In a series of recent articles (3, 4, 5), Dr. Rethlingshafer compares the behavior following the interruption of activities of normal and feebleminded children of the same mental ages, in the attempt to test Lewin's theory of the rigidity of behavior in the feebleminded (2). The basis of the majority of the comparisons made is the difference between the groups in terms of scale values of the classes of behavior distinguished by Dr. Rethlingshafer. The percents of behavior in each of 17 classifications were converted into standard deviation values. This method of converting a classificatory system into a scale of equal units presupposes two conditions: (a) that the distribution is normal, and (b) that the classes are ordered in a hierarchical sequence (1, pp. 97,99-102). The assumption of a normal distribution is the best assumption that can be made for an unknown distribution. The second assumption is the point at issue in the present • instance. There is considerable doubt that the 17 classes are ordered from strong to weak in tendency-to-continue, as Dr. Rethlingshafer contends (3, p. 118). There are two reasons for this doubt. First, the order was determined by the judgment of only one individual, the experimenter. In ranking the order of items on a scale, the judgments of a large group of judges is necessary. The second, and perhaps more important reason for the doubt, lies in the possibility, or probability, that two variables are involved in the behavior classified in the scale. As indications of tendency-to-continue, Dr. Rethlingshafer utilizes behavior in two situations: (a) response to being interrupted, or resistance to interruption, and (b) resumption of the interrupted activity when opportunity was provided. She apparently realized that these two kinds of behavior might not be identical or highly related, since in one report (5) she deals with this question, and concludes that the use of refusal-to-be-interrupted as a criterion of tendency-tocontinue is justified. Her findings do indicate that there is some relationship, but on the other hand they suggest that this relationship is not particularly close. To the present writer the combination of these two kinds of behavior in a single scale is not justifiable, and results in a rank order that is not true for either the combined or double variable nor for either separately. Tendency to continue an activity may be measured by refusal to be interrupted in this activity. Tendency to return to an activity may be measured by its resumption after interruption. These two tendencies may or may not be correlated. It is quite