112 PROCEEDINGS OF THE NATIONAL MUSEUM vor. iii 



families and then as suborders. There is now some doubt that this 

 arrangement can go unchallenged. 



That the Atopetholidae occupies a somewhat intermediate position 

 is evidenced by the fact that Attems placed Saussurobolus in the Trig- 

 oniulidae despite its lack of a gonopod sternite; presumably he was 

 influenced by the overall similarity in other respects. Yet Verhoefi 

 considered the closely related Orthichelus to belong to the other sub- 

 order, and made it the type of a new family allied to the Khinocricidae. 

 It seems possible that both were not far from the truth, and that a 

 division into suborders may have to be drawn along lines other than 

 those now in use. 



The gi-oup that most resembles the Atopetholidae as far as gonopod 

 structure goes is the group typified by the East Asian Sjnrobolellus. 

 The species of this group have the peculiar coxal apodemes, and the 

 posterior gonopods are not strikingly different from those of the 

 atopetholids. Another group in which coxal apodemes occur is the 

 Khinocricidae, which is pretty clearly a well specialized ensemble and 

 has a distinct solenomerite on the posterior gonopods as well as a 

 vesicle in the base of the telopodite. Stabilization of the clypeal 

 foveolae at a constant number of four and the development of scobinae 

 are likewise to be considered specializations. 



The rhinocricid posterior gonopod preserves the same number of 

 segments as are found in the atopetholids, but differs in that the coxae 

 are vertical instead of transverse. It is easy to see the similarity of 

 this appendage in Piedolus and Eurhinocricus. At the present, it 

 seems reasonable to regard the Atopetholidae as a very primitive 

 family whose progenitors doubtless gave rise to the existing families 

 Khinocricidae and Spirobolellidae as well as, perhaps, to the Trigoni- 

 ulidae. There are, unfortunately, no species of spiroboloids which 

 have retained most of the primitive characters. We can assume that 

 the most primitive gonopods are those that most nearly approximate 

 the original walking-leg structure, and on this basis the anterior 

 gonopods of Comanchelus are those that are least modified. Retention 

 of a sternite remnant between the posterior gonopods — obviously a 

 primitive character — is known onl}^ in the TrigoniuUdae, but in this 

 group the gonopods are strongl}" specialized by the presence of coxal 

 vesicles and associated seminal groove and solenomerite. It there- 

 fore seems safe to assume Comanchelus to be, probably, the most 

 primitive of existing spiroboloids as its gonopod structure could not 

 be any more generalized without the subdivision of the telopodite into 

 several segments. 



Within the family, then, the least speciahzed group is the Eurehnae, 

 followed closely by the Atopetholinae. Onychelus, which is in a sub- 

 family of its own, is somewhat aberrant and out of the main Une of 



