229 
ance in autumn, by supposing that it is destroyed by the 
ozonized oxygen given off by the growing chlorophyll, and 
can be again formed when that action becomes impaired or 
altogether ceases. We could also thus explain why it is so 
much more prevalent in the stalks, which contain far less 
chlorophyll, than it is in the leaves themselves. 
On the approach of autumn, before the leaves have 
withered, we have thus in the foliage of different plants an 
exceedingly variable mixture of chlorophyll, xanthophyll, 
and erythrophyll, with the different members of the chryso- 
tannin group, and itis tothe changes which afterwards occur 
in some or all of these substances that the very variable 
tints of autumn are mainly due. The most striking of these 
depend on the alteration of the chlorophyll. So long as it 
remains green, the production of bright reds and yellows is 
impossible, but, when it disappears, the yellow colour of the 
xanthophyll is made apparent; and, if much erythrophyll 
is also present, its colour combined with this yellow gives 
‘rise to scarlet or red. In many cases, however, the chloro- 
phyll does not disappear, but it changes into the dark olive 
modification, easily prepared artificially by the action of acids. 
This gives a spectrum with an absorption-band in the green, 
which does not occur in normal chlorophyll, and may even 
be seen in the light transmitted by such changed leaves, when 
illuminated by condensed direct sunlight. We may thus 
easily understand why the special tints of early autumn are 
yellows and reds, or dull and dark green. In these changes 
the various pale yellow substances of the chrysotannin group 
remain comparatively unaltered, and in some cases actually 
increase in quantity. They, however, soon pass into the 
much darker red-browns of the phaiophyll group, whilst 
the erythrophyll fades, and thus later in the autumn the 
most striking tints are the bright or duller browns charac- 
teristic of different kinds of plants or trees. 
As already named, there are many different species of 
colouring matter belonging to the chrysotannin group, both 
of those which are, and of those which are not, closely re- 
lated to the more typical kinds of tannic acid. The study of 
these has occupied more time than that of all other groups, 
and it seems likely to throw much light on certain branches 
of vegetable chemistry. It would require more space than 
can be devoted to this paper to describe in detail the separate 
substances, and I must, therefore, confine myself to such 
general particulars as are essential to the proper understand- 
ing of the subject before me. One of the most striking 
peculiarities of these compounds is, that they are readily 
