602 CALVIN B. BRIDGES 



It is possible that some of the anomolous appearances of fiiea 

 in the cultures of Morgan and Cattell may be explained in this 

 way. 



The case referred to by Sturtevant ('13 b) is probably one of 

 non-disjunction coming from miniature white stock. 



There are many other cases similar to these, some of which 

 involve other sex-linked factors than those already given. 



The sum of all the Fi counts derived from single females which 

 were themselves matroclinous is as follows: 



matroclinous 9 hybrid 9 matroclinous cf patroclinouscf 



133 2934 2677 138 



This gives the percentage of exceptions as 4.83, or roughly 5 

 per cent. 



At first glance it would seem that these exceptions are the 

 result of partial sex linkage, that is, that we are dealing with a 

 linkage ratio in which sex is one of the characters. That this 

 is not the case, can be seen from the fact that the results are 

 independent of what particular sex linked factor is used, while 

 in linkage, each different combination has a characteristic value, 

 for instance, the percentage of crossover of white and yellow is 

 about 1.2 per cent, of white and vermilion about 33 per cent, of 

 white and miniature about 35 per cent, and of white and rudi- 

 mentary about 44 per cent. If the sex differentiator gives 5 per 

 cent of crossover with white it ought to give 33 + 5 or 33 — 5 

 with vermilion, 35 + 5 or 35 — 5 with miniature and, so on. 

 From the fact that I get no such variation depending on the 

 particular factor used, I conclude that this phenomenon is not 

 one of partial sex linkage. It is obvious, however, that non- 

 disjunction forms an alternative hypothesis for all the cases which 

 have been described as partial sex linkage. In some of these 

 cases it is to be preferred, but no decisive choice can be made in 

 any particular case between the two hypotheses, unless, as in 

 Drosophila, there are present more than one sex linked factor. 



Morgan and Sturtevant have shown that a certain number of 

 sex linked factors can be arranged in a consistent linear series 

 if the percentage of crossovers be used as the measure of the dis- 



