582 WHITE— STUDIES OF INHERITANCE IN PISUM. 



although the data are scant. R and Tl as shown by Vilmorin, 

 Bateson and Pellew are only partially coupled, there being a small 

 per cent, of the combinations rTl and Rtl in Fo. These are inter- 

 preted by Morgan and his students as cross-overs or combinations 

 due to the simultaneous breaking of the chromosomes with respec- 

 tively rtl and RTl at a point between the two kinds of factors and the 

 subsequent union of the parts so as to bring r and Tl, R and tl 

 together. This breaking occurred in about 1.5 per cent, of the total 

 cases as regards the factors R and Tl. S and Bl or W are also in 

 all probability partially coupled, similar to the case just described. 

 The work of Morgan and his students on Drosophila has shown 

 that by assuming that the linked factors of a group are arranged in 

 an end-to-end straight-line series, definite places in the chromo- 

 some may be assigned to each factor, and their relative distances 

 from each other may be given in terms of a standard unit equal to 

 I per cent, of crossing-over. When a large number of the factors 

 of a single chromosome have been studied the relative frequency 

 of the cross-overs of the various factors may be approximately 

 calculated and predicted. 



When the relations in inheritance of the various factors to each 

 other in such a group as Pisum are worked out, a definite basis for 

 predicting correlation between different characters will have been 

 found. On this basis, it will be possible to calculate with compara- 

 tive ease the somatic characteristics of Fo hybrid populations in- 

 volving large numbers of factors, because so many of these char- 

 acters will be associated together by linkage and may be considered 

 as the expression of a single factor. Supposing the inheritance of 

 a hundred factors in Pisum is involved in a cross about which it is 

 desirable to have forehand knowledge. If each is independent of 

 all the others in its inheritance, it is obvious that accurate predic- 

 tions in regard to the combinations would be made with great dififi- 

 culty, but if these are linked together in large groups, predictions 

 can be made with fair accuracy and considerable ease. 



Crossing-over makes predictions regarding Fo hybrid popula- 

 tions somewhat more difficult than if the factor linkage was abso- 

 lute, "but at the same time they bring about new combinations in 

 predictable proportions which, in a system where the coupling was 

 absolute, would not be possible. 



