THE ALTERNATION PROBLEM 381 



for each rat, it is probable that some of these scores are signifi- 

 cant. 7. The introduction of the long delays has tended to 

 disrupt the act for the shorter intervals. The rats were finally 

 tested again for the 50 sec. interval. Much poorer records were 

 obtained than for the initial tests. The group record was de- 

 creased by 10%. Only four of the animals were now able to 

 choose correctly for a score of 67% or better. 



The experiment permits of the following general conclusions. 

 1. The guiding and controlling stimulus to each choice is consti- 

 tuted in part by the sensory aspects of the preceding act. A 

 certain percentage of correct responses was obtained when all 

 possibility of distinctive motor attitudes during the delay was 

 wholly eliminated. Furthermore, any increase of the time 

 interval beyond 60 sec. decreased the percentage of correct 

 responses. 2. The rat may thus establish an associative nexus 

 between a sensory stimulus and an act which are separated by 

 a time interval of 16.5 sec, provided that relatively constant 

 conditions exist during this period. 3. When an association has 

 been established for a period of 16.5 sec, approximately one 

 minute is the maximum time of separation of the stimulus and 

 the response that may be obtained without disturbing their 

 functional relation. 4. The functional efficiency of the co- 

 ordination depends in large part upon the stability of the con- 

 ditions that obtained for the period of delay. This fact sup- 

 ports the hypothesis that the guiding stimulus to each choice 

 is constituted to a large extent by a distinctive motor attitude 

 resulting from the previous act. The proof is not at all con- 

 clusive, however, for it is entirely possible to assume that the 

 disruption of the act was due to the distractive influences of 

 the novel sensori-motor conditions. 5. The efficacy of motor 

 attitudes in the solution of the problem is indicated by the 

 following facts. The relative disturbing effects of an increase 

 of the time interval and the introduction of new conditions 

 during the delay differ with animals. One may infer that some 

 animals rely mainly upon the sensory aspects of the previous 

 act as guides to conduct while other animals rely mainly upon 

 motor attitudes. It is logical to suppose that those animals 

 that place their chief reliance upon motor attitudes will learn 

 the problem with the least effort because of the closer temporal 

 contiguity of the stimulus and the response. As a matter of 



