THE M I 'TA TED GENE 1 1 7 



In this case, we have important information on some decisive 

 points which makes possible a definite picl ure of the phenomenon. 

 1. There are first the facts concerning the lower members of the 

 series. As reported above and discussed from different angles, 

 these lower members produce their effects in a definite way. 

 The lowest member nicked produces no visible effect in homo- 

 zygous condition, according to Mohr. (In Goldschmidt's lines, 

 there was a very small percentage of nicked individuals. The 

 same shift as compared with Mohr's lines applies to other com- 

 binations which explains the small difference between the data 

 in table 7, page 80, and those used for construction of Fig. 

 24. The difference is probably due to different modifiers.) 

 The higher allele notched affects only 42.5 per cent individuals, 

 and it is only with still higher alleles that 100 per cent effect 

 is reached. If we represent these combinations and those com- 

 pounds that range between them (see pages 110 and 111), as is 

 done in Fig. 24, the effect of the allele is considered statistically 

 as a curve of variation, a more or less large part of which is 

 situated beyond the threshold for Wild type. (See also page 

 68, the first case of this type in Lymantria.) If we follow these 

 curves from no/no with 100 per cent scalloping to ni/ni with only 

 a few per cent scalloping, it is obvious that the curves for lower 

 combinations, all within Wild type, ought to show a continuous 

 shift to the right, as represented in Fig. 24. This is of course 

 only an extrapolation; but it can be proved correct. In Fig. 24 

 are also, represented as broken curves the actions of the same 

 genes and compounds, with the addition of the dominigenes 

 (see page 110). These shift the curves of the phenotypes to the 

 left, and they do it, as the figure shows, in a way exactly parallel 

 for the different combinations. Thus, the first heterozygote 

 m/+ is shifted a little beyond the threshold, and all the higher 

 compounds correspondingly more. This shows that there is 

 actually within the Wild type a corresponding arrangement of 

 more or less Wild-type individuals — degrees of Hyperwild, we 

 might say, which thus far could not be distinguished phenotypi- 

 cally. To this series would also belong at their proper place 

 the normal vg/-\- heterozygote (to be put between ni/ni and 

 no/ '+) and the dominigene action upon it, which in Fig. 24 have 

 been put into another group for certain reasons. 



