454 Mr. H. C. Sorby on Analysis of Animal and Vegetable [Apr. 11, 



to have recourse to the position and character of the general absorption. This 

 requires a good deal of care, since a difference in the state of the solution 

 may make the same colour differ more than two quite distinct colours. 

 After trying a number of experiments, I find that the best spectra for com- 

 parison are those obtained by adding first a moderate excess of carbonate 

 of soda, and then a considerable excess of citric acid. Both of these solu- 

 tions change very slowly, and give well-marked spectra ; whereas ammonia 

 often causes decomposition, and weaker alkalies or acids give much more 

 faint spectra, or such as rapidly fade. Closely related colours should be 

 compared together, and made as nearly equal as possible after the addition 

 of the carbonate, and then citric acid added in considerable, and nearly 

 equal, excess. We thus can compare two different spectra ; and even if 

 the position of the absorption is the same in both cases, the relative inten- 

 sity may vary considerably. Very closely allied colours may often be 

 easily distinguished in this manner, and the only great difficulty is when 

 coloured impurities are present. As an example, I give some colours be- 

 longing to subgroup 1, B, aq^am^al^am.^. 



Carbonate of soda. Citric acid. 



Petals of WaMower(C^.^■- 1 21 ... 5 8 .. 9 - - 10- 3| 7 



ranthus Cneiri) \ ^ ^ 



Darkgrapes 2i..--5|...9 10. .11-- 4i..--..8i 



Fruit of Elder C'^^^^" 1 2i . - - 6 . . 9 11... 41..--. .8^ 



bucus nigra) ^ 



The first differs entirely from the latter two, but they are so similar that it 

 requires great care to be sure that they differ essentially. If it were quite 

 certain that such colours were pure, it would not be difficult to distinguish 

 them with confidence ; but since they may contain coloured impurities, 

 we must occasionally be content with results somewhat doubtful in ques- 

 tions of minute detail, which might not be of the least importance in some 

 practical investigations. 



24. Yellow Colours. 



One of the best general methods of distinguishing yellow colours be- 

 longing to subgroup C, aq^ am^ al^ am^, or those with bands which are 

 much alike, is to make them as nearly as possible of the same tint when 

 neutral, and then to add excess of ammonia, which may make them very 

 unequal. For example — 



Neutral. Ammonia. 



Yellow DahHa (D. variabilis) 8 .. 9 - - 10— 3. . .4 - - 4^— 



Yellow Calceolaria ( C. aurea-fioribunda) 7...9--11 — 65. .6|- - 7 — 

 '^zSxorci. {Crocus sativus) 7...8--11 — 7..8--11 — 



The action of ammonia thus shows that they differ very much, but at the 

 same time the Calceolaria might be a mixture of the other two, and this 

 would have to be decided by other facts. 



