THE BEHAVIOR OF COCKROACHES 89 



Marked individual differences were . noted with respect to 

 the time during which the cockroaches retained their newly 

 acquired habit. 



The data given below were gathered by measuring the time 

 elapsing between the last effective shock given and the animal's 

 retreat into the dark part of the glass box. 



No relation is evident between the degree of permanency of 

 the newly acquired habit and the number of shocks necessary 

 to establish it. Though the training influences the behavior 

 of the cockroaches only for a short time, the memory of it lies 

 latent, as is proved by the rapidity with which they relearn. 

 An animal once trained after a few days performs the obligatorv 

 number of turnings-back after only a few shocks have been 

 applied. The following table shows the results of relearning. 



The curves of relearning resemble those of the first series. 

 They indicate rapid progress without noteworthy fatigue. As 

 an example of the curve of relearning, I give that for No. io, 

 Fig. 4, R L, io. 



Before concluding my description of the experiments, I must 

 mention that it was possible to establish the habit in an animal 

 whose antennae had previously been removed. A hundred and 

 twenty-six shocks were required before the animal performed 



