THE GENE 



319 



visible character during development. 

 In trying to formulate an opinion about 

 the production of this phenomenon two 

 primary possibilities are given. The 

 dominant and the recessive gene might act 

 independently, the result being determined 

 by some resultant of both actions. It is 

 evident that in this case the phenomenon 

 ought to be explained on the same basis as 

 the phenomenon of epistasis. For this 

 we have shown in a special case, namely 

 the competing reactions produced by the 

 sex-genes, that the result depends upon 

 the two respective velocities of the 

 competing reactions and a third variable, 

 an independently determined point in 

 development, determined by some or 

 many other genes. In this case the 

 dominant and recessive, or the two epi- 

 static genes, might differ qualitatively or 

 quantitatively. In any case the phenom- 

 enon of dominance or epistasis can be 

 understood only on the basis of a set of 

 differential reactions of definite velocities 

 (for details see Goldschmidt, '2.ja). The 

 second possibility is that in the heterozy- 

 gote the dominant and the recessive gene 

 do not act independently. If we assume the 

 difference between these two genes to be 

 one of quantity, this would mean that, 

 e.g., the allelomorphs of the respective 

 quantities 10 and 6 would have a single 

 combined action based on the quantity 

 16. In our experiments on intersexuality 

 we have produced experimental proofs 

 (see 'x^b, '15a and much unpublished 

 material) that actually this is the case if 

 two different quantities of an allelo- 

 morph are combined (male intersexes 

 built up from three races). If the same 

 applies also to other cases dominance and 

 its different degrees would result from a 

 system of reactions where one, with its 

 specific velocity, is the product of the 

 combined action of the added quantities 

 of the two allelomorphs, the others being 



determined by the other genes controlling 

 the exactly timed reactions necessary for 

 development. How such systems work is 

 discussed in Goldschmidt, '17a, p. 6j ff. 

 Here the argument is also applied to such 

 cases as the dominance within the bar 

 eye series of Drosofhila. It seems there- 

 fore to the present writer that the phenom- 

 enon of dominance lends further support 

 to the idea that genes are connected 

 with timed chains of reactions propor- 

 tional to their quantities. In harmony 

 with this interpretation are the facts 

 known about the results of the combina- 

 tion of two recessives with one dominant 

 gene (Correns, Wettstein, Bridges, Mor- 

 gan). In some cases two recessives 

 dominate one dominant; in others the 

 opposite is true. It seems difficult to 

 explain such facts without assumptions 

 of the foregoing type. 



There might be mentioned also the 

 phenomenon of change of dominance 

 during individual life. The present 

 writer ('17, '2.0b, '24) first analyzed such 

 a case genetically in the already quoted 

 work on the pigmentation of caterpillars. 

 In Fi between a light and a dark race 

 young caterpillars are light and change 

 later towards the dark side. The study 

 of the pigmentation curves in such cases 

 showed that the facts have to be explained 

 on the basis of typical curves of reaction 

 of definite velocities in the parents and 

 intermediate velocity in the hybrid. The 

 phenomenon then again supports the 

 general idea. Honing ('17) and Ford 

 and Huxley ('17) have since analyzed 

 similar cases and accepted our explanation. 



P. THE CASES OF VARIEGATION 



In recent genetic literature the study of 

 a certain type of variegation has aroused 

 great interest and has been used as a starting 

 point for conclusions upon the nature of 

 the gene. Speaking generally these varie- 



