92 THE ENTOMOLOGIST. 



occurs as an occasional variation or aberration in another ; in 

 other words, the M of one area is the M + or M — of another. 



We expected also to find another piece of evidence, if the 

 principle discussed in Part I. had come into play ; and that was 

 that the schemes derived from separate areas should not be 

 normal. Although my material is not sufficient to ensure any- 

 thing more than some degree of probability, the scheme derived 

 from the percentages of forms occurring in a confined area is not 

 a normal one. There is too sharp a transition in numbers from 

 those possessing large ocelli to those with the ocelli replaced by 

 spots. 



Further additions are much needed to confirm or correct the 

 imperfect observations here recorded, both in regard to the 

 species treated, and also other species of this and other genera. 



To recapitulate the whole discussion — the two leading charac- 

 teristics of the species of the genus Erehia are variability and 

 numerical disproportion of sexes. These two characteristics are 

 put into causal connection by the principle that equality in the 

 numbers of the sexes tends towards equilibrium, i.e. constancy 

 of hitherto variable characters, and that inequality when coupled 

 with variability in the preponderating sex tends towards con- 

 tinued fluctuation of variable characters. This is effected in the 

 following ways : — 



When the sexes are equal in numbers constancy of characters 

 is attained by — 



1. Kegular product of variability of one sex into constancy 

 of the other, under undisturbed regulation of the laws of 

 chance. 



2. Eegular product of variability of one sex into variability 

 of the other, under undisturbed regulation of the laws of 

 chance. 



When the sexes are unequal in numbers fluctuation is pre- 

 served by — 



1. Irregular product of variability of one sex into constancy 

 of the other, the laws of chance acting irregularly. . 



2. Ditto, when both sexes are variable. 



The evidence of the existence of this principle as a working 

 factor would be that the individuals of a variable species satis- 

 fying the conditions of principle, and taken from a confined area, 

 do not fall into a normal scheme with respect to the variable 

 characters under consideration, although individuals taken from 

 many areas very likely do so ; and also that the M of different 

 areas would differ both from one another, and from the M of all 

 the areas combined. Observations were then given on E. tyn- 

 'darus, melampus, and ligea-euryale, which tended to show that 

 such evidence was forthcoming. 



