Axillary clearance in node-negative breast cancer patients is performed only for staging and prognostic purposes. The sentinel node concept may provide an alternative conservative approach for these patients. This paper reports on the learning experience with lymphatic mapping involving the use of patent blue dye for the identification of sentinel lymph nodes (SLNs), followed by axillary dissection. The histopathology of the SLNs included serial sectioning and immunostaining for cytokeratin and epithelial membrane antigen, the remaining nodes being processed as usual. Of the 70 mapping procedures, 58 were successful; the surgical performance revealed a well defined learning period. The mean diameter of the successfully mapped tumors was 2.4 cm (ranging from in situ carcinoma to 4.8-cm invasive cancer). The mean numbers of SLNs and non-SLNs were 1.3 (range 1-3) and 19 (range 7-42), respectively. There were 36 SLN-positive cases, 21 of which had metastases only to these nodes. There were 19 node-negative cases, and 3 SLNs were falsely negative. Possible causes of the errors during lymphatic mapping are analyzed in the light of experiences published to date. SLN biopsy seems a good approach to enhancing the selectivity of axillary lymphadenectomy, but the limitations of the procedure must be evaluated and carefully considered.