272 
Transactions of the Boyal Society of South Africa. 
No. 8 is broader than 6 or 7 and absorbs from \4690 to \4680. These 
last three Hnes resemble those of indium. 
The most curious thing about the ruby spectrum is that none of the 
lines can be definitely identified with those of any known colouring 
matter. The colour is due to an inorganic cause, as the ruby can be 
ignited without alteration. It is independent of the anisotropic structure 
also, since the natural variety occurring in nearly isotropic hexagonal 
plates shows the same spectrum. Hair-lines of this character occur only 
in erbium and " didymium " solutions, but none of the latter agree with 
the ruby lines. 
The cause of the dichroism of ruby has been discovered by the author. 
When the gem (best an artificial one cut with two parallel faces) is 
examined with nicol and spectroscope, there is seen, in addition to the 
hair-line spectrum above described, a broad band in the green, which 
appears and disappears as the gem is rotated. The pink-violet stage 
is the one of deep absorption and the yellowish stage that of little 
absorption. 
As the artificial ruby is said to be coloured with chromium, the author 
has examined the spectrum of most of the typical chromium compounds. 
The following have no sharp lines in their spectrum : KCr(S04)2 (either 
solution or solid), Cr2(S04)3 (solid scales), Cr in beads of either borax or of 
NaPOj, KgCrCys in solution. 
On the other hand, a diffuse narrow line at about X6960 + 10 was 
seen in the spectrum of a strong solution of potassium chrom-oxalate 
K3Cr(C02)6» which agrees in position though not in character, with the 
first ruby line. The other ruby lines are completely absent ; neverthe- 
less, the coloration of ruby may possibly be due to chromium, the 
vibrations of the element being modified by its imprisonment in an 
anisotropic medium. 
