Memoirs presented to the Cambridge philosophical society on the occasion of the jubilee of Sir George Gabriel Stokes, bart., Hon. LL. D., Hon. SC. D., Lucasian professor.

ABSORPTION SPECTRA OF SOLUTIONS OF DIDYMIUM AND ERBIUM SALTS. 301
THE SOLUTIONS EXPERIMENTED ON.
These have been chiefly those of salts of didymium and erbium. Most coloured salts
have only very wide absorption bands which fade on either hand very gradually, so that it
is extremely difficult, or even impossible, to recognise small changes in them. On the
other hand, didymium and erbium salts have a great many absorption bands, of various
degrees of sharpness and of intensity, and distributed through a wide range of the
spectrum. No other salts seem so well adapted for my purpose. However, I made a
number of observations on uranous chloride, but found it so prone to chemical change
when in solution that I could not with certainty distinguish the effects of dilution, or of
elevation of temperature, from those due to chemical change. The absorption spectra of
salts of cobalt have already been investigated by Dr Russell, though not exactly from my
present point of view, and they are not as good for my purpose as the salts of the two
metals to which I now confine myself.
Both series of salts had been purified as far as possible, by my assistant Mr Purvis,
by a long series of fractional precipitations. The didymium was spectroscopically free from
lanthanum, but it had not been found possible to get it, or the erbium, so free from
yttrium*. No attempt was made to separate the neodymium from the praseodymium, and
there is no method at present known for separating the various metals of which ordinary
erbium is supposed to be a mixture. Indeed for my purpose there would be no
advantage in doing so; though for a quantitative estimation of the concentration of
absorbent material in the solutions it was important to get rid of an admixture of
unabsorbent salt. In order to obtain solutions of the salts of different acids in equivalent
concentration the metal was precipitated as oxalate, washed, dried, and ignited in air
until it was reduced to oxide. Weighed quantities of this oxide were dissolved in the
several acids, and, in the case of nitric and hydrochloric acids, the solutions evaporated and
excess of acid driven off. The residual salts were then dissolved in measured quantities of
water. The most concentrated solutions of didymium employed contained, respectively, of
the nitrate, 611'1 grams to the litre, and of the chloride the equivalent quantity, namely
462'9 grams of anhydrous chloridet. These each contain 1862 gram-molecules of the salt
* Lanthanum and yttrium cannot be recognised by of three sets in the green and citron of which the brightest
any absorption bands, but when induction sparks are taken begin at X5599 and X5380 respectively, and the third at
from solutions of their salts, each gives a very character- X5173. There is another weaker set in the orange beginning
istic channelled spectrum, by which it is easily recognised at X5865, and two sets in the indigo beginning at X4419 and
in a solution containing one per cent., or even less, of the X4370 respectively. My measures were not made with any
salt. The yttrium channellings are in the orange, the large dispersion and the last figure of the measured wavebrightest of those of lanthanum in the citron and green, length may not be quite correct, but near enough for recogand both fade towards the red. Thalén in his paper (1874) nition of the channellings which are easily seen with a
on the Spectra of Yttrium and Erbium, and of Didymium small spectroscope, especially the two first mentioned.
and Lanthanum, gives the wave-lengths of the sharp, t The (crystalline) didymium chloride in this solution
more refrangible edges of the yttrium channellings, one was dissolved in just about twice its weight of water; the
set beginning at X6131 and the other at X5970'5. He does equivalent solution of nitrate had still less water.
not give those due to lanthanum. These I find to consist