Résumé
Seventy-seven wild sunflower accessions from USA were compared in France (Mauguio) for 13 quantitative characters using multivariate methods, including clustering, principal component and canonical discriminate analyzes to assess the patterns of morphological and climatic variation. Geographic and climatic data from their sampled sites such as latitude, longitude, elevation, rain fall, temperature, number of rain fall days, temperature range, and sunniness were also analyzed and we sought for correlations of climatic variables and morphological traits. Climatic data clustered the accessions into four groups. Principal component analysis was used to summarize the characteristic habits of annual wild sunflower sampled sites in terms of geography and climates, thus we characterized the ecogeographical profiles of wild sunflowers. As we studied the traits in a common environment we speculated whether local adaptation patterns are sustained by our data. Indeed traits shared by accessions that belong to climatic clusters are suggested to be due to local adaptation