W. E. Agar 321 



shows the distribution of the values in the two parental species. The 

 number of males measured in clone H of D. obtusa is only four, but 

 many more were obtained at other periods of the experiment, only not 

 measured under the conditions necessary for inclusion in this table. 

 Their general characters confirm the values given for this small sample. 

 Males were more abundant in other clones of D. obtusa, the values of 

 their mean ratios being given in Table I. The apparent high degree 

 of variability among these four clone H males is not very significant. It 



is plain that when we are dealing with fi'actions, such as the ^ ratio, of 



which the numerators are not very different, but the denominators many 

 times larger in one case {D. pulex and the hybrid) than in the other 

 (D. obtusa), small variations in the small denominators — whether bio- 

 logical or due to experimental error — will produce much larger differences 

 in ratios than variations of equal extent in the larger denominators. 



There is no evidence of progressive increase in diversity in successive 

 generations of the hybrid males, whether measured by the standard 

 deviation or by the extreme range of variation. The numbers in each 

 generation are indeed in most cases too small for a satisfactory deter- 

 mination of the actual value of the standard deviation, but it is plain 

 that it is not increasing progressively. If, in order to obtain larger 

 numbers, generations are amalgamated into two groups, one consisting 

 of the earlier, and the other of the later, generations, we get the following 

 result (Table VII) : 



TABLE VII. 



-^ ratios /or males of the earlier and of the later generations of the 



The difference between the standard deviations of the two groups is 

 only "087 + "065, which is plainly not significant. 



Before leaving this table, it is necessary to draw attention to certain 

 omissions which were made in its compilation. Altogether nine males, 

 all belonging to the hybrid clone, have been left out of the table for the 

 following reasons. 



One complete brood of five males, and one of another brood, giving 

 six in all, when measured at the age of 21 days, were found to be still 

 exhibiting juvenile characteristics, especially of the first antennae, thus 

 indicating that for some reason sexual maturity had been delayed in 



22—5 



