The Behavior of the Chromosomes as Studied through Linkage. 



239 



The figures as given in the diagram represent calculated distances 

 from Y. In every case the observed "distances"' between adjacent loci 

 are used for calculating positions, for reasons stated in my earlier paper 

 and in the next section. Thus, tiie above diagram is based on the 

 crossing over between V and W, W and V, Y and M, M and R, R and Br. 



It will be seen that the observed results agree with expectation 

 to this extent: Y gives a series of values which increases in the expected 

 order — YW < YY < YM < YR < YRr. In fact, in table I there is 

 only one case in which two values differ from eai-h other in the wrong 

 direction: YR should be less than MBr by about 1 percent, whereas it 

 really gave a slightly higher result. There is some reason to suspect 

 that YR may not be very accurate, since it is based on smaller num- 

 bers than most of the combinations, and since the data show a strong 

 differential viabüitv as regards the rudimentarv winged flies. 



Table II. 



T\Tien we compare the calculation with actual observations for given 

 combinations (see table II), it becomes obvious that there is some further 

 complication present. Thus, the calculation for YR is 53 '3, but 42 6 

 was observed. This discrepancy is consistently greater for the longer 

 distances than for the shorter. WM is calculated as 35 '4, and gave 

 33' 1: but for WBr calculation is 56 '6, observation is only 43 '9. It is 

 significant that calculation is greater than observation in all but three 

 cases, and in these the difference is slight. As was pointed out in my 

 former paper this discrepancy is not to be considerad as evidence against 

 the ideas put forward by Morgan but is a corollary which might 

 have been deduced from them. It is due to the phenomenon of ''double 



