STURTEVANT 



employed here (see below). These 

 facts eliminate the possibility of ex- 

 plaining the result as due to accidental 

 genetic or environmental differences. 

 The second possible escape from the 

 conclusion as to the effect of position 

 lies in the assumption of an effect pro- 

 duced by the round allelomorph. This 

 has been tested by determining the 

 effect on facet number of reverted bar 

 and reverted infrabar. Round obtained 

 bv^ reversion from homozygous bar or 

 infrabar stocks cannot carry a normal 

 allelomorph on the view advanced in 

 this paper, unless such an allelomorph 

 is already present in the parent stocks. 

 But such a gene is, almost by defini- 

 tion, not an allelomorph of bar; and in 

 an\' case cannot be supposed to pro- 

 duce the effects here under discussion, 



since the ^ and the — — would both 

 carry it. 



141 



Two different rounds by reversion 

 have been introduced into the inbred 

 stocks, by the .same method as used for 

 BB' and B'B'. Care was taken never to 

 use flies carrying one of these reverted 

 rounds in the same culture with the 

 old round of the inbred stock, so that 

 it is certain that these new rounds are 

 really due to reversion,— if not from 

 the supposed source, then from some 

 of the allelomorphs within the inbred 

 stock, since a new reversion may have 

 occurred during the process of getting 

 the desired modifiers into the reverted 

 round stocks. One of the reverted 

 rounds, called "rev. B," came from the 

 homozygous bar experiments of table 

 1; the other, called "rev. B'," came 

 from the homozygous infrabar experi- 

 ments of table 8. The results of these 

 tests are shown in table 24. 



These data suggest that the rever- 

 sions,— especially the more thoroughly 



Table 24 

 Facet counts to test iiatnre of reversions 



tested one from infrabar,— may give 

 slightly smaller eyes than does the 

 wild-type allelomorph. Certainly they 

 do not give larger eyes. And still more 

 certainly, with BB or BB' they do not 

 give as large eyes as do the correspond- 



B B 



ing types— and— . These data elimi- 



nate the second possible method of ex- 

 plaining away the effect of the position 

 of the bar allelomorphs on facet num- 

 ber. The conclusion must stand as 

 stated on p. 137. 



It seems probable that such an in- 



fluence of the relative position of genes 

 on their effectiveness in development 

 may be interpreted in terms of diffu- 

 sion and localized regions of activity 

 in the cell. This idea is, however, 

 scarcely worth elaborating until more 

 evidence is obtained. It may, however, 

 be pointed out that there is another 

 possible application of the hypothesis 

 of a position effect. 



It has been shown by Bridges ( 1 92 1 ) 

 that in triploid individuals the reces- 

 sive genes brown, plexus and speck do 

 not become dominant to their normal 



