed behind a slit, and the transmitted
light be analysed, it is found to exhibit three remarkable bands of absorption, much
resembling those of blue glass, but closer together, and beginning later in the
spectrum, the first appearing about the place of the orange. These bands are still
better seen in a solution of the colouring matter in weak alcohol. On examining
this medium by the third method, with a lens of shorter focus than usual, and look-
ing down from above, the places of the absorption bands were indicated by tooth-
shaped interruptions in the beam of light reflected from motes. The points of these
teeth were occupied by red dispersed light, which did not appear in the intervening
beams of light reflected from motes, from whence it appears that there is the same
sort of connexion between the absorption and dispersion of this medium as was
noticed in Art. 59, in the case of solutions of chlorophyll and its modifications.

3 u 2

514 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

121. A collection of sea- weeds appeared all more or less sensitive, most of them
highly so. All, or almost all, except the white ones, exhibited in the derived spectrum
the peculiar red band indicative of chlorophyll and its modifications. The transmitted
light also exhibited more or less the absorption bands due to this substance, which
was likewise, in the specimens tried, extracted by alcohol. But the most remarkable
example of sensibility found in sea-weeds occurs in the case of the red colouring
matter contained in orangy red, red, pink, and purple sea-weeds. To judge by its
optical properties, this colouring matter appears to be the same in all cases, but to
be mixed in different proportions with chlorophyll, or some modification of it, and
probably other colouring matters, thus giving rise to the various tints seen in such
sea~weeds. The derived spectrum exhibited by sea-weeds of this kind consists
mainly of a band of unusual brightness, containing some red, followed by orange
and yellow. This band fades away gradually at its less refrangible limit, where it is
separated by a dark interval from the narrow well-defined red band of still lower
refrangibility due to chlorophyll. At its more refrangible limit, however, it breaks
off with unusual abruptness.

122. When the light transmitted through such a sea-weed is subjected to prismatic
analysis, in addition to one at least of the absorption bands due to chlorophyll, there
is seen a band obliterating the yellow, another dividing the green from the blue, and
a third, far less conspicuous, dividing the green into two. The whole of the green
is absorbed more rapidly than the blue beyond, and not merely than the red, which
last is the final tint.

123. The red colouring matter is easily extracted by cold water from certain
kinds of red sea- weed, if fresh gathered ; but when once the plant has been dried, the
colouring matter cannot be extracted in any way that I know of. It is apparently
insoluble in alcohol and ether, and is decomposed by boiling. Cold water extracts
only a trace of it after a long time.

124. A piece of recently gathered red sea-weed, on being mashed with cold water,
readily gave out its red colouring matter. When the residue was treated with
alcohol, the fluid was almost immediately coloured green by chlorophyll, whereas
this substance is only very slowly and sparingly extracted by alcohol from dried sea-
weeds. A dried sea-weed may apparently be assimilated to an intimate mixture of
gum and resin, which it would be very difficult to dissolve, whether it were attacked
by water or alcohol.

125. The solution of the red colouring matter was highly sensitive, exhibiting a
copious dispersive reflexion of a yellowish orange light. The transmitted light was
pink or red, according to the thickness through which the light passed. When this
light was analysed, the same three absorption bands which have been already men-
tioned were perceived. The analysis of the light transmitted by the fronds of various
red sea-weeds had rendered it extremely probable that the faint division in the

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT. 515

green did belong to the red colouring matter; but till I had obtained this matter in
solution I did not feel certain that it might not have been due to chlorophyll^ the
spectrum of which exhibits a division in the green.

126. When this fluid vi^as examined in Sir David Brewster's manner, and the
dispersed beam was analysed^ the spectrum was found to consist of a broad band
like that which has been already described as seen in the derived spectrum given by
a frond of red sea-weed. When the solution, which happened to be very weak, was
examined by the third method, the dispersion was found to be produced chiefly by a
portion of the incident spectrum, having a breadth about equal to that of the interval
between the two principal bands of absorption. To each of these bands corresponded
a maximum of activity. The tint of the dispersed light was nearly uniform ; but by
the fourth method of observation some faint dispersed red could be made out, which
appeared before the main part of the dispersion had come on. This medium affords
a very good example of an intimate connexion between absorption and internal
dispersion.

127. The colouring matters of birds' feathers appeared to be insensible, white
feathers being most sensitive, pale ones next, and dark ones not at all : however, I
have not examined a large collection.

128. Of coloured fruits, such as currants, &c., the colouring matter appeared, in
the very few cases which I have examined, to be quite insensible.

129. A set of water colours were by no means remarkable for sensibility, but rather
the contrary. The inorganic colours appeared quite insensible, except white lead,
the sensibility of which was perhaps due to size, and offered nothing striking, either
as to its character or as to its amount. Some lakes and other organic colours proved
moderately sensitive. But I found one water colour, called Indian yellow, which
stands pretty high among sensitive substances. In its mode of dispersion it much
resembles turmeric, but it does not come up to that substance in the amount of
sensibility. It is said to be composed of urate of lime, but I do not know how far it
may be regarded as chemically pure.

130. Many of the substances used in dyeing, and dyed articles in common use,
furnish very remarkable examples of sensibility. Archil, litmus and turmeric have
been already mentioned ; and I have been recently informed by a friend that the
Mercurialis perennu, in which a striking instance of sensibility was observed, was
formerly employed in dyeing. A piece of scarlet cloth, examined in a linear spec-
trum, gave a copious derived spectrum which was very narrow, consisting chiefly of
the more refrangible red. With a vertical slit the bands H and fixed lines beyond
were seen on a red ground. Paper washed with a solution of cochineal and after-
wards with a solution of alum, when examined in a linear spectrum, displayed a pretty
high degree of sensibility, the derived spectrum consisting in this case of a red band.
If tartaric acid be used instead of alum, the dispersion is a good deal more copious.

Common red tape is another example in which the derived spectrum is very copious^

516 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT,

consisting mainly of a red band. Some red wool^ dyed I suppose with madder^ proved
extremely sensitive. The derived spectrum in this case was pretty broad, but red was
the predominant colour. Green wool, dyed I do not know with what, was also very
sensitive, giving a pretty broad derived spectrum, in which green was the predomi-
nant colour. These examples may suffice, but the reader must not suppose that they
form the only instances in which dispersion was observed among dyed substances.
On the contrary, it is extremely common in this class.

131. Brazilwood, safflower, red sandal wood, fustic and madder, all gave rise to
solutions having a pretty high degree of sensibility. The solutions here referred to
were such as were obtained directly by water, &c., in which the colours which these
substances are capable of producing were not brought out. The beautiful red colour-
ing matters of logwood and camwood appear to be insensible ; for a fresh-made solu-
tion of logwood in water exhibited no perceptible sensibility, and the slight sensibility
exhibited by a similar solution of camwood seemed to have no relation to the red
colouring matter.

132. Paper washed with a solution of madder in alcohol was sensitive in a pretty
high degree, but the sensibility was greatly increased by afterwards washing with a
solution of alum. Accordingly I found that a decoction of madder in a solution of
alum exhibited a very high degree of sensibility, displaying a copious dispersive
reflexion of a yellow light. In this medium the dispersion commenced about the
fixed line D, and continued from thence onwards far beyond the extreme violet, so
that the group of fixed lines n was seen with great ease.

133. Safflower red, examined in the shape in which it is sold on what is called a
pink saucer^ proved highly sensitive, giving a bright and narrow derived spectrum,
which consisted chiefly of the more refrangible red. This substance possesses some
other remarkable optical properties, which however do not belong to the immediate
subject of this paper.

134. Metals proved totally insensible. I have examined gold, platinum, silver^
mercury, copper, iron, lead, zinc and tin. Brass is like simple metals in this respect ;
but if the surface be lackered the lacker displays its own sensibility.

135. The non-metallic elements, carbon, sulphur, iodine and bromine, are in-
sensible.

136. Among common stones I have found dark flint, limestone, chalk and some
others which were sensitive, though only in a low degree compared with organic
substances. To guard against any impurity of the surface, the stones were broken
across, and the fresh surface examined. In the cases mentioned, the sensibility
observed is not to be attributed to the chief ingredient of the stone, for quartz, chal-
cedony, Iceland spar and Carrara marble were insensible.

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OP LIGHT. 517

Compounds of Uranium.

137. Towards the end of last autumn, when the lateness of the season afforded
but few opportunities for observation, I learned from different sources that the kind
of yellow glass which has been already mentioned as possessing in so high a degree
the property of internal dispersion was coloured with oxide of uranium. This ren-
dered it interesting to examine other compounds of uranium ; and I accordingly
procured some crystallized nitrate of the peroxide, which, with a few other com-
pounds formed from it, and some of the natural minerals which contain uranium,
were examined by methods which have been already explained.

138. The crystals of the nitrate were not sufficiently large and perfect to admit of
observation by the methods applicable to fluids and clear solids, but they could be
readily observed by means of a linear spectrum. They proved to be sensitive in a
very high degree, dispersing a green light which had the same very remarkable com-
position that has been already described in the case of the yellow glass. On placing
a crystal in the continuation of the same linear spectrum with the glass, and viewing
the whole through a prism, the five bright bands of which the derived spectrum given
by each of the two media usually consisted, appeared to correspond to one another as
regards their position in the spectrum. With great concentration of light I have
seen an additional band of greater refrangibility in the spectrum of the crystals.

139. Some crystals of nitrate of uranium were gently heated so as to expel a good
part at least of the water of crystallization. The residue after some time became
opake and nearly white. In this state it was still more sensitive than the crystals.
The dispersed light was not exactly of the same tint, but more nearly white ; and the
derived spectrum was found on being analysed to contain, in addition to the bright
bands usually seen in the derived spectrum of the crystals, another blue band still
more refrangible. The fused mass gradually attracted moisture from the air, its
colour changed to that of the crystals, and the most refrangible of the bright bands
disappeared from the derived spectrum. Although when the incident light was very
much concentrated I have seen this band even in the crystals, it was faint com-
pared with the preceding bands, whereas in the case of the whitish mass its intensity
was not very different from that of the others. It appears therefore that the quality
as well as the quantity of the dispersed light was altered by depriving the crystals of
a part of their water.

140. A solution of nitrate of uranium in water is decidedly sensitive, though not
sufficiently so to exhibit much dispersive reflexion. When the dispersed beam is
analysed it is resolved into bright bands. When the solution is examined in a pure
spectrum, the mode of dispersion is found to agree with that of canary glass. The
dispersion commences abruptly at the same part of the spectrum as in the case of the
glass, and after a rather narrow band in which light is copiously dispersed, there
follows a remarkable minimum of sensibility, just as in the glass (see Art. 7^-)} where
the dispersed light is almost imperceptible. After this the dispersion is resumed^.

518 PROFJESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT*

and offers nothing remarkable. The minimum of sensibility occurs at the very same
place in the spectrum^ whether the sensitive medium be a solution of nitrate of ura-
nium or glass coloured yellow by uranium.

141. Yellow Uranite. — This mineral, when examined in a linear spectrum, proved
to be sensitive in an extremely high degree. The derived spectrum consisted, as in the
case of the glass, of bright bands arranged at regular intervals, but in this case six
were seen, a band being visible in the faint red at the extremity of the spectrum
which could not be made out in the case of the glass.

142. Green Uranite^ or Chalcoliie.— According to M. Pbligot the formula of the
yellow uranite of Autun is PhO', CaO, 2(U^0^0), 8H0, and the green uranite
differs from the yellow only in having the lime replaced by oxide of copper =^. Yet a
specimen of green uranite on being examined in a linear spectrum proved totally
insensible. The primitive spectrum showed however a very remarkable system of
dark bands depending on the absorption of light by the mineral. In examining
these bands, the previous prismatic decomposition of the light, so far from being
necessary, is decidedly inconvenient. It is better to dispense with the prisms alto-
gether, using only the lens, and placing the mineral so that the image of the slit is
formed upon it. The bright line thus formed is viewed from a convenient distance
through a prism, the eye being held out of the direction of regular reflexion. The
position of any bands which may appear in the spectrum can then be determined by
means of the fixed lines, which are seen at the same time ; or, if it be desired to see
the latter more distinctly, it will be sufficient to attach a fragment of paper to the
mineral or other substance, placing it so that the image of the slit is formed partly
on the paper and partly on the substance to be examined. I have frequently found
this mode of observation convenient in examining the absorption of light by opake
substances. The manner in which the absorption of the medium comes into play in
this case will be considered in greater detail further on (see Art. 176.).

143. When green uranite was examined in this manner, it showed a very remark-
able system of dark bands of absorption. These bands were seven in number, or at
any rate six, and were arranged with all the regularity of bands of interference. The
first was situated at about 5yF, the second at F; the middle of the sixth fell a very
little short of G; the third, fourth and fifth were arranged at regular intervals be-
tween the second and sixth ; the seventh was situated about as far beyond the sixth
as the sixth beyond the fifth. The spectrum was so faint in the region of the seventh
band as to leave some slight doubts respecting its existence. There would not have
been light enough to see bands further on.

144. Uranite is highly lamellar in its structure, from whence it is otherwise called
uran-mica. The reader may perhaps suppose that the dark bands described in the
last paragraph were bands of interference, which I had mistaken for bands of absorp-
tion, and that they were really of the nature of Newton's rings, or more exactly of

* Annales de Chimie, torn. v. (1842) p. 46.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 519

the bands seen in an experiment due to the Baron von Wrede. There may, it will
perhaps be said, have been a fissure parallel to the first surface, so as to separate a
thin plate ; and the interference of the two streams of light reflected respectively on
the upper and under surface of this plate may have produced the bands observed*
But various phenomena attending these bands are irreconcilable with such a suppo-
sition. Towards the edges of the crystal, where flaws did in fact exist, bands of the
same nature as Von Wrede's were actually observed. But these had an appearance
totally different from that of the others. The dark bands of the interference system
were more intensely black and better defined than those of the other system, and
were very variable, depending as they did upon the thickness of the plate by which
they were formed, whereas the bands belonging to the first system were always the
same. Besides, were these bands due to interference, there is no reason why they
should be confined to one region of the spectrum, and that by no means the brightest.
However, to take away all possible doubts respecting the nature of the bands, I
detached a small scale from the crystal, and having placed it behind a slit in a beam
of sunlight condensed by a lens, I analysed the transmitted light by a prism. Were
the bands really due to absorption, they ought to be more distinct in the transmitted
light, whereas, were they of the nature of Von Wrede's bands, they ought to be faint,
and almost imperceptible. The spectrum of the transmitted light contained however
four dark bands, which were well defined and intensely black. The whole of the
spectrum beyond the place of the next band was absorbed, which is the reason why
four bands only were visible.

145. The absorption bands of green uranite, though they showed great regularity
with respect to their positions, did not appear very regular with regard to their in-
tensities. The second, fifth and sixth seemed to me to be more conspicuous than the
first, third and fourth. I cannot say for certain whether this ought to be attributed
to fluctuations in the absorbing power of the medium, or fluctuations in the original
intensity of the solar spectrum, but I am strongly inclined to prefer the former view.

146. The intervals between the absorption bands of green uranite were nearly equal
to the intervals between the bright bands of which the derived spectrum consisted in
the case of yellow uranite. After having seen both systems, I could not fail to be im-
pressed with the conviction of a most intimate connexion between the causes of the
two phenomena, unconnected as at first sight they might appear. The more I examined
the compounds of uranium, the more this conviction was strengthened in my mind,

147. Yellow uranite exhibits a system of absorption bands similar to those of green
uranite. Nitrate of uranium also shows a similar system. In a solution I have ob-
served seven of these bands arranged at regular intervals. The first absorption band
coincided with F, the fifth with G nearly. The absorption bands may also be seen
by analysing the light transmitted through the crystals. The following arrangement
exhibited at one view the absorption bands and those due to the light which had
changed its refrangibility.

MDOCCLir. 3 x

52Q PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT.

