182 



Prof. (1. G. Stokes. On a remarkable [Feb. 26, 



reflection mentioned by Professor Mills as belonging to the crystals 

 in general at an early stage of their growth. This, however, proved 

 to be quite a different and no doubt a much, simpler phenomenon. 

 The difference was shown by the polarisation of the light, and above 

 all by the character of the spectrum of the light so reflected, which 

 resembled ordinary spectra of interference, and did not present the 

 remarkable character of the spectra of the peculiar crystals. 



15. When, however, the whole was left to itself for a day or so, 

 among the mass of usually colourless crystals a few were found here 

 and there which showed brilliant colours. These colours were com- 

 monly far more brilliant than those of the crystals mentioned in the 

 preceding paragraph, and they showed to perfection the distinctive 

 character of the spectrum of the peculiar crystals. It would have 

 been very troublesome, if possible at all, to examine the twinning of 

 such thin and tender plates as those thus obtained by working on a 

 small scale ; but the character of the spectrum, which is perhaps the 

 most remarkable feature of the phenomenon, as well as the dependence 

 of the colour on the orientation, may be examined very well ; and 

 thus anyone can study these features of the phenomenon, though he 

 may not have access to such fine coloured crystals as those sent me 

 by Professor Mills. 



16. A certain amount of disturbance during the early stages of 

 crystallisation, whether from natural currents of convection or from 

 purposely stirring the solution somewhat gently so as not to break 

 the crystals, seems favourable to the production of the peculiar 

 crystals. When the salt crystallised slowly from a quiet solution I 

 did not obtain them. 



17. As it is easy in this way, by picking out the peculiar crystals 

 from several crystallisations, to obtain a good number of them, the 

 observer may satisfy himself as to the most usual character of the 

 spectrum. It is best studied at a moderate incidence, as it is sharper 

 than when the incidence is considerable. The spectrum most com- 

 monly shows a single intensely bright band, standing out on the 

 general ground of a continuous spectrum of moderate intensity. 



A few cases seem worthy of special mention. In one instance two 

 bright bands were seen, one at each faint end of the spectrum, some- 

 what recalling the flame-spectrum of potassium salts. In another 

 case a red, a green, and a blue band were seen, reminding one of the 

 spectrum of incandescent hydrogen. This crystal in air was nearly 

 colourless at moderate incidences, but showed red at rather high 

 incidences. In another case the crystal was red of intense brilliancy 

 in the mother-liquor, but was colourless when taken out, even at high 

 incidences. Presumably the stratum in this case was so thick that a 

 steeper incidence than could be obtained out of air was required to 

 develop colour. 



