Cotton (Gossypium hirsutum L.) is sensitive to water shortage during its establishment phase. In the Sudan-Sahelian zone, a significant part of rainfall events can be lost through runoff, particularly as the rainy season begins. Using residues from previous crops as soil cover is expected to reduce runoff and thus improve water availability to young cotton plants. To test this hypothesis, we set up a trial in Sikasso, Southern Mali, with a two-year rotation of cotton and sorghum (Sorghum bicolor L.). The conventional cropping system (CS) was compared with two no-tillage, mulch-based cropping systems, where cotton was grown on straw residues of a sorghum crop (DMC1) or sorghum/brachiaria (Brachiaria ruziziensis) intercropping (DMC2). Soils in this region are classified as ferruginous types with hydromorphic charac-teristics at depths below 0.7 m (Plinthic Haplustalfs). Measurements included runoff, soil water content, biomass, cotton yield, and soil surface status (i.e. covered, open with visible macropores, or closed with no visible macropores). On DMC systems, the amounts of residues from the preceding crop kept on the soil surface at the beginning of the cotton cycle ranged from 4.4 to 9.3 × 103 kg ha?1 of dry matter, depending on the year and preceding crop. The soil surface status was more open under DMC1 and DMC2 (38 and 62%, respectively) than under CS (30%). Crop residues reduced runoff at the beginning of the rainy season in both years (?26 and ?56 mm between CS and DMC2 in 2006 and 2007, respectively) and increased soil water availability for cotton (+26 mm and +50 mm, respectively). This gain in soil water did not result in higher cotton yields, probably because 2006 and 2007 rainfall amounts were unusually high, with a low gap between water supply and cotton requirements. Row ridges made perpendicular to the slope and after sowing in CS, created an effective obstacle to runoff and increased soil water content. The effects of crop residues on rainwater inf