the world's emptiest ecospaces [shallow water environments around a maximally isolated group of islands].)

When the reproductive pair in a termite colony dies, secondary reproductives (also called, and spelled, "neoteinics") can be quickly recruited by inducing sexual maturity in larvae or nymphs. In some species, an entire caste (the pseudergates) is kept in a larval state, despite repeated molting, with potential for rapid development into either soldiers or secondary reproductives as the need arises (Lebrun, 1961). The "idling" pseudergates cannot be properly labeled as either progenetic or neotenic, but the speed with which they can be recruited as secondary reproductives suggests a progenetic strategy. Gay (1955) reported a high frequency of paedomorphic secondary reproductives in two species of Coptotermes recently established as successful colonists in New Zealand. Yet the same species in their native territory of Australia very rarely develop secondary reproductives. This latent capacity for rapid recruitment of reproductives has evidently been exploited by the expanding colonists.

Parasites

In defining progenesis, Giard linked its frequent occurrence with parasitism. He cited a general antagonism between generation and growth, arguing that parasites often need rapid and copious generation, but do not require many adult organs: "Thus we see that a very great number of parasitic animals are progenetic" (1887, p. 24). De Beer (1958, pp. 64, 90) maintained this association by dismissing progenesis as unimportant in evolution because it is commonly associated with such "degenerate" developments as the simplified morphology of parasites.

Progenesis among parasites may be recognized by two standards of comparison. Some species are progenetic with respect to normal, free-living relatives; they have "traded" adult morphology for one of the outstanding attributes of parasites—devotion of energy to reproduction. Other species are progenetic with respect to related parasites. These have usually developed a simpler life cycle by maturing in an intermediate host (Baer, 1971, p. 126).

The conditions of parasitic life often impose a regime of r selection. An uninfected host is an unexploited resource; if parasites are small relative to their host, a period of rapid increase in numbers may follow the first infestation. If the host cannot support many parasites (because of their large size or debilitating effect), high r might still be favored in order to produce progeny rapidly for dispersal to other hosts. Any host is an evanescent resource (for it will die a natural