In Maritime Guinea, the interpenetration of upland, lowland and mangrove rice growing ecosystems has found expression in the cohabitation of the two rice cultivated species. Recent changes in cropping practices may lead to the replacement of local varieties by modern high-yielding varieties. In the framework build-up of a strategy for the preservation of local varieties, we analysed the extent, the organisation and the specificities of the rice genetic diversity. One hundred seventy accessions collected in farmers' fields were genotyped with 11 SSR markers and phenotyped with 26 morpho-physiologic descriptors. The general organisation of rice genetic diversity in Maritime Guinea, and its tight relationship with the rice growing ecosystems were similar to the one observed elsewhere. The two major subspecies of O sativa--indica and tropical japonic--as well as the two major ecotypes of O. glaberrima--''floating'' and ''upright''--were present. Moreover, an original genetic compartment was detected, highlighting the occurrence of glaberrima · sativa hybridisation. Allelic diversity was found to be comparable to that noted worldwide for indica and japonica groups of O. sativa, but not as large for O. glaberrima. Given its extent, its original compartment, and its potential for inter-specific and inter-subspecific indica · japonica recombination, the preservation of rice genetic diversity in Maritime Guinea deserves special attention.

Greater insight into the dynamics of genetic resources of crop plants is needed in order to pinpoint detrimental evolutionary patterns and draw up conservation priorities. Temporal evolution of rice genetic diversity was monitored in Maritime Guinea where subsistence-oriented agriculture prevails. Diachronic comparison was performed between samples collected in six villages during the 1979/ 1982-period and in 2003, based on the names and number of varieties inventoried and the polymorphism of microsatellite markers. The number of varieties appeared not to be comparable between the two dates, due to differences in the collection methods. The varietal composition had evolved very substantially between the two collection dates. Many long-duration varieties present in 1979/1982 had been abandoned and several improved varieties had been introduced. The mean number of alleles per locus and per accession was significantly higher in accessions collected in 2003. Pairwise comparisons of the mean number of alleles per locus in 1979/1982-2003 homonymous accession pairs indicated higher intraaccession diversity for the 2003 collections. Genetic differentiation, measured with the F ST values, was very high and significant for more than 80% of these pairs of accessions. The overall genetic differentiation between accessions from the two collections dates was also significant. Significant changes were also observed for allelic composition. However, alleles specific of each collection date had much lower frequency, compared to alleles common to the two collection dates. These results suggest that rice genetic diversity in Maritime Guinea has been maintained or even enhanced. Old collections of crop genetic resources are often not exhaustive enough to undertake perfect diachronic comparison. New methods to utilize this historical data for diversity monitoring are needed.

Rice genetic diversity partitioning between farms, varieties and, within-variety diversity, were analysed in two villages of Maritime Guinea with contrasted agroecological conditions. One thousand and two hundred individual plants belonging to 45 accessions collected in eight farms were genotyped using 10 SSR markers. The molecular variance was evenly shared between and within accessions, while the farm effect was almost nil. Local varieties had a multi-line genetic structure. The number of multilocus genotypes was proportional to the utilisation rate of the variety in the village. The F ST values between different accessions of each variety were significant which indicated low genetic consistency in the variety names. This varietal structure could mainly be explained by the migration phenomenon and the high varietal turnover. Compared to allelic diversity, multilocus genotypic diversity seemed to be the most suitable indicator of the quantitative distribution of diversity at different management scales (accession, farm and village). The within-and between-farm F ST values were in the same order of magnitude. The within-farm diversity was not farm-specific but quantitatively high, i.e. up to 50% of the total genotypic diversity of a given village. Given the relative importance of the within-variety diversity, the in situ approach stands out as the most effective solution. As farms do not host specific diversity the in situ approach could be implemented by working with a small number of farms.

Genetic resource conservation is widely acknowledged as important. The implementation of conservation requires an insight into the distribution of genetic diversity at the scale of small regions or villages. We present an analysis of rice diversity at such a scale, in a region where traditional farming still prevails. Regional allelic diversity was comparable to that noted worldwide for Asian rice (Oryza sativa), but not as high for African rice (O. glaberrima). Each village pooled more than half of the regional allelic diversity. Genetic differentiation between varieties from the same village accounted for 70% of the regional variation. The differentiation associated with lowland and upland rice-growing ecosystems was 23%, while that associated with differences between villages within the same ecosystem was 7%. In the upland ecosystem, geographical distance had a significant effect on the F ST between pairs of villages. In the lowland ecosystem, differences in soil salinity between villages affected F ST . Genetic diversity within a single village may have up to three components: an ancient glaberrima component shared with neighbouring or ethnically related villages; a relatively ancient sativa component which was hardly or no longer shared with other villages due to local differentiation; and a recently introduced sativa component shared with other villages. Genetic resource conservation could be achieved, in terms of allelic diversity, through stratified sampling according to described genetic differentiation factors, whereas current farming systems must be preserved to ensure conservation of the diversity of allelic associations.

We carried out a nationwide survey of elephant Loxodonta africana in Guinea Bissau, a small West African country for which records of elephant are limited. We also investigated parts of western Guinea along the border with Guinea Bissau likely to harbour a transboundary elephant population. Standardized interviews with hunters were held in 110 villages in Guinea Bissau and 60 villages in Guinea, and field surveys were carried out to validate interviewee responses. Results suggest that elephants are mainly restricted to an area between the Corubal River (Guinea Bissau) and the Kogon River (Guinea) and that elephants occur only seasonally in Guinea. Based on the number, geographical localization and interpretation of observed tracks, our estimate of the minimum number of elephants in Guinea Bissau is 4-10 animals. We did not observe any signs of young elephants. The most immediate threat to elephants is a road scheme between Guinea Bissau and Guinea that cuts through elephant range. The future of elephants in this region depends on the capacity of the two countries to manage their common elephant population jointly. In particular, the creation of a transboundary park is urgently needed.