52 INHERITANCE IN GUINEA-PIGS. 



result from the cross, but (2) that a physiological growth stimulus (not 

 hereditary) results in F : from the fact that the zygotes produced are 

 formed by the union of gametes from very dissimilar races, and (3) that 

 the increased F! vigor is largely, but not entirely, lost in F 2 . No evi- 

 dence is found that it persists in full force in any F 2 zygote (with one 

 possible exception) , since the upper half of the range of the F! zygotes 

 is almost completely wanting in F 2 , while the absence of any appreciable 

 increase of variability in F 2 shows that any increased vigor due to the 

 cross which persists into F 2 persists also very generally among the 

 zygotes of that generation, so that practically all are changed in the 

 same sense and in like amount; otherwise increased variability must 

 result, irrespective of whether size inheritance occurs other than by 

 complete blending. 



HYBRIDS OF THE AREQUIPA d* 1002. 



Crosses between the Arequipa cf 1002 and females similar to those 

 of race B, but averaging a little larger, have yielded an extensive and 

 vigorous race for the study of size inheritance. Among the animals of 

 this crossed race the mortality has been comparatively small, so that 

 good numbers are available. The male 1002, sole male ancestor of this 

 race, is still alive, so that his bone measurements are not available; but 

 a female, 1001, secured in the same cabin in Arequipa in 1911, lived 

 until fully grown and her bones are available for comparison. Further, 

 a son of cTl002 and 9 1001 lived until fully grown and his bones also 

 are available. From the measurements of these two and a comparison 

 of the empirical ratio of female to male measurements in the other races 

 studied, it is possible to arrive at estimates of the racial size of the 

 Arequipa stock which it is believed are fairly reliable. These are given 

 in table 32, where it is further assumed that the racial size of the animals 

 mated with cf 1 1002 was substantially that of race B, measurements of 

 the latter being given for comparison. But whether these assumptions 

 are sound or not does not affect the validity of the observations on the 

 F! and F 2 hybrids from this cross, which are valuable as regards their 

 interrelations, for the numbers of adult individuals are considerable 

 (43 F! and 77 F 2 animals) and the mortality among them is small. F! 

 in this experiment (tables 29 to 31, rows 9-12) regularly exceeds 

 the assumed mid-parental measurement, as in the crosses previously 

 considered; F 2 is in all cases close to the mid-parental, being slightly 

 greater in three cases and slightly less in three cases. As regards the 

 relative variability of the two generations, the standard deviations 

 indicate that the F 2 females (as compared with those of FJ are con- 

 siderably more variable in skull length (though scarcely more so than 

 race B) and are slightly more variable in skull width and femur length. 

 The male F 2 hybrids differ very little in variability from the Fj hybrids, 

 the standard deviations being slightly greater in skull measurements 

 but less in femur length. 



