Forty-eight horses with osteochondritic lesions of the femoropatellar joint were treated, 23 of them by an arthrotomy and 25 by arthroscopy. Follow-up information was obtained from either the owner or the referring veterinarian by telephone inquiry. There were no statistical differences between the groups of horses undergoing the two procedures with regard to age, sex, breed, the involvement of one or both limbs, the size of the lesion, and the duration of either the surgery or anaesthesia. However, the horses treated by an arthrotomy spent 14.5 days in hospital after the operation whereas those treated by arthroscopy spent only three days. Nineteen of the horses treated by arthroscopy were able to pursue athletic activities and a further five were expected to enter training in the future, whereas only 12 of the horses treated by an arthrotomy were suitable for their intended use (P < 0.05). When the severity of the lesions was considered, the success rate after an arthrotomy deteriorated with increasing severity, whereas the success rate after arthroscopy remained at a high level. Fourteen of the 16 owners of horses treated by an arthrotomy who replied to the question considered that the horses had a palpable scar, a femoropatellar effusion or both, whereas only two of the horses treated by arthroscopy (8 per cent) were considered to have had a poor cosmetic outcome (P < 0.05).