324 INHERITED LINKAGE VARIATIONS 



were back-crossed to black vestigial (cultures 230 to 240 inclusive). 

 The results are shown in table 22. 



None of these females had both Cm and Cun but it is possible 

 that one of the factors may have been present, especially in the off- 

 spring of female N. These tests show only that the Nova Scotia 

 stock was not homozygous for Cjj „ and probably not for Cn i. No 

 other stocks from northern localities have been tested, so that it is 

 impossible to even guess whether or not these factors occur frequently 

 in Nova Scotia or neighboring regions. 



ANOTHER SECOND-CHROMOSOME LINKAGE VARIATION. 

 Cultures 733 and 734, referred to elsewhere, contained females 

 of the constitution ^7 — r . As was pointed out above, they 



gave an unexpectedly high percentage of crossing-over for black and 



plexus. Culture 812, descended from the same culture that pro- 



*S b 7) c s 

 duced females 733 and 734, contained a female y^ ^ ^ . This 



^IIl ^Ilr 



female produced 72 offspring, of which none were cross-overs between 



S' and Pr, or between c and Sp, but 11 were cross-overs between p, 



C C 



and c. Later descendants of 812, of the constitution £ -, 



pr c ' 



gave this same increased value for p^ c without any increase for b p^. 

 But it was found impossible to fix this increased value, which fluctu- 

 ated between the expected value (less than 1 per cent) and 20 to 30 

 per cent. Several selection experiments have been carried out in an 

 effort to get a stock that would constantly give the high value, but 

 without success. The most recent of these experiments has now been 

 carried through 23 generations of brother-sister matings, always 

 breeding only from those pairs that gave the ''high" value for p^ c. 

 Yet, in the fifteenth generation, occurred a culture that gave only 1 

 cross-over among 130 offspring, and in the twenty-third was a culture 

 that gave Y4~g^ = 3.4 per cent. The latter value, while slightly higher 

 than is usual for Cm Cn „ is much lower than the 20 to 30 per cent 

 now given by most of the "high" selected cultures.^ The nature of 

 this case has not yet been worked out in detail, though culture 812 was 

 counted in December 1915, and the problem has been worked at 

 continuously since that time. 



The following points now seem fairly certain, though they must 

 still be checked and extended. 



(1) The "high" value is due, in large part, at least, to a dominant 

 gene. 



* The Cjji has apparently been lost, by crossing-over, in part of this experiment. But since the 

 values given above are too high for heterozygous Cn f, the discussion given is not affected. 



