and Vegetable Colouring -matters by the Spectrum Microscope. 155 



might easily fancy that all the colour was lost by decomposition ; 

 and an evaporating dish containing it might appear merely filled 

 with brownish alcohol, and yet on evaporation the whole dish might 

 be covered with a fine deep colour. The same change may occur over 

 and over again, the deep-coloured solution first obtained soon fading, 

 and the colour being restored by subsequent evaporation. 



When such a colour is dissolved in a little water and added to alco- 

 hol in an experiment -tube, the colour may at first be very deep, but 

 may fade so rapidly that there is scarcely time to observe the spec- 

 trum before it passes into that molecular state which does not absorb 

 any of the rays of light. The colouring-matter of the flowers of the 

 red Salvia (S. splendens) is an excellent example. Neutral solutions 

 do not undergo this rapid change ; a different condition of acidity 

 is requisite for different colouring-matters ; and some do not change 

 at all. A large excess of citric acid very often restores the intensity 

 of the colour ; and usually the absorption-bands are seen to the 

 greatest advantage when the solution is in that state which rapidly 

 fades ; and by adding too much colour and watching whilst it fades, 

 they may be seen and measured when at their best. This fading of a 

 dark-coloured solution must not be confounded with the change 

 which takes place on diluting some salts, as described by Dr. Glad- 

 stone in his paper on that subject*. 



14. Absorption-bands. 



Though acids and alkalies thus, to a greater or less extent, alter 

 the position of the general absorption, they act very differently on 

 the special, local absorption to which it is very convenient to restrict 

 the term " absorption-bands." Since I shall often have to speak of 

 their being at equal intervals, it would be well to say that I have found 

 it convenient to construct a wedge-shaped piece of quartz, cut parallel 

 to the axis of the crystal, and to use it along with two Nicol's prisms 

 in such a manner that the spectrum may be divided into any requi- 

 site number of equal portions, by interference-bands situated in 

 any requisite position. This of course avoids the errors which so 

 often happen when we compare together measurements that cannot 

 be made with very great accuracy. 



As an excellent illustration I select the colouring-matter of Alkanet- 

 root (Anchusa tinctoria). It is insoluble in water, but is easily dis- 

 solved by alcohol, even when much diluted with water, and gives a 

 clear pink solution. The spectrum is nearly the same when the 

 colour is dissolved in absolute alcohol as when much water is pre- 

 sent, only each of the absorption-bands is situated rather higher. 

 Thus, taking the centres of the bands, we have — 



b. c. d. 



Absolute alcohol 4| 5-J- 7$ 



Very dilute alcohol 4% 5 j 7\ 



The general spectrum of the solution in dilute alcohol will be 

 best understood from the following figure, No. 1 : — 



* Quart. Journ. Chem. Soc. vol. xi. p. 36. 



