50 INDIVIDUAL DIFFERENCES AND FAMILY RESEMBLANCES 



litters and to decrease the family resemblances, while it accounts for 

 the rather large sex difference that has already been discussed. The 

 second litter, resulting from crossing No. 115 with No. 118, gave two 

 males that made averages of 21 and 24 seconds respectively, one fe- 

 male that died, and again a very slow female that made an average 

 of 132 seconds. The family average was 59 P.E. 29.5 seconds. 

 The middle group of seventh generation mice that is given in Fig. 12, 

 resulted from two successful matings of a male, No. 136, with an 

 average record of 9 seconds, with No. 138 whose average was 76 sec- 

 onds. It may be seen that in the first litter, Nos. 190 to 195 inclusive, 

 all the males and one female made good records, while again, as was 

 found in the previous litters of this generation, one female made ex- 

 ceptionally poor records. This was the case of No. 195 that made an 

 average of 265 seconds. The family average amounted to 65 P.E. 

 25.6 seconds. The litter that was subsequently obtained from the 

 same parents gave three males that made good records, 22, 28 and 

 29 seconds, respectively, and three females that made time averages 

 considerably slower than the average for the entire population. One 

 female made an average of 75 seconds, another 91 seconds, and still 

 another 106 seconds. Again it may be noted that the females fell 

 behind the males in quickness of learning. The family average of 

 the second litter is similar to the first, and amounted to 58.5 P.E. 

 12.3 seconds. It is hoped that sufficient data may be later obtained 

 to show whether or not there is a closer resemblance between litters 

 belonging to the same parents than between unrelated litters, and 

 also to determine whether or not mice belonging to the first litters of 

 any two young parents tend to be superior or not to the mice that are 

 produced from these same parents when they are considerably older 

 and have produced several litters of mice. The number of double 

 litters that has been obtained in these experiments is hardly large 

 enough to warrant any conclusion on this point, but it may be noted 

 that several of the cases that have been found show that the family 

 averages of first and second litters overlap when due consideration 

 is given to the size of the probable error that is attached to each 

 average. No. 136 and No. 137, that made very good time records, 

 were mated, and they added two more mice to the seventh generation 

 of the white family; one of them, a male, made an average of 19 

 seconds and the other, a female, took 59 seconds. Their combined 

 average amounted to 44 P.E. 21.2 seconds. 



In order to study another branch of the white family we must 

 again go back to the fifth generation and note what resulted from 

 mating No. 91 with another white female, namely No. 86, that also 



