20 CALVIN B. BRIDGES 



we might calculate how much interference should rise if that 

 change be due to an equivalent looser twist. Thus an experi- 

 mental method is provided for the analysis of the mode of twist- 

 ing and the distribution of chiasmas, not only under the changed 

 condition, but also under the normal condition as compared 

 to the changed condition. It is, however, no small task to 

 secure data for such a study, and in any other material than 

 Drosophila the problem would be wellnigh hopeless of solution. 

 At present the unfavorable case of black purple curved furnishes 

 only enough data to give a suggestion as to the mode followed. 



The data for the calculation of interference in the case of black 

 purple curved are given in table 7 and for the first broods may 

 be summarized as follows: 



B Pr Cv B_| Pr Cv B Pr | Cv B | Pr ] Cv 



3,053 170 669 , 42 



Percentage 77.61 4.32 17.4 * 1.07 



Here the total amount of crossing-over between black and 

 purple is 4.32 + 1.07 per cent, and between purple and curved 

 is 17.4 + 1.07 per cent. The expected percentage of double 

 crossing-over is therefore 5.39 per cent of 18.07 per cent, which 

 is 0.97 per cent. The observed percentage (1.07) of double 

 crossovers was somewhat larger than this. Although the 

 difference is so small that it may be due to chance fluctuation, 

 yet it will be instructive to consider its meaning on the assump- 

 tion that it is not due to chance. The actual increase of 0.1 

 per cent is a relative increase of 11 per cent over the expected 

 0.97 per cent (percentage of coincidence 111). But inter- 

 ference which increases the percentage of doubles is a reversal 

 of the ordinary type and the explanation of this 'negative' 

 interference is as follows: 



The preceding considerations have applied to the case in which 

 the length of the average internode is such that when one node 

 lies between A and B, the next node will most often lie beyond 

 C-D. This relationship between the relative position of the 

 gens and the length of the internode is such that a chiasma 

 between one couple will tend to prevent one between the other 



