102 The American Naturalist. [February, 
of Compositae is determined by the position of the asters, 
Eupatorium, golden-rods and sunflowers. 
AS a result of the divergence of the blooming periods from 
the maximum point of the group we find that plants come 
into competition with species of other groups, but as a rule 
they can stand this better than competition with their own 
allies. 
Trees have such a remarkable influence upon one another 
and upon the herbaceous flora that they should properly, it 
seems, be considered separately. The fact that most of them 
agree in being wind-pollinated is an additional reason for this 
course. Of 488 indigenous insect-pollinated plants, upon 
which my observations are based, only 18 are trees. On ex- 
amining the curve for the insect-loving flora (Fig 1, Plate VIII, 
5 species to the millimetre), it will be observed that the max- 
imum is reached in August. At this time 187 species are in 
bloom, but not a single tree is among them. The flowers of 
trees are so interfered with by their own leaves and the leaves 
of other trees that it is disadvantageous for them to bloom 
after the leaves are fully developed. In the case of wind- 
pollinated trees it is obvious that, if the leaves were developed 
before the flowers, the process of pollination would be greatly 
impeded by the leaves interfering with the free circulation 
of the wind and catching the pollen which is intended for 
the stigmas. This fact makes trees an evident exception to 
Mr. Clarke’s generalization, though they are frequently less 
specialized than their later flowering allies. In the anem- 
ophilous nettle family (Urticaceae) there is a marked con- 
trast between the blooming times of the trees and herbaceous 
species, as stated by Mr. Clarke. Thus the elm, hackberry 
and mulberry are early, while the hop, hemp and wood- nettle 
(Laportea) are late. 
In the case of insect-pollinated trees the conditions are 
similar to those of wind-pollinated ones, and they generally 
3 Unless otherwise specified, the curves given in this paper are on the scale of 
one species to the millimetre, i, e., the height of the curve in millimetres in- 
dicates the number of insects flying, or flowers in bloom at a giventime. The 
details on which the curves are based will be given elsewhere. 
