'^ourna\Z;°n«7a'] RoT/al Microscopical Society. 231 



pink solution gradually faded to pale orange, even wlien hermetically 

 sealed, and showed only a faint trace of the band originally seen in 

 the green. On the whole, therefore, the facts seem to show that 

 when decomposition takes place with much albumen present, both 

 colours are generated, and then the blue is decomposed in such a 

 manner that the pink colouring matter is left by itself; and thus we 

 have a good illustration of the value of studying the changes that 

 occur during slow decomposition, and of the importance of our care- 

 fully examining whether any particular spectrum is due to a single 

 substance or to a mixture of several. 



Since writing the above, Mr. Sheppard has kindly sent to me a 

 specimen of the liquid prepared with casein. This gave a spectrum 

 with the same two absorj^tion-bands as when albumen had been used. 

 Having thus, as I believed, proved that it was a mixture of the two 

 coloured substances, I thought it would be interesting to ascertain 

 whether they could be separated, or one decomposed without chang- 

 ing the other. After trying various experiments, I found that by 

 mixing the liquid with an equal bulk of absolute alcohol a precipitate 

 was formed ; and when this had been removed by filtration, a clear 

 23ink solution was left, which gave exactly the same spectrum as that 

 of the pink colour obtained by slow decomposition, as described above. 

 On evaporating this alcoholic solution at a gentle heat, so that nearly 

 all the alcohol was removed, and a more concentrated aqueous solu- 

 tion left, I was able to examine the spectrum to greater advantage 

 than heretofore. The transmitted light showed the narrow and dark 

 absorption-band at the yellow end of the green, and also another, 

 very faint, nearer the centre of the green, but there was no trace of a 

 band in the orange. The light of fluorescence gave a single, narrow, 

 bright-yellow band. These different bands corresponded in every 

 respect with those characteristic of the pink colour already described, 

 and also with the equivalent bands in the spectra of the mixed liquid. 

 The alcohol seems to decompose the blue colour ; for when the 

 precipitate was redissolved in water, only a turbid brownish-grey 

 Bquid was obtained. 



On agitating the mixed solution with ether, it rose to the top 

 coloured pink, leaving the water blue ; but the separation is unsatis- 

 factory, on account of both solutions being very turbid, and the 

 colouring matter in great measure precipitated. 



Through the kindness of Mr. B. D. Jackson, I have also been 

 able to examine another specimen of the blue fluid, prepared without 

 albumen or casein. This gave, hke the specimens previously exa- 

 mined, a single absorption-band in the orange, but none in the green ; 

 and thus it appears that the production of the pink colour depends 

 on the presence of albumen or casein in a state of decomposition, and 

 whatever difference there may be in the result is apparently due 

 to a variation in the relative amount of the two colours. 



