To the Editor.—The article by Shiu et al1 in the November 1987 issue of the Archives demands some additional comment. The authors have presented a new prognostic factor, the "extent of lymphadenectomy in relation to nodal involvement." To briefly but adequately state the authors' findings, improved survival was noted for N0- and N1-staged tumors when the surgeon included in the resection uninvolved nodes, far from the tumor site. Exactly how the removal of normal tissue can favorably influence a patient's course is not explained. One explanation, however, should be considered. This is the "Will Rogers" effect. (Will Rogers once said, "When the Okies moved to California, they raised the IQ of both states.") A wide lymphadenectomy undoubtedly improves the accuracy of the staging. With a wide resection, those with faraway positive nodes are staged N2 and removed from the pool of those "benefiting" from the added surgery. Conversely,