

INAUGURAL ADDRESS. 9 
system, the arrangement of the vessels in the branchial 
sac, the number and arrangement of the tentacles, and 
so on, to their possessors; and even in the case of such 
apparently trivial characters as the shapes and distri- 
bution of the minute spicules throughout the colonies of 
Leptoclinum amongst Compound Ascidians, I know from 
experience that they affect the hardness and roughness of 
the colony, and so may be of considerable importance in 
repelling the attacks of enemies and in keeping the colony 
free from injurious parasites. 
My contention, then, is that Dr. Romanes’ three 
objections can be got over;* that useful variations would 
not be swamped by intercrossing with the parent forms; 
that specific characters are really useful to their possessors, 
and that consequently natural selection is not merely a 
theory of the origin of adaptations, as Romanes insists, but 
also a theory of the origin of species, as Charles Darwin 
believed. 
It is now time that I should state what Romanes’ theory 
of physiological selection is. If we suppose that in the 
members of an incipient variety some slight change takes 
place in the reproductive system which will render these 
individuals sterile with the rest of the species, while still 
fertile amongst themselves, then the result will be that 
although the varietal characters are useless, and the parent 
species overwhelmingly more numerous, still there will be 
no swamping effect because there can be no intercrossing, 
and our variety will be on the high road to becoming a 
distinct species. 
Romanes gives the following as an illustration :+— 
* See also ‘‘ Physiological Selection,” by Henry Seebohm: London, 1886; 
in which the question is discussed mainly from the ornithologist’s point of 
view. 
+ Romanes, loc. cit., pp. 352, 353. 
