One thousand two hundred twenty-six cases of colorectal cancer were treated surgically between 1956 and 1978. Seven hundred ninety-eight cases (65.08 per cent) were rectal, 74.3 per cent of which were located extraperitoneally. Dukes' C2 cases and cases with distant metastasis constituted 39.15 per cent, and Dukes' A cases, 9.22 per cent. Resectability rate was 75.1 per cent (77.6 per cent for rectal cancer). Two hundred and four cases (32.96 per cent) of rectal excision were restorative resections. Of the combined excisions for extraperitoneal lesions in females, 84.8 per cent were posterior pelvic exenterations. The overall operative mortality rate was 2.93 per cent. The mortality rate for 921 cases of resection was 1.73 per cent, for rectal resection, 0.8 per cent, and for curative rectal resection, 0.63 per cent. The follow-up rate was 94.13 per cent. The five-and ten-year survival rates for rectal resection were 53.08 +/- 2.29 per cent and 47.65 +/- 2.44 per cent; for curative rectal resection. 66.91 +/- 2.54 per cent and 60.27 +/- 3.03 per cent; and for Dukes' A cases, 98.05 +/- 1.35 per cent and 96.39 +/- 2.13 per cent. The five-and ten-year survival rates for colonic resection were 59.79 +/- 2.04 per cent and 52.18 +/- 3.49 per cent; for curative colonic resection, 72.79 +/- 3.39 per cent and 62.06 +/- 4.17 per cent; and for Dukes' A cases, both 100 per cent. Besides the extent of spread and degree of malignancy of a lesion, the local immunologic reaction of the host is also important in prognosis. The more lymphocytic infiltration in and around the cancer, the more follicular hyperplasia and sinus histiocytosis in regional lymph nodes, the better is the prognosis. The problem of anal preservation in radical resection of rectal cancer and the problem of improvement of results in the treatment of extraperitoneal rectal cancer are discussed in detail.