Because of the high variability of cattle infestation, selective breeding has been proposed to control three-host ticks in Africa in the same way as it has been implemented to control the one-host tick Boophilus microplus in Australia. Interbreeding (divergent selection experiment) between Gudali zebu cattle was performed at the Wakwa Research Centre, Adamawa, Cameroon to assess the feasibility and impact of such a selective breeding programme. Fifty cows (chosen out of 129) and 2 bulls (chosen out of 10) were selected according to their greater or lesser attractiveness for the tick Amblyomma variegatum, assessed by their infestation degree (ID, ratio between individual animal infestation and mean herd infestation). Half of the animals displayed a high ID (H) and the remainder a low ID (L). The presence of a suckling calf had a significant effect on the infestation of its dam, the udder of lactating cows being significantly less infested (P < 0.001) because of a tick removal effect by the calves. Observed infestation data were therefore corrected to allow comparison between lactating and non-lactating cows. The ID of the cows was assessed on five occasions from 1992 to 1994: the consistency of the infestation hierarchy was good overall (P < 0.001) despite high variability seen in some animals which would complicate the selection of cattle with low infestation. Four breeding groups were set up (L × L, L × H, H × L, H × H) and 40 calves were born in 1993 and 1994. There was a correlation (P < 0.02-0.001) between liveweight (or age) and ID when the younger calves of the groups were 6 months old, but the correlation disappeared by the time the calves were all older than 1 year. There was no difference between the average ID of the calves from the four breeding groups, and no correlation was observed between dam ID and calf ID. The average ID of the 10 calves born of the dams with the lowest ID was however significantly lower than that of the 10 calves born of t