THE MUTATED GENE 



143 



for this group (the points (J) in the curve and the third section 

 of the table). A glance at the column for X shows that this is 

 not the case. X is mostly zero or even negative. This may mean 

 one of two things: Either B l is not a dosage difference from B + 

 like B (in this case, the consistent results of the second section 

 could not be accounted for) ; or there is an additional disturbing 

 feature in the third section of combinations which has to be 

 analyzed. 



We have already discussed the dominance relation of the Bar 

 series in another chapter, the special features of which had 



17 25 



Temperature in degrees C. 

 Fig. 29. — Semilogarithmic plot of the relation of facet number (DrosophUa 

 melanogaster) to temperature in two sets of homozygous females and the corre- 

 sponding heterozygotes. AA, Infrabar; A' A', Bar; AA', Barinfrabar. (From 

 Hersh, 1934, Amer. Naturalist 68.) 



already induced Zeleny, Goldschmidt, et al. to special discussions. 

 Wright (1929) called attention to the fact that in the combina- 

 tions of Bar and Infrabar, with which we are here concerned, Bar 

 behaves as an almost complete dominant, keeping Infrabar from 

 expression. This is, of course, another description of the non- 

 quantitative results of the third section of Table 16 and of the 

 value for X in terms of dominance. But this description is 

 true only for Sturtevant's experiments at 25°. Luce (1931) 

 showed, as already mentioned, that Infrabar has a different 

 temperature relation (thermophene) from Bar, the number of 

 facets increasing instead of decreasing with temperature. The 



