Non-Adaptive Characters

Abstract

In reply to Prof. Graham Cannon, I neither stated at the British Association, nor have I ever held, that all characters must possess 'selection value' : precisely the contrary, since I took some care to explain that the spread of non-adaptive characters, which certainly exist, cannot be responsible for evolution in wild populations. At the same time I pointed out the danger of stating that any particular character is non-adaptive, since even a 1 per cent advantage can rarely be detectable by the most accurate experiments, though it is considerable from an evolutionary point of view. It is genes, not characters, that must very seldom be of neutral survival value. That is by no means to say that they are never so, but, as stressed at the meeting, such genes cannot spread in a semi-isolated population so as to produce the 'Sewall Wright effect' unless it be numerically very small, of well under a thousand individuals, and such populations are unlikely to be sufficiently permanent units.