148. The sun's light was reflected horizontally by a mi rroi% and condensed by
passing through a large lens. It was then transmitted through a vessel with parallel
sides containing a moderately strong ammoniacal solution of a salt of copper. The
strength of the solution, and the length of the path of the light within it^ were such
as to allow of the transmission of a little green besides the blue and violet. A crystal
of nitrate of uranium was then attached to a narrow slit, and placed in the blue beam
which had been transmitted through the solution, the crystal being turned towards
the incident light. The light coming from the crystal through the slit was then
viewed from behind, and analysed by a prism. A most remarkable spectrum was
thus exhibited, consisting from end to end of nothing but bands arranged at regular
intervals. The interval between consecutive bands appeared to increase gradually
from the red to the violet, just as is the case with bands of interference. Although
this interval appeared to alter continuously from one end of the spectrum to the
other, the entire system of bands was made up of two distinct systems, different in
appearance, and very different in nature. The less refrangible part of the spectrum,
where only for the crystal there would have been nothing but darkness, was filled
with narrow bright bands, due to the light which had changed its refrangibility.
These bands were much narrower than the dark intervals between them, but they
were not mere lines containing light of definite refrangibility. The more refrangible
part of the spectrum was occupied by the system of bands of absorption. The in-
terval between the most refrangible bright band and the least refrangible dark band
of absorption appeared to be a very b'ttle greater than one band-interval, so that had
there been one band more of either kind the least refrangible absorption band would
have been situated immediately above the most refrangible bright band. With strong
light I think I have seen an additional band of this nature.

149. Pitchblende. — This mineral proved to be quite insensible, and exhibited no-
thing remarkable.

150. Hydrate of Peroxide of Uranium,~Some crystallized nitrate of uranium was
exposed to a heat a good deal short of redness, whereby most of the acid was expelled.
The residue was of a deep brick-red colour, and consisted no doubt chiefly of an-
hydrous peroxide. It was quite insensible. In order to remove any undecomposed
nitrate, it was boiled with water, whereby the undecomposed nitrate was dissolved,
and the peroxide converted into a hydrate. This hydrate, after having been washed
and dried at the temperature of the air, was of an extremely beautiful yellow colour,
and was I suppose the hydrate U^0^+2H0 described in chemical treatises. It was
tolerably sensitive, in fact for an inorganic substance extremely so, though the sensi-
bility was much less than that of nitrate of uranium, yellow uranite, or canary glass.
The derived spectrum consisted as before of separate bright bands. A small portion
of the powder was attached by water to blotting-paper, and dried before a fire. The
powder thus obtained on paper was duller than before, and inclined a little more to
orange, though the colour was not much deeper than that of the former hydrate.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 521

From its colour and the circumstances of its formation^ it was probably the other
hydrate U^O'+HO. It proved on examination to be totally insensible.

151. Acetate of Peroxide <^l/ra?^iwm5 prepared by dissolving the yellow hydrate of
the peroxide in acetic acid, and evaporating to crystallize.— This salt is extremely
sensitive, about as much so as the nitrate. The derived spectrum consisted of six
bright bands arranged at regular intervals. It seemed to me that the last five of
these were respectively a little more refrangible than the five bands given by the
nitrate, and then a sixth band was visible in the faint red in the case of the acetate
which was not ordinarily seen in the nitrate. However, this observation has need to
be repeated under more favourable circumstances.

152. Nitrate and acetate of peroxide of uranium, yellow nranite, and canary glass,
are all so highly sensitive as to allow the primary spectrum to be examined with a
prism at some distance. In the first three media the bright bands are narrow,
much narrower than the dark intervals between ; in the glass they appear much
broader than in the other media.

153. Oxalate of Peroxide of Uranium^ prepared in the manner mentioned by
M. Peligot, namely, by adding a saturated solution of oxalic acid to a solution of
nitrate of uranium, washing and drying the precipitate. — ^This salt was sensitive, but
only in a low degree. However, the derived spectrum bore prismatic examination
sufficiently to show three or four bright bands. The absorption of the medium was
examined by spreading some of the powder on glass along with water and allowing it
to dry. The layer was then examined by different methods. The salt exhibits three
very intense absorption bands in the highly refrangible part of the spectrum. The
positions of these bands, by measurement, were FO'Bl G, FO'58 G, FO-85 G.

154. Phosphate of Peroxide of Uranium^ prepared by precipitation from a solution
of nitrate of uranium by adding a solution of common phosphate of soda.— -This ^alt
was sensitive, though not in a high degree. It was a good deal more sensitive than
the oxalate, but I think not so much so as the hydrate of the peroxide. The de-
rived spectrum consisted of bright bands as usual ^.

155. Uranate of Potassa^ prepared by dropping a solution of nitrate of uranium
into a solution of caustic potash, stopping long before the alkali was neutralized. —
This salt was found to be insensible, both in its original state as a gelatinous hydrate,
and in various stages of drying.

156. Uranate hf Lime^ prepared in a similar manner with lime-water.- — ^This salt,
which after drying is of a fine orange colour, was like the preceding found to be
insensible. It seamed interesting to examine these two salts, because the former con-
tains two elements (not counting oxygen) in common with canary glass, and the
latter two elements in common with yellow uranite. Yet the salts are insensible while
the tw^o other media are so remarkably sensitive.

157. Solutions by means of alkaline carbonates. --At is known to chemists that alka-

* See note 0.
3x2

522 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

line carbonates, added in solution to a solution of nitrate of uranium, give yellow pre-
cipitates which are redissolved in an excess of the precipitant. The solutions thus
obtained with the carbonates of potassa and soda, which were of a greenish yellow
colour, were found to be totally insensible. They exhibited however four of those
singular absorption bands so characteristic of salts of peroxide of uranium. Of these
the third fell a little short of G,its more refrangible edge nearly coinciding with that
fixed line ; the first and second were situated between F and G, the distance of the
first beyond F being somewhat greater than the interval between two consecutive
bands. The fourth, which was situated beyond G, was fainter than the others.
The second and third were the most conspicuous of the set.

158. The absorption bands due to peroxide of uranium afford an easy mode of
detecting that substance in solution. For this purpose the solutions mentioned in
the preceding paragraph are much preferable to the nitrate, for they produce much
stronger bands when only a small quantity of uranium is present. The absorption
bands of nitrate of uranium are visible, as might have been expected, in presence of
a large quantity of nitrate of copper =^.

Optical Tests of Uranium in Blowpipe Experiments.

159. When a bead of microcosmic salt is fused with oxide of uranium, and brought
to its highest state of oxidation, it is yellow by transmitted light. Such a bead is
sensitive in a very high degree, quite as much so as canary glass. When the light
falls sideways on it, and it is held against black cloth or a dark object, it exhibits
plainly the green colour due to internal dispersion. When properly examined by
means of sunlight its sensibility is evident at once, and when the dispersed light is
viewed through a prism it is resolved into bright bands. One of the most convenient
modes of examining such minute objects consists in reflecting the sun's light hori-
zontally through a large lens, intercepting by means of absorbing media all the rays
except those of very high refrangibility, placing the object to be examined in the
condensed beam, and viewing it through a prism. So delicate is this test when
applied to uranium, that on one occasion, when engaged in examining a bead
coloured green by chromium, which had be6n fused in the exterior flame, I observed
the appearance given by uranium. This turned out to be actually due to uranium,
of which a mere trace was accidentally present without my knowledge.

160. The green communicated to microcosmic salt by uranium after exposure to
the reducing flame has a very peculiar composition, by means of which the presence
of uranium may be instantly detected. For this purpose it is sufficient to view
through a prism the inverted image of the flame of a candle formed by the bead, the
latter being so held as to be seen projected on a dark object. The observation is
perfectly simple, and occupies only a few seconds. The spectrum exhibits an isolated
band at the red extremity, followed by a very intense dark band of absorption. A

* See note D.

PROFESSOR STOKES ON THE CHANGE OP REPRANGIBILITY OP LIGHT. 523

similar dark band, but not quite so intense, occurs in the green: beyond the green
there is usually but little light seen. As the absorption progresses the iirst dark band
invades all the space from the red to the green, and the spectrum consists of an iso-
lated red band and a green band divided into two. In its mode of absorption, the
medium has a strong general resemblance to chlorophyll. The green due to copper
or to chromium shows nothing remarkable when viewed through a prism, and could
not possibly be confounded with the green due to protoxide of uranium. The absorp-
tion bands due to this oxide are not completely brought out till the bead is cold.

161. Uranium produces the same effects with borax as with microcosmic salt, but
they are less distinct, or at least less easily produced.

162. When the uranium contained in a bead of microcosmic salt is thoroughly
oxidized, and the bead is gently heated, so as Just to be self-luminous, the light
which it gives out is not red, like that of most substances at a low heat, but green,
or rather greenish white.

163. Solutions of protoxide of uranium have a very remarkable effect on the spec-
trum, resembling more or less that of a bead of microcosmic salt coloured green by
uranium. Of course the absorption can be observed much better by means of a solu-
tion than by a mere bead. I have observed several bands of absorption in such
solutions, but the cases which I have hitherto examined are too few to justify me in
entering into detail. Besides, the absorption bands due to protoxide of uranium do
not belong properly to my subject, the compounds of this oxide, so far as I have ex-
amined, being insensible.

Appearance of highly Sensitive Media in a Beam from which the Visible Rays are

nearly excluded.

164. When a large beam of sunlight is reflected horizontally into a darkened
room, and transmitted through an absorbing medium, placed in the window, of such
a nature as to let pass only the feebly illuminating rays of high refrangibility and the
invisible rays beyond, various sensitive media have a very strange and unnatural
appearance when placed in the beam, on account of the peculiar softness of the di-
spersed light with which the media appear as it were self-luminous, and the almost
entire absence of strong light reflected from convexities. Among substances emi-
nently proper for this experiment, may be mentioned a solution of the bark of the
horse-chestnut, or of sulphate of quinine, or of stramonium seeds, a decoction of
madder in a solution of alum, and above all, ornamental articles of canary glass.
The appearance of a specimen of yellow uranite was curiously altered by this mode
of examination. By daylight the mineral appeared much of the same colour as the
stone in which it was imbedded, but when placed in a beam such as that above men-
tioned the uranite was strongly luminous, while the stone remained dark.

524 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

Natural Crystals.

165. Of natural crystals I have hitherto examined only a small number. For a
long time I was occupied almost exclusively with vegetable products^ the mineral
kingdom not appearing promising. However, I have found internal dispersion in
certain specimens of apatite, arragonite, chrysoberyl, cyanite, and topaz. In all
these cases the dispersion appeared due, as in the case of fluor-spar, to some sub-
stance accidentally present in small quantity ; so that yellow uranite is at present the
only natural crystal to the essential constituents of which the property of internal
dispersion has been found to belong.

166. Among the minerals just mentioned apatite was the most sensitive, though it
fell very far short of yellow uranite. That the sensibility was not due to phosphate of
lime, was plain from the circumstances that a colourless specimien was insensible, and
that the amount of sensibility was found to be different in different parts of the same
sensitive specimen. With the exception of the colourless crystal already mentioned,
all the specimens of apatite examined were of a greenish colour, and all were sensi-
tive. The dispersed light was something of an orange colour, but was not homo-
geneous orange. In one specimen it consisted of three distinct bright bands at
I'egular intervals. The mode in which the sensibility of this crystal was connected
with the refrangibility of the incident rays was very peculiar. In arragonite di-
spersion was found in the transparent specimens examined ; the translucent speci-
mens were found to be insensible. The dispersed light was of a brownish white
colour. In the same crystal some parts were insensible and others more or less sen-
sitive. The portions of equal sensibility were arranged in plane strata, just as in the
case of fluor-spar, as has been noticed by Sir David Brewster. In a specimen which
had been cut for showing conical refraction, the strata were in some places perpen-
dicular to the plane of the optic axes, and in other parts parallel to the line bisecting
the axes, and inclined to their plane at such an angle that the two directions of the
strata must have been parallel to two of the commonest lateral faces. Another spe-
cimen showed strata parallel to an oblique terminal face. The strata are plainly due,
as Sir David Brewster has remarked with reference to fluor-spar, to some substance
taken up during crystallization. Accordingly, they preserve a sort of history of the
growth of the crystal. In a twin crystal of fluor-spar, the direction of the strata in
that part of the mass which was common to the geometrical forms of both crystals,
showed to which crystal it really belonged. In fluor-spar the strata are parallel to
the faces of the cube, at least in the specimens which I have examined, and the same
has been observed by Sir David Brewster.

In chrysoberyl, cyanite and topaz, the dispersed light was red or reddish, and was
too variable to allow of its being attributed to the essential constituents of the
crystals. In these cases the sensibility was but slight ; indeed in cyanite there was
only a trace of dispersion when the crystal was examined under great concentration
of light.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 525

Coloured Glasses.

167. Besides canary glass, I have examined the common coloured glasses, including
that coloured by gold, but with one exception have not met with any example in
which the sensibility observed appeared to have any connexion with the colouring
matter. The paler glasses exhibited a little internal dispersion, because the colour
was not sufficiently intense to mask the dispersion which a common colourless glass
would exhibit.

168. The exception occurred in the case of the pale brown glass, which has been
already mentioned in connexion with my first experiment. This glass dispersed a
red light under the influence of the highly refrangible rays. The colour of the light
was not pure prismatic red, but red was predominant. A similar dispersion, due
apparently to the same cause, was observed in the case of one of the common reddish
brown German wine bottles. The sensibility of these glasses appears to be due to
an alkaline sulphuret. A bead purposely coloured in this manner was in fact found
to disperse a red light like the glasses. Moreover, in the confused masses obtained
by fusing sulphate of soda and sulphate of potash on charcoal before the blowpipe,
certain portions were found which dispersed a red light, and that pretty copiously
for an inorganic substance. A similar dispersion was observed among the products
obtained by fusing together sulphur and carbonate of potash, while other parts of the
confused mass exhibited dispersion of a different kind. It seems plain that among
the combinations of sulphur with the alkalies sensitive compounds exist, but what
they are I have not examined.

Cautions with respect to the discrimination between true and false internal dispersion.

169. In the early part of this paper certain tests were given for distinguishing
between true and false internal dispersion in a fluid. But it requires some experience
in observations of this kind to be able readily to decide, and a too rigid adherence
to one of the tests to the exclusion of the others might lead to error.

The first test relates to the continuous appearance of a truly dispersed beam. But
sometimes solid particles exist in mechanical suspension, which are so fine and so
numerous, that this test alone might lead the observer to mistake a falsely for a
truly dispersed beam. On the other hand, if a fluid which itself alone exhibits no
internal dispersion, true or false, hold solid particles in what is obviously mere
mechanical suspension, we must not immediately conclude that the medium, taken
as a whole, is incapable of changing the refrangibility of any portion of the light
incident upon it. For we have seen that the fluid state is not in the least degree
essential to the exhibition of sensibility, and of course a fluid will serve as well as
anything else for the mere mechanical support of a sensitive substance.

170. Thus lycopodium is very sensitive, as appears by examining the powder in a
linear spectrum. Accordingly, I found that when a little lycopodium was mixed
with water, and the whole medium was examined by the fourth method, it displayed

526 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

its sensibility^ although the beam of light which had changed its refrangibility was
plainly discontinuous. When Indian yellow was used instead of lycopodium, the
whole medium exhibited its sensibility when it was examined by the fourth method.
In this case the suspended particles were so fine that the beam of light which had
changed its refrangibility appeared to be continuous, though of course it was not really
so. In observing with muddy fluids like these^ it is almost necessary to employ
absorbing media, since otherwise the effect of the light scattered at the surfaces of the
prisms and large lens might lead the observer to conclusions altogether erroneous.

