Mixed Colouring Matters with the Spectrum Microscope. 127 



before it gave that band no darker than in the other solution. Ex- 

 pressing this by figures in front of the letters, we may say that the 

 spectra were as follows : — • 



The original solution A + B 



The part soluble in ether . . . . 2 A + B 

 The part soluble in water . . . . A + 2 B 



We have here very good proof of the existence of two substances. 

 If there had been only one, however much we might have divided it 

 into different portions, the relative intensity of the two bands would 

 have remained the same, that is to say, in solutions of equal strength, 

 in all the above cases we should have had simply A -f- B. Of 

 course the extent of this separation might be more or less than what 

 I have supposed, but on detecting any such difference we should 

 endeavour to effect a still further separation, and may perhaps ulti- 

 mately succeed in obtaining both colouring matters in a comparatively 

 pure state. The same principles are equally applicable in the case 

 of substances insoluble in water, but soluble in alcohol, which can be 

 more or less completely separated by the use of bisulphide of carbon. 

 For example, I found that on agitating an alcoholic solution of 

 chlorophyll from green holly leaves with excess of bisulphide of 

 carbon, the part carried down in solution in the bisulphide did not 

 give the same spectrum as that left in the alcohol. Making certain 

 absorption-bands seen at the blue end, equal in both cases, other 

 bands at the red end were so much darker in the part carried down 

 in the bisulphide, that the solution had to be diluted about eight 

 times, before they became equal. There being thus evidence of a 

 mixture, I tried the experiment over and over again, in order to 

 ascertain whether a more comj)lete separation could be effected, and 

 found that, if the leaves were well crushed and then heated in alcohoL 

 the solution rapidly filtered, and treated with the bisulphide as soon'' 

 as cold, it could be separated into a deep green, carried down by the 

 bisulphide, and an almost clean yellow, left in the alcohol, which 

 corresponded to the xanthophyll of yellow leaves. The separation 

 was indeed so complete, that for an equal amount of xanthophyU 

 this yellow solution did not contain above g^th the amount of 

 chlorophyll contained in the other solution, so that a mixture of 

 these two colours in ordinary green leaves was placed beyond aU 

 doubt, as also shown by Professor Stokes in the paper already cited. 



Such a case as this leads me to the consideration of the spectra 

 of substances giving several absorption-bands. For example, sup- 

 pose the original solution showed four bands. A, B, C, and D, and 

 when partially separated, as described above, the spectra became as 

 follows, — 



Original solution A + B + C + D 



One product 2A + 2B+C +D 



Another A+B+20 + 2D 



