&NA;The experience of Karolinska Hospital with colorectal carcinoma over a 30‐year period, 1950 through 1979, was reviewed and analyzed. Altogether, 1061 patients were seen; 69 per cent were resected with the intention of cure. The proportion of curative resections did not change with time. Operative mortality decreased from 12 to slightly over 1 per cent. Five‐year survival rates were 34 per cent overall and 47 per cent in curable cases. After correction for operative mortality, the latter rate increased from 47 to 58 per cent over the study period (no significance). Sex and age were not correlated with survival. Of tumors resected for cure, 28 per cent were Dukes' A, 39 per cent Dukes' B, and 33 per cent Dukes' C. Corresponding five‐year survival rates were 82, 44, and 21 per cent. Though the proportion of Dukes' A tumors increased from 23 to 33 per cent over the 30‐year interval, the impact on overall prognosis was insignificant. A possible overdiagnosis of Dukes' A is discussed. The proportion of sphinctersaving procedures for rectal carcinoma increased moderately; excisional and restorative procedures were equal with respect to mortality and survival. Some patients underwent local excisions for Dukes' A rectal carcinomas with a close to 90 per cent five‐year survival.