ATOPETHOLID MILLIPEDS — HOFFMAN AND ORCUTT 101 



second article in such forms usually extends distad beyond the rounded 

 apex of the genae. There seems to be no significant variation with 

 respect to the size and distribution of setae and pubescence on the 

 antennae, and all known species have four terminal sensory cones. 

 No other sensory structures or areas have been detected. 



Mandibles: The basal joint of the mandibles is somewhat enlarged 

 on the outer surface and appears elongated in the dorsoventral direc- 

 tion; the collum is correspondingly cut away on the anterior edge to 

 accommodate this convexity. There is therefore no provision for an 

 antennal groove such as occurs in the Spirobolidae, the antennae 

 being held instead in a shallow depression on the outer face of the 

 stipital joint of the mandible. This surface m^ay be either smooth 

 or moderately transversely striate and may be shown by future 

 studies to be useful in the recognition of species, although the surface 

 seems to vary within the limits of a genus. The free ventral edge 

 may or may not be set off by a fine marginal ridge. 



Gnathochilarium: The appearance of the gnathochilarium is 

 essentially the same in all genera of the family and seems not to differ 

 from that in most spiroboloid families. It may be noted in passing, 

 however, that the base of the mentum is produced on each side into a 

 distinct lobe possibly homologous to the cardine, which is said by 

 Attems to be missing in the Spirobolida. This development is readily 

 seen in Eurelus soleafus (see fig. l,a). 



Hypopharynx: The shape of the hypopharynx was not investi- 

 gated by previous workers for its possible utility in classification. In 

 the belief that some differences may exist in the structure at least at 

 the family level, we introduce a dramng (fig. l,b) of the hypopharynx 

 in Eurelus soleatus for comparison with species in other families. 



Collum: In general appearance the collum is similar in all members 

 of the family in being smooth, evenly acuminate laterad, and with the 

 anterior margin excavated, but there is considerable minor variation 

 reflecting specific differentiation. Particular reference is made to the 

 shape of the lateral ends of the coUum, and the submarginal anterior 

 groove on each side. In Centrelus kerrensis (fig. 5,c), for instance, the 

 groove is very deep and distinct and sets off the edge as a sort of swollen 

 ridge, which is extended ventrad into a rounded projection more 

 pronounced than in any other atopetholid. In Arinolus torynophor 

 (fig. 10,6) the caudal edge is somewhat concave just before the end 

 so that the extreme tip seems to be produced caudoventrad, a peculi- 

 arity not observed in A. apacheUus or A. citrinus. In species of 

 Watichelus the condition is even more accentuated and thus provides 

 a means of separating them from the closely related species of Atopetho- 

 lus. In some forms the submarginal groove is nearly parallel to 

 the anterior edge; in others (fig. 7,(1) the groove becomes more remote 



