Aim To investigate the formation of nestedness and species co-occurrence patterns at the local (sampling station), the intermediate (island group), and the archipelago scale.
Location The study used data on the distribution of terrestrial isopods on 20 islands of the central Aegean (Greece). These islands are assigned to two distinct subgroups (Kyklades and Eastern islands).
Methods The Nestedness Temperature Calculator was used to obtain nestedness values and maximally nested matrices, the EcoSim7 software and a modified version of Sanderson (2000) method were used for the analysis of species co-occurrences. Idiosyncratic temperatures of species and the order of species placement in the maximally nested matrices were used for further comparisons among spatial scales. The relationships of nestedness values with beta-diversity, habitat diversity and a number of ecological factors recorded for each sampling station were also investigated.
Results Significant nestedness was found at all spatial scales. Levels of nestedness were not related to beta-diversity or habitat diversity. Nestedness values were similar among spatial scales, but they were affected by matrix size. The species that contributed most to the nested patterns within single islands were not the same as those that produce nestedness at the archipelago scale. There was significant variation in the frequency of species occurrence among islands and among spatial scales. There was no direct effect of ecological factors on the shaping of patterns of nestedness within individual islands, but habitat heterogeneity was crucial for the existence of such patterns. Positive associations among species prevailed at all scales when species per station were considered, while negative associations prevailed in the species per island matrices. All associations resulted from the habitat structure of sampling stations and from particularities of geographical distributions.
Conclusions There was no clear-cut distinction between nestedness patterns among spatial scales, even though different species, and partially different factors, contributed to the formation of these patterns in each case. There was a core of species that contributed to the formation of nested patterns at all spatial scales, while the patterns of species associations suggested that biotic interactions are not an important causal factor. The results of this study suggest that locally rare species cannot be widespread at a higher spatial scale, while locally common species can have a restricted distribution.