THE REPRODUCTIVE FACTOR IN EVOLUTION. 309 



bodies, but the processes of nutritive differentiation, and other 

 individual developments, are secondary, not primary. 



Thus it is the central generalisation of botany that, despite 

 the individual differentiation of fern, selaginella, cycad, conifer, 

 and flower, these turn out, on deepest analysis, to be but the 

 surviving phases of a continuous and definite increase in the 

 subordination of the sexual parents to their asexual offspring 

 (see pp. 201, 211). 



Or if we take in particular the origin of the flower, which 

 all botanists agree in regarding as a shortened branch, the 

 natural selectionist explanation (did the theory trouble itself 

 with such questions) would seem to be, that the flower had 

 arisen by selection from the two other alternatives of lengthened 

 and unshortened axes. But this is at once excluded by the 

 physiological explanation that shortening of the axis was inevit- 

 able, since the expense of the reproductive functions necessarily 

 checks the vegetative ones, for it is evident that we cannot 

 speak of selection where the imaginable alternatives are physi- 

 cally impossible. So too the shortening of the inflorescence 

 from raceme to spike or flowerhead, or still further into the 

 hollowed form of a fig, with the corresponding reduction in the 

 size of the flowers, is again the result of the check imposed by 

 reproduction on the growth of axis and appendages. 



The same simple conception of a continuous checking of 

 vegetation by reproduction, unlocks innumerable problems of 

 floral structure, large and small alike, from the inevitable 

 development of gymnosperm into angiosperm by the continuous 

 subordination of the reproductive carpellary leaf, to the varia- 

 tions of cabbages as seen in the transitions between leafy kale 

 and cauliflower. Or again, the origin of floral colour, as 

 primarily an inevitable consequence of the same principle of 

 vegetative subordination through reproductive sacrifice, was 

 long ago pointed out by Spencer, and admits of detailed elabora- 

 tion without attaching more than secondary importance to 

 selection by insects. 



In another way, the antithesis between reproduction and 

 nutrition may be illustrated among the existing orders and 

 species of flowering plants. Just as the lilies, for instance, 

 range on the one side towards the characteristically vegetative 

 grass, or on the other towards the reproductive orchid, so it is 

 with the main variations of every natural alliance. Thus, the 

 Ranunculacese have their grassy and their orchid-like types in 



