The experimental manipulation of separate, but originally identical, populations of Ixodes ricinus (L.) by applying three conventional tick control measures in different enclosures on naturally infested moorland in Wales allowed the elements of the tick-host interaction to be analysed quantitatively and the effectiveness of the control methods to be compared. From the relationship between the sheep stocking density and the numbers of questing ticks picked up by fortnightly blanket-dragging in each enclosure, the death rate of ticks during their activity season and the rate of contact between sheep and ticks were calculated. From this, it was possible to investigate the effect of different stocking densities on the feeding success of ticks. A major factor determining the much lower contact rate for larvae than for nymphs was the different spatial distribution of questing ticks, clumped for larvae and random for nymphs. The non-random use by sheep of the three different vegetation zones in the paddock resulted in the highest contact rate between sheep and ticks in the pasture area, but tick survival was apparently highest in the bog area. Combining these factors resulted in the prediction that the bracken area was the least favourable habitat for ticks. In the two enclosures where the sheep were not treated with acaricide the mean tick loads on the sheep were similar, but the lower overall use of the pasture by the sheep in the low stocking density enclosure (2/ha) resulted in slightly lower tick loads there compared with those in the high stocking density enclosure (4/ha). The numbers of ticks counted in the second year showed that pasture spelling was the most effective control method, acaricide treatment was less effective, and the benefits of halving the stocking density were marginal. The implications of these results for control strategies are discussed.