120 PHYSIOLOGICAL GENETICS 



condition, whenever the threshold for Wild type is not trans- 

 gressed. This threshold, however, is determined independently 

 by other reactions, another variable. In the present case, the 

 facts of development show that the threshold is identical with 

 the time in development at which the differentiation of the wing 

 An luge in the pupa is finally determined. If up to this moment, 

 which is controlled independently from the vg-gone, the growth 

 substance has been available in minimum quantity, the normal 

 Wild-type wing will appear. It is finally evident that domini- 

 genes or environmental conditions may shift one or the other 

 of these two known variables — as well as unknown others — 

 with the effect of changed dominance conditions. It is also 

 evident that it will be most difficult to accomplish such a shift- 

 ing effect if the Wild-type gene is involved in the heterozygote. 

 The reason is that the curve of this gene is so far above the 

 threshold that only in the heterozygote with the lowest, almost 

 lethal member of the series some individuals will remain below 

 the threshold. 



This best studied example (next to the Bar case in Drosophila, 

 which belongs to another chapter, see page 143) then shows 

 dominance as a function of the interplay of gene-controlled 

 reactions and thus furnishes further evidence toward an under- 

 standing of gene action. 



We have mentioned the different types of behavior in regard 

 to dominance found in series of multiple allelomorphs. The 

 question arises whether all these types may be interpreted on the 

 same basis as the type case of vestigial. Let us consider first 

 the cases in which each member of a series is dominant over the 

 lower one. It is obvious that such a case cannot be explained 

 with intermediate reactions and one threshold value for Wild 

 type. We must rather expect some type of threshold conditions 

 at each level. Such may easily be produced if the final reaction 

 is not of a plus-minus type with all grades in between but of a 

 strictly alternative type. Let us assume that a case of pigmenta- 

 tion is involved, as in certain actual cases (see Stern, 1930). 

 We remember from the work of Scott-Moncrieff that usually 

 the higher oxydized or methylated types are dominant. There is 

 nothing intermediate between one or two methyl residues. If, 

 then, two alleles differ by producing the final addition of one or 

 two methyl residues to the pigment molecule, one must neces- 



