The aim of this study was to determine the efficacy and long term results of straight colo-anal anastomosis (CAA) after resection for rectal carcinoma as described by Parks. From January 1986 to June 1989, 40 patients underwent this operation: 27 men and 13 women with a mean age of 63.5 years (range 37-81). In 36 cases, the indication was for carcinoma of mid and low rectum and in 4 cases for carcinoma of the upper rectum associated with a low rectal benign tumour (3 Dukes A, 19 Dukes B, 13 Dukes C, and 5 Dukes D). A diverting colostomy was constructed in all cases. Operative mortality was one patient (2.5 per cent) by pulmonary embolism. Anastomotic dehiscence occurred in four patients. None of these patients required reoperation and all colostomies have been closed. 6 patients presented a local recurrence (15.4 per cent) 6 to 34 months after CAA, of whom two were treated by abdomino-perineal resection. 5 patients died 6 to 34 months after CAA from local recurrence (2 cases) or distant metastasis (3 cases) and one patient has liver disease. All others patients are alive free of disease with a mean follow-up of 21.7 months (range 3-46 months). Actuarial survival is 77 per cent at 40 months. Functional results were assessed in the 26 patients followed up more than one year. The mean stool frequency was 2.4 per day (range 0, 3-6). All patients are continent, with a good discrimination gas-stool. 4 patients (15.4 per cent) suffer from soiling, 5 (19 per cent) from stool frequency, and 2 (7.7 per cent) from urgency. In conclusion, CAA is a good alternative of abdominoperineal resection for some mid and low rectal carcinomas. Functional results might be improved by the construction of a colonic reservoir.