In vitro multiplication is often reported as the main cause for triggering the production of episomal infectious BSV particles in banana hybrid species harbouring some or all of the B (Musa balbisiana) genome, through the activation of BSV endogenous pararetrovirus (EPRV) sequences integrated into the B genome. Nevertheless, mass production of vitroplantlets remains the most widely used method for the diffusion of either Musa germplasm or new improved hybrid species. A better understanding of the effects of in vitro culture on BSV EPRV activation is thus necessary to evaluate the risks of spreading BSV through the wide diffusion of banana plantlets. In order to evaluate this risk, CIRAD and INIBAP are developing a collaborative project aimed at answering the following questions: Does BSV EPRV activation occur through in vitro culture in all inter-specific hybrid species and in " natural " plantain species too? Is there a correlation between the duration of in vitro subculture steps and the percentage of plantlets exhibiting BSV episomal particles ? In order to answer these questions, three different accessions were used: two natural triploid plantains (AAB) - 'Kelong Mekintu' (KM) and 'Black Penkelon' (PK)- and the tetraploid hybrid (AAAB) - CRBP 39. Virus-free suckers were selected and mass propagated using standard in vitro budding methods. During the successive multiplication (proliferation) subcultures, at least 40 shoots were randomly picked and screened for the presence of episomal BSV particles, using standard immuno-capture PCR based detection methods. BSV episomal particles were detected during in vitro culture in both natural plantains and CRBP39 hybrid, with BSV-Ol being the predominantly detected BSV strain . Both natural plantains and CRBP39 displayed similar patterns of activation. Percentages of plantlets indexed positive for BSV-Ol rapidly increased after the first subculture cycles. Depending on cultivars, maximum percentages of BSV-Ol posit