13897] BEES AND THE DEVELOPMENT OF FLOWERS 105 
Those flowers of the original race which were cross-fertilised should 
have survived rather than those self-fertilised. The passage from 
perfect adaptation for cross-fertilisation to perfect adaptation to 
self-fertilisation is a long one, and must have been spread over 
many generations in each of which the latter was an advantage, if 
it is to be accounted for on the principles of natural selection. 
But it has been suggested that the change has been due to the 
absence of insects and that thus only those plants which were able 
to fertilise themselves survived. If, however, we suppose that in the 
ancestral orchid the apparatus for cross-fertilisation was as perfect 
as in many species at the present day, it would be incapable of 
self-fertilisation, and therefore die out in the absence of insect 
visits. Even if it could in a few cases fertilise itself, how could its 
fewer and weaker progeny compete with the stronger seedlings of 
nearly related and cross-fertilised species probably occupying the 
same station? If, however, in the ancestral orchid the arrangement 
for fertilisation was such that self-fertilisation usually took place 
in default of insect visits, then no benefit would arise from change 
of form to perfect adaptation for self- and avoidance of insect- 
fertilisation. With regard to the benefits resulting from cross- 
fertilisation generally, Professor Henslow points out that orchids, 
the most remarkably adapted of all plants for cross-fertilisation by 
insects, “set the least amount of seed even when fully exposed to 
insects.” 
Another fact established by Darwin in relation to cross-fertilisa- 
tion is that the offspring of the cross is more vigorous when between 
slight varieties of the same species, or between individuals grown 
under slightly different conditions. This fact is also adverse to the 
theory of the development of a species of flower by the selective 
action of the bee. For among the offspring of the crosses affected 
by it those will be strongest which occur between varieties, or be- 
tween plants grown at a distance, and therefore likely to differ 
slightly from each other. But these are precisely the individuals 
in which the incipient characters tending to the formation of a 
new species will be least marked. Hence the action of the bee is 
rather to retard development; and Darwin himself has remarked 
that frequent in-crossing tends to give uniformity to species varying 
slightly as they do in a state of nature. 
A brief allusion to Professor Henslow’s amendment of the 
Darwinian insect selection theory will suffice. Apart from the 
extreme improbability—as shown by recent research—that such 
acquired characters as the lengthening of the petal of a flower by 
the weight of an insect stretching it, or the coloration caused by 
the irritation of an insect’s feet, are transmitted, Professor Hens- 
low’s theory splits on the same rock as the older one. For, like the 
H 
