3 12 - 



THE QUARTERLY REVIEW OF BIOLOGY 



otherwise it could hardly be understood 

 that hybrids between light and dark races 

 have the same curve as pure intermediate 

 races (The effect of the allelomorphs 

 At At = effect o£Ad + AT). 



The work of Ford and Huxley ('2.7) does 

 not deal with multiple allelomorphs but 

 with ordinary genes controlling eye-color 

 in Gammams. But as eye colors in other 

 animals (Drosofhila) furnish typical 

 examples of multiple allelomorphism it 

 might be quoted at this point. The 

 authors found in Gammarus chevreuxi races 

 with different rates of velocity in regard 

 to the darkening of eye-pigment, as shown 

 in empirical curves. It was shown that 

 the hereditary difference of the two races 

 was that of a simple Mendelian gene, 

 thus demonstrating that certain genes are 

 causing definite rates of reaction. In 

 this case there was no way visible to 

 connect the different rates of the specific 

 developmental process with different gene 

 quantities. On bringing together facts 

 of this type with such ones as reviewed 

 before the present author has been led to 

 consider whether or not the majority of 

 mutations consist in changes of the 

 typical quantity of the gene ('17, '2.0b, 



'2-3 a > ?7 a )' 



As mentioned before a direct demon- 

 stration that multiple allelomorphs are in 

 fact different quantities of the same gene 

 can be found in Sturtevant's ('2.6) work on 

 the bar eye of Drosofhila. It is known 

 that bar is a dominant mutation from 

 normal, reducing the number of facets. 

 A further reduction was found by 

 Zeleny in his mutation ultrabar, which 

 behaved genetically as a multiple allelo- 

 morph to bar and normal. Sturtevant 

 however proved in a series of exceedingly 

 beautiful experiments that ultrabar is the 

 product of unequal crossing-over resulting 

 in the location of two bar genes in one 

 chromosome. Thus he was able to build 



up individuals with 1, z, 3 and 4 such 

 bar genes and to study the effects on 

 facet number. He further found another 

 mutation in the bar series, infrabar, 

 having an intermediate effect between 

 normal and bar. But this gene has some 

 special features in regard to the pheno- 

 typical effect as well as in its behavior in 

 temperature experiments . Therefore Stur- 

 tevant does not consider it as a member 

 of the otherwise quantitative series. 

 The present writer is of different opinion, 

 but for the present purposes a discussion 

 of this point might as well be left out of 

 consideration. Here then we have a 

 series of genotypes behaving experimen- 

 tally like multiple allelomorphs. But it 

 is an experimental fact that the majority 

 of the members of the series are formed by 

 different quantities of a gene. This series 

 is linked with a series of facet numbers 

 exactly parallel to these quantities. An 

 analysis of the following tables computed 

 from Sturtevant's data will illustrate our 

 point. If B is the gene for normal, By 

 for infrabar, B 2 for bar the combinations 

 shown in table 1, page 313, have been 

 produced. 



We can arrange this table so that the 

 different possible types of heterozygous 

 forms parallel the respective homozygous 

 forms and get the table (the numbering 

 of the combinations is the same as in 

 the first table), shown in table z, page 313. 



This table seems to the present writer 

 very impressive. With the single excep- 

 tion of No. 13 all the heterozygous 

 combinations closely parallel the homo- 

 zygous series in regard to decreasing 

 numbers of facets. Sturtevant has now 

 proved conclusively that No. 4 has four 

 quantities of the gene B u No. 5 two 

 quantities of B x and B 2 each, No. 6 four 

 quantities of B 2 , etc. This series 1 to 6 

 affects increasingly the number of facets 

 in negative order. Among the members 



