322 INHERITED LINKAGE VARIATIONS 



For this combination also no second-brood data is available. Coinci- 

 dence seems to be of approximately the value that is usual, but can 

 be satisfactorily studied only in the series that may have C///, //. 



ri 



The ^r~ L ratios are clearly not very different from those obtained 



Wlr 



with the "usual' ' second chromosome. 



NO TESTS OF HOMOZYGOUS Cm. 



No tests were made of females homozygous for Cm, because it 

 was hoped that a cross-over would occur that would give a S f Cm 

 chromosome, and thus make possible a test of the region in which 

 GUI is located. A few attempts were, it is true, made to get a pure 

 stock of Cm', but no careful records were kept, and these attempts 

 were all unsuccessful. Recent tests show that there is now a lethal 

 gene in the Cm chromosome that is being studied, so that it will 

 probably be impossible to obtain homozygous Cm- It is not certain 

 whether this lethal represents a recent mutation or not. 



TESTS SHOWING NO CROSSING-OVER IN MALES. 



Very few counts have been made from heterozygous males; but no 

 crossing-over in males has been assumed throughout the work, and 

 has been depended on frequently in keeping stocks and in producing 

 many of the more unusual combinations of GUI and C// r . These 

 matings have never produced flies that seemed to result from crossing- 

 over in males, and have always given in later generations results that 

 are consistent with the view that such crossing-over does not occur. 

 Taking this evidence in connection with the counts given below 

 (table 21), and with the evidence that shows crossing-over not to 

 occur in males of Drosophila in any of the chromosomes under any 

 known circumstances, 1 we may safely conclude tnat Cm and C// r 

 do not cause exceptions to the general rule. 



CONSTITUTION OF THE NOVA SCOTIA STOCK. 



The original Nova Scotia female had in her second chromosome two 

 factors for decreased crossing-over. It would be of some interest to 

 find out whether or not this condition was widespread in the stock from 

 which she came. Unfortunately the original stock was lost before it 

 was discovered that two factors, instead of one, are responsible for 

 the result. The following tests are therefore not entirely satisfactory. 

 Three females, from the Nova Scotia stock, were mated to curved 

 speck, and 4, 4, and 1 daughters, respectively, were back-crossed to 

 curved speck. Only a few offspring were counted from each, but 

 enough to show that all 9 females were giving at least 20 per cent of 



1 Except the curious case of "somatic crossing-over" recorded by Muller (1916). 



