A reconnaissance survey and participatory varietal selection trials (PVS) were conducted in four major pearl millet-growing countries of the Sahel between 2001 and 2003. The studies aimed to identify farmers' preferences in improved pearl millet varieties, increase awareness, test new varieties and enhance farmers' access to the improved varieties. Farmers selected five out of 10 tested varieties, with preferred characteristics, namely, maturity cycles of 80–90 d in the Sahel and 90–100 d in the Sudanian agro-ecozones, acceptable grain yield, compact and long (30–100 cm) panicles, a large number of tillers with panicles, adaptation and an acceptable taste. Farmers indicated that their local varieties were of superior adaptation and taste. They mentioned that hindrances to uptake and sustained use of improved varieties were due to lack of awareness, traditional values, seed unavailability, early maturity, bird damage and lack of fertilizer. The strong genotype × environment interactions in the Sahel suggests that breeding should be directed towards producing varieties adapted to specific zones rather than for wide adaptation. Notably, since farmers often cultivate pearl millet without any soil amendments, it may be advisable to disseminate varieties as a package (with fertilizer and agronomic instructions) rather than as varieties alone in a PVS programme, in order to achieve the full potential of improved varieties. The PVS trials are synergistic to plant breeding in identifying varieties suitable for harsh environments, which are difficult to duplicate in the research station. However, in the absence of formal distribution seed systems in the trial countries, village- or community-based seed production of varieties selected by farmers appears critical to the sustainable adoption of selected varieties.