116 T. H. MORGAN 



gives 26.5 per cent. Similarly the totals based on the experi- 

 ments with 1:1 ratios gives 25.2; and those based on the females 

 alone of the 2:1 ratios give 27.5. The difference in these is not 

 greater than is expected. The lower percentage for the males 

 taken alone is again probably due to viability. 



Summary of linkage data of Part II 



In the following table all the data available in the preceding 

 pages (Lethal ii) bearing on each linkage value have been brought 

 together in a grand total: 



per cent 



total cross-overs cross-overs 



White (or eosin) lethal II 8,152 812 10.0 



Lethal II, miniature 6,407 974 15.2 



Lethal II, vermilion 1,604 313 19.5 



White (or eosin) miniature 36,022 11,048 30.7 



Eosin (or white) vermilion 6,023 1,612 26.8 



In the preceding table there are some discrepancies that appear 

 when the percentages of crossing-over of the different counts of 

 the same experiment are compared. Discrepancies like these 

 may be due sometimes to insufficient numbers, sometimes to 

 variations in differential viability (including the lagging behind 

 of some of the male classes) ; in the case of lethals the absence of 

 the 'contrary' classes (since one class dies) makes it impossible 

 to check up viability in the male classes. Special external and 

 internal conditions may at times affect the degree of interchange 

 between the homologous chromosomes, which, by changing the 

 gametic ratio, will affect the realized classes. 



To determine the loci of factors accurately special studies and 

 corrections are necessary. The present data pretend no more 

 than to give the approximate positions of the loci, and, in this 

 sense, the results are as consistent as is to be expected under the 

 conditions of the experiments that were carried out with another 

 end in view. 



