708 
MR. W. CROOKES ON RADIANT MATTER SPECTROSCOPY. 
fusing the nitrates will effect the purpose, but I have not found the converse to hold 
good. The potassic sulphate method cannot separate the last traces of didymium 
from samarium, for the didymic double sulphate, not being quite insoluble, would wash 
out along with the first portions of the samaric salt. I have found no method better 
than fractionation with ammonia, and Professor Cleve tells me that is his experience. 
Towards the end of the operations, when the samaria is getting pure, it is useful to 
precipitate it as a double sulphate with potassic sulphate, and wash it well for some 
time, to remove any traces of earths of the yttria and other groups which might have 
been present and become accumulated with the samaria (123). 
134. During fractional precipitation with ammonia an experienced eye can judge 
roughly what is the preponderating earth present, by the appearance of the precipitate 
as it comes down. When much samaria is present, with but little didymia and 
lanthana, the precipitate forms immediately. When there is much didymia, and little 
samaria and lanthana, the precipitate forms almost as quickly as in the first case, but 
does not settle so rapidly. With much didymia, and a fair quantity of lanthana, the 
precipitate forms more slowly than before and settles sluggishly. When there is 
much lanthana and little didymia the precipitate takes a long time to settle, the liquid 
remaining opalescent for days. These peculiarities are due in great measure to the 
varying basicity of the elements, samarium being the least basic and lanthanum 
being the most basic, didymium occupying an intermediate position. 
In freeing samarium from the last portions of didymium the only test available to 
detect the presence of the latter metal is the absorption spectrum. The best plan is 
to provide a strong solution of the samaric nitrate in a flask, to act as a lens, and to 
concentrate the light of a gas-flame by its means on to the slit of a low dispersion 
spectroscope. Long after the light colour of the ignited oxide shows that the 
didymium is getting small in quantity, its absorption bands will be so strong as almost 
to obliterate the fainter samarium spectrum. 
Fig. 
o 
135. The fractionation should be persevered in till no didymium bands are seen m 
the absorption spectrum. After this point is reached I prefer to keep on fractionating 
for some time longer, if the material will hold out, so as to make assurance doubly 
