230 CHARLES ZELENY 



of both the affected spermatozoa. In view of the small per- 

 centage of spermatozoa that sm^vive, such an explanation, how- 

 ever, is improbable. 



12. Are mutations at the har locus more or less frequent than 

 those in accessory factors affecting eye-facet numbers? Where 

 more than one germinal difference affects the same somatic 

 character it has been found in studies of other characters than bar 

 that these differences are more frequently at various loci than at 

 a single one. While the data on accessory factors affecting facet 

 number are to be given elsewhere, it may be -stated here that all 

 the observed mutations in the accessory factors together are not 

 as numerous as the changes at the bar locus. In this respect 

 bar differs from the general rule as stated above. On the prin- 

 ciple that there are several loci outside of bar and therefore 

 several materials whose change affects facet number, it might 

 have been supposed that there would be that much more oppor- 

 tunity for change in some one of them than in the single material 

 at the bar locus. The material at the bar locus therefore has 

 a higher degree of instability than the materials at the loci of 

 the different accessory factors. 



13. Do the three genes at the bar locus form a quantitative series? 

 This question naturally arises in connection with a discussion of 

 the origin of mutations. Ultra-bar may presumably be con- 

 sidered as due merely to the addition of another bar factor to the 

 previous one. In a former paper (Zeleny, '20) it was pointed out 

 that a comparison of the facet values of 1) full, 2) full x bar, 

 3) bar, 4) ultra-bar x bar, and 5) bar females should throw some 

 light on the correctness of the quantitative theory. The five 

 kinds of females should, according to this theory, differ by equal 

 amounts of a single germinal material and there should be some 

 constant relation between their facet values. No such regular 

 gradation in facet values is found. Furthermore, ultra-bar 

 males should have a value approximating that of bar females and 

 bar males a value similar to that of heterozygous females. This 

 also is not true. Therefore, it seems necessary to assume that the 

 two steps in the series are qualitatively different or at least are 

 not of the same order. 



