Lymph node involvement is the most important prognostic factor for patients who have undergone radical surgery for colorectal carcinoma. An accurate examination of the surgical specimens is mandatory for the correct assessment of the lymph node status of the tumor. The risk of understaging is particularly high for patients with tumors classified as Dukes B (TNM stage II). The aim of this study was to determine if a specified minimum number of lymph nodes examined per surgical specimen could have any effect on the prognosis of patients who had undergone radical surgery for Dukes B colorectal cancer. Between 1988 and 1995 a total of 140 patients underwent radical resection of Dukes B colorectal cancer by the same surgeon (C.C.). The relation between clinicopathologic variables and survival was estimated using the Kaplan-Meier method. The Cox proportional hazard regression model was used to identify the variables that can independently influence survival. A median of 12 lymph nodes (range 3-38) was examined per tumor specimen. The 5-year survival rate of Dukes B patients who had had eight or fewer lymph nodes examined after surgery was 54.9%, whereas the survival rate for those who had had nine or more lymph nodes examined was 79.9% (p < 0.001). Cox regression analysis identified the number of lymph nodes as the only independent prognostic factor (p = 0.01). Seventy patients with one to four metastatic lymph nodes (Dukes C patients) who had been operated on during the same period were included in the survival analysis for comparison. The 5-year survival rate of the Dukes B patients with eight or fewer lymph nodes examined was similar to that of the 70 Dukes C patients (54.9% and 51.8%, respectively). Examination of eight or fewer lymph nodes in Dukes B colorectal patients may be considered a high risk factor for missing positive lymph nodes in the surgical specimens. Our results suggest that harvesting and examining a minimum of nine lymph nodes per surgical specimen may be sufficient for reliable staging of lymph node-negative tumors.