70 A.A. Girault: 



segmentation, the venation is similar, the wings are similar in every- 

 thing but discal ciliation, tlie body is nearly alike, differing slightly 

 in size. The only diff erence between tlie two genera is this : In Brachi- 

 Stella the discal ciliation of the fore wings is normal, that is, dense 

 enough so that no regulär lines of ciliation are observable; but in 

 Abbella this ciliation is noticeably less dense so that a half 'dozen or 

 more regulär and long lines become prominent, the rest of the ciliation, 

 irregularly placed in more or less dense patches between these lines 

 around the wing's apical margin. If one will examine Trichogramma, 

 a genus typical of those bearing the radiating arrangement of the 

 ciliation, although it will be seen that the alignment of the ciliation is 

 more distinct and prominent, yet the same irregulär patches of the 

 ciliation is more distinct and prominent, yet the same irregulär patches 

 of the ciliation are between the lines near the apex. Around the apex 

 of the fore wing in Abbella, the straight lines are hardly discernible 

 from the rest of the ciliation. Thus, the change from normal ciliation 

 to that of radiating lines is merely a change of density in ciliation and 

 we have reason to doubt the value placed upon it for taxonomic pur- 

 poses, since it is so gradual, continous and noncorrelated. In many 

 genera of the allied Aphelininae the relative density of the discal 

 ciliation is a specific character and this is true for genera of the My- 

 maridae. 



Though it is a general principle in taxonomy that ,,the less any 

 part of the Organization is concerned with special habits, the more 

 important it becomes for Classification" or the less adaptive a part, 

 the more its importance (since it would of necessity be less variable 

 and more constant, not changing with a change of habit), yet there are 

 numerous exceptions and this variability of the wing ciliation is one 

 of them. Discal ciliation of the fore wing is certainly not an adaptive 

 character, since it may be entirely absent or very sparse (species of 

 Oligosita) without change of habit and because we are unable to see 

 what particular use it m.ay be in this case, as an adaptive character. 

 Surely, because Trichogramma has a different arrangement of hairs 

 on its fore wings than has Tumidifemur we would not be justified merely 

 upon that ground, in inf erring that the two genera differ in habit; 

 experience would not lead us to connect the two, since it would soon 

 shew that the occurrence of one kind of ciliation was not always followed 

 by the occurrence of one kind of habit ( Tumidifemur and Trichogramma- 

 tella have opposite sets of ciliation but both are parasitic upon the 

 same host). Moreover, it is a general taxonomic principle that organs 

 of high physiological importance are of great value, but again with 

 this principle, if it may be so called, there are numerous exceptions. 

 We are lead, therefore, to the conclusion that the Classification of any 

 group whatsoever is dependent not upon characters valued in reference 

 to thoir function or their form but upon characters in reference to 

 their relationship and broadly speaking, this will be found to depend 

 on variability or nonvariability, since the theory of descent makes 

 it necessary that stable characters are those which have been handed 



