54-6 
Fruits and Flowers. [September, 
case of an increased demand for nourishment (as in 
hive-bees and hummle-bees*), a more or less general 
utilisation of flowers whose honey lies exposed is 
observed. . .. , , 
a. Visitors with short trunks prefer flowers ol light 
colours,— white or yellow. Species with long trunks 
prefer dark-coloured flowers,— blue, red, violet. Spe- 
cies very eager for food visit a more promiscuous 
circle of colours. 
To these propositions may be added a fourth, which, 
though rendered probable by Muller, was pot demonstrated, 
— that a disharmony between the adaptation of flowers and 
the sphere of action of the agents in crossing may appear if 
the circle of visitors is subsequently modified, i.e., by the 
distribution of plants in a new territory. Low observed 
that the species of Apis flying in the Botanical Gaiden 
made a different selection among exotic flowers from 
amongst native species. They select, indeed, the classes ol 
flowers of the Mediterranean plants in the same serial order 
as those of Central Europe, but their preference for “ bee 
and hummle flowers ” — consequently for flowers with dark 
colours — is increased about 20 per cent. Very diffeient is 
the proportion among American species. Heie the light- 
coloured flower associations seem to attract bees, most 
strongly. This behaviour may be explained by the circum- 
stance that among the North American plants cultivated in 
the Botanical Gardens the yellow Composite predominate, 
whilst among the Mediterranean species bee and humm e 
flowers ” are most numerous. ... 
A11 important result of these researches is that the selec- 
tion of flowers and of colours is not, as Muller supposed, 
determined absolutely by the length of the trunk. This is 
proved by the fatf: that different Apidae, with trunks of equal 
length, make a quite different selection. Low terms this 
phenomenon hcterotropism. He terms species which give an 
exclusive preference to certain flowers ohgotropic, and those 
which proceed otherwise polytropic. The most distinttly 
oligotropic species is Antliidiuin manicatum , and the most 
poly tropic is Apis mellifica. 
In almost all genera of bees there can be shown a factor 
which, in conjunction with the length of the trunk, deter- 
mines the peculiar selection of flowers. As such may be 
* The b commonly figuring in hum&le bee is an error. The h also is rightly 
aspirated, as the word has no connection with hiwiilis • 
