236 

sky would be similar in composition to that of dilute solutions of 
copper, which acquire their light blue tint by a partial suppres- 
sion of the extreme red.* There is no doubt that the colour 
equations are dependent on the character of the light, as may 
easily be proved by taking an observation looking all the time 
through a layer of coloured liquid. It is not, however, the 
most brilliantly coloured solutions that cause the most distur- 
bance, for anything like a complete stoppage of all the rays which 
are capable of exciting one of the primary colour sensations 
would affect both the mixtures to be compared in nearly the 
same manner, putting the observer in fact very much into the 
positions of a colour-blind person. Those liquids will be most 
efficient which have a different action on parts of the spectrum 
allied in colour. For instance, an aqueous infusion of litmus 
has a strongly marked action on the yellow ray, stopping it with 
great energy, even in rather dilute solutions. It is easy to trace 
the effect of looking through this on most of the colour equations. 
Consider, for example, the fifth equation of July 20 (that from 
which the blue is absent) wherein red and green are matched 
against black, white, and yellow. The red and green will for 
the most part escape absorption, but the white and yellow will 
be shorn of a part of their yellow rays. The match supposed to 
have been adjusted without the litmus must evidently be spoiled ; 
the red and green mixture becoming strongly yellow in com- 
parison with the other. Inorder to restore equivalence the 
yellow must be considerably increased. On trial I found, 124 black 
+ 19 white + 49 yellow matclied 121 red + 71 green. 
It is only the impurity of the colours on the discs that prevents 
the effect being still more strongly marked, for with the pure 
colours of the spectrum the most violent alterations are possible. 
When a match is made between the simple yellow and that com- 
pounded of pure red and green, almost any coloured liquid acts 
unequally on the two parts and destroys the balance. The 
simple yellow, of course, retains its colour under any absorbing 
influence, and can only be changed in luminosity. Chloride of 
copper extinguishes the red component of the compound yellow, 
which accordingly becomes green. Litmus would leave the 
compound colour nearly unchanged, while it extinguishes the 
simple yellow. It is needless to multiply instances. 
Before leaving the compound yellow, of whose very existence 
many are incredulous, I will mention an easy way of obtaining 
it, which is the more desirable as the use of the pure spectral 
colours is not very convenient. In order to isolate the red and 
green rays of the spectrum by means of absorption, the first 
thing is to find a liquid -capable of removing the intermediate 
yellow and orange. With this object we may fall back on the 
alkaline solution of litmus, whose opacity to the yellow, and 
particularly to the orange rays is so marked. The next 
step is to remove the blue and bluish green, for which nothing is 
more convenient than the chromate of potash. A mixture of 
these two liquids in proper proportions, easily found by trial, 
isolates the green and extreme red rays with considerable per- 
fection, and exhibits in ahigh degree the phenomenon of Di- 
chromatism. According to the thickness traversed by the light 
the red or the green predominates, and there is no difficulty with a 
given thickness in arranging the strength of the solution so as to 
give a full compound yellow. It is worth notice in confirma- 
tion of the opinion expressed as to the character of the sky-blue, 
that when a cloud seen through the liquid appears a full yellow, 
or even orange, the former, if at all intense, acquires a decided 
green colour. A window backed by well-lighted clouds, when 
looked at across a room through the liquid and a prism, has a 
very splendid appearance, the red being isolated on one side, 
and the green on the other ; while the intermediate space, where 
the two overlap, shows the compound yellow in great perfection, 
Another liquid, in some respects preferable, which answers the 
same purpose, may be made by mixing chloride of chromium and 
bichromate of potash. Through either of them the sodium flame 
is invisible, though they may easily be made to correspond with 
it in colour very closely. I tried to obtain a liquid capable of 
isolating the pure yellow ray, but only with partial success. The 
* Direct observations, made since the above was written, show that there 
is no peculiar deficiency at the red end of the spectrum, but a general falling 
off as the refrangibility dimimshes from one end to the other. If lights from 
sky and cloud are of equal intensity at the line C inthe red the first will be 
somewhere about twice as bright as the other at Bin the green. This is 
for a well-developed blue light taken from the zenith ; but, even with a large 
allowance, enough difference remains to account for the discrepancies in the 
two sets of colour disc observations. I have lately found from theory that 
the power of very small particles to scatter the rays belonging to different 
parts of the spectrum varies as the inverse fourth power of the wave 
length 
NATURE 


