Predictions for the use of autosomal translocations for pest control have been made on the assumptions that translocation heterozygotes are semi-sterile and translocation homozygotes are fully fertile and have viabilities which are either equal to or slightly less than that of the wild type. These parameters were assessed in controlled crosses using a translocation in Drosophila and they showed good agreement with expectation. It was expected that in a mixed population of such translocations and wild types, the karyotype at the higher frequency would be favoured by selection. Two cage experiments were initiated with mixed populations of translocation homozygotes and wild types at a frequency of 9 : 1 in favour of the translocation. Contrary to expectation, the translocation was eliminated from both populations. During this process, there was a reduction in population fertility. Computer studies showed that the results were consistent with a reduction in fitness of the translocation homozygote to about 0.5, i.e. about the same as the semi-sterile heterozygote, so that negative heterosis did not exist and hence frequency dependent selection could not occur.