171. The next test relates to the polarization of a falsely dispersed beam. Being
engaged on one occasion in examining the effects of acids and alkalies on a weak
solution of a sensitive substance, employing sunlight which had been merely reflected
through a small lens, I met with a beam which had every appearance of having
been only falsely dispersed, but on viewing it from above through a doubly refracting
prism I was surprised at first by finding it unpolarized. It soon occurred to me that
the beam must have been due, not to solid motes, but to excessively small bubbles
of carbonic acid gas, the existence of which was thus revealed, though they were too
small to be seen directly. The light being incident on these bubbles at an angle of
about 45°, which is very little less than the angle of total reflexion, the reflected
light would be almost perfectly unpolarized =*.

172. Water which had been merely boiled in a test tube gave a similar result.
The unpolarized beam of falsely dispersed light was of course due in this case to the
air which had been held in solution. This shows why long-continued boiling should
be necessary, in order to free water from air. It is not that the afiinity of water for
air is so great as to be only gradually overcome, but that the air, immediately
expelled from solution when the temperature rises sufficiently, is still retained in a
state of mechanical mixture, forming excessively minute bubbles, the terminal velo-
city of which is insensible. Accordingly it is not till larger bubbles are formed, by
the casual meeting of a number of these small bubbles, that the air rises to the sur-
face and escapes.

173. With respect to the test of true dispersion depending on the change of
refrangibility, it has been already remarked that in some cases the change is so
slight, that if this test alone were applied, the observer might mistake true dispersion
for false. However, it is only in rare cases that there is any danger of being deceived
in this manner in the application of the test; but on the other hand, in observing a
muddy fluid or a translucent solid by the fourth method, the observer, if not on his
guard, might easily be deceived by the effect of scattered light, and be led to mistake
false dispersion for true. Thus suppose the medium to be water holding in suspen-
sion particles of an insensible water colour, and the small lens to be placed a little
beyond the commencement of the violet. Two beams of light would enter the lens,
namely, a regularly refracted beam of violet, and a scattered beam of white light,

* See note E.

PROFESSOR STOKES ON THE CHANGE OF REPRANGIBILITY OF LIGHT. 527

Of these the latter would be insignificant compared with the former, were it not
that the illuminating power of the colours belonging to the middle of the spectrum
is so very much greater than that of the violet. When the dispersed beam was
analysed by a prism, it would be decomposed into a violet beam of definite refrangi-
bility, followed by a dark interval, and then a broad band containing the colours of
the brighter part of the spectrum in their natural order. This is what is constantly
seen in cases of true dispersion; but the polarization of the beam, and its behaviour
under the action of absorbing media, would reveal the counterfeit character of the
dispersion.

On the Colours of Natural Bodies.

174. By this expression I mean to inclutie only the colours to which it is usually
applied, namely, those of leaves, fliowers, paints, dyed articles, &c., which form the
great mass of the colours that fall under our observation. I do not refer to colours
due to refraction, such as those of the rainbow, or to diffraction, such as those of the
coronae seen about the sun and moon, or to interference, such as those seen in the
clear wings of small flies, or to the colours which accompany specular reflexion,
which last are usually but slight, though sometimes pretty intense.

In some few instances, as for example in the case of fluor-spar, various salts of
peroxide of uranium, acid solutions of disulphate of quinine, &C.5 colours are observed,
sufficiently strong to arrest attention, which have a remarkable and hitherto unsus-
pected origin. But I am not now speaking of colours arising from a change of
refrangibility in the incident light. In the vast majority of cases these colours are
far too feeble to form any sensible, portion of the whole colour observed. The colours
which dyed articles give out under the influence of the highly refrangible rays usually
agree more or less nearly with those of which such substances commonly appear,
and it is possible that the colour arising from a change of refrangibility may contri-
bute in some slight degree to the brilliancy of the tint observed. If, however, the
effect be sensible I am persuaded that it is but slight; and very brilliant colours may
be produced without a change of refrangibility, as for example in the case of biniodide
of mercury. For the present I shall neglect the light which may have changed its
refrangibility.

175. Few, I suppose, now attach much importance to the bold speculations in
which Newton attributed the colours of natural bodies to the reflexion of light from
thin plates. Sir David Brewster has shown how extremely different the prismatic
composition of the green of the vegetable world is, from what it ought to be, accord-
ing to Newton's theory, and what Newton supposed that it was. It is now admitted
that the various colours of natural bodies are merely particular instances of one
general phenomenon, namely, that of absorption. Absorption is most conveniently
studied in a clear fluid or solid, but it does not the less exist in a body of irregular
structure, such as a dyed cloth or a coloured powder.

MDCCCLII. 3 Y

528 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT.

The green colouring matter of leaves affords an excellent example of the identity
of the effect produced on light by natural bodies and of ordinary absorption ; for the
same very peculiar system of absorption bands which are displayed by a clear solu-
tion of the colouring matter may be observed directly in the leaf itself. However, it
is needless to bring forward arguments to support a theory now I suppose universally
admitted; my present object is merely to point out the mode in which the colours
which bodies reflect, or more properly scatter externally, depends upon the absorb-
ing power of the colouring matter, so as to justify the conclusions deduced in Art.
142, from observations made in the manner there described.

176. Let white light be incident on a body having an irregular internal structure,
such as a coloured powder. A portion will be reflected at the first irregular surface,
but the larger portion will partly enter the particles, partly pass between them, and
so proceed. In its progress the light is continually reflected in an irregular manner
at the surfaces of the particles, and a portion of it is continually absorbed in its
passage through them. For simplicity's sake, suppose the light incident in a direc-
tion perpendicular to the general surface, and neglect all light which is more than
once reflected. Let t be the thickness of a stratum which the light has penetrated,
I the intensity of the light at that depth, or rather the intensity of a given kind of
light, so that the whole intensity may be represented by/I^o, f^ being the refractive
index in some standard substance. In passing across the stratum whose thickness
is dt^ suppose the fraction q^dt of the light to be absorbed, and the fraction rdt to be
reflected and scattered in all directions, then

dl='^{q+r)ldt.
Integrating this equation, and supposing lo to be the initial value of I, when ^=0,
we have

I=I,e-^^-^^')^ (a.)

For the sake of simplicity, suppose the body viewed in a direction nearly perpen-
dicular to the general surface ; and of the light reflected and scattered in passing
across the stratum whose thickness is dt^ suppose that the fraction n would enter the
eye if none were lost by absorption, &c. Then the intensity of the light coming from
that stratum would be nrldt. But in getting back across the stratum whose thick-
ness is t, the intensity is diminished in the ratio of lo to I. Hence if V be the inten-
sity of the light actually entering the eye,

dV=nrl-'rdt=nrl,e-'^'''''^'dL
If we suppose the thickness of the body sufficient to develope all the colour which
the body is capable of giving, the superior limit of t will be oo, and we shall have

^— %T^^«- . • • (b.)

177. The colour which accompanies ordinary reflexion being usually but slight, I
shall neglect the chromatic variations of r. It is q which is subject to extensive and
apparently capricious variations, depending upon the refrangibility of the light.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 529

Imagine two curves drawn whose abscissae are proportional to [Jb, and ordinates pro-
portional to the ratio of I to lo for the first, and the ratio of V to lo for the second.
These curves will serve to represent to the mhid the composition of the light trans-
mitted through a stratum of the body having a thickness t, and of that reflected
from the body when seen in mass. It is plain that the maximum and minimum
ordinates in the two curves will correspond to the same abscissae ; but unless t be
very small, so small as to be insufficient to bring out the colour of the medium seen
by transmission, the maxima and minima will be much more developed in the first
curve, whose ordinates vary as e~^^, than in the second, whose ordinates vary as
{q+ry\ If, then, the absorbing power be subject to fluctuations depending on the
refrangibility of the light, the bands of absorption may be observed either in the
reflected or in the transmitted light, but they admit of being better brought out in
the latter.

178. If the nature of the substance be given, q will be given. If now the body be
of a loose nature, as for example blue glass reduced to a fine powder, r will be con-
siderable. Hence, in accordance with the expression (b.), the quantity of light
scattered externally will be considerable, but the tint will be but slight. If the
powder be now wetted, the reflexions at the surfaces of the particles will be di-
minished, r will be diminished, and, as appears from (b.), the quantity of light scattered
externally will be diminished, but at the same time the tint will be deepened, since
the chromatic variations of I' are increased. If the body be compact and nearly
homogeneous, r will be small, and therefore very little light will be returned, except
what is regularly reflected at the first surface. The tint of the small quantity of
light which is reflected otherwise than regularly, will be somewhat purer than before,
inasmuch as the chromatic variations of V tend to become the same as those of q~\

On the naturae of False Dispersion, and on some applications of it.

179. It has been already stated that a beam of falsely dispersed light seen in a
fluid has generally more or less of a sparkling appearance, indicating that it owes
its origin merely to motes held in mechanical suspension. Sometimes, however, no
defect of continuity is apparent. This is especially the case when two fluids are
mixed together, of which one contains in solution a very small quantity of a sub-
stance which we might expect to be precipitated by the addition of the other, or
when a slightly viscous fluid has remained quiet for a long time. If some part at
least of a falsely dispersed beam be plainly due to motes, that does not of course
prove for certain that there is no part which may have a different origin, and may
be essentially connected with true dispersion ; nor do the theoretical views which I
entertain of the cause of the latter lead me to regard it as at all impossible that a
beam polarized in the plane of reflexion, and having the same refrangibility as the
incident light, may be a necessary accompaniment of true dispersion. However,
observation, I think, points in a contrary direction ; for although more or less of

3 y2

530 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT*

false dispersion is almost always exhibited along with true dispersion, the quantity of
the former seems to have no relation to the quantity of the latter, but does seem to
have relation to the greater or less degree of clearness which we should be disposed
to attribute to the fluid.

180. The phenomenon of false internal dispersion seems to admit of being applied
as a chemical test to determine whether or not precipitation takes place. Thus, if a
little tincture of turmeric be greatly diluted with alcohol, and then water be added,
a yellow fluid is obtained which appears to be perfectly clear, exhibiting no sensible
opalescence ; but the occurrence of a copious false dispersion when the fluid is
examined by sunlight, reveals at once the existence of suspended particles, though
they are too minute to be seen individually, or even to give a discontinuous appear-
ance to the falsely dispersed beam. Although such a precipitation could not, I
suppose, be used as a means of mechanical separation, it might still be useful as
pointing out the possibility of an actual separation under different circumstances as
to strength of solution, &c.

181. One of the best instances of false dispersion that I have met with, best, that
is, in forming a most excellent imitation of true dispersion, occurred in the case of
a specimen of plate-glass which was made, as I was informed, with a quantity of
alkali barely sufficient. This glass, which was very slightly yellowish brown, when
viewed edgeways by transmitted light, had a bluish appearance when viewed pro-
perly, strongly resembling that of a decoction of the bark of the horse-chestnut,
diluted with water till the dispersed light is no longer concentrated in the neigh-
bourhood of the surface. But when the glass was examined by sunlight, the polariza-
tion of the dispersed beam, and the identity of its refrangibility with that of the
incident light, showed that this was merely an instance of false dispersion. Another
very good example of false dispersion is affbrded by chloride of tin dissolved in a
very large quantity of common water.

182. When a horizontal beam of falsely dispersed light is viewed from above, in a
vertical direction, and analysed, it is found to consist chiefly of light polarized in
the plane of reflexion. It has often struck me, while engaged in these observations,
that when the beam had a continuous appearance, the polarization was more nearly
perfect than when it was sparkling, so as to force on the mind the conviction that it
arose merely from motes. Indeed, in the former case, the polarization has often
appeared perfect, or all but perfect. It is possible that this may in some measure
have been due to the circumstance, that when a given quantity of light is diminished
in a given ratio, the illumination is perceived with more difficulty when the light is
uniformly diff'used than when it is spread over the same space, but collected into
specks. Be this as it may, there was at least no tendency observed towards polariza-
tion in a plane perpendicular to the plane of reflexion, when the suspended particles
became finer, and therefore the beam more nearly continuous,

183. Now this result appears to me to have no remote bearing on the question of

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OP LIGHT. 531

the direction of the vibrations in polarized light. So long" as the suspended particles
are large compared with the waves of lights reflexion takes place as it would from a
portion of the surface of a large solid immersed in the fluids and no conclusion can
be drawn either way. But if the diameters 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 perpendicular to the vibrations in the incident ray. Let us suppose for
the present, that in the case of the beams actually observed, the suspended particles
were small compared with the length of a wave of light. Observation showed that
the reflected ray was polarized. Now all the appearances presented by a plane-
polarized ray are symmetrical with respect to the plane of polarization. Hence we
have two directions to choose between for the direction of the vibrations in the
reflected ray^ namely, that of the incident ray, and a direction perpendicular to both
the incident and the reflected rays. The former would be necessarily perpendicular to
the directions of vibration in the incident ray, and therefore we are obliged to choose
the latter, and consequently to suppose that the vibrations of plane-polarized light
are perpendicular to the plane of polarization, since experiment shows that the plane
of polarization of the refliected ray is the plane of reflexion. According to this
theory, if we resolve the vibrations in the incident ray horizontally and vertically,
the resolved parts will correspond to the two rays, polarized respectively in and per-
pendicularly to the plane of reflexion, into which the incident ray may be conceived
to J3e divided, and of these the former alone is capable of furnishing a reflected ray,
that is of course a ray reflected vertically upwards. And in fact observation shows,
that, in order to quench the dispersed beam, it is sufficient, instead of analysing the
reflected light, to polarize the incident light in a plane perpendicular to the plane of
reflexion.

Now in the case of several of the beams actually observed, it is probable that many
of the particles were really small compared with the length of a wave of light. At
any rate they can hardly fail to have been small enough to produce a tendency in the
polarization towards what it would become in the limit. But no tendency what-
soever was observed towards polarization in a plane perpendicular to the plane of
reflexion. On the contrary, there did appear to be a tendency towards a more com-
plete polarization in the plane of reflexion.

M. Babinet has been led by the same reasoning to an opposite conclusion
respecting the direction of the vibrations in polarized light, resting on an experiment
of M. Arago's, in which it appeared that when light was incident perpendicularly on
the surface of white paper, and the reflected or rather scattered light was viewed in
a direction almost grazing the surface, it was found to be partially polarized in the
plane of the sheet of paper^. But the actions which take place when light is inci-
dent on a broad irregular surface, like that of paper, bounding too a body which is
so translucent that a great part of the light must enter it and come out again, appear

* Comptes Rendus, torn. xxix. p. 514.

532 PEOFiSSOR STOKES ON THE CHANGE OF REFRANGIBIMTY OF LIGHT.

to me to be too complex to allow us to deduce any conclusion from the result respect-
ing the direction of vibration. Besides^ the result itself admits of easy explanation,
by attributing it to the light which has entered the substance of the paper and come
out again, which might be expected to be polarized by refraction.

Effect of Heat on the Semibility of Glms^ ^c.

184. The sensibility of glass is temporarily destroyed by heat. The glass may be
heated by holding it in the flame of a spirit-lamp, as a heat much short of redness is
sufficient. This takes place even with glass coloured by oxide of uranium, which is
in general so highly sensitive. The sensibility returns again as the glass cools. A
bead of microcosmic salt, containing uranium in its highest state of oxidation, is very
sensitive when cold, but insensible when hot. The sensibility gradually comes on as
the bead cools. A solution of nitrate of uranium in water on being heated has its
sensibility impaired, very much so by the time the temperature reaches the boiling-
point. The sensitive compounds, whatever may have been their precise nature,
obtained by fusing the sulphates of soda and potassa on charcoal before the blow-
pipe, were insensible while hot. The few vegetable solutions which I have
examined with this object did not seem to have their sensibility affected by being
heated.

Eff^ect of Concentration and DilMti(m.

185. In investigating the change of refrangibility produced by a sensitive substapce
in solution, it is almost always convenient to have the solution weak. This however
is by no means merely a matter of convenience, for the quantity of light which the
medium is capable of giving back with a changed refrangibility is often materially
diminished by increasing the concentration of the solution. Thus a solution which,
when in a concentrated state, exhibits no sensible dispersive reflexion, will often
exhibit when much diluted a very copious appearance of that nature. On the other
hand, the dilution may of course be carried too far, so as to render imperceptible the
peculiar properties of the substance dissolved. Yet it is wonderful what a degree
of dilution a highly sensitive solution will bear before its sensibility ceases to be
perceptible.

