92 



GERT llONNIEU 



to conclude from lliosc fi^'urcs whether the percentages from the 

 tables 7, 8 and 9 belong,' to dillerent categories or not. But at least in 

 the case of the tables 8 and 9 the difference between the percentages is so 



TABLE 9. 

 Offspring of scute cosin non-disjiinctiondl (laughters from A61 table 6 

 and their exceptional female descendents when mated to various males. 



great when compared with the mean error as to make it very possible 

 that it is not due solely to errors in sampling. 



It was, however, necessary to make an extensive number of 

 further crosses in order to have more available data on the problem. 



The same principles for testing the problem were used at this 

 time, but it seemed to be more convenient to arrange the experiments 

 in such a way that the non-disjunctional females which were to be 

 examined had inherited one whole X and that there only in the other 

 sex-chromosome were present greater parts of other Ä':s. For one line 

 of cultures, however, (table 16) one X-chromosome was used in which 



