30 The Graspedosomatidse of North America. 



< Second joint rounded oval, basal joint flask-shaped E 



( much longer than wide, the first not attenate F 



< Third and fourth joints suhequal, narrow Pseildotreiuia 



( much longer tlian fourth, thicker than second BactropUS 



J First joint gradually attenuate distad Underwoodia 



( distally broader, notched and toothed Caseya 



( Second joint with a small chitinous knob, genitalia simple G 



( smooth ; the first attached to posterior genitalia. . . . Conoty la 



\ Second joint clavate; eyes present; segments 28 Tricliopetalum. 



( obclavate ; eyes wanting ; segments 30 Zygonopus 



It should be remembered, however, that these genera rest on 

 other characters, and that for the present, at least, the ninth legs 

 are not a necessity to the generic classification here adopted. 



American species of this family have been referred to the fol- 

 lowing genera not here recognized as valid or as properl}' ap- 

 plied. It uxsij be worth while to notice briefly the different rea- 

 sons for such treatment. 



Spirostrephon, Brandt (1840) was erected to contain Lysio- 

 petalum lactarium (Say), then looked upon as one of the lulidse. 

 To use the name for species distinct from lactarium by diflfer- 

 ences of famil}^ importance was entirely unjustifiable. 



Campodes of C. L. Koch (1847), was referred to this family by 

 its au.thor, and was not heard from afterward until Mr. Bollman 

 videntified Gryptotrichus cdesioannulatus (Wood) Packard with it. 

 It was clear to us on reading Koch's description with specimens 

 of csesioannulatus in hand that the description could not possi- 

 bly apply to that species, and the idea suggested itself that 

 Koch's description portrayed an animal astonishingly similar to 

 lulus virgatus Wood. On comparing specimens of that species 

 we found the agreement with the description to be complete, 

 down to the smallest detail mentioned. Both the varieties of 

 coloration which Koch described as species are also common, so 

 that the two can be united. Koch's specific name (1847) ante- 

 dates that of Wood (1864), and the species must stand as lulus 

 fiavicornis (Koch). The genus Campodes never had any reason 

 for its existence, and can be dismissed from the subject. It 

 might be well to add that we have found sexually mature I. fiavi- 

 cornis with 30 segments. This species seems to be very nearly 

 allied to the European /. pusillus Leach, described long before 

 the time of Koch, but with which he did appear to be familiar. 



Craspedosoma of Leach and Rawlins (1814), is the oldest genus 

 of the family and contains a considerable number of European spe- 



