354 Action of the Genetic Material 



presence of ci^. Thus, nos. 2 and 3 are expected to be as they are, 

 and no interference is needed. No. 1 presents the real problem. If 

 dosage alone were involved, it should be abnormal. Thus the Minute 

 M(4) =0 must be responsible for counterbalancing the inhibiting 

 effect of ci^^'. Once before we had to come to a similar conclusion, 

 which I prefer to the idea of mutual interference. I wonder how 

 ci^/+ would look in the presence of Minutes in other chromosomes. 



The third set of facts in Stern's analysis has to do with dosage 

 effects when one of the alleles is the position effect of a rearrangement 

 R, either R( + ) or R(ci). We considered this when studying the 

 Dubinin effect, which is just this position effect, meaning that R( + ) 

 acts as if + had mutated to ci, but only in the combination R( + )/ci, 

 not as R( + )/R( + ) and R( + )/0. We analyzed the basic facts (see 

 I 3 C c cc) and tried to explain the distinguishing features of the 

 Dubinin effect without considering the dosage problem and Stern's 

 interpretation. We turn now to Stern's analysis. On the basis of the 

 foregoing data, he states that the peculiarities of the Dubinin effect 

 are of the same type as the relations between the alleles just studied. 

 (I may add that this would be expected if mutants are invisible 

 position effects; but our former explanation of the Dubinin effect was 

 completely free of dosage ideas and based only upon specific features 

 of the position effect at the ci locus. ) According to Stern, the specific 

 results are an expression of interference phenomena, which make 

 combinations of alleles less effective in terms of venation than single 

 doses of alleles by themselves. The degree of interference varies with 

 the different R( + ) used. Among seventeen R( + ) tested, interference 

 was so strong in four of them that R( + )/ci was more deficient than 

 ci/ci. In the others, R( + )/ci was equal to ci/ci or less deficient. One 

 of the R( + ) alleles, R-( + ) produced a normal phenotype if hemi- 

 zygous, but venation like ci/ci in the homozygous condition. I used 

 these facts previously for an explanation in terms of influence of 

 heterochromatin upon the strength of position effect, while Stern sees 

 here the working of interference between the alleles. I prefer my ex- 

 planation, because it fits into all the other facts of position effect with- 

 out requiring the dosage concept. This means that all the blame for un- 

 expected results is laid upon the Dubinin position effect, whatever 

 the other allele does. 



The next point is the position effect R(ci), the workings of which 

 are found also in the enumeration above. In accordance with the 

 explanation of standard position effect, we might expect such a re- 



