178 



INHERITANCE IN RATS. 



from the actual proportions of the four possible kinds of gametes which 

 the FI parents produced. With no linkage these gametes should be of 

 four sorts, all equally numerous, viz, RP + Rp + rP + rp. Linkage 

 would tend to increase the proportion of the two middle classes (Rp 

 and rP, the original combinations) at the expense of the extremes (RP 

 and rp, the double dominant and double recessive classes) . The latter 

 may be called " cross-over" classes, the former "non-cross-over." In 

 producing the 85 F 2 yellow rats which were tested, twice that number 

 of gametes were concerned, viz, 170. From the demonstrated genetic 

 constitution of the tested animals, we can estimate how many cross- 

 over and how many non-cross-over gametes entered into each. 



The estimated proportion of cross-over to non-cross-over gametes is 

 seen to be 29 : 141 or 1 : 4.8. In the terminology of Bateson and Pun- 

 nett this would be a reduplication series lying between 1:4:4:1 and 

 1:5:5:1; in the terminology of Morgan, 17 per cent of the gametes 

 formed by FI individuals are cross-over gametes. 



We can test this linkage theory in another way. If linkage exists 

 it should modify the proportions of the apparent classes in F 2 as well 

 as of the real classes, which we have just been considering. The 

 apparent classes are three, viz, black-eyed, red-eyed, and pink-eyed, 

 with observed frequencies of 162 : 90 : 72. If no linkage exists the 

 expected frequencies are 182 : 61 : 81, which deviate considerably from 

 the expected frequencies. But if linkage exists, it will lessen the discrep- 

 ancies. Linkage of 17 per cent strength will change the expectations to 

 164 : 79 : 81 . This alteration shows agreement almost perfect in the case 

 of the black-eyed class, a much reduced discrepancy in the case of the 

 red-eyed class, and no change in the pink-eyed class on the whole a 

 much improved agreement between expected and observed frequencies. 



Sturtevant has called attention to the fact that double recessives 

 could occur among our F 2 animals only as a result of cross-overs occur- 

 ring simultaneously in the gametes of both parents, a fact which Wright 

 and I considered too obvious to demand comment in our preliminary 

 paper, but recognized in our calculation by counting two cross-over 

 gametes for every double recessive zygote. Sturtevant has questioned 

 the adequacy of our tests in the case of these doubly recessive indi- 

 viduals because apparently he had formed the idea, from studies made 

 on insects, that crossing-over could occur only in the gametogenesis of 



