55 
evidently due to a yellow colouring matter contained in it. 
It gave an abundant yellow precipitate with lead acetate, 
and a dark green precipitate with ferric chloride. It also 
contained a considerable quantity of tannin, since it yielded 
a thick curdy precipitate with gelatine and an abundant 
deposit on the addition of a mineral acid. On again evapo- 
rating the solution in the water bath some decomposition 
evidently took place, for on adding water to the residue a 
quantity of matter in the form of brown powder remained 
undissolved. It is almost certain that it is the same pro- 
cess of decomposition going on in the yellow leaf on exposure 
to air and moisture that causes the colour to change to 
brown. I think it probable that the process is one of oxi- 
dation, and that it affects the tannin of the leaf or the solu- 
tion rather than the yellow colouring matter, for it is well 
known that watery solutions of tannin undergo decompo- 
sition, accompanied by change of colour from light to dark, 
on exposure to air, especially when the solutions are hot. 
This simple experiment shows that the yellow colour of 
faded elm leaves is due partly, perhaps chiefly, to a yellow 
colouring matter soluble in water, partly to a yellowish 
green substance consisting essentially of chlorophyll. The 
former is, in my opinion, the true xanthophyll. 
In order to gain, if possible, a little more insight into the 
process whereby the colour of green leaves changes to 
yellow, I took an alcoholic extract of fresh grass, which 
was of the usual bright green colour, and exposed it in a 
window to the action of the sun and air. After some days’ 
exposure it had undergone the well known and frequently 
described transformation, i.e. the bright green colour had 
changed to a greenish-yellow and the solution from being 
opaque even in thin layers had become transparent in con- 
sequence of the oxidation of the chlorophyll contained in it. 
Now this liquid, though not quite so yellow as the alcoholic 
extract of faded elm leaves, was found closely to resemble 
