Screwworm (Cochliomyia hominivorax) case incidence in the southwestern USA suggests a failure in the sterile male method of suppression in 1972–1976. Estimates of sterile mating frequencies from 1954 to 1981 in Curacao, Florida, Texas, Tamaulipas, Chiapas and Veracruz were examined for evidence of premating isolation between released and wild flies. Sterile fly release indices and native fly population densities accounted for 67% of the variance in the sterile mating frequencies. Analysis of covariance was used to correct class variables of fly strain and region for the continuous variables. Statistically significant heterogeneity according to region was detected. No significant strain effects were detected. The data do not support the idea that the prolonged screwworm epizootic was related to genetically-determined factors of strain deterioration or assortative mating.