386 FAUNA HAVVAIIENSIS 



D. perkinsi and its near ally D. gernmniis have no connection with the other 

 Dromaeoliis found in the Hawaiian Islands, but I do not separate them as a distinct 

 genus because of the unsatisfactory condition of the taxonomy ol the subfamily. 



I have recognised t^'^ species of Dromaeoliis in the Hawaiian Islands. It is 

 probable that all are precinctive. The material before me for dealing with this division 

 is very inadequate. Leaving out of the question the two isolated species mentioned 

 above we have less than loo specimens, an average of about three to a species. As 

 some of the species of Eucnemini are known to be variable, this material is clearly 

 inadequate to enable me to decide on the specific limits with any certainty, and my 

 attempt to define the species must be considered as merely tentative. 



I find that the phenomenon of fiightlessness exists, as some of our most peculiar 

 species have nothing but mere rudiments to represent the wings. Others have the 

 organs of flight fully developed. Owing to the scanty material before me I am unable 

 to deal with this question satisfactorily, and in many cases I am only able to guess as to 

 whether the insect has wings or not. I am not aware that any flightless Eucneminae 

 have been detected elsewhere. 



The following key is therefore merely a temporary device and is perhaps not 

 even accurate. 



Group 1. Large insects, elongate and narrow, fully winged, with short coxal 

 laminae (all the other forms have longer coxal laminae). Species i and 2. 



Group 2. Winged forms, slender in build and elongate in form ; usually of small 

 size. Species 3 — 16. 



Group 3. Clumsily built forms of large or moderate size, most of which I believe 

 to be winged forms. But this is in several cases a mere guess. Species 17 — 2"]. 



Group 4. Small forms with thora.x large in proportion to the after body, which is 

 always reduced in length. Most, if not all, of the species are pretty certainly wingless. 

 This is however specially doubtful in the case of nos. 28 and 30, D. cephalotes and 

 D. haivaiiensis. Species 28 — t^Z- 



Dromaeolus Kiesenwetter. 

 Dromaeoliis Kiesenwetter, Nat. Ins. iv. 2, 1858, p. 197. 



Group 1. 



( 1 ) Dromaeoliis perkinsi, sp. nov. 

 Elongatus, angustus, nigerrimus, nitidus, pube albidescente sparsim vestltus, protho- 

 race canalicula brevi. Long. 10 — 15 mm., lat. 3 — 2,\ mm. 

 Plate XIII. fig. II. 

 Entirely black, except that there is a white pubescence- — scanty on the upper surface, 



