88 CASEY 



the taxonomic assignment of those forms represented by single exam- 

 ples, although of course mistakes under these circumstances may 

 readily occur. Difficulties of this kind, nevertheless, scarcely war- 

 rant such carelessly sweeping opinions as that put forth by Crotch, 

 in surmising that all the immaculate forms allied to maculativentris, 

 constitute a single species, for this is certainly very far from being 

 true. It is also most improbable that there can be the transitions be- 

 tween lineata and maculipennis hinted at by LeConte, as apparently 

 proved, at any rate, by the large series at hand. 



The genus Buprestis, as represented by octoguttata, which was sug- 

 gested by Kerremans as the type to be definitively adopted and to 

 which no valid objection can apparently be advanced, includes a 

 wide range of bodily habitus, but the elytra are regularly striate 

 throughout, without the faintest indication of such irregularities as 

 the depressions or areolae of the preceding genera, and this would 

 seem to be the most important general distinguishing character of 

 Buprestis within this group having the outer antennal joints elongate, 

 more narrowed toward base, inferiorly punctate and with a distinct 

 antero-internal sensory fossa. The frontal margin is always broadly 

 sinuato-truncate, the antennas inserted in triangular cavities which are 

 generally accentuated internally by the moderate though clearly de- 

 fined obhque ridges; the latter, however, become obsolete in langi, 

 rufipes and related forms. The labrum is usually wholly pale and 

 coriaceous, but in langi and allies it becomes partially corneous and 

 metallic in lustre; it seemed at first as though this disparity in labral 

 structure might form the basis for subgeneric division, but as it would 

 have placed two such very similar species as rufipes and elongala in 

 different subgenera, it evidently fails to have any such value. The 

 pronotum generally has a smooth and punctureless, more or less 

 embossed median line, but in langi and other similar forms it is wholly 

 devoid of this character, the surface being feebly impressed and more 

 punctured along the median line. The prosternum and first ventral 

 segment may be impressed to strongly sulcatc, or without trace of 

 impression and with pronounced even convexity. The legs are slen- 

 der, the basal joint of the hind tarsi more or less considerably elongate, 

 and the anterior tibiai of the male may have a strong reflcxed spine 

 internally near the tip or betray no vestige of any such structure. 

 The abdominal apex is more or less truncate in both sexes, though, as 



