INTELLIGENCE OF EARTHWORMS 349 



reversed and backed into the exit tube. The posterior end 

 seemed to "recognize" the tube. Thruout this trial the "tail" 

 manifested unusual initiative. 



The worm lost itself in the third, and again in the fifth trial. 

 Concerning the latter the note-book affords the following com- 

 ments. "The behavior contrasts markedly with that of pre- 

 vious trials. There is little initiative for forward movement. 

 The worm finally had to be driven into the exit tube." 



May 6 was evidently a "bad" day, due probably to an un- 

 favorable physical condition. 



The worm did not burrow in earth for three weeks after the 

 operation. On May 21, it was found completely covered with 

 loose earth. The regenerated portion at this time appeared 

 clearly segmented and complete, but not yet as large as were 

 the amputated segments. 



The tendency to climb the walls of the T which was so con- 

 spicuous a feature of the behavior immediately after the opera- 

 tion gradually disappeared as the process of regeneration pro- 

 gressed. Within three weeks, the worm pushed along close to 

 the floor of the apparatus, as it did before the operation. 



For the first time, the worm was found naturally embedded 

 in a well-formed earth-burrow on May 28. The regenerated 

 segments were nearly full-size and differed from the others 

 only in their lighter color. On this date the worm wandered 

 more than formerly. It exhibited increased initiative and a 

 proportional increase in mistakes. Whereas immediately after 

 the operation the subject was automatic in its reaction to the 

 apparatus, it now exhibited varied response. 



From June 7 until July 4, a period of four weeks, the subject 

 was permitted to rest in order that the persistence of the habit 

 might be tested. 



Previously the effect of rest had been observed by the sus- 

 pension of training from December 19 until January 8, three 

 weeks, with the result that the trials immediately after the 

 "rest" were better than those before it. No indication of loss 

 of the habit by the normal worm within three weeks appeared. 



From the outset the trials with the regenerated worm, fol- 

 lowing upon a four-week's interval of rest, were marked by 

 mistakes. On July 4 not a single trial was correct; on July 5, 

 only two out of five; on July 8, only one of five (table 8). 



