Staging of malignant melanoma now relies routinely on the sentinel node (SN) technique. On average, 2.1 SNs are removed per patient. Nevertheless, despite the success of the SN technique, approximately 10% of patients with negative SNs experience metastatic recurrence. Because a prior theoretical analysis using Poisson and Bayes probability models suggested that limited sampling of SNs could cause false-negative results, we undertook this study to see whether the subset of patients with negative SN and only 1 or 2 nodes examined have a shorter time to recurrence than patients with 3 or more nodes examined and found to be negative. Our study cases comprised 178 melanoma cases with SN biopsy: positive SN, 47; negative SN and fewer than 3 nodes examined, 68; and negative SN and more than 2 nodes examined, 63. Patients with negative SNs and fewer than 3 examined had disease-free survival intermediate between patients with positive SNs and those with negative SNs and more than 2 examined (P = .013). These results suggest that among patients with negative SNs, those with fewer than 3 nodes examined have greater risk for recurrence.