Diurnally foraging planktivorous fishes are exceptionally dense in marshes worldwide, and larvae of most species of marsh crabs largely may escape predation in time and space by hatching during nocturnal maximum amplitude high tides (NMAHT) when they are rapidly transported to deeper waters under the cover darkness. Species that have long spines or other well-developed larval defenses may be less constra~ned to release larvae during this safe period. Three species of diurnally foraging planktivorous fishes (Menldja menidia, Fundulus heteroclitus, F. ma~alis) were very abundant and comprised nearly all of the fishes collected in Flax Pond salt marsh In New York, USA, during the summer of 1994. Four species of crabs (Sesarma retjculatum, Uca pugnax. U. pugdator, Dyspanopeus say13 also were abundant there and released larvae from mid J u n e to September. These crabs did not release larvae randomly, rather larval release peaked during NMAHT when strong e b b tides transported 97.7 of the larvae from the marsh by the follo\ving day. Larvae that were released during nocturnal minimum amplitude high tides were transported from the marsh less effectively (84.5",4). Only 0.37 % of larvae remained in the marsh following the molt to the second instar. Larval release by D. say1 was weakly synchronized with the tidal amplitude cycle, and these larvae were transported from the marsh least effectively. Although D. say1 larvae that were released during NMAHT were transported from the marsh nearly as effectively as were the other species (95.6% vs 98.0% of S. reticulatum and 99.9% of Uca spp.). more of them were released near nocturnal minimum amplitude high tide when transport was least effective (58.2% vs 86.0",, of S. retfculatum and 87.5% of Ucd spp.). Transport of D. sayilarvae from the marsh may have been slowed further by the tidal vertical migrations undertaken by these larvae. In feeding trials conducted in the marsh, M, menidia and E heteroclitus ate about half as many newly released D, sayi larvae as S. rehculatum and U, pugnax larvae (57.2% vs 26 1 % of S. reticulatum and 32.9% of U. pugnax). Long spines may have deterred predation on D sayilarvae and may reduce selection for rapid transport of these larvae from the marsh. Recruitment to Flax Pond occurred during flood tides, especially strong flood tides, at night when transport to adult habitats was maximal and predation by fishes was minimal. Megalopae of all study species recruited regularly in the same relative proportions that they were released throughout the entire reproductive season, even though larvae of some of the study species (S. reticulatum, D. say^) likely develop entirely in adjacent Long Island Sound whereas others (Uca spp.) disperse through the sound and onto the continental shelf. This suggests that reproductive and larval behaviors largely overcame mortality during the planktonic phase of the life cycle and coupled production to recruitment, regardless of whether or not recruits primarily originated from local populations. Thus, pred...