PART LY. | 3 GENERAL RETROSPECT. 593 
gradual attamment of their special. characters. I may now close 
this book with a summary of my chief conclusions. The good effect 
of cross-fertilisation may be recognised not only in the structure of 
insect-fertilised flowers, but also in the water-fertilised and wind- 
fertilised plants which preceded them. Even in the lowest aloze we 
find cross-fertilisation, to effect which two individuals move towards 
each other in the water by means of a cilium. Ina higher grade we 
find the two individuals differentiated, one losing its motility, and 
increasing in size to form an ovwm, while the other (sperm-cell 
or antherozooid) retaining its primitive motile form swims about in 
quest of the ovum. - This mode of cross-fertilisation is retained not 
only: by all cellular cryptogams (except the Floridee or red sea- 
weeds, whose tailless antherozooids are moved passively in the 
current) but also by all vascular cryptogams. With the change to 
dry localities, where even occasional moisture sufficient for the 
migration of the antherozooids is not attainable, the vascular 
cryptogams seem to haye developed wind-fertilised _ unisexual 
flowers; thus, first the Gymnosperms, and from these afterwards 
the Angiosperms have arisen. 
Finally from the wind-fertilised Angiosperms, entomophilous 
flowers arose, as insects came first accidentally and afterwards 
regularly to seek their food on flowers, and as natural selection 
fostered and perfected every change which favoured insect-visits 
and thereby aided cross-fertilisation. With the transition to 
insect-fertilisation caine, on the one hand, great economy of pollen, 
but, on the other hand, the uncertainty of insect-visits made it as 
a rule necessary that self-fertilisation should remain possible. 
Thus, though descended from unisexual (anemophilous) ancestors, 
entomophilous flowers are usually hermaphrodite, and are capable 
to a great extent of fertilising themselves when insect-visits fail. 
But in the course of further development, many of them have so 
increased their means of attracting insects (by colour, perfume, 
honey, etc.), that the power of spontaneous self-fertilisation has 
become superfluous and finally has been lost. 
Insects in cross-fertilising flowers endow them with offspring 
which in the struggle for existence vanquish those individuals of 
the same species which are the offspring of self-fertilisation. The 
insects must therefore operate by selection in the same way as do 
unscientific cultivators among men, who preserve the most pleasing 
or most useful specimens, and reject or neglect the others. In 
both cases, selection in course of time brings those variations to 
perfection which Serrespond t to the taste or to the needs of the 
Q Q 
