124 Nature of the Genetic Material 



to dosage. Now breakage or position effect acts like a mutant in 

 the region, and therefore the same argument is good for the baroid 

 position effects and for the duphcation effect Bar. This means that 

 we cannot predict from the knowledge of the quantity of the effect 

 of two homozygous breaks (Bar) versus two heterozygous breaks 

 (heterozygous Bar) and three homozygous breaks (double Bar) 

 whether the quantity of action of two homozygous breaks will be 

 the same or different in the three heterozygous breaks, the double 

 Bar heterozygote. It is possible that the four breaks in two chromatids 

 have a greater action than three breaks in one chromatid, or it may 

 be the other way around. In both we would have Sturtevant's position 

 effect (in favor of one or the other combination). This, the original 

 position effect, has nothing to do with position but is a phenomenon 

 of dominance, proportional or not proportional to the number of 

 breaks. 



Actually, if we calculate as a measure of dominance the per- 

 centage of facet number decrease from normal (i.e., as loss of 

 inhibiting action upon facet formation) in the heterozygote (i.e., 

 in this case the improvement toward normalcy), the value for B/+ = 



Bb B^ I 



„j, is 20 per cent; but for BB/+ = B^| it is 50 per cent. Therefore 



I B^l 



it is impossible to predict that the latter value should be the 



same as that of BB = ^b Db ^^^ ^^ attribute the actually found dis- 

 crepancy to the different positions. The Sturtevant position effect 

 has ceased to exist, meaning, of course, the explanation in terms of 

 location of two "genes" in one or two chromosomes. The Bar effect 

 itself remains a standard position effect in the sense of breakage 

 effect (i.e., change of pattern) and as such acts like any mutant. 

 This includes also dosage effects, visible in the Bar-double Bar re- 

 lation and demonstrated further by Rapoport (1939, 1941), who 

 inserted a number of Bar duplications into a chromosome with 

 increasing effect. (The problem of dosage effect will be discussed 

 later with genie action.) It is interesting to see how defenders of 

 the old concept of position effect in terms of gene neighborhood 

 account for this situation, which as such cannot be denied. Lewis in 

 his review (1950) concludes a discussion of the facts with the 

 sentence: "These results then point to the conclusion that there are 

 at least two separate components of the Bar phenotype to be con- 



