Twenty-seven sesame varieties from Sudan and U. S. A. were grown in replicated trials in the Sudan central rainlands from 1979 to 1982. Regression coefficients were calculated giving response to environmental indices based on location and year mean yields. Entries with high mean yields generally had regression coefficients exceeding unity and were equal to or better than entries having smaller response even in poor environments. Few lines were identified as desirable varieties which combined relatively high yield and stability of performance, thereby deserving a place in both commercial production and future breeding programmes.