Résumé
Groundnut contamination with aflatoxin constitutes a great threat to human and animal health in West Africa, with subsequent negative effects on agricultural production and commercialization. An efficient management of the risk due to aflatoxin contamination in semi-arid zones could be achieved through a better understanding of the interactions between the plant, the soil water status and Aspergillus flavus. For this purpose, two field experiments were conducted in 2002 and 2003 in a split plot experimental design at Bambey, Senegal. Three varieties (55-437, 73-30 and Fleur 11) were submitted to two watering regimes: well watered (WW), and water stressed (STR) during the grain filling phase ie beyond 60 days after sowing. The results showed that a high fraction of transpirable soil water (FTSW > 0.8) or notably a low FTSW (< 0.4) during this phase increased the risk of aflatoxin contamination. For all water regimes, only 55-437 exhibited a low aflatoxin contamination rate compared to Fleur 11 and 73-30. While under conditions of water excess (FTSW >1), 73-30 was more contaminated than Fleur 11. The ability of 55-437 to minimize aflatoxin contamination could be related to its drought responses. In fact, the lower leaf area index (LAI) of 55-437 compared to Fleur 11 resulted in limited water losses under water deficit condition, resulting in better leaf water status and consecutively in higher photosynthetic activity. This drought response of variety 55-437 combined with lower grain size trait allowed a better grain filling and maturation, warranty for a low aflatoxin contamination risk. (Texte intégral)