The results of a study of removing the subject from the situation in which no aversive stimulation was employed also are rather equivocal so far as contiguity theory is concerned. Seward (1942) compared the effects of three treatments, removing each rat from the apparatus as soon as (1) a bar press had been emitted, (2) the animal had seized the food pellet delivered when he pressed the bar, or (3) one minute had passed. Mere removal following the bar press was less effective than food reinforcement, but more effective than noncontingent removal, in terms of the latency of the first response on each training day and the number