No. 587] VARIABILITY AXD AMPHIMIXIS 



673 



tibia in the cross-bred forms (Table X) the average 

 variability of the three inbred groups is 68 per cent, 

 greater than that of the cross-bred group. Consequently, 

 the results decidedly support the facts in the present 

 paper. 



The remaining papers are those of Jennings ( '11 and 

 '13) in a study of Paramecium. In the first paper the 

 breeding experiments are summarized as follows: 



other respects, than the progeny of the equivalent nnn-rnnju-miis. 



It seems difficult to account for this conclusion if one 

 subjects the data to a critical review. So far as a "pure 

 race" is concerned the non-con jugants and their progeny 

 were decidedly more variable than the conjugants and 

 their progeny (Table 28, p. 94), although the small number 

 utilized March 31 for the statistical work (42 and 34) is 

 not sufficient to justify a conclusion in either direction. 

 Even in a "wild culture" (Table 32, p. 99) the evidence 

 is too conflicting to justify a definite expression of 

 opinion. Of the seven comparisons here made among 

 the progeny, five showed an excess variability for the 

 conjugants, but in only one case did the difference exceed 

 three times the probable error, while two cases showed 

 an excess variability for the non-conjugants, the differ- 

 ence in one case exceeding twice the probable error. 

 Data from numbers so small (22-95) can scarcely be con- 

 sidered reliable. The comparison of the variability of 

 "all pairs" and "all impairs" on June 22 and June 23 

 denotes an excess variability for those completing con- 

 jugation at the beginning of the experiment. 



In the second paper Jennings concluded (p. 363) that 

 conjugation increased the variation in the rate of repro- 

 duction. The variation was increased, but the explana- 

 tion of such increase seems comparatively simple when 

 it is noted that among the conjugants there were many 

 with a low rate of fission with death occurring. As com- 



