THE NATURE OF THE GENE 375 



This discovery does not alter the fact that a fly homozygous for Bar, 

 with four of the segments arranged two in each chromosome, has a 

 larger eye than "a heterozygous double Bar" fly with four segments 

 arranged three in one chromosome and one in the other. It seems that 

 the order of the chromosome sections may have a developmental effect, 

 and Muller^ suggested that this may be the reason for the commonly 

 observed fact that phenotypic effects are associated with chromosom.e 

 rearrangements in which no gross losses or gains of chromatin can be 

 shown to be involved. This explanation turned out to be satisfactory, 

 and the phenomenon, knov/n as the position eff"ect, is now generally 

 accepted. A classical case is that of cubitus interruptus, a IVth chromo- 

 some recessive studied by Dubinin and Sideroff.^ The position effects 

 consist in alterations of the degree of dominance of the normal allelo- 

 morph in translocations of the IVth to the X, Y and Ilnd chromosomes. 

 MuUer^ has also described some very striking examples in which 

 different breakages near the locus of the gene scute altered the pattern 

 of bristles which were affected by the gene. Moreover, the same break- 

 age turned up on two or more different occasions, and was always 

 found to have the same effect. 



We must conclude that the expression of a gene may be affected by 

 the nature of the genes in its neighbourhood; it is not known exactly 

 how far along the chromosome the influence extends, but its range is 

 certainly short.>The most usual interpretation of this phenomenon is 

 that the connections between genes are not purely physical, but involve 

 chemical changes in the united genes. In a sense, then, the whole 

 chromosome should be considered as a single chemical unit. We can- 

 not, however, exclude the alternative and perhaps more commonsense 

 view which attributes the position effect not to a direct interaction 

 between the genes, but to reactions between the primary products of 

 gene action. If, for instance, two genes both act on the same substrate 

 during development, their rates of reaction may well not be the same 

 when they are lying side by side in the same chromosome as when they 

 are widely separated in different regions of the chromatin. 



Muller"* has drawn attention to the fact that inversions and trans- 

 locations may occur which are too small to be detected by cytological 

 or crossing-over investigations. Their position effects would then be 

 falsely interpreted as gene effects. Several cases are already known 



1 Muller 19326. 2 Dubinin and Sideroff 1934. 



' Muller and Prokofieva 1935, cf. Gnineberg 1936. 



* Cf. Muller and Prokofieva 1935, Muller, Prokofieva, and Raffel 1935. 



