1867.] Colouring-matters hy the Spectrum Microscope. 453 



Such spectra are at once seen to differ most decidedly when compared side 

 by side ; and that the colouring-matters are entirely different is proved by 

 other facts. If the absorption-bands agree very closely, we ought to com- 

 pare other spectra before concluding that the substances are the same. 



22. Mixed Colours. 



Of course, if any impurity is present which absorbs that part of the 

 spectrum where the characteristic bands occur, it may be difficult, or even 

 impossible, to determine the nature of the substance ; but the rest of the 

 spectrum may be obscured, and the general colour entirely changed, without 

 the least difficulty being thereby produced. Thus, for example, on adding 

 a solution of Saffron {Crocus sativus) to that of the blue Lobelia, the 

 colour is changed from blue to a curious olive, and the spectrum be- 

 comes — 



* * 



Lobelia and Saffron 2|— 3|. . .3|— 4f 6^ . . 7- — 



Lobelia 2i— 3|...3|— 4f 11... 



Difference , • • 7 



If we did not know it, we might thus infer that they were the same sub- 

 stance, and only differed because one contained a yellow colour ; and this 

 conclusion would be borne out by adding to each citric acid and sulphite of 

 soda, which make the Lobelia colourless, and leave the residual yellow 

 colour, 6| . . 7- — , in the case of the mixture. The petals of very many 

 flowers do really contain more or less of such a yellow, which appears to 

 be that developed to a much greater extent in the stamens, &c. ; and 

 though this often modifies the general colour and the spectra, its presence 

 may be recognized in a similar manner. Different species of Diantkus, 

 various kinds of Roses, and Digitalis purpurea are good examples of one 

 general colouring-matter m_odified in this manner. Its normal charac- 

 ter is 



* 



1, A, aq^ am^ al^ am^ (l|--2|...4i 11 . .). 



In studying mixed colours, so much depends on their special characters, 

 that it would be difficult to give any other general rules ; and particular 

 cases do not form part of the plan of the present paper. 



23. Spectra with no Bands. 

 The principal difficulty to be contended with in this method of qualita- 

 tive analysis, is in the case of the subgroups where no decided absorption- 

 bands can be developed by any of the reagents. They can be easily 

 divided into the groups A, B, and C, but the difficulty is to distinguish the 

 separate colours, if we are not sure that they are pure and simple. Some- 

 times special facts may be of use ; but, as a general rule, we are compelled 



