A formal deduction from Hull's theory of extinction could lead us to some strange predictions. As training is continued, with reinforcement every trial, habit strength approaches an asymptote but conditioned inhibition continues to accumulate. Eventually our S should cease to respond. So runs the redactio ad abswdzcm presented by Gleitman, Nachmias, and Neisser ( 2 ) . Some simple tinkering with Hull's theoretical structure might set the matter right, but in a recent article Calvin and his students ( 1 ) troop to the rescue for Hull unsullied. Their proof turns out to be stranger than Hull's fiction. White rats were trained to traverse an elevated runway for a food reward, and, after a number of trials, the authors' crucial finding was that all of the Ss extinguished! (1, p. 53). If this finding were general, the end of the animate world would be in sight, for all learned behavior could be expected to cease in like manner. Before we brace ourselves for death by inaction, however, we might well consider some of the special conditions employed by Calvin and his associates. First, there is good reason to suspect that the rats were not very hungry. They were given 30 trials per day, with 10 sec. of eating per trial, 5 min. in all; as a result, Ss devoured virtually all of their ration while in the goal (1, p. 52). (This may be a slight exaggeration, since the low drive rats had 20% more to eat than the high drive rats, and presumably none of the latter ran out of food.) The total daily rations were 10 grams of dry food, mixed with water, for the high drive group, and 12 grams for the low drive group. These are large rations, more than some rats can be expected to eat promptly, in my experience, and whatever portion was not eaten during the experimental session was left in the cage, to be consumed possibly as late as a few minutes before the next day's session. The authors' defense against the charge of low drive is magnificently irrelevant. They calculate the median trial within the daily session on which . . . Ss actually (1, p. 53). But some Ss may already have been low in hunger at the beginning of the day's session. And even if they were more likely to refuse to run late trials than early trials, the session was ended when they stopped running early trials, and they had no opportunity to demonstrate their unwillingness to run subsequent trials. Therefore, the median stopping trial is a meaningless value. In the second place, wet food mash, apparently in considerable volume, was available in the goal box as the reinforcement. In other studies using wet mash it has been observed1 that .many animals scoop up large quantities of