HABIT FORMATION IN FROGS 317 



it by throwing out the tongue and shaking olT the caterpillar 

 in the usual way. In about 5 seconds the frog snapped up the 

 caterpillar again and immediately released it. About two 

 minutes later the frog snapped for the third time but missed 

 the caterpillar. The feeding reaction seemed to be slightly 

 inhibited. Seven minutes later the caterpillar crawled back 

 and forth on a walnut leaf stalk stripped of leaflets located 7 cm. 

 above the surface of the water into which the frog had in the 

 mean time jumped. The frog oriented in the water to face 

 the caterpillar, and then raised the head, watching the cater- 

 pillar for about 15 seconds. Then a house fly alighted on a 

 piece of board 12 cm. away from, and 45 degrees to the left of, 

 the frog. Immediately the frog oriented, and after a half second's 

 examination, jumped and caught the fly. 



This day's observations show that the habit of leaving hairy 

 caterpillars alone was well fomied by the experiences of August 

 9, for the unusual state of hunger occasioned by very scanty 

 feeding since August 9 must have put the habit to a severe 

 test. Nevertheless, two of the frogs, Rana clamata (medium) 

 and R. sylvatica, withstood the test. 



Unavoidable circumstances made it necessary to terminate 

 this series of experiments with these observations. 



Five months later certain experiments were carried on at 

 Knoxville, Tennessee, with the Rana clamata (medium) of 

 the preceding experiments, in a cage similar to that employed 

 at Cold Spring Harbor. (The other frogs mentioned had es- 

 caped during transportation.) The observations from this 

 second set of experiments confirm the conclusions drawn from 

 the first series, and in a few instances new features of behavior 

 are brought out. During the five months' interval various 

 kinds of food were fed at irregular intervals. Flies, leaf hoppers, 

 earthworms, and Pieris rapae caterpillars were the chief articles 

 of diet. 



In this series of experiments some food organisms which are 

 usually eaten with great readiness were altered with chemicals 

 and with electrical stimuli to make them disagreeable. Although 

 such artificial changes are easily accomplished, the method 

 is not as productive of results in feeding experiments as is the 

 natural history method, chiefly because artificial stimuli, espec- 

 ially those proceeding from chemicals, are usually not as easily 



