Chromosomes and Genes 141 



fluenced much by the break of weak action, and therefore ci/+ 

 and R(ci)/+ should be essentially alike, though a little higher effect 

 of the latter might be observable. Nos, 1, 2, and 10 show this to be 

 true. Therefore, R( + )/— and R(ci)/— should also be alike, which is 

 true, as nos. 7 and 12 show. In R(ci)/ci the small shifting effect of 

 the break might have a better chance of becoming visible, and a 

 phenotype more extreme than ci/ci could result, which actually 

 occurs (no. 11). R(ci)/R( + ) should be ci; if R(+) acts as much 

 below ci as R(ci) does above, the effect should resemble ci/ci, and 

 ff R( + ) is above that level the effect should be beyond that of ci/ci. 

 The latter is the case (no. 14). Only one combination remains, which 

 at first sight does not fit at all, no. 12, R(ci)/— . If the break near the 

 mutated ci adds to the ci action (like producing a more extreme 

 allele) R(ci)/— should be still more extreme than ci/— with the 

 exaggeration effect, but it is almost normal. It would be almost hope- 

 less to explain this result by the same theory as the others were it 

 not for the fact that Stern (1948) found that R(ci) may have a 

 break to the right of the ci locus; that is, the piece of heterochromatin 

 with the break is not involved. Here it is not the Dubinin effect that 

 is expected but a standard position effect. This may lead to a still 

 more extreme ci effect, or may not influence the already present ci 

 effect of one mutant locus, or may exercise a dominant position 

 effect, that is, ci with R acting like +. The situation would be similar 

 to that studied by Oliver and Green (see 1 S C c bb). This appar- 

 ently aberrant result would have to be separated clearly from the 

 real Dubinin effect, which involves the heterochromatic section left 

 of ci. This is the weakest point in our entire analysis, (We shall re- 

 turn to these problems in the chapter on dosage.) 



dd. Position effect and point mutation 



As the foregoing chapters on position effect have indicated, 

 there is no difference between position effect and point mutation 

 except that in point mutation no chromosomal rearrangement can be 

 made visible with the means available at present. The obvious con- 

 clusion is that point mutations are rearrangements at the submicro- 

 scopic level. What this means for the theory of the gene will be 

 discussed later. In this chapter a number of specific facts will be 

 assembled which point in the same direction. 



aaa. Deficiencies and Minutes. — For a long time geneticists (e,g., 

 Muller, Stadler) were impressed by the fact that in radiation-induced 

 mutation no clear line of demarcation could be drawn between point 