That the sensibility will be diminished, and will at last become imperceptible, if
only the dilution be carried far enough, is nothing more than might have been pre-
dicted with the utmost confldence. In such a case the light passes completely
through the fluid long before it has produced all the effect which it is capable of pro-
ducing. But that concentration should be an obstacle to the exhibition of the phe-
nomenon is not perhaps what we should have expected, and deserves an attentive
consideration.

186. Imagine a given sensitive substance to be held in solution, in a vessel of which
the face towards the eye is plane, and the breadth in the direction of vision as great
as we please ; and suppose the solvent, or at least the fluid used for diluting the solu-

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT. 533

tion, to be itself colourless and insensible. Suppose the fluid to be illuminated by
light of given intensity and given refrangibility entering at the face next the eye, and let
the eye E from a given pomtion look in the direction of a given point P in the nearer
surface of the vessel. In short, let everything be given except the strength of the
solution. For the sake of simplicity regard the eye as a point, and make E the vertex
of an indefinitely thin conical surface surrounding the line EP. Call this conical
surface C, and let cbe the surface within the fluid generated by right lines coinciding
with the refracted rays which would be produced by incident rays coinciding with
the generating lines of the surface C. This latter surface we may if we please regard
as eylindrical, since we shall only be concerned with so much of the fluid contained
within it as lies at a distance from P less than that at which the light entering the
eye in consequence of internal dispersion ceases to be sensible; and in the cases to
which the present investigation is meant to apply this distance is but small compared
with PE. Let the fluid within c be divided into elementary portions by planes
parallel to the surface of the fluid at P, and at distances from P proportional to the
strength of the solution. It is evident that an element of a given rank, reckoned
from P, will contain a constant number of sensitive molecules, and the incident light
in reaching this element has to pass through a thickness of the medium such that a
plate of the same thickness, and having a given area, contains a given number of
sensitive or absorbing molecules. The same is true of the dispersed light which
proceeds from the element and enters the eye. Now it seems natural to suppose that
if the strength of a solution be doubled, trebled, &c., or reduced to one-half, one-third,
&c., the quantity of light absorbed will be the same provided the length of the path
of the light be reduced to one-half, one-third, &c., or doubled, trebled, &c. This
comes to the same thing as supposing that each absorbing molecule stops the same
fractional part of the light passing it, whether the solution be more or less dilute.
We should similarly be inclined to suppose that each sensitive molecule would give
out the same quantity of light, when influenced by light of given intensity, whether
it belonged to a stronger or a weaker solution. If we admit these suppositions, it is
plain that the quantity of dispersed light which reaches the eye from the element
under consideration will be independent of the strength of the solution. This being
true for each element in particular will be true for the aggregate effect of them all,
and therefore the quantity of light exhibited by dispersive reflexion will be indepen-
dent of the strength of the solution. It may be readily seen that the result will be
the same if we take into account the finite size of the pupil.

187. Now this is by no means true in experiment. On examining in a pure spec-
trum a highly concentrated solution of sulphate of quinine, a copious dispersion was
observed to commence a little below the fixed line Q. It remained very strong as
far as H, and beyond. In the weak solution first mentioned in this paper, it will be
remembered that the dispersion seemed to come on about G^H. The reason of this,
or at least one reason, is evident, and was very prettily shown by the form of the

534 PROFESSOE STOKES ON THE CHANiJE OF REFRANGIBILITY OF LIGHT.

space to which the dispersed light was confined. On looking down from above, so
that this space was seen in projection, it appeared in the case of the weak solution
to have approximately the form of the space contained between one branch of a rectan-
gular hyperbola, one asymptote, and a line parallel to the other, the first asymptote
being the projection of the anterior surface, and the line parallel to the other being the
course of the least refrangible of the active rays which were capable of producing a
sensible quantity of dispersed light. The breadth of the illuminated space, which among
the most highly refrangible rays was almost insensible, continually increased, until
the space ended in a blue beam which went quite across the vessel. But in the case
of the strong solution the illuminated space had throughout an almost insensible
breadth^ except just close to its lower limits that is, the limit corresponding to the
least refrangible of the active rays, where it ended in a sort of tail or plano-concave
wedge, which penetrated to a moderate distance into the fluid. Hence one reason,
though perhaps not the only reason, why the strong solution showed a copious
dispersion from G to G|H, where the weak solution showed hardly any, is plain
enough. But in the region of the invisible rays beyond the violet, the dispersion
was plainly more copious with the weak than with the strong solution. It appears
then that in such a case the sensitive molecules do not act independently of each
other, but the quantity of light emitted by a given number of molecules is less, in
proportion to the light (visible or invisible) consumed, than when a solution is more
dilute. We should expect h priori that when a solution is tolerably dilute further
dilution would make no more difference in this respect. This seems to agree very
well with experiment. For when a pretty dilute solution and one much more dilute
are compared with respect to the quantity of dispersed light gi?en out in a given
portion of the incident spectrum^ they appear to be alike. I suppose the comparison
to be made with respect to such a portion of the incident spectrum, or in the case of
solutions of such strength, that the dispersed light is confined to a space extending to
no great distance into the fluid in either solution. Under these circumstances the
comparison may be made easily enough.

188. In the actual experiment, the elementary portions of light coming from the
elementary strata of fluid situated at different distances from the anterior surface
enter the eye together. Let us however trace the consequences of the very natural
supposition, that in passing across a given stratum of fluid the quantity of light
absorbed, as well as the quantity given out by dispersion, is proportional, ccefem
paribus, to the intensity of the incident light. The incident light is here supposed
to be homogeneous, and to belong indifferently to the visible or invisible part of
the spectrum. In crossing the elementary stratum having a thickness dt, let the
fraction qdt of the incident light be absorbed, and the fraction rdt dispersed in such
a direction as to reach the eye; and of the latter portion let the fraction sdt be
absorbed in crossing a stratum having a thickness dt, s being different from g on
account of the change of refrangibility. Then by a very simple calculation similar to

PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT. 535
that of Art. 176, we find for the intensity I' of the dispersed light which enters the eye

lo being the intensity of the incident light. Since a sensitive fluid is in general coloured,
and the dispersed light is in general heterogeneous, s will in general be different for
the different portions into which the dispersed light would be decomposed by a prism.
However, if the fluid be colourless, or all but colourless, as is the case with a solution
of sulphate of quinine, s will be insensible, so that V will be proportional simply to
rq~^. Hence from the observed variations in F, arising from variations in the strength
of the solution, we may infer the corresponding variations in rq~\

If, then, we represent by the ordinate of a curve the ratio of the quantity of light
given out to the quantity of light absorbed by a given number of active molecules,
the abscissa being the ratio of the quantity of diluting fluid to the quantity of the
sensitive substance in solution, it appears that the curve will be concave towards the
axis of the abscissae, and will have an asymptote parallel to that axis.

On the Choice of a Screen.

189. We have seen that white paper, the substance commonly employed as a
screen on which to receive the spectrum, gives back with a changed refrangibility a
portion of the light incident upon it. This might in some cases lead an observer
not aware of the circumstance to erroneous conclusions. Since the colour of dispersed
light depends upon its refrangibility, which is different from that of the active light,
the colours of a spectrum received on white paper must be somewhat modified. In
truth the intensity of the light dispersed is so small compared with the intensity of
the light scattered, that the modification is quite insensible except in the extreme
violet. But beyond the extreme violet the spectrum seems to be prolonged with a
sort of greenish gray tint, which belongs neither to that nor to any other part of the
true spectrum. In experiments on absorption, if instead of receiving the light
directly into the eye it be found convenient to form a pure spectrum on a screen of
white paper, then, if the absorbing medium be placed in the path of the incident
light, the scattered light forming any part of the spectrum cannot be cut off or
weakened without at the same time cutting off or weakening the dispersed light coming
from the same part of the screen. But if the absorbing medium be held in front of
the eye, its effect on the spectrum will sometimes be very sensibly different from what
it would be were the screen to send back none but scattered light.

It is true that the quantity of light dispersed by white paper is so small that this
substance may very well continue to be used as a screen, without any danger of the
observer's being deceived, if only he be aware of the fact of dispersion, so as to be on
his guard. Still, it is not unreasonable to seek for a substitute for paper, which may
be free from the same objection.

190. A porcelain tablet appeared to be unexceptionable in this respect, for it exhi-

MDCCCMI. 3 z

536 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT.

bited no perceptible sensibility^ even when examined by a linear spectrnm. However,
the translucency of the substance gave the spectrum a blurred appearance, and the
fixed lines were not shown so well as on paper.

Chalk scraped smooth is well adapted, from its fineness, its whiteness and its opa-
city, for showing the most delicate objects. The finest fixed lines are beautifully seen
on it, decidedly better than on paper. Its sensibility too, though not absolutely null,
is much less than that of most kinds of white paper. Indeed, it would be an unne-
cessary refinement to seek for anything better, were it not that a piece of suflicient
size might not always be at hand. From what I have seen, I believe that the best
kind of screen will be obtained by the use of some white inorganic chemical precipi-
tate, but my experiments in this department have not yet been sufliciently extended
to authorize me in recommending any particular process.

191. The object of the observer may however be altogether different, and he may
wish to extend the spectrum as far as possible, for the purpose of viewing the fixed
lines belonging to the invisible part beyond the extreme violet, or making experi-
ments on the invisible rays. For this purpose it would be proper to employ a clear
and highly sensitive solid or fluid. A weak solution of sulphate or phosphate of
quinine would do very well, or a weak decoction of the bark of the horse-chestnut
(no doubt a solution of pure esculine would be better), or an alcoholic solution of the
seeds of the Datura stramonium. But perhaps the most convenient thing of all would
be a slab of glass coloured yellow by oxide of uranium. This would be always ready,
and in point of sensibility the glass does not seem to yield to any of the solutions above
mentioned, at least so far as relates to those rays which are capable of passing through
glass ^.

192. In making experiments on the invisible rays, it is well to get rid, as far as
possible, of the glare arising from the bright part of the spectrum, and therefore a
clear solid or solution is preferable to an opake screen. If it be desired to show the
fixed lines in the visible and invisible parts of the spectrum at the same time^ a screen
may be employed consisting of paper washed with a moderately strong solution of
sulphate of quinine, or an alcoholic solution of stramonium seeds. Turmeric paper
is not, I think, quite so good for showing the fixed lines of very high refrangibility,
but is at least equally good for the extreme violet and for the rays a good distance
further on, especially if it has been washed with a solution of tartaric acid. It is
likely that many other acids would do as well. Very excellent screens might pro-
bably be prepared by washing paper with a solution of esculine, or even of the bark of
the horse-chestnut-f, or by covering pasteboard with yellow uranite reduced to fine
powder, and made to adhere by a weak solution of pure gum Arabic; but these I
have not tried.

* See note F. t See note G.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 537

Application of internal dispersion to demonstrating the course of rays.

193. Solutions of quinine have already been employed for this purpose^ and a weak
decoction of the bark of the horse-chestnut appears to be decidedly better. But the
effect is immensely improved by using absorbing media to cut off all the rays belong-
ing to the bright part of the visible spectrum. A deep blue glass will answer very
well for this purpose if its ftices be even, so as not to disturb the regularity of the
refraction. The appearance of the general pencil refracted through a rather large
lens, with its caustic surface, its geometrical focus^ &C.5 is singularly beautiful when
exhibited in this way, on account of the perfect continuity of the light, and the deli-
cacy with which the different degrees of illumination belonging to different parts of
the pencil are represented by the different degrees of brightness of the dispersed light.
The solution should be contained in a vessel with plane sides of glass, and ought to
be very weak, or else only the part of the pencil which lies near the surface by which
the light enters will be properly represented.

Application of internal dispersion to the determination of the absorbing power of media
with respect to the invisible rays beyond the violet y and the reflecting power of surfaces
with respect to those rays.

194. Hitherto no method has been known by which the absorbing power of a me-
dium with respect to these rays could be determined for each degree of refrangibility
in particular, except that which consists in taking a photographic impression of a pure
spectrum, the light forming the spectrum having been transmitted through the sub-
stance to be examined. It is needless to remark how troublesome such a process is
when contrasted with the mode of determining the absorption which media exercise
on the visible rays. But the phenomenon of internal dispersion furnishes the philo-
sopher, so to speak, with eyes to see the invisible rays^ so that the absorbing power of
the medium with respect to these rays may be instantly observed. For this purpose
it is sufficient to form a pure spectrum, using instead of a screen a highly sensitive
fluid or solid^ such as one of those mentioned in Art. 191^ and to hold before it the
medium to be examined, or else to place the medium over the whole or a part of the
slit.

195. In this way the transparency of glass coloured yellow by oxide of silver with
respect to the violet rays and some of those still more refrangible, which has been
remarked by Sir John Herschel'^, may be at once observed. A set of green glasses
were found to be very variable in the mode in which they absorbed the invisible rays,
some absorbing the more refrangible of the rays capable of affecting a dilute solution
of sulphate of quinine and transmitting the less refrangible, others absorbing the less
and transmitting the more refrangible, and others again absorbing them all. These
rays were absorbed by solutions of chromate and bichromate of potash so weak as
to be almost colourless. A thickness of about a quarter of an inch of sulphuret of

* Philosophical Transactions for 1840, p. 39.
^3z 2

538 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT.

carbon was sufficient to absorb all the raj^s beyond H^l, so that a hollow prism filled
with this fluid would be useless in experiments on these rays. It should be remarked
that the sulphuret of carbon employed was not yellow from dissolved sulphur^ but
apparently as colourless as water.

196. To determine qualitatively the reflecting power of a polished surface with
respect to the invisible rays of each particular degree of refrangibility, it would be
sufficient to form a pure spectrum as usual, reflect the rays sideways before they
come to the focus of the larger lens, place a sensitive medium to receive them^ and
compare the effect with that produced on the same medium when the rays are
allowed to fall directly upon it.

Effect of different Flames.

197. Want of sunlight proved to be such an impediment to the pursuit of these
researches that I was induced to try some bright flames, with the view of obtaining
some convenient substitute. Candle-light is very ill adapted to these experiments.
The flame of a camphene-lamp proved no better, perhaps rather worse, for it abounds
so much in rays belonging to the bright part of the spectrum that the glare of the
light prevents all observation of faint objects ; and the flame does not appear to be
rich in invisible rays in anything like the proportion in which it is rich in visible
ones. The flame of nitre burning on wood or charcoal produced a very good effect,
exhibiting, when the combustion was most vivid, a copious dispersive reflexion in a
weak solution of sulphate of quinine contained in a bottle held near it. The tint of
the dispersed light appeared to be not quite the same as that given by daylight, but
to verge a little towards violet. However, I do not place very strong reliance on the
judgment of the eye under such circumstances. A still stronger dispersive reflexion
was produced by a flash of gunpowder. The tint in this case appeared to be the
same as that seen by daylight.

198. While engaged in some of these experiments on bright flames, I was surprised
by discovering the strong effect produced by the flame of a spirit-lamp, the illumi-
nating power of which is so feeble. When this flame was held close to a bottle con-
taining sulphate of quinine, a very distinct dispersive reflexion was exhibited. The
same was the case with several other sensitive solutions. However, the full effect of
the flame is not thus exhibited, because a considerable portion of the rays which it
emits is stopped by glass. It is best observed by pouring the solution into an
open vessel, such as a wine glass or tumbler, holding the flame immediately over it,
and placing the eye in or very little below the plane of the surface. In this way
nothing is interposed between the flame and the fluid, except an inch or two of air,
the absorption produced by which, it is presumed, is insensible ; and the plane strata,
parallel to the surface, into which the illuminated portion of the fluid may be con-
ceived to be divided, are all projected into lines, whereby the intensity of the blue
light is materially increased. It is to be observed further, that if the eye be held a

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 539

little below the plane of the surface, there enters it, not only the light coming directly
from the blue stratum itself, but also that coming from its image formed by total
internal reflexion. This mode of observation has already been emph)yed by Sir John
Herschel in the case of sunlight. As it is frequently useful in these researches it
will be convenient to have a name for it, and I shall accordingly speak of it as the
method of observing by superficial projection.

