i;;s 



/'// YSlol.oCKM. GENETICS 



(L. V. Morgan). This discovery makes it possible to compare the 

 Bar gene in four different quantities with Normal, without any 

 other locus being involved. The effect of Bar is to reduce the 

 number of facets and may therefore be expressed in facet counts. 

 Table 16 is compiled from Sturtevant's data (B + = Wild, 

 B = Bar, BB = Infrabar) : 



Table 14 

 (From Sturtevant) 



This and other work have led Sturtevant to conclude that no 

 Wild-type allelomorph to B exists. The chromosome which 

 was known to have lost the Bar-gene by unequal crossing over 

 had no other influence upon Round eyes than the normal Wild- 

 type chromosome. The series then would give the actual effect 

 of one to four doses of B in terms of reduction or of production of 

 facet number. The reducing effect obviously increases with 

 dosage of the B gene. (The difference between B/B and BB has 

 been explained as so-called position effect, see page 304.) In an 

 analysis of these results, Goldschmidt (1927c) pointed out that 

 the results might be interpreted on the basis of reaction velocities 

 of a process responsible for facet formation. If we assume that 

 it is the number of divisions of the primary retinula-forming 

 cells that is decreased by the reaction controlled by the Bar gene, 

 we may visualize this reaction as the production of a division 

 promoting substance, which either reaches the necessary thresh- 

 old too late and therefore acts for too short a period or is not 

 produced in sufficient quantity to last long enough. We might 

 also visualize the reaction as producing something that stops these 

 divisions prematurely or as producing an insufficiency which 

 destroys already formed cells (as in the vgr-case). Embryological 

 facts will have to decide between such possibilities (see the work 

 of Margolis discussed on page 76). But in any case it is perfectly 



