THEORETICAL HABITAT TEMPLETS, SPECIES TRAITS, AND SPECIES RICHNESS - AMPHIBIANS IN THE UPPER RHONE RIVER AND ITS FLOODPLAIN

1. The purpose of this study was to test predictions of the habitat te mplet and the patch dynamics concept by analysing the relationship bet ween either the species traits or species richness of amphibians and t he spatial-temporal variability of eight habitat types of the Upper Rh one River and its floodplain, France. 2. The information on species tr aits of the twelve amphibian species was accumulated primarily from th e literature; that on habitat utilization was based on field surveys. The information was 'fuzzy coded', analysed by ordination techniques, and finally linked to the spatial-temporal variability of habitat type s elaborated elsewhere in this issue. 3. After elimination of the vari ance caused by the differences between urodeles and anurans, correspon dence analysis of the matrix of species traits demonstrated that the f lexibility of the temporal pattern of reproduction was the major sourc e of variation among amphibian species. A less significant source of v ariation within anurans was related to traits usually linked to the co ncept of r-K selection (potential number of descendants per reproducti ve cycle and ratio of egg size to female size). 4. A correspondence an alysis of habitat utilization by amphibian species separated four grou ps of habitat types, corresponding to different degrees of influence t hat the active channel had on adjacent waters in the floodplain (i.e. from habitats that are most frequently disturbed by floods, to habitat s of braided sections, old oxbow lakes, and the more terrestrialized h abitats). 5. A co-inertia analysis of the relationship between species traits and habitat utilization separated two habitat types (terrestri alized former meanders and oxbow lakes) from the others. Species of te rrestrialized habitats had a reproductive display, a low number of des cendants per reproductive cycle, and a high ratio of egg size to femal e size. Species of oxbow lakes were early or synchronous breeders, and laid a single clutch. 6. Neither the predictions of the habitat templ et concept nor those of the patch dynamics concept were validated. Spe cies traits were not significantly correlated to the axes of the spati al-temporal variability templet. The highest species richness was obse rved in habitat types with the highest temporal variability; in additi on, richness did not peak at highest spatial variability. 7. Deviation from the predictions of the river habitat templet could be explained by: (i) the over-simplification of this model with regard to the evolu tion of the complex life cycles of the amphibians; and (ii) an underes timation of the importance of phylogenetic constraints and the evoluti on of community interactions.