A set of 12 cotton genotypes were evaluated during the main growing seasons, from 1985 to 1990 at Abobo , which is characterised by low elevation (530 meters above sea level) and a sub-humid climate. The same genotypes were tested in all years. Combined analysis of yield data indicated significant differences among genotypes and environments. Genotype-environment interaction was highly significant, indicating differential performance of the cultivars in an array of environments. For each cultivar a linear regression of yield on the mean yield of all cultivars for each year was computed to measure cultivar adaptation. The cultivars with the highest mean yield exhibited a similar degree of adaptation to different environments with regression coefficient close to 1.0. For example, the breeding lines, Acala 1517Br x AMS1 (74)#11, A-333-57, showed average stability with regression coefficient (b) approximating 1.0, and all produced above average yield in all years indicating general adaptability. On the other hand, Albar 637 and Reba B-50 also had regression coefficient (b) approximating 1.0, but they consistently produced below average yields in different environments which indicated that they were poorly adapted to different environments. The check cultivar ( L-299-10), typical of cultivars specifically adapted to high yielding environments, was characterised by the highest regression coefficient (b=1.2), while Lafrego bract 2, with regression coefficient (b=0.85), typified cultivars specifically adapted to low yield. (African Crop Science Journal: 2000, 8(2): 153-158)