242 NATURAL SCIENCE Go chaie 
selection only securing that the variation is a good one. The 
natural selection of Darwin works in the same way: that is, it 
isolates beneficial variations by killing off the others. It is not so 
much natural selection as natural elimination; for, as Professor 
Lloyd Morgan has pointed out,’ it is not by the survival of the 
fittest, but by the elimination of the least fit, that new species are 
made. Isolation by elimination must always tend to preserve 
variations which are useful to the competing individuals, while 
isolation by other means may preserve not only useful, but also 
indifferent and even injurious, variations, as in the case of domesti- 
cated animals and cultivated plants. 
Isolation by Selection—or natural selection in a restricted 
sense—implies an outside agency as selector for whose benefit the 
isolation is made. These cases of true natural selection are limited 
to selection by insects. Several kinds of beetles, domesticated by 
ants, have become blind, and some of them are unable to feed them- 
selves. These variations are evidently injurious to the domesticated 
animals, but useful to the ants, as they prevent the beetles from 
running away. The sticky secretion of aphides must also be in- 
jurious, if we may judge by the eagerness with which they allow the 
ants to remove it; and we must, therefore, suppose it to be due to 
selection by the ants. But how the selection was made we do not 
know. 
The structural growths which, in many flowers, necessitate the 
visits of special insects to fertilise them, are also probably due to 
natural selection in its narrow sense, for it is very doubtful whether 
they are useful to the plants. In the first place the plants which 
have the most elaborate apparatus for securing fertilisation by 
certain insects only are uniformly rare; while self-fertilising and 
anemophilous plants are abundant.” Secondly, very few annual 
plants, which must set seeds every year, have complicated flowers, 
and some of these—such as the annual peas and beans—are also 
self-fertilising. Thirdly, many perennial plants with elaborate 
flowers have resorted to other means to secure fertilisation in case 
insects fail to visit them. Fourthly, the great number and abund- 
ance of plants, whose inconspicuous gamopetalous flowers show that 
they have reverted from insect fertilisation, is a sufficient proof that 
they have not suffered any harm by doing so. We must therefore 
conclude that the elaborate flowers found in many of the so-called 
entomophilous plants are quite unnecessary for their well-being, and, 
indeed, must be sometimes harmful, for they render fertilisation 
1<«¢ Animal Life and Intelligence,” 2nd ed., p. 791. 
2 Henslow, ‘‘On the Self-Fertilisation of Plants,” Zrans. Linn. Soc., 2nd series 
(Botany), vol. i., p. 317. 
