154 Dr. A. Philippi on the genus Serpula. 



ville. The characters on which these separations are founded are of 

 different value. Vermilia and Galeolaria differ from Serjmla solely 

 by the structure of the operculum ; according to Lamarck, Serpula 

 possesses an ' operculum pedicellatum infundibuliforme aut cla- 

 vatum [corneum) '; for some lines further he says, " cette opercule, 

 par consequent, n'estjyoint calcaireJ' (2nd ed. An. sans vertebres,v. 

 p. 361.) Vermilia, on the contrary, has an 'operculum testaceum 

 orbiculatum, simplex^; and further on, 'a dos convexe, leplus souvent 

 conique.' (Ibid. p. 368.) Galeolaria, lastly, is said to possess an 

 'operculum testaceum compositum,^ which, according to my obser- 

 vations however, does not consist of five to nine but of fifteen 

 pieces ; the number however may differ in the various species ; at 

 all events, the di-awing in the 'Diet, des Sciences Naturelles' is 

 decidedly bad. Filograna, Berkeley, is said to possess constantly 

 two opercula, which has likewise been observed exceptionally in 

 other species. Protula, Risso, and Spiromella, Blainv., have no 

 operculum : Cuvier refers them curiously enough to Sabella. 



The genera Cymospira and Spirorbis have been established 

 according to the number of filaments into which the branchiae 

 are divided and according to their arrangement. In Cymospira 

 the branchiae are on each side divided into numerous filaments and 

 rolled up spirally; in Spirorbis they consist only of three filaments ; 

 but these characters are of vciy slight value. The different spe- 

 cies of Serpula which I have observed with the animal have 3, 

 4, 6, 7, 8, 10, 11, 13, 18, 30, 40, and more filaments to each 

 branchia, and the larger their number the more requisite is it for 

 them to adopt a spiral arrangement. I have likewise found in 

 Vermilia triqvetra and Pomatoceros tricuspis (see below), that the 

 filaments of the branchiae describe a spiral of one convolution of 

 the kind represented ^ when they are expanded. It appears 

 therefore to me that no very accurate limits exist between spiral 

 and non-spiral branchiae. According to Blainville, the branchial 

 filaments of the Vermilia have cirrhi only on one side, which I 

 look upon as an error. 



The mode of growth has likewise been taken into consideration, 

 and those species with a sjiiral growth have been referred to Spi- 

 rorbis; yet the hkewisc rcinarkaljly spirally wound S. cereolus,\he 

 animal of which is still unknown, is excluded. One of the prin- 

 cipal results of my observations is, that no relation exists between 

 the nature of the animal and the shell, except perhaps in the 

 genus Galeolaria ; this indeed is a sad result : thus, for instance, 

 we have a three-ridged shell in three different sections, an 07'bi- 

 cular shell in still more ; in one division there are smooth orbicular, 

 orbicular with longitudinal bands, triangular, quadrangular, &c. 

 From what has been stated above, the structure of the oper- 

 culum is the best character upon which to form the subdivisions 