{ Fan. 19, 1871 

best was a mixture of bichromate and permanganate of potash 
with a salt of copper (sulphate or chloride). The first removes 
the blue and violet, the second the green, and the third the red, 
and thus the yellow is isolated in considerable purity. This 
liquid is very unstable. The comparison of the simple and com- 
pound yellow (which nearly matched) was interesting. One was 
transparent to the sodium flame, the other completely opaque to 
it. When the two are brought together so that the light has to 
traverse both, almost complete darkness results, even when the 
brightest clouds are used. I should mention that it is only when 
the light is strong that any of these liquids give yellow in full 
perfection ; otherwise the colour is more nearly described as 
brown, which is, in fact, identical with a dark yellow or orange. 
The best natural yellows, such as chrome, are partly simple and 
partly compound, returning all the light which falls upon them 
except the blue and violet. It is clear that neither a purely 
simple nor a purely compound yellow can rival them in brilliancy, 
Impartial observers, unprejudiced by the results of mixing pig- 
ments, or, on the other hand, by experiments on the spectrum, 
see, so far as I can make out, no connection between the four 
principal colours—red, yellow, green, and blue. It seems to 
them quite as absurd that yellow should be compounded of red 
and green, as it most unquestionably is, as that green should be 
a compound of blue and yellow, though many have accepted the 
latter alternative on the authority of painters, and some have 
even worked themselves into the belief that it is only necessary 
to look at the colours in order to recognise the compound nature 
of green. My own prejudice would be on the other side, the 
result of experiments on the compound yellow, which is seen so 
easily to pass into green on the one side or red onthe other. The 
most impartial opinion that I can form is that there is no real 
resemblance between any of the four, and if this be so it is 
certainly a most remarkable, if not unaccountable, fact. The 
difficulty is not so much that we are unable to analyse the com- 
pound sensation, as to explain why our inability is limited to 
yellow (and white). For everyone, I imagine, sees in purple a 
resemblance to its components red and blue, and can trace the 
primary colours in a mixture of green and blue. Sir John 
Herschel even thinks that our inability to resolve yellow leaves 
it doubtful whether our vision is trichromic or tetrachromic, but 
this seems to me to be going much too far. Surely the fact that 
the most saturated yellow can be compounded of red and green, 
deprives it of any right to stand in the same rank with them as 
primary colours, however little resemblance it may bear to them 
and blue. Besides, if yellow is to be considered primary, why 
not also white, which is quite as distinct a sensation as any of 
the others? Undoubtedly there is much that is still obscure 
in the mutual relations of the colours—why, for instance, as 
mentioned by Sir John Herschel, a dark yellow or orange sug- 
gests its character so little as to be called by a new name (brown), 
while a dark blue is blue still. But difficulties such as these 
should make us all the more determined to build our theories of 
colour on the solid ground that normal vision is threefold, and 
that the three primary elements of colour correspond nearly 
with red, green, and blue.* J. W. Strutr 


SCIENTIFIC SERIALS 
THE Quarterly Fournal of Science for January commences a 
new régime under the sole editorship of Mr. W. Crookes. As will 
be seen from the following summary of its contents, all the papers, 
with only one exception, refer to some department of Physical 
Science to the exclusion of Natural History. The articles are as 
follows :—1. ‘‘ Double Spectra,” by W. M. Watts. A résumé 
of the facts known to the present time to modify the conclusions 
drawn from the earlier spectrum researches of Bunsen ana Kirch- 
hoff, from which it was concluded that the spectrum of each 
element was one and invariable. A plain and a coloured litho- 
graphic illustration show the spectrum of copper chloride when 
volatilised undecomposed, as contrasted with that of the metal ; the 
different spectra of barium, strontium, and calcium obtained at 
different temperatures ; the three spectra of hydrogen, Ho, Hf, 
and Hy, probably due to differences in temperature; the two 
spectra of aluminium ; the two of nitrogen ; and the four spectra 
which are all probably due to incandescent carbon vapour. 2. 
“«The Great Pyramid cf Egypt, from a modern scientific point 
of view,” by C. Piazzi Smyth, Part I. 3. ‘‘On the Theory of 
* This paper was read before Section A of the Liverpool Meeting of the 
British Association, ; 
