Communities of larval helminths in fish: co-occurrence patterns across host species depend of the structure of the compound community.

JUAN T. TIMI; ANA LANFRANCHI; PAOLA BRAICOVICH; ANA JULIA ALARCOS

Viterbo, Italia

Simposio; VIIth International Symposium on Fish Parasites; 2007

Null models based on randomization tests have been of fundamental importance in community ecology and biogeography; among them, the nested subset pattern analysis has been widely used to identify non-random patterns of species composition in insular biotas, including parasite communities. Two main approaches to the study of nestedness have been commonly used in parasite ecology, one of them to investigate the way in which parasite species are distributed among host individuals in a host population, and the other to analyze co-occurrence of species across component communities. From the results of each kind of studies, it is evident that, whereas nestedness is rare across individual hosts within component communities of parasites, it is fairly common across component communities of a host species. This is probably because component communities can be assimilated as real islands, where extinction and colonization events by parasites (the main causative agents of nestedness in insular biotas) can be considered as evolutionary events. In the present work, nested subset analyses were preformed beyond the limits of host species, being applied across host species within a compound community. The program BINMATNEST was applied to parasite species of low specificity, common in the study area, and shared by several fish species, in many if which some of them dominate infracommunities (larval anisakid nematodes, larval trypanorhynch cestodes and juvenile polymorphid acanthocephalans). A matrix of presence absence of 12 larval parasite species harboured by 13 fish species with different ecological habits inhabiting the waters off Buenos Aires province, Argentina  was analyzed. The matrix temperature was calculated and compared with 1,000 randomly generated matrices. The probability that the temperature of a random matrix being lower than or equal to that of the observed assemblage was calculated according the null model 3. Parasites displayed a nested subset pattern across the component communities (matrix Tº= 12.5º, P