1895.] The Philosophy of Flower Seasons. 111 
and all having late maxima. These late Compositae have few 
competitors outside of their family and so are favorably situ- 
ated, although the insect fauna has begun to decline. We will 
now leave the Sympetalae with the general statement that the 
late preponderance of the irregular flowers is explained in 
connection with the late preponderance of the higher bees, and 
that of the regular flowers is accounted for in the late maxima 
of the highly specialized long-tongued insects. 
We have reviewed the principal groups of insect-pollinated 
plants and have noted a correspondence, more or less well 
marked, between their blooming seasons and the seasons of the 
insects upon which they depend. In different positions we find 
bumble-bee flowers and, although they all occur within the 
time of flight of these insects, it is not easy to explain why one 
of these flowers comes at one time and another at another time. 
Under the maximum of the buttercup family (Fig. 9, Plate VII) 
we find a bumble-bee flower in the larkspur (Delphinium tri- 
corne) and under the maximum of Leguminosae (Fig. 15, Plate 
VIII) another in a tick-trefoil (Desmodium canadense). We may 
say that the larkspur comes earlier because it had its origin in 
an earlier group. The flight of the bumble-bees, however, can- 
not be left out of consideration. It is obvious that a bumble- 
bee flower cannot arise at a time when the attentions of 
bumble-bees cannot be secured, so that the flight of the bees 
determines the time within which these flowers may have their 
origin. When a flower undergoing modification to suit 
bumble-bees changes its characters so that it no longer comes 
in competition with its allies, it becomes a competitor of other 
bumble-bee flowers. A point at which many of these are in 
bloom simultaneously would naturally be an unfavorable time, 
unless the new form should early offer more inviting attrac- 
tions. Ifthe blooming time were long, the attentions of the 
bees would be likely to be most constant at the point where 
there were the fewest competitors, and so finally the blooming 
time would tend to be limited to this point. Or if the earlier 
flowers were better tended, so that they became the most effect- 
ually fertilized, the blooming time would tend to become 
earlier. Some flowers we find far from the “tension” points 
