Habitat selection of Anurans in small ponds around the Victoria Lake, Kenya

The shores of Lake Victoria have been largely altered by human activities, mainly from subsistence agriculture. In the region of the Yala swamp (Kenya), the original marsh vegetation was reduced to favour plantations of vegetables for domestic consumption (cassaba, cabbage, seet potatoe, etc.). An ecological survey on the amphibians was conducted in order to compare the species assemblage in natural and artificial ponds and to assess the factors influencing the species distribution. Amphibians were studied during February and March 2004 between the eastern shore of Lake Victoria and a small satellite basin, Lake Sare (00°01’-00°03’S / 34°01’-34°03’E; Kisumu district). Amphibians were sampled at 15 different (10 artificial and 5 natural) ponds that differed in terms of several factors: hydro-morphology (size, depth, bottom type, turbidity), vegetation (aquatic vegetation, surrounding vegetation, land use), and location (distance from lakes and from papyrus stands). The presence of amphibians within and around the ponds was assessed by Visual Encounter Survey method during diurnal and nocturnal samplings. PCA and Cluster analysis were performed to detect similarities among ponds and to assess specific pond occupancy patterns. The following species were collected in the study area: Xenopus laevis, Phrynobatrachus sp., Ptychadena mascareniensis, Hemisus marmoratus, Hoplobatrachus occipitalis, Kassina senegalensis and Bufo sp. Due to the exiguous number of records, K. senegalensis and Bufo sp. were excluded from the analysis. The five amphibian species showed different patterns of occupancy in using the 15 ponds. X. laevis was found only in artificial ponds, whereas Phrynobatrachus and Ptychadena species were observed in both artificial and natural ponds. H. marmoratus and H. occipitalis showed an exclusive preference towards natural ponds. The PCA allowed to spatially arrange the amphibian species according to three main groups: two groups were monospecific (i.e., X. laevis and Phrynobatrachus sp.) and another group consisted of the remaining three species, i.e. P. mascareniensis, H. occipitalis and H. marmoratus. The artificial ponds were populated by a lower number of species than the natural ones. Based on hydro-morphological, vegetational and location parameters a UPGMA dendrogram was generated which showed two main clusters, consisting respectively of artificial and natural ponds. A further cluster analysis, based on data of amphibians pond occupancy, classified the ponds in the same groups generated by the previous analysis, thus evidencing a high relationship between the degree of human pond exploitation and the fauna composition. The impact of human activities on small basins had a significant effect in terms of ponds morphology, vegetation into and surrounding the water bodies, and consequently of amphibians composition. The artificial ponds and those modified for irriguous purposes showed a reduced species diversity and a clearly different species composition from natural ponds. Some species (i.e. X. laevis) could gain advantage from the changes made by humans, because of their euriecity and their capability to tolerate low oxygen levels and high organic load that are conditions often present in the ponds utilized for watering purpose.