In a prospective study of 197 patients with resected colon carcinoma treated between 1974 and 1985, we explored the relationships between pathologic parameters, and the effect of the latter on survival, to identify the parameter whose systematic measurement would improve the predictive capacity of pathologic staging. Prognostic characteristics were studied by univariate analysis. The results showed significant relationships between the location and number of lymph nodes involved, blood vessel invasion, depth of tumor penetration, and metastases. The five-year survival rates were 45 percent and 17 percent (P < 0.001) for patients without and with apical lymph node involvement, respectively, and 44 percent and 6 percent (P < 0.05) for those with four or less nodes involved and more than four involved, respectively. Among the patients treated by incomplete resection, the respective survival rates of those resected for metastases and of those resected for apical lymph node involvement did not differ significantly. We conclude that the involvement of apical lymph nodes has a significant effect on prognosis and suggest systematic pathologic examination of these nodes to allow simpler and more reproducible selection of patients for treatment by incomplete resection who are at high risk of disease-related death.