250 Sturtevant. 



usually show about 25 to 30 '^/o crossing- over gave, in some crosses, no 

 evidence of linkage. Tanaka ('13) finds in silkworm moths that the 

 same pairs of genes may show complete linkage in some crosses, about 

 10 to 15 "/o crossing over in others. 



Thus it appears certain that linkage is a phenomenon which is 

 extremely variable. It may seem that this fact makes the sort of cal- 

 culations discussed in this paper dangerous and untrustworthy. I can 

 only answer that the range of variability in chromosomes I and II of 

 Drosophila females seems to be rather narrow (with the exception 

 of the race discussed above) and that when large numbers are ob- 

 tained under varying circumstances the inequalities should even up 

 in the end. At any rate, the scheme as here presented works, and 

 successfully stands the test when used to predict the results of untried 

 combinations. It should be pointed out that the "distances" spoken of 

 are not conceived as representing actual relative spatial relations. There 

 is no means of knowing whether or not crossing over is more frequent 

 in one part of a chromosome than in another. If there are such differ- 

 ences, then our diagrams do not show the relative distances between 

 loci, but I do not think that decreases their value as diagrams. That 

 a unit of distance as here adopted represents the same length of 

 chromatin thread in the different chromosomes seems very unlikely, since 

 even in the same chromosome a given segment may cross over more 

 often under some conditions than under others. The fact that Vg and 

 Sp show 36"/o of crossing over in some females, about l^/o in others, 

 and probably C/o in males, does not show that these genes occupy 

 different loci at different times, but is to be interpreted rather as 

 showing that the amount of crossing over between the two loci varies 

 to that extent under the different conditions. The explanation is prob- 

 ably to be sought in a difference in behavior of the chromosomes 

 during maturation. 



Independence of the Three Chromosomes. 



There have been very few data published from which one may 

 judge accurately as to the degree of independence shown by genes lying 

 in different chromosomes in Drosophila. Morgaj^ ('lib, '12a, '12 e etc.) 

 has published a large number of F2 counts from crosses involving V 

 (chromosome I) and P (chromosome III), and also Y (chromosome I) and 

 B (chromosome 11). Several Fa counts in which the independence of 

 various autosomal genes from the sex differentiating gene may be seen 



