ABSTRACT
So far I have been more concerned to identify the common function of the various modes of cheap, conservative reproduction than to distinguish between them. But if each mode of reproduction is indeed selected to serve a different function then it should be easy to distinguish between automixis and apomixis because of their contrary effects on heterozygosity. Most previous authors seem to have burked the issue. The most popular line of argument, to be found in most textbooks of botany, is that automicts (in this context, obligate selfers such as cleistogams) lose the capacity for evolutionary change, even though automixis creates a short-term advantage, so that an optimal frequency of selfing will be set by the balance of short-term and long-term selection (see Mather, 1973). On the other hand, Haldane (1932) imagined that selection will proceed faster in selfed than in outcrossed organisms, because genes are selected more rapidly when homozygous. In the case of facultative or cyclical automixis or autogamy, it has often been argued (following Darwin, 1888) that the ability to self-fertilize is favoured because it ensures reproduction when mating partners are not available. These theories have a crucial failing in common, which is that if they are theories of automixis they are equally theories of apomixis. Indeed, they are not really theories of automixis at all, but are rather theories of any conservative, economical mode of reproduction. Because none of them distinguish adequately between automixis and apomixis, they can all be regarded as special cases of the historical hypothesis: granted that selection will favour parthenogenesis, none of them provides any grounds for predicting whether automixis or apomixis should evolve. Williams (1975) has explicitly defended this historical interpretation of thelytoky.
