Abstract
This is a reply to the comments of Patterson (^W26:^n 5920) and Diamond (^W26:^n 5913) on an original paper by Webb and Lemmon (^W24:^n 4983). "Most of their discussion seems to deal with problems of the conventional analysis of variance situation, in which the means of the groups show no trend, but are randomly related to each other . . . . The attitude of the present authors is simply summarized: In each and every application of a statistic, the interpretation of this statistic is inherently dependent upon a logical analysis of the experimental situation involved. Adherence to this dictum does not seem too clearly a part of the comments of Diamond or Patterson."
