Résumé
Banana breeding programs rely on the use of diploid clones. Intensive breeding efforts can only been invested upon a small number of accessions at any one time. Thus, a rational basis is needed to identify clearly the diploid material of the greatest potential within the natural populations for their use in further breeding. This project aims at obtaining preliminary data on diploid cultivars. Evaluation data include information on basic agronomic characters, fruit characteristics, and resistance to diseases for 200 diploids. Laboratory characterisation of the diploids, through flow cytometry and the use of molecular markers supplements evaluation. The study has been conducted on four locations: CIRAD-FLHOR Guadeloupe for agronomic value and molecular characterisation, WIBDECO St Lucia for resistance to Yellow Sigatoka, CIRAD-FLHOR New Caledonia for resistance to Black Sigatoka and QDPI Australia for resistance to Panama Disease. (Résumé d'auteur)