Seedlessness is a major criterion for the citrus fresh fruit market. Triploidy breeding therefore appears very promising for the selection of new citrus cultivars. CIRAD bas developed a new method for triploid synthesis, achieved by somatic hybridisation between diploid and haploid lines. This method makes it possible to add, without recombination, a haploid genome to the whole diploid genome of high organoleptic quality cultivars. Somatic hybridisation is done by electrofusion of protoplasts isolated from haploid embryogenic callus lines of clementine, obtained from haploid embryos generated by induced gynogenesis, and callus lines or mesophyll protoplasts of nine diploid cultivars. Regenerated plants or calli are studied by flow cytometry and molecular markers in order to assess their ploidy level and their nuclear genome origins, respectively. Triploid hybrids, as well as some tetraploid and pentaploid hybrids, have been obtained for each combination. This ploidy polymorphisrn could be due to the presence, detected by flow cytometry, of diploid and triploid cells in the haploid callus. More than 200 polyploid hybrid plants are currently being propagated in greenhouses. Triploid hybrids should be exploited directly after evaluation, while tetraploid hybrids will join the pool of allotetraploids for further sexual breeding with diploids.