Structure in temperate predaceous diving beetle communities (Coleoptera: Dytiscidae)

Collection records of adult dytiscid beetles (Coleoptera, Dytiscidae) from 312 sites within the province of Alberta, Canada, were analyzed for patterns of similarity in species distribution. Sites compared on the basis of species occurrence (Jaccard coefficient of similarity and group average cluster analysis) separated into 12 principal clusters, interpreted as communities. However, cluster inclusion occurred at very low levels of similarity and 30% of sites were unclustered outliers. Ecological charade us tics of sites that were recognized as varying between dusters were salinity, productivity, stability, water temperature, substrate type, flow and vegetation. An index relating these qualitatively defined variables to species numbers is presented. An analysis of co-occurrence patterns of species was conducted as an alternative approach to definition of multispecies associations. Each possible combination of co-occurrence between pairs of species occurring in more than 5 collections each was tested for significance (p = 0.005) by means of Yates corrected χ² tests. Results of tests were expressed in binary terms, blank – no significance, x - significant. Species were then clustered on the basis of patterns of significant co-occurrences as above. This analysis was used as the basis for ordination of species in a symmetrical matrix with species ordered so as to maximize density of points of significant co-occurrence (x) around the principal diagonal. Areas in the matrix showing high densities of co-occurrence were interpreted as representing communities. The cluster analysis of species produced a better cluster pattern than obtained for sites but while generally similar, the species clustering pattern was less fine. The principal species groups consisted of those of saline water, alpine/subalpine lotic sites, other lotic habitats, and lentic habitats with two subgroups, species in forest areas and grassland species. The ordination, while reflecting this grouping of species, demonstrated that each group graded into others through loss of certain species and acquisition of others and thus showed the continuum nature of these communities. Restricting ordination to congeneric species indicated that these do not usually show identical patterns of co-occurrence but neither are patterns coincident with mutual exclusion. Within genera, considerable overlap in occurrence between species is the norm but species tend to occupy slightly different positions relative to one another along qualitatively defined ecological gradients. Relative position of species in these ordinations shows good correlation with patterns of geographical distribution lending support to the interpretation of ordination patterns reflecting relative ecological segregation. The dytiscid beetle fauna of north temperate regions is especially rich. Also, the large number of significant co-occurrences observed between species shows that species packing in many habitats of this region is dense. It is speculated that this species richness is possible because of the seasonality of habitats at these latitudes which reduces competition and predation from other groups and produces seasonal pulses of high productivity permitting a species rich fauna to develop because of abundant resources. It is likely that factors other than interspecific competition for resources are important in shaping dytiscid communities in habitats with strong nutrient pulses.