BEHAVIOR OF RACCOONS 



243 



(i), In normal behavior the responses occur during the presen- 

 tation of the third card which is common to both series. The 

 responses must then be determined in part by the preceding 

 cards which are no longer present to sense. Hence the stimidat- 



Figure 2. Gregg and McPheeters' experiment table and apparatus 



ing effects of the first two cards of each series must be carried 

 over in some manner until the response does occur. The ques- 

 tionable assumption is now made that this "carrying over" 

 function can be subserved only by an imaginal process. (2), 



