Four privately owned herds (25 to 49 animals per herd) were used in a five-year trial designed to evaluate the effect of anthelmintic schedules on the incidence of colic. These herds had been treated bi-monthly with non-ivermectin, non-benzimidazole drugs for two years before the trial. Prior parasitological studies showed that they had substantial pre-treatment faecal egg counts (900 to 2200 eggs per gramme), and that they were infected with benzimidazole-resistant cyathostomes. In Years 1 and 2 of the trial, all herds (A, B, C, D) were treated bi-monthly with non-ivermectin anthelmintics (Schedule I). In Years 3 to 5, Herd A received monthly non-ivermectin anthelmintics (Schedule II) and Herd C was treated with bi-monthly ivermectin (Schedule III). Herd B was treated with Schedule II in Years 3 and 4 and Schedule III in Year 5. Herd D was maintained on Schedule I throughout the study period. In each herd, the incidence of colic, while on Schedule I, was compared to the incidence while on Schedule II or III. The risk of colic for horses on Schedule I was 2.61 to 13.04 times that of the same horses while on Schedule II and 2.27 to 9.64 times that of the same horses while on Schedule III. In Herd D, treated according to Schedule I for five years, the incidence of colic did not vary significantly throughout the study period. More effective anthelmintic treatment schedules decreased the incidence of colic in these herds, supporting the conclusion that bimonthly treatment with non-ivermectin drugs may not maximise horses' health in all management conditions. Further, the data suggest that a high proportion of colics seen in this study had a parasitic aetiology, and that cyathostomes, rather than large strongyles, were the primary cause of disease.