CYDNIDAE OF THE WESTERN HEMISPHERE — FROESCHNER 339 



considered diagnostic of all genera except Scaptocoris and the extra- 

 limital genus Stibaropus. The presence of a fringe of close-set 

 bristles on the apices of the middle and posterior coxae and the location 

 of the spiracle in the membranous anterior part of the second sternite 

 are nearly unique within this superfamily, being shared only with the 

 corimelaenids. As brought out later in this paper, the leg armature 

 may be simply an adaptive feature for the fossorial habits of these 

 insects and not at all an indicator of phylogenetic relationships. 

 This character shows considerable variation from genus to genus. 

 The same criticism may be valid for the coxal bristles, which are 

 present only on the ventral or exposed side of the structure where they 

 may be functional in preventing sand and grit from entering the 

 articulation during burrowing. Therefore, even though these features 

 furnish good recognition characters theu- actual value as indicators of 

 phylogeny within the Pentatomoidea is open to question. This 

 uncertainty in accepting proposed characters for separation of the 

 groups included in the Pentatomoidea once again emphasizes the need 

 for a very thorough study of the higher classification of the group. 

 Until such a study is carried on by someone with access to collections 

 containing goodly representation of all parts of the Pentatomoidea, 

 I feel free to follow my usual tendency to be a "splitter" at the family 

 level when the breaks in the morphology and biology of the groups 

 permit a distinct and independent biologic-taxonomic concept to be 

 formed. 



The problems in the classification of this group have not been 

 confined to the family level. Instead, they are evident at all levels. 

 Previously only two subfamilies have been recognized, whereas at 

 least five are strongly evident in the material at hand. 



Most authors have considered the genera from one of two extremes — 

 either with the idea that any prominent or unusual feature (regardless 

 of its value as a phylogenetic indicator) automatically serves for the 

 estabhshment of a genus, or, from the other extreme, that the limits 

 of previously erected genera must constantly be expanded to take 

 in new forms that appear regardless of the relationships of the species 

 involved. The former approach has resulted in too many mono- 

 typic genera (i.e., Colobophrys Horvdth, Cryptoporus Uliler, Pachy- 

 meroides Signoret, Psedrocephalus Van Duzee and Syllohus Signoret 

 to mention some from the Western Hemisphere) as characters of no 

 more than specific value often have been used, while the second 

 method has resulted in a few "catch-all" genera (i.e., Aethus and 

 Geotomus as accepted by most recent authors) that have worldwide 

 distribution and consequently little or no zoogeographic significance. 

 I believe that if a genus is to consist of a group of "closely related" 

 species, consideration must be given not only to the characters which 



