VARIATION IN LINKAQE 145 



function of their nature. It may be well to point out that 

 in the only cases where the evidence suffices to give an 

 answer to such a question, that answer is very clearly 

 against such a view. For instance, if we determine the 

 linkage between two factors A-M and then exchange one 

 of the intermediate genes for its allelomorph, we find that 

 in general the change has no effect on crossing over 

 between A and M. If we exchange factors outside of 

 A and M — either near them or far away — still no effect on 

 crossing over between A and M is observed. K we sub- 

 stitute one allelomorph for another, in cases where more 

 than two are known, we find no change in the crossing 

 over for that level. This and other evidence shows that 

 crossing over is quite independent of such genes, never- 

 theless there are other specific genes, as shown above, 

 whose sole effect, or main effect at least, is to change the 

 crossing-over values. 



One highly important and significant result of Sturte- 

 vant's work on crossing-over factors should be noticed. 

 The order of the factors is not in any way changed by 

 the ^^ shortening" process, as shown by the experiments 

 in which three or more loci are followed at the same time. 



The most remarkable fact connected with crossing 

 over is that no crossing over at all takes place in the 

 male of Drosophila, and this applies not only to sex- 

 chromosomes (XY) but also to the other pairs or auto- 

 somes. When the absence of crossing over was discovered 

 for sex-linked genes, it seemed probable that this was due 

 to the presence of only one X-chromosome in the male, for 

 at this time Steven's work had led us to conclude that the 

 male Drosophila, like some other insects, is XO. Later, 

 when failure to cross over in the male was found in other 

 chromosomes as well, it was evident that some more gen- 

 eral relation was behind the phenomenon in these chromo- 

 somes at least. It is true that other genetic evidence 

 has shown that the F-chromosome is ''empty'' {i.e., con- 

 tains no genes dominant to any of the mutant genes as yet 



10 