199. The opacity of a solution of sulphate of quinine appears to increase regularly
and rapidly with the refrangibility of the light. Hence we may form an estimate of
the refrangibility of any light by which the solution may be affected, by observing
the degree in which the illumination is concentrated in the neighbourhood of the
surface. For this purpose it is essential to employ a weak solution, since otherwise
streams of invisible light of various degrees of refrangibility produce each their full
effect in strata so very narrow, that they cannot be distinguished by the breadth of
the stratum. Now to judge by the great concentration of the illumination produced
by a spirit-lamp, even in the case of an extremely weak solution, as well as by the
considerable degree in which the active rays were intercepted by glass, these rays,
taken as a whole, must have been of very high refrangibility, such as to place them
among the most refrangible of the fixed lines represented in the map, or perhaps
even altogether beyond them. In making observations on the solar spectrum, it
was plain that the prisms were by no means transparent with respect to the rays
belonging to the group p of fixed lines. Yet these rays, before they produced their
effect, had to pass twice through the plate-glass belonging to the mirror (except so
far as regards the rays reflected at the first surface), then through three prisms^
though to be sure as close as possible to the edges, then through a lens by no means
very thin, and lastly, through the side of the vessel containing the fluid. Such a
train of glass would be sufficient materially to weaken, if not even wholly to cut off
the active rays coming from the flame of a spirit-lamp.

200. The flame of naphtha produces nearly the same effect as that of alcohol. The
flame of ether is not so good ; but whether this arises solely from its richness in
visible rays, which only produce a glare, or likewise from a comparative poverty in
highly refrangible invisible rays, it is not easy to say. The flame of hydrogen pro-
duces a very strong effect. The invisible rays in which it so much abounds, taken
as a whole, appear to be even more refrangible than those which come from the
flame of a spirit-lamp. In making some observations with the flame of hydrogen,
when the gas was nearly exhausted, so that the flame was reduced to a roundish
knob no larger than a sweet pea, and giving hardly any light, it was found still to
produce a very marked effect when held over the surface of a solution of sulphate of
quinine. The flame of sulphuret of carbon produces on most objects a much stronger
effect than that of alcohol. It exhibits distinctly the blue light dispersed close to the
surface of a solution of guaiacum in alcohol, which the flame of alcohol does not .It
appears then that the flame of sulphuret of carbon is rich in invisible rays of such u

540 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

refrangibility as to place them among the groups of fixed lines m^ n^ or a little beyond^
since when a solution of guaiacum is examined in the solar spectrum, it is found that
that is the region in which the blue dispersed light is produced. The blue light
dispersed by a solution of guaiacum may also be seen by using the blue flame of
sulphur burning feebly. The poverty of the flame of a spirit-lamp, not only with
respect to visible rays, but also with respect to invisible rays, except those of very
high refrangibility, accounts for the circumstance that it does not exhibit, or at least
hardly at all exhibits, the blue light dispersed by fluor-spar.

Mode of determining^ by means of the light of a spirit-lamp^ the transparency of'
bodies ivith respect to the invisible rays of high refrangibility.

201. If the body be a solid, and be bounded by parallel surfaces, its transparency
with regard to these rays is easily tested. For this purpose it is sufficient to hold
the flame of a spirit-lamp a little way above the surface of a weak solution of sulphate
of quinine contained in an open vessel in a dark room, and then, placing the eye so
as to see the dispersed light in projection, alternately to interpose and remove the
plate to be examined.

202. On examining in this way various specimens of glass, I found none which
did not show evident defects of transparency. The purest specimens of plate-glass
appeared, I think, to be the least defective. I cannot say whether the observed
defects of transparency were due to the essential ingredients of the glass, or to acci-
dental impurities. It is possible that glass made with chemically pure materials
might be transparent^. I believe that a mere trace of peroxide of iron, or of sul-
phuret of soda or potassa, would be sufficient to impair materially the transparency
of glass with respect to these rays, and such impurities are very likely to be present.
Quartz, however, appeared to be perfectly transparent, the active rays passing
through the thickness of one or two inches, whether parallel or perpendicular to the
axis, without any perceptible loss. The contrast between quartz and mica was very
striking, for a plate of mica no thicker than paper produced a very sensible diminu-
tion in the illumination.

203. For the purpose of observing fluids, I procured two vessels consisting of sec-
tions of a wide glass tube, about an inch long, closed at one end with a disc of quartz.
I shall call these for brevity quartz vessels, though of course the bottom is the only
part in which there is any occasion to use quartz. When a fluid is to be examined
it is poured into a quartz vessel, and then the vessel with its fluid contents is ex-
amined in the manner of a solid plate, as described in Art. 201. On account of the
perfect transparency of quartz, the fluid is as good as suspended in air. When a

=^ Some specimens of glass belonging to Dr. Faraday's experiments, whicli from the absence of colour and
of internal dispersion seemed hopeful, could not be examined for transparency, on account of their irregular
figure ; and as they were only lent to me by a friend, I did not feel myself at liberty to get them cut and
polished.

FROPISSOR STOKES ON THE CHANGE OP REPRANGIBILITY OP LIGHT- 541

quartz vessel was partly filled with water^ the addition of a very small quantity of
nitrate of iron was sufficient to cause the absorption of the active rays. The solution
was so weak as to be almost colourless when viewed through the thickness through
which the rays would have to pass. A solution of perchloride of iron had a similar
effect. These fluids I had specially examined by sunlight^ and had not found in them
the least trace of internal dispersion. When a fluid exhibits internal dispersion^ it is
almost always very opake with regard to rays of high refrangibilityj as is shown^
without any special experiment^ in the course of the observations by which the internal
dispersion is exhibited; but it by no means follows conversely^ that when a fluid is
very opake with regard to these rays, though nearly transparent with regard to the
visible rays, it exhibits the phenomenon of internal dispersion.

204. I have little doubt that the solar spectrum would be prolonged, though to
what extent I am unable to say, by using a complete optical train in every member
of which glass was replaced by quartz. Such a train would be rather expensive, but
would not involve any particular difficulty of execution. If solid prisms of quarts
were used, half of the incident light would be lost, on account of the double refraction
of the substance, unless the prisms were cut in a particular manner, which however
would seem likely to involve some difficulties, both in the execution and in the ob-
servations. But hollow prisms holding fluids might be employed, having the two faces
across which the light has to pass made of quartz plates. For a reason already men-
tioned, sulphuret of carbon cannot be employed for filling the prisms, and the disper-
sive power of water is very low, but there appears to be no objection to the use of a
solution of some colourless metallic salt. At least saturated solutions of sulphate of
zinc and of acetate of lead, the only salts I have tried with this view, showed no
defects of transparency when examined in quartz vessels by means of the flame of a
spirit4amp and a solution of sulphate of quinine*.

Effect of Hydrochloric Add^ 8fc. on Solutions of Quinine. Optical evidences ofcombi-'

nation in other instances.

205. Sir John Herschel, in his interesting paper already so often referred to, ob-
serves that it is only acid solutions of quinine which exhibit the peculiar blue colour,
and that among different acids the muriatic seems least efficacious (page 145).

For my own part I have tried solutions of quinine (not disulphate) in dilute sul-
phuric, phosphoric, nitric^ acetic, citric, tartaric, oxalic, and hydrocyanic acids, and
also in a solution of alum. In all these cases the blue colour of the dispersed light
was plainly seen by ordinary daylight, especially when the fluid was examined by
superficial projection. It was not easy to say which solution answered best, but I
am inclined to think that in which phosphoric acid was used.

206. But when quinine was dissolved in dilute hydrochloric acid the blue colour
was not exhibited, not even when the fluid was held in the sunlight, and examined
by superficial projection. Certain theoretical views led me to regard this as an evi-

* See note H.

542 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT.

dence of a more intimate union between quinine and hydrochloric acid than between
quinine and the acids first mentioned, and to try whether the addition of hydrochloric
acid to the solutions mentioned in the preceding paragraph would not destroy the
blue colour. On trial this proved to be actually the case, so that even sulphuric acid
is incapable of developing the blue colour in a solution of quinine in hydrochloric
acid.

207. That the quinine was not decomposed when the blue colour due to sulphate
of quinine was destroyed by hydrochloric acid, but only differently combined, was
shown by adding a solution of carbonate of soda, which produced a white precipitate ;
and when this was collected on a filter, washed, and redissolved in dilute sulphuric
acid, it exhibited the blue colour as usual.

208. The addition of a solution of common salt, instead of hydrochloric acid, to
the solutions mentioned in Art. 205, likewise destroyed the blue colour. In the case
of sulphuric acid this is only what might have been confidently anticipated; but we
should not perhaps have expected that quinine in combination with a weak acid,
such as citric, would decompose hydrochlorate of soda, giving rise to citrate of soda
and hydrochlorate of quinine ; yet this appears to be the nature of the reaction.

209. It might perhaps be supposed that the sulphuric acid was only partially ex-
pelled from sulphate of quinine by hydrochloric acid, and that the salt in solution
was really a sort of double salt, in which the same base, quinine, was combined with
sulphuric and hydrochloric acids in atomic proportion. But if so, it is probable, though
not certain, that the same salt would be formed on adding hydrochloric acid to a
solution of disulphate of quinine, even though the quantity were not sufficient to
combine with the whole of the disulphate. On this supposition, if hydrochloric acid
were added by small quantities at a time to a solution of disulphate of quinine, the
blue colour ought not to be developed ; and when acid enough had been added it
ought to be incapable of being developed by the addition of sulphuric acid ; whereas,
if the whole of the sulphuric acid be expelled by hydrochloric acid, the blue colour
ought to be first developed, by the conversion of a portion of the disulphate of quinine
into a sulphate, and then destroyed, on the addition of more acid, by the conversion of
the sulphate into a hydrochlorate. On trying the experiment with a solution of
disulphate of quinine in warm water, it was found that the blue colour was actually
first developed and then destroyed.

210. A practical conclusion which seems to follow from these results is, that in
the employment of quinine in medicine it is of little consequence whether the sulphate,
phosphate, acetate, or hydrochlorate be used, since the first three salts would be
immediately converted by the common salt in the body into the hydrochlorate,
and the small quantity of a neutral salt of soda resulting from the double decompo-
sition could hardly, one would supppose, be worth considering. However, the com-
mon quinine is associated with cinchonine, the reactions of which may be different.
According to Sir John Herschel, the latter alkaloid does not exhibit the blue colour,
and therefore the optical tests do not apply to it. If it be desired to obtain a soluble

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 543

salt of quinine which shall not be converted by common salt, by double decompo-
sition, into a hydrochlorate, it must apparently be sought for among the combina-
tions of quinine with very weak acids, the affinity of which for soda does not much
help that of hydrochloric acid for quinine. It seems likely enough that such salts
may exist; for though acetate or citrate of quinine decomposes hydrochlorate of
soda, hydrochlorate of quinine is decomposed by carbonate of soda ; and it is probable
that many vegetable acids behave like the carbonic in this respect.

211. The blue dispersion of a solution of sulphate of quinine is destroyed by
hydrobromic and hydriodic acids just as by hydrochloric. In the experiment,
solutions of bromide and iodide of potassium were used ; but as a considerable ex-
cess of sulphuric acid was purposely added to the solution of quinine, the potassa
introduced would merely remain inert in the solution as a sulphate, without impeding
the observation. The same experiment was tried with phosphate of quinine with the
same result.

212. It is stated in Turner's Chemistry, that the play of colours observed in
solutions of polychrome (i. e. esculine) is destroyed by acids, and heightened by
alkalies. The destruction, or at least almost complete destruction, of the blue colour
due to dispersed light in a decoction of the bark of the horse-chestnut, which is
produced by acids, is readily observed ; but I could not perceive that the addition
of alkalies in the first instance to a fresh solution made any difference one way or
other. If the blue colour had previously been destroyed by an acid, it was restored
by the alkali. If the horse-chestnut had never been examined chemically, these
observations alone would indicate that in all probability the principle to which the
blue colour was due was capable of entering into firm combination with acids, but
did not combine with alkalies. It is, in fact, as we know, a vegetable base.

213. A solution of nitrate of uranium in ether is insensible, as if some of the
elements of the ether entered into firm combination with the oxide of uranium. In
connexion with this circumstance, it is rather remarkable, that although the ether
passes off by evaporation when the solution is left to itself in an open vessel, if heat
be applied chemical action sets in, and the residue consists chiefly of a salt which
has all the appearance of oxalate of uranium. This salt, when washed and ex-
amined in the moist state, without very great concentration of light, was found to
be insensible^.

214. It is rare to meet with solutions so highly sensitive as those of quinine and
esculine, but similar observations may be made on a great number of solutions, by
employing suitable methods. The most searching method consists in forming a
bright and tolerably pure spectrum, by transmitting the sun's light through a very
broad slit, or even leaving out the slit altogether. It is desirable to use a lens of
only moderate focal length in connexion with the prisms. The solution having been
placed in the spectrum, the acid, or other agent whose reactions it is desired to study,

* See note L
MDCCCLII. 4 A

544 PBOFISSOE STOKES ON THE CHANGE OP REFRANGIBILITY OP LIGHT.

is to be addedj and the effect, if any, observed. It is usually advantageous to cover
the slit vrith a blue glass, or similar absorbiog medium; but sometimes effects take
place in the bright part of the spectrum, which is intercepted by such a medium.
When false dispersion abounds, it is well to look down on the fluid through a
Nicol's prism, so as to stop all light which is polarized in the plane of reflexion.

Wegative results with reference to a mutual action of the rays incident on

sensitive solutions,

215. The antagonistic effects of the more and less refrangible rays, which have
been observed in certain phenomena, induced me to try whether anything of the
kind could be perceived in the case of internal dispersion. The following arrange-
ment was adopted for putting this question to the test of experiment.

A tumbler was filled with a very dilute solution of sulphate of quinine, and placed
in a pure spectrum. As usual, the illuminated portion of the fluid consisted of two
distinct parts, one the blue beam of truly dispersed light, corresponding to the
highly refrangible rays, the other the beam reflected from motes, exhibiting the usual
prismatic colours, and corresponding to the brighter of the visible rays. The fluid
was nearly free from motes, so that the first beam was by far the brighter of the two ;
and the second beam, without being bright enough at all to interfere with the ob-
servation, was useful as serving to point out where the red, yellow, &c. rays lay. A
flat prism, having an angle of about 130^, was then held in front of the vessel, with
its edge vertical, and situated in the more refrangible part of the visible rays. The
rays forming the two beams were thus bent in opposite directions, and the beams
made to cross each other within the fluid ; and by turning the prism a little in both
directions in azimuth, that is, round an axis parallel to the incident rays, it was easy
to make sure that the beams did actually cross. But not the slightest perceptible
difference in the blue beam was made by the passage of the red and other lowly
refrangible rays across it.

216. Certain theoretical views having led me to regard it as doubtful whether the
intensity of light internally dispersed was proportional to the intensity of the incident
rays, other circumstances being the same, I was induced to try the following experi-
ment.

