72 A. A. Girault: 



their eggs into a covered and concealed host and hence need a longer 

 ovipositor than usual ; secondly, it is not constant, not occurring with 

 a large number of species; and thirdly, of itself, its length varies greatly; 

 and fourthly it is present in very unlike genera, which obviously are 

 not closely related. The third character, the presence or absence 

 of the antennal funicle, I consider second in importance of the four 

 because the antennae are of high physiological importance, the pre- 

 sence or absence of the funicle is not adaptive in natura, is constant 

 for a large number of closely related species and for the group is not 

 very variable; moreover, the antennae themselves, as regards their 

 segmentation, are not variable with species. In regard to the discal 

 ciliation of the fore wings, another unadaptive character, I think we 

 may conclude with safety that it ranks third in order of importance, 

 for it is variable in closely allied genera, one kind of ciliation is present 

 in very unlike genera, an arrangement of genera on its basis leads to 

 parallelism where divergence it seems would be expected, in some 

 genera it is indeterminate and even disappears (in Oligosüa it may 

 be either dense and normal, sparse, faint and in lines, or totally absent, 

 thus becoming a specific character); more than this, its Variation is 

 gradual. This character, therefore, must be valued as less than the 

 presence or absence of the antennal funicle, since it is less constant 

 and accordingly we must not assign to it more than generic importance. 

 The first character thus takes precedence ; f irstly, the kind of venation 

 is constant for a largest number of like genera; secondly, venation 

 itself is nonadaptive; and thirdly, it shews no continuous or gradual 

 variations between the two kinds; finally, arrangement of genera 

 with its aid gives divergence and fulfils the expectations. Still more 

 than these is the impression given of natural evolutionary processes, 

 development from straight venation to curved. Thus, of the genera 

 of the group so far known, it appears that the kind of venation is con- 

 stant for the largest number, the segmentation of the antennae for 

 the next largest and the other characters — exserted ovipositor and 

 discal ciliation of the fore wing — being subordinate and characteristic 

 only of unlike genera. We must thus reduce the latter character to a 

 generic basis. 



There is still one other character in the Trichogrammatidae, of 

 rare occurrence, of importance neither physiologically nor adaptively 

 but which at first thought one might take to possess relatively high 

 taxonomic value. This character is a median sulcus of the meso- 

 thorax ^), common in eulophid genera but rare with this group. Of the 

 knoWn genera, it occurs only in Ittys Girault and two Australian 

 genera just described, perhaps in Trichogrammatoidea lutea and in 

 one species of Oligosüa (a very minute Australian species). These 

 genera are all closely related; all bear straight marginal veins and the 

 antennal funicle is present. But of these genera, excepting Oligosüa, 



*) In Ittys, one of the Australian genera and the Oligosita at least, 

 running along all of the thorax to the end of the long phragma. 



