470 THE ANNELIDA POLYCHAETA. 



ee. With protruding uncinigerous tori; uncini not forming a girdle on abdomen. 



/. A ventral groove on the posterior somites Euchone Malmgren. 



//. No ventral groove on posterior somites. 



g. Inferior thoracic notopodial setae lanceolate Megachone Johnson. 



gg. Inferior thoracic setae clavate or spatulate, distally broad. 



h. Inferior thoracic setae with an acute tip Chone Malmgren. 



hh. Inferior thoracic setae distally rounded, wth no acute tip or mucro. 



Melachone Bush. 

 dd. Branchiae free, not united over most of length by a weak membrane. 

 e. Dorsal thoracic setae of two kinds. 



/. Abdominal uncini pectinate Dialychone Claparede. 



//. Abdominal uncini avicular Bispira Claparede. 



ee. Dorsal thoracic setae of one kind. 

 /. Thoracic somites five. 



Collar none or rudimentary; abdominal uncini few, pectinate. 



Oriopsis Caullery and Mesnil. 

 //. Thoracic somites normally at least eight. 



g. With a collar Oriades, nom. nov. 



gg. Collar lacking Fabricia Blainville. 



II. First seven or eight somites (thoracic) with notopodial capillary setae only (or rudimentary crochets 

 on the first) no alimentary tract in adults. 

 Crochets of avicular type on three, or rarely four, somites beginning on eighth or ninth somite 

 in dorsal position, on these and following somites capillary neuropodial setae. 



Caobangiiniae, subfam. nov. 

 Caobangia Giard. 



The general features of the grouping indicated in this key were estabhshed 

 by Langerhans (Zeits. wiss. zool., 1880, 34, p. 111). The principles of the 

 classification have in the main proved satisfactory and have been developed 

 particularly by the excellent work of St. Joseph (Ann. sci. nat., 1894, ser. 5, 17, 

 p. 246 et seq.), whose system in its essentials is here adopted. I have endeav- 

 ored so far as possible to include in the key all genera that have been proposed, 

 although recognizing the slight basis upon which several of them appear to rest. 

 It has not been possible to include Demonax Kinberg, which is not described in 

 sufficient detail in the original account and of which species have not been rec- 

 ognized by subsequent workers. The same is true of that author's Parachonia, 

 which may prove to be identical with some other known genus having thoracic 

 setae of two types; but as the uncini are not described as to form or arrange- 

 ment identification seems impossible at present. Anamobaea Kroyer has also 

 not been identified and is not included. Its type, A. oerstedi, sl West Indian 

 species, has a rudimentary collar and two types of thoracic notopodial setae, 

 with branchiae united. 



Synonymy of Genera. 



I have felt much hesitation in deciding on the proper application of the 

 generic name Bisi)ira. It was proposed first by Kroyer in 185G (K. Danske 