The sun's light was reflected horizontally through a large lens, which was covered
by a screen containing two moderately large round holes, situated in the same hori-
zontal plane, and a good distance apart. The beams coming through the two holes
converged of course towards the focus of the lens, and at the same time contracted in
width, and became brighter from the concentration of the light. For our present
purpose, they may be regarded as cylindrical beams converging towards the focus of
the lens. When they had approached each other sufficiently, they were transmitted
through a blue aramoniacal solution of copper, contained in a vessel with parallel
sides. The object of this was of course to absorb all the bright visible rays, which

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. ■545^

would not only be useless for exciting the solution which it was meant to try, but
would materially hinder the observation by the glare which they would produce.
The beams were then admitted into a vessel containing a decoction of the bark of
the horse-chestnut, greatly diluted with water. In passing through the fluid they
produced two blue beams of truly dispersed light, which converged towards a point a
little way outside the vessel. A flat prism, with an angle of about 150% was then
held in front of the vessel, with its edge vertical, and situated between the incident
beams^ The blue beams of dispersed light were thus made to cross within the fluid ;
and by moving the prism in azimuth, it was easy to make one beam either fall above
the other, cross it, or fall below it. Now on looking down from above with one eye
only, and moving the prism backwards and forwards in azimuth, I could not perceive
the slightest difference of illumination, according as the blue beams actually crossed
each other, or were merely seen projected one on the other. In this experiment, then,
it appeared that one beam of incident rays produced as much additional dispersed
light in a portion of fluid already excited by the other beam, as it was capable of
producing in a similar portion of fluid not otherwise excited.

Effect of an electric spark. Nature of its phosphorogenic rays.

217. For the use of the apparatus with which the following experiments were
made, I am indebted to the kindness of Professor Cumming.

An electric spark produces an internal dispersion of light in a very striking manner
in the case of an extremely dilute solution of sulphate of quinine. Having prepared
a solution so weak, that when it was examined by superflcial projection by the light
of a spirit-lamp, nothing was seen but a pale gleam of light extending a good way
into the fluid, and not only not confined to the surface, but not even showing any
particular concentration in the neighbourhood of the surface, I placed it so as to be
illuminated by the sparks from the prime conductor of an electrifying machine, which
passed at no great distance over the surface. A very marked internal dispersion was
produced, but the natore of the effect depended in a good measure on the character
of the spark. A feeble branched spark, giving but little light, and making little
noise, produced an illumination extending to a considerable depth, and very much
stronger than that occasioned in the same solution by the flame of a spirit-lamp.
The rays by which this was produced passed in a great measure through a plate of
glass interposed between the spark and the surface of the fluid. But a bright linear
spark, making a sharp crack, produced an illumination almost confined to an ex-
cessively thin stratum adjacent to the surface of the fluid ; and the rays by which this
was produced were cut off^ by glass, though transmitted through quartz. The same
was the case with the discharge from a Leyden jar, which produced a bright light
almost confined to the surface*.

218. The opacity of a solution of sulphate of quinine appears to increase regularly

* See note J.
4a2

646 PROFESSOR STOKES ON THE CHANGE OP REFRANGIBILITY OF LIGHT.

and rapidly with the refrangibility of the rays incident upon it. Hence we are led
to the conclusion that a strong electric spark is excessively rich in invisible rays of
extremely high refrangibility. Glass is opake with respect to these rays, but quartz
transparent.

219. It is known that the phosphorogenic rays of an electric spark, at least those
which affect Canton's phosphorus, pass very freely through quartz, but are stopped
by a very moderate thickness of glass. This alone, after what has been already
mentioned, would lead us to suppose that the phosphorogenic rays coming from
such a spark are merely rays of very high refrangibility. If so, they ought to be
intercepted by a very small quantity of a substance known to absorb such rays with
energy.

After having made some experiments on the production of phosphorescence in
Canton's phosphorus by means of an electric discharge, and observed how the in-
fluence of the discharge was transmitted through quartz and stopped, or almost
entirely stopped, by glass, I felt confident that my own observations were comparable
with those of others. A small portion of the phosphorus was then placed on card,
covered by an empty quartz vessel, and had the discharge of a Leyden jar passed
over it. The phosphorescence was powerfully excited, being visible in a room which
was by no means quite dark ; and when the card was carried into a dark place, the
phosphorescent light remained plainly visible for a good while. The experiment was
then repeated with a fresh portion of the same phosphorus, the vessel this time con-
taining water. The phosphorescence was produced as before, though not I think so
copiously. But on taking a fresh portion of the phosphorus, and substituting for
water a very dilute solution of sulphate of quinine, the influence of the spark was
arrested, and the phosphorus was not rendered luminous. It was found that a solu-
tion containing only about one part of quinine in 10,000, with a depth of half an inch,
was sufficient to prevent the generation of phosphorescence.

220. This result, it seems to me, would be sufficient, were proof wanting, to show
that no part of the efffect is attributable directly to the electrical disturbance. The
efffect produced when the phosphorus is at the distance of an inch or so from the
points of the discharger seems exactly the same as when it is nearer, being merely
somewhat weaker, as would naturally be expected, whatever view were taken of the
nature of the influence. But at the distance of an inch, the influence of the spark,
though it passes freely through quartz and water, is cut off by adding to the water
an excessively small quantity of sulphate of quinine. It cannot be supposed that the
electrical relations of the medium, or its permeability to electrical attractions and
repulsions, are utterly changed by such an addition ; while, on the other hand, the
result is in perfect conformity with what we know respecting the stoppage of radia-
tions by absorbing media. However, the principal object of the experiment was not
to confirm the view which makes the influence of the spark to consist in the rays
which emanate from it, a view which I suppose is pretty generally adopted, but to

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 547

investigate more fully the nature of these rays. Enough has, I think, been adduced
to show that they are merely rays which there is no reason to suppose are physically
different from those of light, but quite the contrary, and which are of very high
refrangibility, and are therefore invisible, since they fall far beyond the limits of
refrangibility within which the retina is affected. Indeed, it seems very likely that
the highly refrangible rays never reach the retina, but are absorbed by the coats of
the eye ^. Hence the phenomena relating to the phosphorescence produced by an
electric discharge afford no countenance to the supposition that it is possible to
divide rays of a given refrangibility into phosphorogenic, chemical, luminous, &c.
Of course the most unexceptionable mode of determining the refrangibility of the
phosphorogenic rays would be by actual prismatic decomposition, but this would
require the employment of a quartz train.

Points of resemblance and contrast between internal dispersion and phosphorescence,

221. As the term phosphorescence has been applied to several different phenomena,
I must here explain that I mean the spontaneous exhibition of a soft light, inde-
pendently of chemical changes, which some substances exhibit for a time after
having been exposed to the sun's rays, or to an electric discharge, or to light from
some other sources.

In many respects the two phenomena have a strong resemblance. Thus, the
general features of internal dispersion cannot be better conceived than by regarding
the sensitive medium as self-luminous while under the excitement of the active rays.
Again, it is well known that the rays of the solar spectrum by which the phosphores-
cence of Canton's phosphorus, sulphuret of barium, and other phosphori, is pro-
duced, are those of high refrangibility, as well as the invisible rays beyond ; and
these are precisely the rays which in the great majority of cases are most efficient in
producing internal dispersion. I do not however know how far it may be true that
when phosphorescence is excited by homogeneous light the refrangibility of the inci-
dent light is a superior limit to the refrangibilities of the component parts of the light
emitted. Indeed, according to Professor Draper, when the phosphorescence of Can-
ton's phosphorus is excited by the rays from incandescent lime, the active rays belong
to the red extremity of the spectrum-}-. If this result be confirmed, it follows that
the most striking law relating to internal dispersion is not obeyed in the case of
phosphorescence.

In the same paper Professor Draper remarks, " Some time ago I determined the
refrangibility of the rays of an electric spark which excite phosphorescence in sul-
phuret of lime ; they are found at the violet extremity of the spectrum." In what
way Professor Draper determined the refrangibility of rays with respect to which
glass is so opake, he does not give the least hint. Being perfectly in the dark as to
the evidence on which the conclusion is based, I cannot accept it in contradiction to
* See note K. t Philosophical Magazine, vol. xxvii. (Dec. 1845) p. 436.

548 FEOFESSOE STOKES ON THE CHANGE OF EEFEANGIBILITY OF FJGHT.

my own experiments. Perhaps, however^ ^^ at the violet extremity" may mean no-
thing more than somewhere in the highly refracted region beyond the visible rays.
If SO5 Professor Drapre's statement is in accordance with my own conclusions.

222. When one part of a phosphorus has been excited^ the phosphorescence is
found gradually to extend itself to the neighbouring parts. In this respect a sub-
stance which exhibits internal dispersion presents a striking contrast. The finest
fixed lines of the spectrum are seen sharply defined, whether in a solution^ or in a
clear solid, or on a washed paper,

223. Of course, theoretically, there ought, to a certain extent, to be a communication
of illumination from one part of a sensitive fluid to another, on account of the light
which is twice, three times, &c. dispersed. This however must be excessively small ;
for the mean refrangibility of the dispersed light is usually much lower than the
refrangibility of the active light, perhaps lower than that of any light capable of ex-
citing the solution. However, generally some few of the dispersed rays would have a
refrangibility sufficiently high to be dispersed again. But practically the intensity
of the light twice dispersed in this manner would be so very small that it may safely
be altogether disregarded.

224. But by far the most striking point of contrast between the two phenomenaj
consists in the apparently instantaneous commencement and cessation of the illumi-
nation, in the case of internal dispersion, when the active light is admitted and cut
off. There is nothing to create the least suspicion of any appreciable duration in the
effect. When internal dispersion is exhibited by means of an electric spark, it
appears no less momentary than the illumination of a landscape by a flash of light-
ning. I have not attempted to determine whether any appreciable duration could be
made out by means of a revolving mirror.

225. There appears to be no relation between the substances which exhibit a change
of refrangibility and those which phosphoresce, either spontaneously, or on the appli-
cation of heat. Thus the sulphurets of calcium and barium, on being examined for
internal dispersion, were found to be insensible, as was also Iceland spar. The last
substance phosphoresced strongly on the application of heat. So far as was examined,
the minerals which did exhibit a change of refrangibility showed no special dispo-
sition to phosphoresce. Sir David Brewster has remarked, that a specimen of
fluor-spar which exhibited a blue light by internal dispersion, exhibited when
heated a blue phosphorescent light; but this appears to have been merely a casual
coincidence*.

On the Cause of True Internal Dispersion^ and qf Absorption,

226. In considering the cause of internal dispersion, we may I think at once
discard all supposition of, reflexions and refractions of the vibrations of the lumini-
ferous ether among the ultimate molecules of bodies. It seems to be quite contrary

* Report of the Meeting of the British Association at Newcastle in 1889, p. 11.

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 549

to dynamical principles to suppose that any such causes should be adequate to
account for the production of vibrations of one period from vibrations of another.

All believers^ I suppose, in the undulatory theory of light are agreed in regard-
ing the production of light in the first instance as due to vibratory movements
among the ultimate molecules of the self-luminous body. Now in the phenomenon
of internal dispersion, the sensitive body, so long as it is under the influence of the
active light, behaves as if it were self-luminous. Nothing then seems more natural
than to suppose that the incident vibrations of the luminiferous ether produce vibra-
tory movements among the ultimate molecules of sensitive substances, and that the
molecules in turn, swinging on their own account, produce vibrations in the lumini-
ferous ether, and thus cause the sensation of light. The periodic times of these
vibrations depend upon the periods in which the molecules are disposed to swing,
not upon the periodic time of the incident vibrations.

227. But in the very outset of this theory an objection will probably be urged, that
it is quite as much contrary to dynamical principles to suppose the periodic time of
the ethereal vibrations capable of being changed through the intervention of ponder-
able molecules as without any such machinery. The answer to this objection is, that
such a notion depends altogether on the applicability of a certain dynamical principle
relating to indefinitely small motions, and that we have no right to regard the mole-
cular vibrations as indefinitely small. The excursions of the atoms may be, and
doubtless are, excessively small compared with the length of a wave of light ; but
it by no means follows that they are excessively small compared with the linear
dimensions of a complex molecule. It is well known that chemical changes take
place under the influence of light, especially the more refrangible rays, which would
not otherwise happen. In such cases it is plain that the molecular disturbances
must not be regarded as indefinitely small. But vibrations may very well take place
which do not go to the length of complete disruption, and yet which ought by no
means to be regarded as indefinitely small. Furthermore, it is to be observed that if
in the cases of indefinitely small molecular displacements the forces of restitution be
not proportional to the displacements, the principle above alluded to will not be ap-
plicable however small the disturbance may be ; and if in the expressions for the
forces of restitution the terms depending on first powers of the displacements (sup-
posed finite), though not absolutely null, be very small, the principle will not apply
unless the molecular excursions be extremely small indeed. In consequence of the
necessity of introducing forces not proportional to the displacements, it would be
very difficult to calculate the motion, even were we acquainted with all the circum-
stances of the case, whereas we are quite in the dark respecting the actual data of
the problem. But certainly we cannot affirm that in the disturbance communicated
back again to the luminiferous ether none but periodic vibrations would be produced,
having the same period as the incident vibrations. Rather, it seems evident that a
sort of irregular motion must be produced in the molecules, periodic only in the

550 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

wsense that the molecules retain the same mean state ; and that the disturbance which
the molecules in turn communicate to the ether must be such as cannot be expressed
by circular functions of a given period, namely, that of the incident vibrations.

228. It is very remarkable with what pertinacity a particular mode of internal
dispersion attaches itself to a particular chemical substance. Thus the singular
dispersion of a red light exhibited by the green colouring matter of leaves is found
in a green leaf, or in a solution of the green colouring matter in alcohol, ether, sul-
phuret of carbon, or muriatic acid. The dispersion exhibited by nitrate of uranium
is found in a solution of the salt in water, as well as in the crystals themselves, which
are doubly refracting. In all probability therefore the molecular vibrations by which
the dispersed light is produced are not vibrations in which the molecules move
among one another, but vibrations among the constituent parts of the molecules
themselves, performed by virtue of the internal forces which hold the parts of the
molecules together. It is worthy of remark that it is chiefly among organic com-
pounds, the ultimate molecules of which we are taught by chemistry to regard as
having a complicated structure, that internal dispersion is found. It is true that
peroxide of uranium furnishes many examples of internal dispersion ; but then the
anhydrous peroxide is itself insensible, it is only some of the compounds into which
it enters that are so remarkably sensitive ; and the chemical formulae of these com-
pounds, so far as they are known, are not by any means extremely simple, although
it is true that they may not be more complicated than formulae relating to other
oxides. Why this particular oxide should be disposed to enter into tottering com-
binations I do not pretend even to conjecture ; but it seems not a little remarkable
that peroxide of uranium, which is so peculiar with respect to its optical properties,
should also present some singularities in its mode of chemical combination, which
led M. Peligot to regard it as the protoxide of a compound radical.

229. We are, I conceive, at present far from an explanation of the phenomena of
internal dispersion in all their details. They appear to be associated with the
inmost structure of chemical molecules, to such a degree as to throw even the
phenomena of polarization into the shade. In this respect, indeed, absorption seems
superior to polarization, since most of the phenomena of polarization refer rather
to the state of crystalline aggregation of the molecules than to their constitution ;
but the phenomena of internal dispersion appear to be much more searching than
those of absorption. There is one law however relating to internal dispersion so
striking and so simple, that it seems not unreasonable to look for an explanation of
it; I allude to that according to which the refrangibility of light is always lowered
in the process of dispersion. I have not hitherto been able altogether to satisfy
myself respecting a dynamical explanation of this law, but the following conjectures
will not perhaps be deemed altogether unworthy of being mentioned.

230. Reasons have already been brought forward for regarding the molecular
vibrations as performed under the influence of forces not proportional to the dis-

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 561

placements. For simplicity's sake, let us suppose for the present the parts of the
forces of restitution depending upon first powers of the displacements to be abso-
lutely null. Then, when a molecule is disturbed, its atoms will be acted on by
forces depending upon the second and higher powers of the displacements. These
forces must tend to restore the atoms to their mean positions ; otherwise the equi-
librium would be unstable, and the atoms would enter into new combinations, either
with one another, or with the atoms of the surrounding medium ; so that, in fact,
such compounds could never be formed. The condition of stability would require the
parts of the forces depending upon squares of the displacements to vanish, but this
is a point which need not be attended to, all that is essential to bear in mind being,
that we have forces of restitution varying in a higher ratio than the displacements.
If the parts of the forces of restitution which depend upon first powers of the dis-
placements, though not absolutely null, be very small, the remaining parts must still
be such as to tend to restore the atoms to their positions of equilibrium ; otherwise
the stability of the molecule, though not mathematically null, would be so very
slight, that such compounds would probably never form themselves, but others of
more stability would be formed instead. Or, even were such unstable compounds
formed, they would probably be decomposed on attempting to excite them in the
manner in which sensitive substances are excited in observing the phenomena of
internal dispersion ; so that whether they exist or not, they may be set aside in con-
sidering these phenomena.

