The introduction of new techniques has enabled more patients with low rectal carcinoma to undergo sphincter-saving resection (SSR). This policy, however, has provoked controversy concerning the adequacy of excision and fear of increased rates of recurrence compared with abdominoperineal resection (APER). We have therefore compared our results of SSR for low and mid-rectal cancers performed between 1978 and mid 1982 with those obtained with APER before this period. One hundred and fifty-nine patients presented with tumours between 3 and 12 cm from the anal verge; 153 (96 per cent) underwent resection, of whom 46 (30 per cent) had disseminated disease. One hundred (68 per cent) underwent SSR, 33 (22 per cent) had an APER and 16 (10 per cent) had a local procedure. Operative mortality was 4 per cent following APER and 7 per cent after SSR (n.s.). Patients were followed for a mean of 4.6 years (range 2-6 years), only four (2.5 per cent) being lost to follow-up. The incidence of recurrence after radical SSR (n = 74) was compared with the historical control group which underwent radical APER, the two groups being matched for Dukes' stage and height of the lesion. After 2 years, local recurrence cumulative rates were 13.6 per cent after SSR and 18.8 per cent after APER. Distant recurrence rates were 14.5 and 20 per cent respectively. Ten patients (13.5 per cent) died of their disease within 2 years of a radical SSR, 15 (15 per cent) after radical APER. Corrected 5-year survival and recurrence rates were similar. Thus, SSR does not appear to carry an increased risk of recurrent disease compared with APER after an equivalent follow-up period.