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



examination. The number of distinct coloured compounds met with in 

 different plants is so great, that some such classification is imperative. In 

 the first place, we cannot do better than divide them according as they are 

 soluble in water or alcohol. This may be looked upon as a chemical divi- 

 sion, and is very useful in practice. Thus — 



Soluble in water and not precipitated by alcohol .... Division 1 . 



Soluble in water and precipitated by alcohol „ 2. 



Insoluble in water but soluble in alcohol „ 3. 



Insoluble both in water and alcohol „ 4. 



Of course cases occur which cannot be unhesitatingly classed with any one 

 of these ; but they often form good practical divisions, and necessarily 

 modify the methods requisite for further examination. 



19. Method and Order of Experiment. 



If a colour belongs to division 1, a small quantity, sufficient for three 

 or four experiments, should be exposed to the vapour of ammonia in a 

 watch-glass, until there is certainly no longer any free acid, and then 

 gently evaporated, so that all excess of ammonia may be lost. If not thus 

 made neutral we might be entirely misled ; for some pink colours are blue,"- 

 reddened by an acid. A small quantity should then be dissolved in water 

 in one of the small experiment-tubes and the spectrum observed. If too 

 little colour has been added to give the characteristic spectrum, more 

 should be introduced ; but if any part is entirely absorbed, the cell should 

 be turned sideways, in order to see whether or no some narrow absorption- 

 band occurs there ; and then it may be desirable to remove some of the 

 solution, and fill up the cell with water. As a general rule, so much colour 

 should be added as to make the darkest part of the spectrum decidedly 

 shaded, but yet not so black as to hide any narrow bands ; and if any 

 occur, the solution should be made of such a strength as to show them to 

 the greatest advantage. This can easily be done, after a little practice, 

 and is made much easier by being able to turn the tubes sideways. Having 

 noted the spectrum of the neutral solution, a very small quantity of am- 

 monia should be added, and then a decided excess, the spectra being 

 examined to see if there be any difference ; for this is very often the case 

 and very important facts may be overlooked if too great an excess be added 

 at first. The addition of a small bit of sulphite of soda then shows at 

 once whether or no a colour belonging to group A is present ; and on 

 adding excess of citric acid we may also determine whether it chiefly 

 belongs to groups B or C. Ammonia should then be added in excess, 

 which may or may not restore it to the same state as before the addition 

 of the acid. To another portion of colour carbonate of soda should be 

 added, and then excess of citric acid, both spectra being carefully observed ; 

 and finally sulphite of soda, which definitely shows whether or no there 

 is any other colour than one belonging to group C. Combining the results 