231. Now when vibrations are performed under the action of forces which vary
in a higher ratio than the displacements, the periodic times are not constant, but
depend upon the amplitudes of vibration, being greater or less according as the
amplitudes are less or greater. Suppose the molecular and ethereal vibrations
already going on, and imagine the amplitudes of the former kept constant by the
application of external forces. According to the value of the epoch of the vibrations
of a particular molecule, the ethereal vibrations will tend, in the mean of several
successive undulations, to augment or to check the vibrations of the molecule. For
some time there will be a tendency one way, then for some time a tendency the
other way, and so on, the opposite tendencies balancing each other in the long run.
The lengths of the times during which the tendency lies in one direction, will depend
upon the periodic times of the molecular and ethereal vibrations, being on the whole
greater or less according as the two periodic times are more or less nearly equal.
But since no external forces actually act to keep the amplitudes constant, when the
ethereal vibrations are favourable to disturbance the molecule is further disturbed,
and therefore its periodic time is diminished; and when they are favourable to qui-
escence the disturbance of the molecule is checked, and therefore its periodic time is
increased. If, then, the ether be vibrating more rapidly than the molecule, when the
action is favourable to disturbance the periodic time of the molecular vibrations is
rendered more nearly equal to that of the ethereal vibrations, and therefore the time

MDCCCLII. 4 B

552 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

during which the action is favourable to disturbance is prolonged ; but when the
action is favourable to quiescence, the effect is just the reverse. Hence, on the
whole, there is a balance outstanding in favour of disturbance. But if the ether be
vibrating more slowly than the molecule, it appears from similar reasoning that there
will be a balance the other way. Hence it is only when the periodic time of the ethereal
vibrations is less than that of the molecular, that the latter vibrations can be kept
going by the former.

232. But it will probably be objected to this explanation, that when a periodic
disturbing force affects the mean motion of a planet, the mean motion is a maximum,
not when the force tending to augment it is a maximum, but at a time later by a
quarter of the period of the force, namely, when the force vanishes in changing sign ;
and that in a similar manner the change in the periodic time of the vibrations of a
disturbed molecule will affect equally the duration of the time during which the
action is favourable to increased disturbance, and that during which it is favourable
to quiescence, or more exactly will not alter either, since the effects in the first and
second halves of those times will neutralize each other. The answer to this objection
is, that we must not treat a molecule as if it were isolated, like a heavenly body,
since it is continually losing its motion by communication, perhaps to neighbouring
molecules, but at any rate to the luminiferous ether; for without a communication of
the latter kind there would be no dispersed light. Hence we must consider the imme-
diate tendency of the disturbing forces rather than their tendency in the long run.

233. When a molecule itself vibrates in an irregularly periodical manner, the vibra-
tions which it imparts to the ether are of course of a similar character. The resolu-
tion of these into vibrations corresponding to different degrees of refrangibility,
involves some very delicate mathematical considerations, into which I do not propose
to enter. But without this it is evident that when the ether is agitated by the vibra-
tions of an immense number of molecules, in all possible states as regards amplitude,
and consequently periodic time of vibration, the disturbance of the ether must con-
sist of a mixture of periodic vibrations, having their periods comprised between the
greatest and least of those belonging to the molecular vibrations ; and corresponding
to these different periods there will be portions of light of different degrees of refran-
gibility found in the dispersed beam. These refrangibilities will range between two
limits, an inferior limit equal to the refrangibility corresponding to the periodic time
of indefinitely small vibrations, and a superior limit equal to the refrangibility of the
active light.

234. This theory seems to accord very well with the general character of dispersed
beams, as regards the prismatic composition of the light of which they consist.
When analysed by a prism, these beams are sometimes found to break off abruptly
at their more refrangible border, but I do not recollect ever to have met with an
instance in which a beam broke off abruptly at the opposite border, except when the
whole beam was almost homogeneous. This is just as it ought to be according to

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 553

the above theory, because the amplitude of vibration decreases indefinitely in ap-
proaching the less refrangible limit. In the case of a solution of chlorophyll, we
may suppose that the part of the molecular forces of restitution depending on first
powers of the displacements is considerable, on which supposition, the effect ought
to approach to what would take place were there no other part.^But were the forces
of restitution strictly proportional to the displacements, the vibrations would be
isochronous, and could only be excited by ethereal vibrations having almost exactly
the same period, but would be powerfully excited by such. Accordingly, in a solu-
tion of chlorophyll the dispersion comes on very suddenly ; a large part of it is pro-
duced by active light of nearly the same refrangibility as the dispersed light; and
the latter, by whatever active light produced, has nearly the same refrangibility that
it had at first. This supposition, combined with the preceding theory, accounts also
for the transparency of the fluid with respect to rays of less reirangibility than the
first absorption band, for the great intensity of that band, for tfie rapidity with which
opacity comes on at its less refrangible border, and the comparatively slow resump-
tion of transparency on the other side. A difference of the same nature on opposite
sides of a maximum of opacity seems to be a very common phenomenon in absorption.
On the other hand, in those numerous cases in which the dispersion comes on gra-
dually, in the manner described in Art. 44, we may suppose the part of the forces of
restitution depending on first powers of the displacements to be but small.

235. It may appear at first sight to be a formidable objection to the theory here
brought forward, that in the experiment mentioned in Art. 216, the intensity of the
dispersed light did not appear to be more than doubled when the intensity of the
incident disturbance was doubled ; and that in the experiment described in Art, 215,
the rays of low refrangibility did not appear to exercise any protecting influence. But
the difficulty may, I think, be got over by a very reasonable supposition. It seems
very natural to suppose that a given molecule remains for the greater part of the
time at rest, or nearly so, and only now and then gets involved in vibrations. On
this supposition, it is only a very small per-centage of the molecules that at a given
instant are vibrating to an extent worth considering. Conceive now a stream of light
consisting of the highly refrangible rays to be incident on a sensitive medium, and to
cause 1 per cent, of the sensitive molecules to vibrate considerably, the rest
vibrating so little that they may be regarded as at rest. Now imagine a second
stream, similar in all respects to the first, to influence the medium which is already
under the influence of the first stream. Of the 1 per cent, of the molecules already
vibrating, many are vibrating, we may suppose, nearly with their maximum ampli-
tude, and consequently are not much affected. Besides, it is a great chance if the
epoch of the ethereal vibrations belonging to the second stream is such as to pro-
duce any great tendency either towards quiescence or towards disturbance in a mole-
cule just for the short time that it is vibrating strongly under the influence of the
first stream. But of the 99 per cent, of quiescent molecules 1 per cent, are made to

4b 2

554 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP MGHT.

vibrate. Hence the effect of the two streams together is very nearly the same in
kind as that of one alone, but double in intensity.

236. The apparent absence of a protecting influence in the less refrangible rays
seems at first more diflficult to account for, but perhaps the following reasoning may
be thought satisfactory. We ought not to attribute more influence in the direction
of protection to a second beam of rays of low refrangibility, than in the contrary
direction to a second beam of rays of high refrangibility. Now if the eff'ect of a
beam of rays of high refrangibility be to throw 1 per cent, of the molecules into a
state of vibration, it would be a commensurate eflFect in a beam of rays of low refran-
gibility to stop the vibrations of I per cent, of the molecules, if they were all vibrating.
But since only 1 per cent, are actually vibrating, the real protecting efffect amounts
to no more than stopping the vibrations of one molecule in every 10,000, an effect
which may be regarded as insensible.

237. The simple consideration that work cannot be done without the expenditure
of power, shows that when light incident on a medium gives rise to dispersed light,
a portion at least of the absorption which the medium is observed to exercise must
be due to the production of the dispersed light. If the dispersed light really arises
from molecular disturbances, and for my own part I think it almost beyond a ques-
tion that it does, it follows that in these cases light is absorbed in consequence of its
being used up in producing molecular disturbances. But since we must not need-
lessly multiply the causes of natural phenomena, we are led to attribute the absorp-
tion of light in all cases to the production or augmentation of molecular disturb-
ances, unless reason be shown to the contrary. It might seem at first sight that the
production or non-production of dispersed light establishes at once a broad distinc-
tion between different kinds of absorption. I do not think that much stress can be
laid on this distinction. In the first place it may be remarked, that we have no
reason to suppose that vibrations which are of the same nature as those of light are
confined to the range of refrangibility that the human eye can take in. If, there-
fore, no dispersed light be perceived, it does not follow that no invisible rays are
dispersed. If the incident light belong to the visible part of the spectrum, the
dispersed rays (if any), being of lower refrangibility than the incident light, can only
be invisible by having a refrangibility less than that of red light, and would manifest
themselves solely or mainly by their heating effect. However, though invisible rays
of this nature are in all probability emitted by the body in consequence of the
absorption of visible light, we are not bound to suppose that in their mode of emission
they precisely resemble the visible rays observed in the phenomena of internal disper-
sion. In most cases, perhaps, they are more nearly analogous to the visible rays
emitted by solar phosphori. It is possible to conceive, and it seems probable that
there exist, various degrees of molecular connexion from mere casual juxtaposition
to the closest chemical union. A compound molecule may vibrate as a whole, by
virtue of its connexion with adjacent molecules, or it may vibrate by itself^ in the

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 555

manner of an isolated vibrating plate or rod, and between these extreme limits we
may conceive various intermediate modes of vibration. Hence, without departing
from the general supposition that the absorption of light is due to the production of
molecular disturbances, we may conceive that the modes in which the ether commu-
nicates its vibrations to the molecules, and the molecules in turn communicate their
disturbances to the ether, are very various.

I do not bring forward the idea that the absorption of light is due to the produc-
tion of molecular disturbances as new, though possibly the communication of the
ethereal vibrations to the molecules may hitherto have been supposed necessarily to
imply the existence of synchronous vibrations among the molecules. The change in
the periodic time of vibrations which takes place in the process of internal dispersion
would hardly have been suspected, had it not been for the singular phenomenon
which pointed it out,

238. The only theory of absorption, so far as I am aware, in which an attempt is
made to deduce its laws from a physical cause is that of the Baron Von Wrede, who
attributes absorption to interference^. The Baron's paper is in many respects very
beautiful, but it has always appeared to me to be a fatal objection to his theory that
it supposes vibrations to be annihilated. It is true that two streams of light may
interfere and produce darkness, but then to make up for it more light is produced in
other quarters. Light is not lost by interference, but only the illumination differently
distributed. Were the disappearance of light in the direction of a pencil admitted
into a medium merely a phenomenon of interference, the full quantity of light admitted
ought to be forthcoming in side directions. Were a series of vibrations incident on
a medium, without producing any progressive change in its state, or any disturbance
issuing from it, it would follow that work was continually being annihilated. But
we have reason to think that the annihilation of work is no less a physical impossi-
bility than its creation, that is, than perpetual motion.

List of highly sensitive substances,

239. For the sake of any one who may wish to make experiments in this subject,
I subjoin a list of the more remarkable of the substances which have fallen under
my notice. It will be seen that most of these substances were suggested by the
papers of Sir David Brewster and Sir John Herschel.

Glass coloured by peroxide of uranium : yellow uranite : nitrate or acetate of the
peroxide. Probably various other salts of the peroxide would do as well. The
absorption bands of the salts, whether sensitive or not, of peroxide of uranium ought
to be studied in connexion with the change of refrangibility.

A solution of the green colouring matter of leaves in alcohol. To obtain a solu-
tion which will keep, it is well previously to steep the leaves in boiling water. The
alcohol should not be left permanently in contact with the leaves, unless it be wished

* Poggendorff's Annalen, B. xxxiii. S. 353 ; or Taylor's Scientific Memoirs, vol. i. p. 477.

556 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITf OF LIGHT,

to observe the changes which in that case take place^ but poured off when the strength
of the solution is thought sufficient. Also^ the solution when out of use must be kept
in the dark.

A weak solution of the bark of the horse-chestnut.

A weak solution of sulphate of quinine, i e. a solution of the common disulphate
in very weak sulphuric acid. Various other salts of quinine are nearly if not quite
as good.

Fluor-spar (a certain green variety).

Eed sea-weeds of various shades : a solution of the red colouring matter in cold
water. If a solution be desired^ a sea-weed must be used which has never been
dried. Sometimes even a fresh sea-weed will not answer well.

A solution of the seeds of the Datura stramonium in not too strong alcohol.

Various solutions obtained from archil and litmus (see Arts. 65 to 72).

A decoction of madder in a solution of alum.

Paper washed with a pretty strong solution of sulphate of quinine, or with a solution
of stramonium seeds, or with tincture of turmeric. The sensibility of the last paper
is increased by washing it with a solution of tartaric acid. This paper ought to be
kept in the dark.

A solution, not too strong, of guaiacum in alcohol.

Safflower-red, scarlet cloth, substances dyed red with madder, and various other
dyed articles in common use.

Many of the solutions here mentioned are mixtures of various compounds. Of

course if the sensitive substance can be obtained chemically pure it will be all the

better.

Conclusi(m.

240. The following are the principal results arrived at in the course of the
researches detailed in this paper : —

(1.) In the phenomenon of true internal dispersion the refrangibility of light is
changed, incident light of definite refrangibility giving rise to dispersed light of
various refrangibilities.

(2.) The refrangibility of the incident light is a superior limit to the refrangibility
of the component parts of the dispersed light.

(3.) The colour of light is in general changed by internal dispersion, the new
colour always corresponding to the new refrangibility. It is a matter of perfect
indifference whether the incident rays belong to the visible or invisible part of the
spectrum.

(4.) The nature and intensity of the light dispersed by a solution appear to be
strictly independent of the state of polarization of the incident rays. Moreover,
whether the incident rays be polarized or unpolarized, the dispersed light offers no
traces of polarization. It seems to emanate equally in all directions, as if the fluid
were self-luminous.

PEOFESSOE STOKES ON THE CHANGE OP EEFEANGIBILITY OF LIGHT. 557

(5,) The phenomenon of a change of refrangibility proves to be extremely common,
especially in the case of organic substances such as those ordinarily met with, in
which it is almost always manifested to a greater or less degree.

(6.) It affords peculiar facilities for the study of the invisible rays of the spectrum
more refrangible than the violet, and of the absorbing action of media with respect
to. them.

(7.) It furnishes a new chemical test, of a remarkably searching character, which
seems likely to prove of great value in the separation of organic compounds. The
test is specially remarkable for this, that it leads to the independent recognition of
one or more sensitive substances in a mixture of various compounds, and shows to a
great extent, before such substances have been isolated, in what menstrua they are
soluble, and with what agents they enter into combination. Unfortunately, these
observations for the most part require sunlight.

(8.) The phenomena of internal dispersion oppose fresh difficulties to the supposi-
tion of a difference of nature in luminous, chemical, and phosphorogenic rays, but are
perfectly conformable to the supposition that the production of light, of chemical
changes, and of phosphoric excitement, are merely different effects of the same cause.
The phosphorogenic rays of an electric spark, which, as is already known, are inter-
cepted by glass, appear to be nothing more than invisible rays of excessively high
refrangibility, which there is no reason for supposing to be of a different nature from
rays of light.

NOTES ADDED DUEING PUNTING.

Note A. Art. 23.

