Of the four species of Bulinus found on Madagascar, three species: B. obtusispira, B. liratus and B. bavayi are endemic while the fourth, B. forskalii, is probably a recent introduction from the African mainland. The evolutionary relationships of these species with Bulinus species from Africa were studied by phylogenetic analysis of DNA sequence variation at two mitochondrial loci: cytochrome oxidase subunit I (COI) and large ribosomal subunit (LSU) or 16S. The observed levels of nucleotide divergence within Bulinus were substantial but may underestimate the true levels as there was evidence of 'saturation' of transitional substitutions at both loci. A putative secondary structure model for the sequenced segment of the 16S was developed. Subsequent phylogenetic analysis using transversional changes only for both loci, showed that there were contrasting levels of divergence within the four species groups. B. obtusispira was consistently placed within the B. africanus group, appearing ancestral to this group and was closest to the basal node within Bulinus. Together with B. bavayi, the two species appear to have been isolated on Madagascar for a long time, contrasting with both B. liratus and B. forskalii that appear more recent colonisers; however, estimate of exact times of divergence is problematic. A PCR-RFLP assay was developed to enable identification and discrimination of B. obtusispira and B. liratus using discriminatory variation within the COI. To enable population genetic analysis within B. obtusispira, microsatellite markers were developed using an enrichment method and 8 primer pairs are reported. Laboratory infection experiments using Madasgacan S. haematobium from the Mahabo area showed that certain populations of B. obtusispira, B. liratus and B. bavayi were compatible.

P. 2005. The influence of genetic factors and population dynamics on the mating system of the hermaphroditic freshwater snail Biomphalaria pfeifferi . Á/ Oikos 108: 283 Á/296.Although self-fertilization and its evolutionary consequences have been widely studied, the relative influence of genetic and environmental factors on the determination of mixed-mating systems remains poorly known. In 1999 and 2000, we surveyed the mating system, the population dynamics and some life-history traits of four populations of the freshwater snail Biomphalaria pfeifferi , the major intermediate host of Schistosoma mansoni in Africa, in two areas of Madagascar (Itasy and Antananarivo). We confirmed that B. pfeifferi is a predominant selfer, with selfing rates ranging between 80 and 100%. Temporal and geographical variation of the selfing rate was observed at both local and large spatial scale. Historical processes of colonization and invasion of B. pfeifferi in Madagascar could explain the geographical variation of the mating system observed at regional scale. Pure selfing has probably evolved in the two populations of Antananarivo area as a reproductive assurance strategy in a metapopulation where extinction is frequent and migration rare. The reproductive assurance hypothesis does not explain the spatio-temporal mating system variations observed in Itasy area. However genetic factors including inbreeding depression-the expression of which can be environmentally mediated-and metapopulation dynamics could influence the mating system in both populations sampled in Itasy and lead to different levels of evolutionary stable intermediate selfing rate in this region. Our results therefore highlight the influence of environmental heterogeneity and stochasticity on mating system.

Recent characterization of nuclear ribosomal small subunit (SSU) genes has shown that variant nucleotides within this region could be useful for species and species group identi®cation within the genus Lymnaea (Gastropoda: Lymnaeidae). This study aimed to characterize a range of populations of Lymnaea natalensis Krauss, 1848 on Madagascar, and addressed two related questions. First, is there any evidence of intraspeci®c variation of the SSU and, if so, what might be its signi®cance? Secondly, might this variation jeopardize the use of SSU for lymnaeid taxonomy and phylogeny? Lymnaea natalensis (n = 212) was collected from 17 sampling localities, spanning the northern and southern ends of the island. Variation within a selected region of the SSU known to vary between species, the V1 and V2, was assayed by polymerase chain reaction (PCR) linked restriction fragment length polymorphism (RFLP) and denaturing gradient gel electrophoresis (DGGE) analysis. The PCR-RFLP pro®les indicated a striking dimorphism across populations at two restriction site loci (CfoI & MspI) within the E10-1 helix of the V2 region. The observed RFLP variation was con®rmed by direct sequencing and by genomic digestion with subsequent hybridization. Putative heterozygotes were also encountered and in these individuals the SSU arrays composed of two distinct types approximately 1% divergent. A severe departure from Hardy±Weinberg equilibrium with a highly statistically signi®cant ( P < 10 -5 ) heterozygote de®ciency was found and genetic variation among populations was highly structured (F st = 0.53). The geographic distribution of the variants was mapped, revealing that one variant was restricted to higher, predominately colder environments and was thought to be an adaptation. The molecular basis of the SSU variation was caused by single nucleotide polymorphisms (SNPs). To test for the possibility of cryptic taxa, an analysis of individuals representative of the SSU variant types with isoenzyme analysis (ISA), randomly ampli®ed polymorphic DNA (RAPDs) and PCR-RFLP analysis of the ribosomal Internal Transcribed Spacer (ITS) was performed. Little variation was revealed and none that correlated to the groups suggested by SSU, con®rming that the SSU variation was intraspeci®c. The levels of intraspeci®c divergence of the V1 and V2 within Lymnaea were not appreciably different (1%) from interspeci®c and would therefore question the validity of these data for lymnaeid taxonomy and phylogeny.