180 BLASTOGENIC VARIATIONS. 



parental correlation showed no marked difference from 

 those obtained in sexual reproduction, just as in the 

 case of his daphnia observations. If anything, there 

 was a decrease in the correlation, on passing from 

 sexual to parthenogenetic forms, rather than the in- 

 crease we should expect. However, the variability of 

 the individuals of a brood was found to be only about 

 60 per cent, of the racial variability; -i.e., distinctly less 

 than in sexual reproduction. Also the mean coefficient 

 of fraternal correlation for aphis and daphnia was .66, 

 or considerably higher than the mean value of .45 ob- 

 tained by Pearson * for fraternal correlation among 

 sexual forms. Warren's general conclusion may be 

 summed up in the words: " The question as to whether 

 we have a real difference between parthenogenetic and 

 sexual offspring can only be decided by further investi- 

 gation both on aphis and other forms." In the light of 

 Weismann's observations, which were carried on for 

 such a number of generations, we seem entitled to con- 

 clude that probably a real difference will be found to 

 exist between them. 



Arguing partly from Warren's observations, and 

 partly from others of his own, Professor Pearson f has 

 come to the conclusion that, " whatever be the function 

 of sex in evolution, it is not the production of greater 

 variability." Thus he says that the individual contains 

 in itself a variability which is 80 to 90 per cent, of the 

 variability of the race, and which it can exhibit quite 

 independently of sexual union, e. g., as in this case of 

 parthenogenesis. As instances of individual varia- 



* Phil. Trans. 1901. A. p. 285. 

 f " Grammar of Science," p. 474. 



