II. STATEMENT OF THE PROBLEM 



THE plan of the experimental work presented in this investiga- 

 tion is to measure individual differences in behavior, to determine 

 the extent to which the animal that departs from the average in one 

 direction will depart in others, to measure the resemblances in fami- 

 lies and in lines of descent, and to determine whether kinds of con- 

 duct can be established in family lines by selection. In a previous 

 publication, 4 of which this thesis is a continuation, it was found that 

 individual differences occurred in the ability of various strains of 

 mice to learn a simple maze, and also that a family resemblance 

 existed among mice of the same litter, that amounted to a coefficient 

 of correlation in the neighborhood of 0.50. Certain mice, and even 

 whole lines of descent, showed marked variations from the average, 

 some taking more than twice the average time to learn a given task. 

 These differences were well beyond the limits of the probable error. 

 The results were obtained from testing 90 mice, as determined by 

 the time required to find their way through a maze. Since then, 

 these mice and their offspring have been tested in other ways, and 

 further experiments are now in progress with the F 7 and F 8 genera- 

 tions. In the present investigation there are described the individual 

 differences and family resemblances of 93 mice, in addition to the 

 90 already reported on in the previous article. These mice have 

 been tested in the same maze as were the previous ones ; in addition, 

 they have been given an interference test, a retention test and have 

 been studied in a second maze, as described below. 



* Bagg, ' ' Individual Differences and Family Resemblances in Animal Be- 

 havior," The American Naturalist, April, 1916. 



The present material was submitted for publication in June, 1918. 



