Chromosomes and Genes 137 



position efiFects any type of di£Ferent actions known for sets of multi- 

 ple alleles. One of the known causes of a recognizable difference 

 between mutant and position effect phenotype is the location of the 

 second break. Raffel and Muller (1940) showed this, and found also 

 that the position effects of different rearrangements were more alike 

 if the second breaks had occurred all in the same region of a chromo- 

 some. A special type of allehc effect (already discussed) is the 

 variegation in the presence of a break in the chromocentric hetero- 

 chromatin as cause of position effect. This, in our former analysis, 

 proved to be a less intensive action of the break varying around a 

 narrow threshold. We saw that this effect could be modified; for 

 example, in the position effect for light (Schultz, 1947) the eye 

 would be more normal — more red spots — in the absence of a 

 Y-chromosome, more mutant — many light spots — in the presence of an 

 extra Y. 



Another group of facts to be remembered in the present discussion 

 deals with so-called isoalleles ( Stern ) : alleles which produce a change 

 which cannot be detected phenotypically, because it is within the 

 zone above the threshold for full normal action, a notion which 

 had been widely used in my analysis of genie actions ( 1920 ff . ) . If 

 a normal allele controls, for example, a full-length bristle, many 

 isoalleles may be present acting above the threshold for full length, 

 and therefore not changing the actual length, which for embryologi- 

 cal reasons cannot go beyond "full length." The apparent absence of 

 a position effect may in fact be due to an action of the isoallelic 

 type, which could be proved only in favorable cases, as by comparing 

 different heterozygotes. A probable example is a case found re- 

 peatedly: only the locus c shows a position effect, where the order 

 break a-b-c-d occurs, the intervening loci a and b showing no pheno- 

 typic change. An isoallelic position effect would be the most plausible 

 explanation. Altogether, a position effect may produce a series of 

 different phenotypes: at one end of the spectrum, an effect beyond 

 the extremest known effect of a point mutation at a locus; at the 

 other end, no apparent phenotypic action due to isoallelic action. 

 Moreover, all these effects are modifiable by extraneous action upon 

 developmental processes, as the plus-minus shifting of extra hetero- 

 chromatin clearly demonstrates. With all this in mind we return to 

 the Dubinin effect. 



The mutant ci causes a gap in the fourth wing vein, the length 

 of which can be used for a quantitative measurement of the effect. 

 Though the ci position effect is in general identical with other 



