Résumé
Species are identified as bioindicators when their presence or abundance within sites of a particular habitat or a conservation area are high. Dufrêne and Legendre (1997), Ecological Monographs 67: 345–366, proposed the use of a flexible asymmetrical approach to detect indicator species and species assemblages characterizing groups of sites, the indicator value for each species being subsequently tested by randomization. Although the advantages of these methods in comparison to other multivariate or two-way indicator species analyses have been demonstrated, relatively few ecological studies have made use of them. The aim of the present study was to demonstrate the advantages of the indicator species test using a data set generated by traditional fisheries activities at four areas in the Bonifacio Strait Marine Reserve, Corsica, France. ANOVA, Kruskal-Wallis and another randomization test were also performed on the same data set. Of the 54 species caught in the fishing nets, the indicator species analyses identified 19 species for which the indicator value was significant for a given area, whereas the a posteriori Tukey test for ANOVA detected only 12 species for which the mass capture per unit effort was significantly greater for a given area. The efficiency of this new approach within the framework of conservation and marine reserve studies is demonstrated by focusing on three species (Sciaena umbra, Palinurus elephas and Maja verrucosa).