Shoetly after the preceding paper was forwarded to the Royal Society, I found M. Edmonb
Becqubbbl^s map of the fixed lines of the chemical spectrum, which is published in the 40th
volume of the ^ Bibliotheque Universelle de Geneve^ (July and August 1842). I had seen in
MoiGNO^s ^Repertoire d^Optique Moderne,^ that the map had been presented to the French Aca-
demy, and naturally felt anxious to obtain it ; but not finding any further notice of it either in that
work or in the ^ Comptes Rendus,^ I supposed that it had not yet been published. The principal
lines in this map I recognized at a glance, M. Becqubrbi,^s broad band I is my I; his group of
four lines M with the preceding band forms my group m ; his group of four lines N forms the first
four of my group n ; his line O is my n. It is only in the last group that there can be any doubt
as to the identification ; but I feel almost certain that M, Bbcqubrel's P is my o, and the next two
lines, the last in his map, are the two between o and j^. It is difficult at first to believe that the
strong line I? should have been left out, while the two faint lines between o and j? are represented,
but the difficulty is, I think, removed by considering the feeble photographic action in that part of

558 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

the spectrum. M. Becquerel expressly states that lines were seen beyond the last he has repre-
sented, though they were hardly distinct; and on comparing together his map, Mr. Kingsley's
photographs, and my own map, I think hardly any doubt can remain as to the identification.

I take this opportunity of referring to another very interesting paper of M. Becquerel^ s, entitled
^ Des eiFets produits sur les corps par les rayons solaires,^ which is published in the Annales de
Chimie, tom. ix. (1843) p. 257^ with which I was not acquainted till lately, or I should have referred
to it before. This paper contains, among other things, an investigation of the effects of trans-
parent and coloured screens on the luminous, chemical, and phosphorogenic rays, in which it is
shown, that, notwithstanding the great difference in the action of a given screen on the three classes
of rays, when we study the effect of the incident rays as a whole, its action is the very same when
we confine our attention to rays of any one refrangibility. Among the media employed by M. Bec-
querel, are some whose absorbing effect I have mentioned in the present paper, as having been
determined by methods depending upon the change of refrangibility. In such cases my own results,
as might have been anticipated, are in perfect harmony with those of M. Becquerel. With
respect to the results at which I have arrived regarding the nature of the phosphorogenic rays of an
electric spark, which are mentioned towards the end of the paper, I have been in a good measure
anticipated by M. Becquerel. Yet I do not think that even he was aware that so much of the
effect of the spark was due to rays of such high refrangibility.

Note B. Art. 105.

I have since succeeded, by a particular arrangement, in seeing so far into the " lavender ^^ rays as
to make out the groups of fixed lines m, ?^, p by means of light received directly into the eye, and
even to perceive light beyond that.

As to the colour of these rays when they are well isolated, I think the corolla of the lavender
gives as good an idea of it as could be expected from the circumstances. They seem to me to M^ant
the luminousness of the blue and the ruddiness of the violet. No doubt much error and uncertainty
has hitherto existed both as to the colour and as to the illuminating power of these rays, because
the gray prolongation of a spectrum formed on paper by projection has been mistaken for the
lavender rays.

Note C. Art. 154.

On adding common phosphoric acid to a solution of nitrate of uranium no effect seemed to be
produced, but on examining the vessel some days afterwards, a precipitate was found to have fallen.
This precipitate proved to be sensitive in a very high degree.

Note D. Art. 158.

I have since observed in a mineral solution a system of absorption bands so remarkable, and so
closely resembling in many respects those found in the salts of peroxide of uranium, though they
occur in a totally different part of the spectrum, that I think no apology is needed for mentioning
the circumstance. The medium referred to is a solution of permanganate of potassa, in fact, red
solution of mineral chameleon. In order to see the bands, it is essential to employ a dilute solution,

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 559

or else to view it in small thickness, since otherwise the whole of the region in which the bands
occur is absorbed. The bands are five in number, and are equidistant, or at least very nearly so.
The first is situated at about three-fifths of a band-interval above D ; the last coincides with F, or,
if anything, falls a little short of it. The second and third are the most intense of the set. I have
carefully examined the solution for change of refrangibility, and have not found the least trace.
Ferrate of potassa shows nothing remarkable.

By means of the bands just mentioned, the colour of permanganate of potassa may be instantly
and infallibly distinguished from that of certain other red solutions of manganese, the colour of which
some chemists have been disposed to attribute to permanganic acid (see a paper by Mr. Pears all
^ On red Solutions of Manganese,^ Journal of the Royal Institution, New Series, No. IV. p. 49).

Note E. Art. 17I.

If we suppose the angle of incidence exactly equal to 45°, assume | for the refractive index of
the fluid, and apply Fresnel^s formula to calculate the ratio of the intensity of light reflected
at the exterior surface of a bubble, and polarized in a plane perpendicular to the plane of inci-
dence, to that of light similarly reflected and polarized in that plane, we find 0*228 to 1, a ratio
which certainly differs much from one of equality. But in order to render the two intensities equal,
it is sufficient to increase the angle of incidence by only 3° 35'; and in fact, as a matter of con-
venience, the position of the observer was usually such that the deviation of the light was somewhat
greater than 90"^, and therefore the angle of incidence somewhat greater than 45"^.

Note F. Art. 191,

I have since received a slab of glass of the kind here recommended, which has been executed for
me by Mr. Darker of Lambeth, and which answers its purpose admirably, the medium being emi-
nently sensitive. Besides its general use as a screen, this slab, from its size and form, has enabled
me to trace further than I had hitherto done (Arts. 75, 76) the connexion between certain fluctua-
tions of transparency which the medium exhibits and corresponding fluctuations of sensibility.

Note G. Art. 192.
Paper washed with a mere infusion of the bark of the horse-chestnut is quickly discoloured; but
a piece washed with a solution which had been purified by chemical means remained white, and
proved exceedingly sensitive.

Note H. Art. 204.

I have since ordered a complete train of quartz, of which a considerable portion, comprising among
other things two very fine prisms, has been already executed for me by Mr. Darker. With these
I have seen the fixed hues to a distance beyond H more than double that of j?; so that the length
of the spectrum, reckoned from H, was more than double the length of the part previously known
from photographic impressions. The light was reflected by the metallic speculum of a Silbbr-
MANN^s heliostat, which I have received from M. Duboscq-Soleil. With the glass train the
group p was faint, but with the quartz train there was abundance of light to see not only the groups,
but the fixed lines as far as Hpl, or thereabouts. From the group n to about the middle of the new
region, the lines are less bold and striking than in the region of the groups H, /,m, ^, but the latter

MDCCCLII. 4 C

560 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT.

part of the new region contains many lines remarkable both for their strength and for their arrange-
ment. I hope to make a careful drawing of these lines as shown by the complete train with a
summer^s sun.

I have some reasons for believing that the photographic action of these highly refrangible rays is
feeble, perhaps almost absolutely null. In the second of the papers referred to in Note A. (p. 300),
M. Bbcquerel describes an experiment in which a prism of quartz was employed to form a spec-
trum ; and yet the impressed spectrum formed by rays which had traversed the quartz alone was
hardly longer than that formed by rays which, in addition to the quartz, had traversed a screen of
pure flint-glass a centimetre in thickness. It is possible, I am inclined to think probable, that glass
made with ^perfectly pure materials would be transparent like quartz, but all the specimens I have
examined were decidedly defective in transparency. Besides, M. BscauEREL, who maybe allowed
to be the best judge of his own experiments, considered the result just mentioned as a proof that the
impressed spectrum formed by rays which had traversed quartz only did not extend, except a very
trifling distance, beyond that formed by his train of glass ; and yet his map, formed by means of the
latter, does not take in the line 'p.

However, among the multitude of preparations capable of being acted on by light, it is probable
that there may be some which are acted on mainly by rays of unusually high refrangibility, and which,
on that very account, would not be suitable for the ordinary purposes of photography. With these
it is possible that the new region of the solar spectrum might be taken photographically.

Note I. Art, 213.

I have since examined the salt, or product, whatever it may be, in the dry state, and under more
favourable circumstances, and have found it sensitive, though not by any means in a high degree.
It exhibits also the absorption bands which seem to run through the salts of peroxide of uranium.

In connexion with the insensibility of a solution of nitrate of uranium in ether, it seems interest-
ing to mention a fact which I have since observed, namely, that the sensibility of a solution of
nitrate of uranium in water is destroyed by the addition of a little alcohol.

Note J. Art. 217.

On repeating this experiment on a subsequent occasion, I could not satisfactorily make out the
difference of character of a strong and of a weak spark from the prime conductor, perhaps because
the machine was in less vigorous action ; but the difference between the effects of a mere spark and
of the discharge from a Leyden jar was plainly evident. I would here warn the reader, that in order
to perform the experiment in such a manner as to obtain a striking and perfectly decisive result, it
is essential to employ an excessively weak solution. The reason of this is evident.

A severe thunder-storm which visited Cambridge on the evening of July 16^ 1852, afforded me a
good opportunity of observing the effect of lightning on a solution of quinine, and other sensitive
media. From the copiousness of the dispersed light, it was evident that the proportion of the active,
and therefore highly refrangible rays to the visible rays was very far greater in the radiation from
lightning than in daylight. A difference of character was observed between the effects of a weak
distant flash, and of a bright flash nearly overhead, similar to that which has been described with
reference to the effects of a spark from a machine, and of the discharge from a Leyden jar. In

PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 561

artificial discharges^ the stronger the spark the more the rays of excessively high refrangibility seem
to abound, in proportion to the whole radiation. Now a flash of lightning is a discharge incom-
parably stronger than that of a Leyden jar. It might have been expected, therefore, that the
radiation from lightning would be found to abound in invisible rays of excessively high refrangibility.
Yet I could not make out in a satisfactory manner the absorption of the rays by glass, even by
common window-glass. I do not wish to speak positively regarding the result of this observation,
for of course observations with lightning are more difficult than those made with a machine which is
under the control of the observer. Yet it did seem as if the spark from a Leyden jar was richer
than lightning in rays of so high a refrangibility as to be stopped by glass. If this be really true,
it must be attributed to one of two things, either the non-production of the rays in the first instance,
in the case of lightning, or their absorption by the air or clouds in their passage from the place of
the discharge. If they were not produced, that may be attributed to the rarity of the air at the
height of the discharge, that is, at the height of the thunder-cloud. No doubt the metallic points
of the discharger belonging to the electrical apparatus may have had an influence on the nature of
the spark; but I am inclined to think that this influence, so far as it went, would have acted in the
wrong direction, that is, would have tended to produce rays of lower, at the expense of those of
higher refrangibility.

NoteK. Art. 220.

My attention has recently been called to a paper by M. Bruckb (Poogendorff's Annalen, B. v.
(1845) S. 593), in which he describes some experiments which show that the different parts of the
eye, and especially the crystalline lens, are far from transparent with respect to the rays of high
refrangibility. The eyes employed were those of oxen and some other animals ; and the inquiry
was carried on by means of the effect which light that had passed through the part of the eye to be
examined produced on a film of tincture of guaiacum that had been dried in the dark. Of course
the phenorftena described in the present paper afford peculiar facilities for such an inquiry, and I
had frequently thought of entering upon it, but have not yet made any observations. Independently
of the facility of the observations, and the advantage of being able to examine readily light of each
degree of refrangibility in particular, the results obtained by means of sensitive media seem to be
more trustworthy on this account, that it would be possible to employ fresh eyes. The experiments
of M. Brijcke necessarily occupied a considerable time, and it may be doubted whether the eye,
especially after dissection, might not have changed in the interval, and whether the results so
obtained are applicable to the eye as it exists in the living animal.

562 PROFESSOR STOKES ON THE CHANGE OF REPRANGIBIUTY OF UGHT.

INDEX TO THE PRECEDING PAPER.

N.B. The figures refer to the articles, the letters to the notes.

Absorbing and reflecting power, determination of, with

respect to invisible rays, 194-196, 201-204.
Absorption, on the cause of, 237, 238 ; connexion of, with

internal dispersion, 59, 60, 63, 71, 76, 120, 126, 146,

148.
Air expelled from water, 172.
Appearance of highly sensitive media, 2"], 29, 164.
Archil, 65-71.
Canary glass, 73-'77> F.
Chemical applications of internal dispersion, 67-70, 205-

214.
Clearness of sensitive fluids accounted for, 86.
Coloured glasses, fundamental experiment with, 7 ; nature

of the blue reflexion in orange glasses, ^^ ; examined

for dispersion, 167, 168.
Colourless glasses, internal dispersion in, 78; defect of

transparency of, 22, 202, 217-219, A.
Colours of natural bodies, 1 74-1 7B.
Concentration and dilution, efl^ect of, 185-188.
Crystals, natural, 165, 166. (See Fluor-spar, Uranium.)
Datura stramonium, solution, 43; paper, 95; capsules,

117.

Dispersion, true and false, defined, 25 ; distinguished, 26,

29; cautions, 169-173; usual features of true, 44-46 ;

on the cause of true, 226-236; nature of false, 179;

instances, 181; applications, 180, 182.
Epipolic dispersion, 1 ; explanation of, 6.
Esculine, 31, 212.
Explanation of terms, 21-30, 101.
Eye, opacity of the, 220, K.
Fixed lines of the invisible rays, exhibited, 16 ; description

of the, 21-23, A, H; eff'ect of viewing through a prism

the fixed lines shown by means of sensitive media, 17,

89-93, 100.
Flames, effect of various, 197-200.
Fluor-spar, 32-36.
Groundsel, petals of the purple, 120.
Guaiacum, solution, 37-41 ; washed paper, 98 ; solution

of, used as a test object, 200.
Heat, effect of, on the sensibility of glass, &c., 184.
Horse-chestnut, 31, 212.
Illuminating power of the highly refrangible rays, 104,

105, B.

Internal dispersion, 2.
Lavender rays, 105, B.
Leaf-green, absorption, 47-^52; internal dispersion, 53-61,

97, 118, 121.
Lightning, effect of, J.
Linear spectrum, 107-109; results obtained with a, 113-

136, &c.
List of highly sensitive substances, 239.
Litmus, 72.

Mercurialis perennis, 62-64.
Methods of observation, general, 13, 17, 110, 112.
Mutual action of incident rays, negative results with re-
ference to the, 215, 216, 235, 236.
Permanganate of potassa, absorption, D.
Phosphorescence compared with internal dispersion, 220-

225.
Phosphorogenic rays of an electric spark, nature of the,

219-221, A.
Polarization, absence of, in dispersed light, 1, 15; of the

incident rays a matter of indifference, 20.
Polarized light, direction of the vibrations in, 183.
Quartz, transparency of, 202, 217-219; train, 204, H.
Quinine, absorption of the violet by a solution of sulphate

of, 11; internal dispersion in a solution of, 14-20;

strong afiinity of, for hydrochloric, hydrobromic and

hydriodic acids, 205-211.
Rays, course of, exhibited, 193.
Reflecting power (see absorbing power).
Refrangibility of dispersed fight, lower than that of the

incident, 80, 102, 230-236; nature of the, 81, S2;

illustrated by a surface, 84, 85.
Results, principal, 240.
Screen, on the choice of a, 189-192, F, G.
Sea-weeds, red, 121-126.
Strata of equal dispersion in crystals, 167.
Test objects, 110, 114, 200.
Triangle, experiment with a paper, 58.
Turmeric, tincture of, 42; paper, 87-91.
Uranium, salts of the peroxide, &c., 157-162, 213, C, I

(see also canary glass); delicate test of, 159; absorption

of light by salts of protoxide of, 160, 163.
Washed papers, 87-98.

Ika.TransM^ZZ'OM.Tlaie JW.

J^ia.l ^^Ljced- hhes o/^ the dotar speciruni ui ihie ejrt/'e/ns yzolet,ajtd zn i/ie mvLsi'Sle region beyond.

b: -k

X

-v

qraup I. group m.

Tt^. 2. drt. 8j..

Fig. 4. ..Art. 86.

.J2^ =_

r\

^A^

JS^

JX^

JX

rmr

^OVf>

grov^ p.

F7^.3.-^rt. 82.

M

^?^. 6. y^rLzop.

E.

6-- G. stokes del.

J.BeLsire sc-