1898.] GROTE — SPECIALIZATIONS OF LEPIDOPTEROUS WING. 31 



thus assent to the conclusion expressed by Prof. Comstock in Evo- 

 lution and Taxono?ny, that we find in the Nymphalidae an even 

 greater specialization of the wings than exists in the Pieridae, we 

 admit that the point of view from which this is regarded may influ- 

 ence any conclusion, while the unequal presentation of the changes 

 in the wings renders a just weighing of the differences a matter of 

 some difficulty. It will be sufficient for my present purpose if the 

 impression left on the mind of the reader is that rank is a relative 

 conception and that it is owing to the constitution of our minds 

 that we, are impelled to string one natural object after another, 

 while we are apt to fortify a classificatory preference for a special 

 group out of several lying nearly abreast, by reasons which, suffi- 

 ciently telling as far as they go, are apt to reflect only one side of a 

 complex subject, I think, then, we may believe that the specializa- 

 tion of the " brush-footed " butterflies is more apparent in the feet 

 than in the wings, and that, if we are not inclined to give them pre- 

 eminence on that account in our sequences, we shall not be induced 

 to do it upon the statement of Prof. Comstock herein discussed and 

 illustrated. 



Phylogenetic Lines Among Pierid Genera. 



I have previously shown that coincidence in the number of the 

 radial branches in reduction does not determine common descent, 

 but that a three-branched condition of the originally five-branched 

 radius has been reached independently, not only in different fami- 

 lies, but on different generic lines within the same group. It may 

 be assumed that three-branched species, differing otherwise unes- 

 sentially, are correctly associated by this character ; but to use this 

 character anywhere alone for taxonomic purposes, or to assign it a 

 commanding value, would be plainly to go wrong. It is probable, 

 for instance, that the three-branched radius correctly indicates that 

 the species of Thecla (^in sensu iniht, with the type given by Scud- 

 derj are monophyletic and that the four-branched Zephyrini stand, 

 at least constructively, as representing the original condition of 

 their ancestors. 



Under these views we may sort out several different lines of prob- 

 able descent in the holarctic Pieridae, in which the examples of 

 extreme reduction have been independently developed. It is clear, 

 since nature does not proceed by jumps, that the missing stages 

 between the five-branched ancestors and the three-branched de- 



