The effects of two parasitoids on the life history and metapopulation structure of the intertidal snail Littoraria filosa in different-sized patches of mangrove forest

We describe differences in life history of the intertidal arboreal snail, Littoraria filosa, among patches of mangroves ranging in size from isolated trees to large stands several square kilometres in area. Recruitment of L. filosa occurred from mid spring (October) to early winter (June), recruits grew rapidly and copulating adults were found during the following September-April. Populations within large patches of forest were annuals; all or most individuals died between October-January (spring-midsummer). In contrast, those in smaller peripheral patches were more likely to survive over the summer but survival differed among patches and years. These differences in life history were caused by a parasitoid fly (genus Sarcophaga) that attacked L, filosa 10 mm and longer and was present in all large patches, but absent from, or rare, in smaller peripheral patches. Experimental introductions to isolated trees confirmed that the fly could kill L. filosa. Another sarcophagid parasitoid that attacked L. filosa from 4 to less than 10 mm long was also found in every patch. The combined effects of these parasitoids appear to determine the metapopulation structure of L. filosa. Most adults in large patches were killed by the larger fly during early summer. Summer recruits were often killed by the smaller fly within a month of settlement and when this happened effective recruitment of L. filosa was reduced to autumn. The planktotrophic larval stage of L. filosa lasts less than 1 month, so the source of autumn recruits to all patches must have been adults that survived the early summer, most of which were in small patches or on isolated trees. Consequently these "peripheral sources" are likely to be important for persistence of the metapopulation of L. filosa. The results of this study demonstrate that metapopulation structure may be determined by complex interactions and that common models cannot be assumed to apply in all habitats.