Isofemale lines of ladybirds produced from a selected stock varied widely in female preference for melanic males. An attempt to repeat these results failed. The ladybirds used in the repeat experiment were collected from a wild population in which melanic frequencies had dropped from 35 to 10 per cent in 3 years. In this repeat experiment, the overall mating rate was twice that of the original experiment. The failure of the repeat experiment can be attributed either to a loss of preference or to the preference not being expressed at high mating rates. New lines were set up from a sample collected in Glasgow where high levels of preferential mating occur in the wild population. Mating tests on these lines strongly corroborated the results of the original isofemale line experiment: some lines showed high preference, others intermediate or no preference. The overall mating rate was low -one-quarter of that in the unsuccessful repeat experiment. Samples collected from areas high in melanics showed high preference. Lower levels of preference were found where melanics are less common. Where no melanics occur, preference is absent. These results provide further evidence of genetic preference in ladybirds.