136 Nature of the Genetic Material 



This group of facts shows us that the definition of position efiFect 

 with which we worked does not cover all cases. We spoke of position 

 effect as a term which actually means that a rearrangement break 

 near a known locus makes this locus behave as if it had mutated. 

 Later we saw that many dominant position effects were known in 

 which a neighboring locus was not available. This might have meant 

 that by chance such a locus had not yet been found. It might also 

 have meant that a specific locus is not essential, that the rearrange- 

 ment alone acts, the type of action depending upon the type of 

 rearrangement or the general region of the breaks. Now we see 

 position effects in the total absence of the locus which is supposed 

 to act, only the general region in which the break occurs being of 

 importance. These facts and their interpretation outside the theory 

 of the gene will play a decisive role when we try to draw conclusions 

 about the nature of the genie material. 



A special, much discussed example of position effect is the so- 

 called Dubinin effect (Dubinin and Sidorov, 1934; Dubinin, 1936). 

 The basic facts are not different from the usual position effect, 

 though the way of describing them may make the effect appear 

 different. (Note here, as in so many instances of scientific interpre- 

 tation, the power of semantics.) The usual description is that trans- 

 locations involving a break near the cubitus interruptus locus in the 

 fourth chromosome (ci) reduce the dominance of this normal locus 

 over the ci mutant in the homologous chromosome. If we express the 

 same facts by saying that the rearrangement break near the ci locus 

 produces a position effect making the normal locus act as if it were 

 mutated, we have, thus far, the same type of position effect as 

 studied before. Nevertheless, one big difference makes the Dubinin 

 effect rather distinctive. While R( + )/ci [where R( + ) means the 

 rearrangement break near the normal +ci locus] acts as if R( + ) 

 were ci, homozygous R( + )/R( + ) and hemizygous R( + )/— are nor- 

 mal. In the cases studied above (e.g., yellow inversion), these com- 

 binations showed the mutant effect. 



Before we continue with the specific facts we should remember 

 certain details. Theoretically, a position effect might be indistinguish- 

 able from the parallel point mutant effect and often it is. However, 

 it might also be a little different after the manner of a multiple allele, 

 and this seems to be the more frequent situation. If position effect 

 results in one or another allelic action, and if many "point mutants" 

 and breaks affecting the same locus are known, the effects look like 

 those of a series of multiple alleles. We can expect for a group of 



