22 INDIVIDUAL DIFFERENCES AND FAMILY RESEMBLANCES 



subsequent learning in B easier. ' ' The average curve in Fig. 5 indi- 

 cates that there is practically no increase in learning between the 

 seventh trial and the twenty-fifth. A fairly uniform decrease in 

 time occurs from the first to the seventh day, when an average speed 

 of twenty-six seconds was made. The only other performance that 

 bettered this record was made on the twenty-fourth day, when an 

 average time of twenty- four seconds was recorded. The number of 

 trials in the multiple choice test could have been much less and still 

 sufficient for the average mouse to learn the maze. It was impossible 

 to determine that more trials were given than were necessary until 

 the experiment was well started. In fact, the experimenter arbi- 

 trarily chose twenty-five as the number of trials because he thought 

 it would take the average mouse longer to learn the second task than 

 the first. As explained above, the results of the experiments dis- 

 approved this assumption. 



In the multiple choice test, the curve based on the median record 

 for each day, as in the previous tests, is considerably below the cor- 

 responding curve for the average. The curves here are essentially 

 similar, and it may be noted that the curve for the median, after the 

 second trial, never rises higher than twenty seconds or lower than ten. 



