The objective of this study was to evaluate the possibility of using the fraction of transpirable soil water (FTSW) under field conditions, to analyse genotypic differences in plant responses to soil water deficit. Two years of field experiments were carried out on a sandy soil under sub-sahelian conditions in Senegal. Five cotton cultivars (Gossypium hirsutum L.), with similar phenology but different yield responses to drought, were compared under two irrigation treatments differentiated after flowering. Despite large differences of soil water content between years and irrigation treatments, FTSW was closely related to the predawn leaf water potential measured on the same day. Plant responses to soil water deficit were analysed with leaf water potential, relative water content (RWC), stomatal conductance (gs), and crop water stress index (CWSI) measured during the crop cycle. The relationships between plant variables and FTSW, over two years of measurements and contrasting soil water profiles, were adjusted to typical logistic functions, previously used in other species. Leaf water status, gs and CWSI did not change appreciably until FTSW reached 0.4-0.5. Significant genotypic differences were found in the relationships of RWC and CWSI with FTSW, which allowed the ranking of the five cultivars for dehydration avoidance. The higher dehydration avoidance of one of the cultivars (STAM F) is not linked to stomatal regulation, but probably to osmotic adjustment. Calculation of FTSW from soil water content measurements provided an efficient way to conduct genotypic comparison of plant response to drought in field conditions over two years of contrasted rainfall pattern.