A classical conditioning procedure was used to determine the degree of conditioning which could be achieved with frogs. An experimental and a pseudoconditioning control group were used. The CS consisted of bub- bles and the UCS was a mild electric shock. The results showed that after 75 trials, some conditioning was achieved which, for the experimental groups, stabilized at a level of 30% CR's, while the controls reached a level of less than 15% CR's. ** * Studies of learning in frogs, toads, and axolotls in an instrumental learning situa- tion (io e., learning to make a response that is instrumental in satisfying a need), show that frogs learn to cease snapping at inedible substances which at first attracted them. In some instances only one trial was necessary for toads to learn to avoid the simu- lated food (Thorpe, 1956)o There are, however, no convincing reports of classical conditioning (i. e,, learning that takes place when a conditioning stimulus is paired with an unconditioned stimulus as developed by L P. Pavlov) in amphibia. Bajandurow and Pegel (1932) reported that they conditioned breathing and jumping responses in frogs using the notes of a whistle or a flash of light as the Conditioned Stimulus (CS), and mild electric shock as the Unconditioned Stimulus (UCS). However, conditioning was quite unstable, and it was necessary to present the stimuli concurrently. In addi- tion, complete failure resulted if the interstimulus interval was more than three or four seconds. All in all, one is not really certain that conditioning actually occurred. Noble (1954) and Thorpe (1956) have reviewed studies that demonstrate instru- mental learning in amphibians. However, we know of no other reports of classical conditioning in amphibia, and a good many hours have been spent in the laboratory at Kansas State University unsuccessfully attempting to obtain learning with classical conditioning procedures. In examining our methods of study and the behavior of frogs and other amphibia, it occurred to us that possibly the several types of stimuli that had been used as a CS (these included lights, tones, buzzers, air puffs, and rotation of the stand) were not neutral; that is, these were possibly stimuli that tended to in- hibit movement. In other words, the problem might be one of obtaining a stimulus situation more relevant for purposes of classical conditioning, a factor that has re- ceived relatively little attention. Preliminary experiments, using a CS- air bubbling through water- that evoked movement and hence was non-inhibitory, gave evidence of learning in frogs. This raised the question as to the degree of learning that could be demonstrated in frogs using classical conditioning. This paper reports an experi- ment to determine what degree of learning could, in fact, be demonstrated. The experimental apparatus consisted of a clear lucite cylindrical tank, five inches in diameter, four and a half inches high, and filled with one and three-fourths inches of water. The two electrodes, opposite each other, were connected to an Applegate Stimulator, Model 230. Four plastic tubes, about one-eighth inch in diameter, were taped to the inside of the tank equidistant from each other. The tip ends of the tubes were about a fourth of an inch under water. The tubes were connected to a Cosmic