Selection of diploid potato genotypes that produce high frequencies of 2n-pollen grains in different environments is an important requirement for analytic breeding as well as for true potato seed production. In this context, quantitative estimates of 2n-pollen formation were made in two type of diploid potato populations . One population set consisted of the progeny from diallel crosses of five well-known diploid potato clones (A = US-W 5293 .3, B = US-W 5295 .7, C = US-W 5337 .3, D = US-W 7589 .2 and E = 772102-37) . Another population set was obtained by pollinating three genotypes of Solanum phureja, IvP35, IvP48 and IvP101 with clones A, C and E . It was observed that the behaviour of the parental clones with respect to 2n-pollen grains formation in different environments, such as in Europe and China, was stable . The correlation coefficient between means of 2n-pollen grains frequency of combinations and percentage of genotypes with more than 5% 2n-pollen grains in a given combination were estimated and that between mean and maximum frequency of 2n-pollen grains formation in each combination . The results showed that the former coefficients were significant whereas the latter were not . The combining ability analysis revealed variation in 2n-pollen grains formation between families . Specific combining ability (SCA) effects contributed much more than general combining ability (GCA) to the variation in 2n-pollen grains formation . The broad-sense heritability of 2n-pollen grains formation was estimated to vary between 0 .45 and 0 .53 and the narrow-sense heritability was calculated to be 0 .15 and 0 .20 in the two types of populations . This means that additive variance was relatively low . Two to four genes were estimated to control the inheritance of 2n-pollen grains formation .