THE GORGONIANS OF THE BRAZILIAN COAST. 401 



in 400 fathoms, west of Chili, and referred to this species by the same authors 



- « « « A « A m 



• TV 



is identical seems to me extremely doubtful, for at that depth the water is v 

 cold, while in the shallow water off Bahia it is tropical. Probably a careful 

 study of the spicules would show valid distinctions. 



Off Bahia, in 10 to 20 fathoms, Challenger Expedition. Type. The nam* 

 purpurea has been applied to several other species of gorgonians, mostly belong 

 ing to Leptogorgia. The original Gorgonia purpurea, of Pallas, was a •lender, 

 virgate species apparently the common purple variety of the South Carolinian 

 species usually called L. virgulata Lam. (purple variety); L. viminea Ellis and Sol 

 (yellow variety) ; and by other names. L. purpurea is the earliest. 



The following is a free translation of the description of Pallas: 



Gorgonia subdichotomous with the branches divaricate virgate; eu-nenchynm 

 violaceous, subverrucose. Axis terete, black, smooth; tips yellow, flagellate 

 Stalk unequally dichotomous; branches few, divaricate, ascending. Coenen- 

 chyma thick, violaceous, rough (scaber); pores alternate, oblong, scattered, a 

 little prominent. Base much expanded, fuscous gray, attached to an oyster 

 American Sea. 



From the above it is evidently a purple Leptogorgia, with slender virgat< 

 branches, like L. virgulata of South Carolina, and not at all like the Brazilian 

 species referred to it. 



Family ELLISELLID^) Gray - GORGONELLID^) Verrill and others. 



Gorgonellacees Val., 1855. M.-Edw., Corall., vol. I, p. 182, 1857. 



Elliselladce Gray, Proc. Zool. Soc. London, 1859, p. 480. Catal. Lithophytcs, p. 21, 1^70. 



GorgonellidcB Verrill, Proc. Essex Inst., Salem, Mass., vol. IV, pp. 148, 189, 1805. 



Although it may be necessary hereafter to reduce this group to a subfamily 

 of Gorgonidae, as stated under the latter (p. 390), I do not wish to discuss the 

 question here, for the group is scarcely known from Brazil. It is abundant in 

 the "Blake" Alcyonaria from the Antilles. 



Numerous species have been well figured and the genera wilt be revised in my 

 forthcoming report on that collection. 



At present it is separated from the Gorgonidae only by the more calcareous 

 axis, — a notably variable character in the Gorgonidse, as well as in the PIcxauridte 

 and some other families. 



Besides this character, it may be added that in addition to the small, double- 

 spindles, there are some short forms of double-heads, double-clubs, double-cones, 

 etc., that are densely warted and often very characteristic. 



Elliselladse has priority over Gorgonellidse as the family name. Since it 

 seems necessary to adopt the genus Ellisella Gray (restricted), having E. elongate 

 as the type, there is no longer any reason for discarding Gray's family name 

 The genus Gorgonella (type G. sarmentosa) is perhaps the least characteristic 

 member of the family. Its calicles are quite immersed, its spicules nearly as in 

 Leptogorgia, and its axis is often only feebly calcareous, much as in L. rathbuni 

 and Gorgonia braziliensis , so that it is very close to Leptogorgia. 



26 JOURN. ACAD. NAT. SCI. PHILA., VOL. XV. 



