Between 30 and 70 per cent of horses at 18 stables had their anthelmintic treatments withdrawn for a year. Three of the stables (group 1) had used no anthelmintics for many years, 10 (group 2) had treated their horses with anthelmintics approximately twice a year, and five (group 3) had used anthelmintics five to seven times a year. A total of 1068 samples of faeces were collected from the 267 untreated horses and faecal egg counts were carried out four times during the year. Overall 50·2 per cent of the samples contained less than 200 epg. When the four counts from individual horses were collated, 35·6 per cent of the horses were low egg-shedders (LES). However, the previous control programme had a marked effect on both the mean log10 epg and the proportion of horses in the stables that were LES. The proportions of LES among the horses in the stables in groups 1 and 2 were similar, although the log10 epg was significantly lower in group 2. There was a markedly higher proportion of LES and a lower mean log10 epg in the untreated horses at the stables in group 3 than in either of the other groups. The horses at the stables in groups 2 and 3 that remained on the anthelmintic programme previously used at the stables had lower egg counts than the untreated horses at the same stables, but the difference was usually not statistically significant. At the stables in group 2 the proportions of horses that were LES in the treated and untreated groups were not significantly different. At the stables in group 3 the proportion of LES was significantly lower in the treated horses than in the untreated horses.