l86 BUSH 



cate membrane or web ; in the Parasabellas these change to less dis- 

 tinguishable four-sided ones, and the web is but slightly developed 

 or disappears, while in the Eudistylias they become distinctly three- 

 sided, rounded outwardly. They may also be simple, or many times 

 divided or split, as in the Schizobranchias . 



It has also been found that, although so many valuable facts have 

 been so comprehensively presented by Saint -Joseph, there are still some 

 genera of which little is known, owing principally to the too broad 

 application by their authors, as evinced by the variety of forms referred 

 to them. This confusion has been greatly increased by subsequent 

 writers, none having restricted the genera to any one of the species 

 as a type, nor published figures as an aid toward a possible correct 

 interpretation. This is especially true of the genera Sabellastarte and 

 Demonax. 



Sabellastarte was proposed by Savigny (1809) as a group or divi- 

 sional name for Sabclla-like forms having the branchijE arranged in a 

 double series. It was adopted as such by Grube and Quatrefages, but 

 Saint-Joseph, following Kroyer, used it as a generic name, without 

 presenting any additional facts in regard to the branchial lobes, form 

 of the collar, or form and arrangement of the setce. The two species 

 — Sabella indica Savigny and Sabella 7nag7tijica Shaw — apparently 

 agree only in having very long and numerous branchise arranged in a 

 double series. The numerous figures given by Shaw show an interest- 

 ing and easily noted character, i. e., the absence of pinnae on the slen- 

 der banded rachises. Neither Qiiatrefages (1S65) nor Marenzeller 

 (1884) mentions such a peculiarity as belonging to S. indica^ thus 

 giving emphasis to the small importance of the arrangement of the 

 branchiae as the only generic character. 



Marenzeller describes S. indica as having from 60 to 84 (in differ- 

 ent individuals) very long branchise arranged in a double series, and 

 equal to about half the entire length of the body, which consists of 

 from 196 to 227 segments and measures from 80 to 135 mm. in length. 

 Qiiatrefages gives the sette as lanceolate in form, avicular uncini only 

 in the tori and the collar as four-lobed. It is proposed to restrict the 

 genus to this species as type. The genus Eurato Saint-Joseph (1894) 

 differs in not having the branchiae arranged in a double series. Seven 

 species are included in this ivithout mentioning any special one for a 

 type. 



Kinberg (1866) placed five species in his genus Dc?f?onax^ the 

 first (Z>. krusensterni) and the last {D. cooki) being the only two 

 that from the descriptions appear to be at all alike. Therefore the 



