Résumé
Citrus are most important fruit crop worldwide. Seed1essness is a key characteristic for the fresh fruit market and the development of triploid hybrids is one strategy developed by several groups over the world. Indeed, triploid hybrids are generally sterile and produce seedless fruits and do not pollinate other varieties. Triploid citlllS hybrids can be obtained by several strategies, including hybridization betwe6n diploid parents. Mechanism of 2n gamete formation and its implication on parental heterozygosity restitution is a main parameter determining the genetic and phenotypic structure of the triploid population. In the case of Citrus it has been shown that the 2n gametes are of maternaI origin. It has been proposed that the origin of 2n gametes is from the second division restitution (SDR) in Clementines and from the first division restitution (FDR) in sweet oranges. No data is available for other genotypes and particularly 'Fortune', a mandarin hybrid producing very high rate of triploids in 2x x 2x crosses and massively used to create triploid progenies. The aim of this work was to analyse the mechanism of 2n gamete formation in 'Fortune' mandarin genotype. One hundred and five triploid hybrids from the crosses between 'Fortune' as female diploid parent and 'Murcott' or 'Mandarino ComÛIl' as male diploid parents were genotyped for twenty- four codominant molecular SSRs (Simple Sequence Repeat) markers using a capillary genetic fragment analyzer. Estimation of ailelic doses from relative peaks area ailowed inferring the female and male gamete structures and thus the heterozygosity restitution in the 2n gametes.; this demonstrated that ail triploid arise from 2n megaspores. The unimodal distribution of heterozygosity restitution in the 2n megaspores among the analyzed genotypes suggests that an these 2n gametes arise from a same mechanism. Restitution of maternaI heterozygosity for the used rnarkers makes suppose that underlying mechanism in the 2n gamete formation is SDR. Indeed there are six markers with less than 50% of maternaI heterozygosity restitution, which is incompatible with FDR hypothesis. SDR hypothesis is coherent with the results published in case of the clementine, which is one of the parents of the 'Fortune' variety. Under this hypothesis, the relatively high global heterozygosity restitution level (60,95%) should indicate that a majority of the analyzed markers are far frorn the centromeres. This genetic struc turation will soon be confronted with phenotypic variability and compared with structuration obtained with other triploid creation strategies such as 2x x 4x hybridization. (Texte intégral)