A process of continuous selection and testing for fusarium wilt resistance of cotton is described and its efficiency discussed. It essentially consists in individual selection, progeny tests, mass selection in the best progenies and testing of the corresponding MS in the third year. Individual selections may come from the breeding programs as well as from crop fields and artificial inoculation in green-houses. All the tests are made in field conditions. Data of 1969 to 1971, obtained in two cycles of selection, showed that the process described was efficient in increasing both tolerance and resistance to wilt. The increase was larger in the case of originally susceptible material than in that of resistant one. This difference was smaller in years of more intense wilt. Also the actual values of increases were smaller in such years. The results suggest that fusarium wilt resistance, in cotton, may be a matter of concentration of minor genes in the population, rather than of the presence of a few major genes.