THE GENETICS OF HABROBRACON JUGLANDIS ASHMEAD 



proportionate dif f erential vlabil ity , and linked 

 lethals or semilethals. 



It is obvious that v/hat has been said as to 

 the relations of a and b may also apply to the 

 relations of either a or b with a third factor 

 c, or a fourth factor, d. It is sometimes nec- 

 essary to make comparisons of pairs of frequen- 

 cies as regards a and b, separately for groups 

 CD, cD, or Cd since either c or d may affect 

 phenotypic ratios of a or b. 



Table I has been prepared in order that the 

 reader may readily estimate the statistical 

 significance of various pairs of frequencies 

 occurring in the summaries of linkage data in 

 Habrobracon literature. By dividing the lower 

 frequency by the sum (n) a percentage is ob- 

 tained deviating more or less below 50 per cent. 

 (P = .90) are those percentages at or below 

 which 10 per cent of the samples of a given 

 size (n) will be expected to fall , while (P = .99) 

 are those percentages at or below which one per 

 cent of the samples of corresponding size will 

 fall. The reader may then be 90 per cent sure 

 that there is a significant difference from 

 equality between the two frequencies if the 

 percentage is as low as or lower than (P = .90) 

 and he may be 99 per cent sure if it is as low 

 as or lower than (P = .99). 'Whether these sta- 

 tistically significant deviations indicate so- 

 matic overlapping, dif f erential viability , linked 

 lethals, linkage of the two genes under consid- 

 eration, or some combination of these condi- 

 tions must be judged in the individual case 

 from the nature and variability of the traits 

 and comparison of the different percentages 



116 



