Larvae of the tropical sand dollar Encope michelini (Agassiz) have a novel feeding strategy in which they can develop all larval structures and survive in the absence of external food, but require external food (greater than 1 mu m) to build the juvenile rudiment and undergo metamorphosis. E. michelini larvae are ''enhanced'' planktotrophs. They use maternal reserves to develop further than planktotrophs in the absence of external sources of nutrition, Yet they are different from lecithotrophs or facultative planktotrophs, because they cannot develop consistently to metamorphosis in the absence of external sources of nutrition, Three experiments were conducted to examine this novel feeding strategy. The first tested the effect of varying the quantity and quality of food, The second tested the effect of varying the timing of feeding, and the third tested the effect of culturing the larvae in sterile, autoclaved seawater. In the first experiment starved larvae survived as well as fed larvae for the first nine days of development and persisted for up to 48 days, which was five times the development time of fed larvae. Larvae that were cultured with enrichment [either 50 cells .mu l(-1) of Dunaliella tertiolecta (Butcher) or Isochrysis galbana (Tahitian) (Green)] had longer bodies, metamorphosed earlier, and resulted in larger postmetamorphic juveniles. Natural seawater was a better culture medium than filtered seawater, because a higher proportion of larvae cultured in natural seawater survived, had larger larval bodies, and metamorphosed. In the second experiment providing food during the early stage of development influenced the size of the larval body early in development and the proportion of larvae that metamorphosed. Providing food during later stages in development, when juvenile structures were being built, resulted in larvae with larger bodies, a higher proportion of metamorphosis, and larger postmetamorphic juveniles, A higher proportion of larvae that were cultured in sterile, autoclaved seawater survived compared to larvae cultured in only filtered seawater, presumably because deleterious bacteria and viruses were not present. Larvae cultured in sterile, autoclaved seawater tended to be smaller (in arm length and body length) than larvae cultured in filtered seawater. The ability of E. michelini larvae to grow to full size under starvation conditions is unique for a species with feeding development. If larvae can still become competent after starvation, then this is a mechanism by which this species of larvae can extend the larval period if food is not available. The feeding strategy employed by E. michelini larvae is a novel intermediate which most closely resembles planktotrophy. Such an intermediate is predicted by theoretical models to be evolutionarily unstable. Perhaps it is, but alternatively, the models may be incomplete because they only consider selection on gamete size and larval survival. Revision of the models to include selection on juvenile success may reveal different evolutionary patterns, because selection for large eggs, in planktotrophic species, may be based on juvenile success rather than on larval success. If this revised theory is correct, then this type of feeding strategy may occur in other marine invertebrate taxa.
