Genie Control of Development 361 



ingredients of the total effect are different. This shows how diflBcult it 

 is to consider multiple factor effects as dosage effects and to draw 

 from them conclusions concerning genie action. The same would apply 

 also to cases in which a definite part effect could be assigned to 

 individual multiple factors. I studied a number of cases (see Gold- 

 schmidt et al., 1951; Goldschmidt, 1953a) in which it was possible to 

 isolate (more or less) the individual multiple loci by chromosome 

 replacement and to measure their individual share in the total effect. 

 Some had a major share, others only a small one. All of them must 

 have affected a definite process of growth and differentiation (homoe- 

 otic transformation of wing anlage in Drosophila in one case, number 

 of extra bristles in the other) in the same general direction, although 

 not alike quantitatively. However, there is no way of deciding 

 whether they acted upon the same set of reactions directly or only in 

 some indirect way via necessary but different conditions for these 

 reactions. Thus it is hardly possible to use such material as indicative 

 of dosage actions of genie material. 



This would be different, however, ff Mather's blocks of polygenes 

 were a reality, whether they were represented by clusters of multi- 

 plied identical loci or simply by different quantities of intercalary 

 heterochromatin, as I prefer to assume. In both cases we should be 

 dealing with actual quantities of the same genie material and there- 

 fore with dosage experiments. Parallelism of the effects to dosages 

 would give us the same type of information on genie action as gained 

 in the discussion of the genuine dosage experiments. Unfortunately, 

 these polygenic blocks are still hypothetical. 



The discussions and interpretations presented thus far are inde- 

 pendent, I think, of the general theory of multiple factor inheritance, 

 which has been under much discussion since Mather's work on "poly- 

 genes." We have discussed (see I 2 C d dd) only that part of the 

 theory which deals with so-called blocks of polygenes of supposed 

 heterochromatic nature. But there is another aspect of Mather's theory 

 which I have difficulty in combining mentally with the one feature dis- 

 cussed formerly (see Mather and Harrison, 1949). From his selection 

 experiments on bristle number, Mather concludes that each quantita- 

 tive character is determined by many "polygenes," each of small effect 

 upon a single character and distributed over all chromosomes. These 

 polygenes are balanced within each chromosome because they act 

 partly in a plus direction and partly in a minus one. Crossing over then 

 may produce much genetic variability and selection becomes possible. 

 This part of the problem, selection, does not concern us here; neither 



