Royal Sociely. 451 



yet in no case are they so considerable as to prevent our recognizing 

 at once that the spectra are all due to didymium. It is also impor- 

 tant to notice that the amount requisite to give a most splendid 

 spectrum when the bead is crystalline, will scarcely show any trace of 

 bands when it is in a vitreous condition, dissolved in the borax. This 

 is analogous to what occurs in the case of solid and powdered crys- 

 tals of sulphate of didymium ; for the absorption-bands in the spec- 

 trum of the light transmitted by a thin layer of the fine powder, 

 strongly illuminated from the other side, are as distinct as in that 

 transmitted by a many times greater thickness of solid and transpa- 

 rent crystal. We may very conveniently take advantage of this fact in 

 studying the spectra of such substances when the amount of material 

 at command is otherwise too small. This seems to be because the 

 transmitted light does not simply pass through the crystals, but is in 

 great measure reflected from them backwards and forwards, and 

 thus, as it were, passes through a greater thickness. It is also to 

 a considerable extent similar to that reflected from the powder when 

 illuminated from above, as may be clearly proved by what occurs in 

 the case of uranic salts. These when in a state of moderately fine 

 powder transmit light, giving a spectrum showing not only the ab- 

 sorption-bands in the blue, which alone are met with in that trans- 

 mitted by a clear crystal, but also the bands in the green, which 

 depend on fluorescence, characteristic of that reflected from the 

 powder*. These two kinds of bands can be easily distinguished by 

 means of a plate of deep-blue cobalt glass, which has an entirely 

 different action accordingly as it is placed below or above the object 

 when the bands are due to fluorescence, but has no such effect when 

 they are due to ordinary absorption. It would perhaps be well to 

 mention here that I have in this manner proved that the abnormal 

 bands seen in the spectra of the compounds of zirconia with the 

 oxides of uranium, described in this paper, are due to genuine ab- 

 sorption, and not to fluorescence. 



The remarkable spectrum of some jargons has been already de- 

 scribed by me in the 'Chemical News 't, and in the Proceedings of 

 the Royal Society J. One of its most striking peculiarities is, that 

 when light passes in a direction perpendicular to the principal axis of 

 the crystal, and the spectrum is divided by means of a double-image 

 prism into two spectra having the light polarized in opposite planes, 

 though some of the absorption-bands are of equal intensity in both 

 images, yet others are comparatively absent, some in one and some 

 in the other ; whereas in the case of other dichroic crystals which 

 give spectra with absorption-bands, they are usually all more dis- 

 tinct in one image than when the light is not polarized, and all fainter, 

 or even comparatively absent, in the other. No sooner had I observed 

 this spectrum (No. 5, given below), than I made various experiments 

 in order to ascertain whether uranium was present or not ; and the 

 then known tests that could be applied to the amount of material at 

 my command seemed to show that it was absent. This was quite in 



* See Stokes's papers, Phil. Trans. 1852, p. 463, and 1853, p. 392. 

 f Vol. xix. p. 122. % Vol. xvii. p. 511. [Phil. Mag. Jan. 1870, p. 65.] 



2G2 



