330 The Botanical Gazette. (September, | 
its light many puzzles have been solved, and many facts 
hitherto inexplicable have been made plain. We now know 
what relationship means, and we have given a fuller meaning — 
to the natural system of classification. From the new point 
of view a natural classification is not merely an orderlyar — 
rangement of similar organisms. It is an expression of gem 
etic relationship. Furthermore in the light of evolutionwe 
now see the meaning of many reduced structures whose sig 
nificance was not at all or but vaguely understood. We have 
become familiar with the fact that degradation isa prominent — 
factor in the vegetable kingdom. Evolution has by no meats — 
always involved an advance in structural complexity. Often 
this catagenesis is a result of parasitism or saprophytism, 6 
is so well illustrated in the fungi, where the degradation has 
gone so far that their relationship has to a great degree beet 
obscured. .. 
But there are many cases of a catagenesis not due toa de 
pendent habit, in which we have evidence of a ina 
from a more complex structure. Thus in the willows 
poplars, where we have a raceme of very simple flowers, cat 
consisting of a single ovary, or one to many stamens, it t 
readily seen that this simplicity is not primitive. The vo 
ies are not single carpels, but are composed of two oF NU” 
united. The flower of a willow is simple by 4 pee 
from a higher type, probably a tricarpellary of pen 
lary type, by the loss of its floral envelopes an 
pistils. : «tt acttae 
Every naturalist should be as familiar with these illus 
tions of evolution by simplification as 
lution by complication. In the growt 
life, while the development has been most 
ward direction so that the great body of the 
above its point of beginning, there are yet m 
and branchlets which droop downward. i 
P cientific mem, 
d stamens and 
evolution as an hypothesis, for we know it a@5 "4 
cal fact, about whose existence there 1s NO 
A natural classification will conform strict: f the suce 
evolution, it will be in fact a clear exposition 0 the prit 
steps in its progress. In such a classification se 
forms will precede the derived ones, and the 
‘oreover, 
in such 4 
latter will be positively indicated. M : 
