A prospective study of postoperative wounds was carried out in West Dorset to determine the incidence of infection, describe the time distribution of presentation before and after discharge from hospital and identify possible contributory factors. There were 702 consecutive patients admitted to the study (600 in-patients and 102 day cases). Fifty one became infected (47 in-patients and 4 day cases), corresponding to an overall infection rate of 7.3%. Over 50% of infections presented during the first week after operation, and almost 90% were diagnosed within 2 weeks of surgery Twenty-eight (55%) wounds that became infected presented after hospital discharge. Of 23 specific aetiological variables studied, four (age, preoperative stay, shaving and the surgeon) were shown to have a statistically significant association with the development of wound infection. A strong association between the individual surgeon and the development of a wound infection was demonstrated and this supports the need for routine surgical audit.