24 INDIVIDUAL DIFFERENCES AND FAMILY SE 'SEMBLANCES 



amounted to -f- 0.85. A similar coefficient of correlation of -f- 0.82 

 was found to hold between the last 23 trials of the multiple choice 

 test and the errors made in that task. These results are what one 

 might naturally expect to find, namely, that the animal that made 

 many errors required more time. The conclusion need not hold, 

 however, for the type of behavior that this investigation deals with, 

 because it might very well be that a stupid animal was one likely to 

 sit in a corner of the maze and make a poor time record, but a good 

 error record. These two high correlations are thus seen to give a 

 measure of the activity of the animals in each task, and they show 

 the value of the tests that were used, in that each was adapted to 

 the instinctive behavior of the animals tested. 11 The native equip- 

 ment of the mice no doubt adapted them to finding their way through 

 narrow passages and forcing their bodies through small openings. 



A low positive correlation of 0.11 was found to hold for the rela- 

 tion between the initial learning period in the maze test and the last 

 23 trials in the multiple choice test. As previously discussed, the 

 time values in the two tasks that were correlated are not comparable, 

 and this fact may account for the lowness of the coefficient of corre- 

 lation. 



A correlation was made between performance at the beginning 

 and close of the trials, and in this case the first group of five trials 

 was correlated with the last group of five. For the maze test this 

 correlation amounted to +0.46, which represented a fairly strong 

 correlation. A similar correlation between the first group of five 

 trials and the last group of five trials in the multiple choice test gave 

 a lower correlation of -(-0.25. 



The correlation between the initial learning period in the maze 

 test and the retention test for the same task was found to be -f- 0.35. 

 This indicated a fairly close relation between proficiency in a given 

 task and the amount of association retained. A similar correlation 

 between the same initial learning period in the maze test and the first 



11 Guinea pigs have been used as subjects of another investigation and 

 when tested in the maze, that apparatus was found to be very well suited to their 

 instinctive type of behavior. They have not as yet been tried in the multiple 

 choice apparatus which was used in these experiments. The maze test, and the 

 multiple choice as well, were found to be unsuited to the average native equip- 

 ment of a litter of seven English Bulldog puppies that also have been tested. 

 Four out of the seven dogs failed completely to learn the task, sitting in a corner 

 of the maze and barking disconsolately at the gates. They held back from pass- 

 ing through an opening through which an average mouse would at once venture. 

 Two of the dogs learned very slowly, while one made a good record. The dogs 

 were more interested in looking for the experimenter and responding to the 

 slightest noise he would inadvertently make, than in paying attention to the 

 maze problem itself. 



