ALCYONARIA. 7 



specimens in the collection with those obtained by the ' Challenger,' and consider 

 they belong to the same species, although there are some points of difference between 

 them. As in < '. <l< licotulu, there is no main stem in these specimens, the colony 

 arising frum a tangled mass of branches in a sponge. The main stem may have been 

 lost, or a colony starting in a sponge may not need, and therefore not form, the stout 

 supporting main stem ; but in either case its absence in these specimens does not 

 necessarily demand the constitution of a new specific group. 



In addition to the sp< -'.-linens referred to above, there are specimens from two other 

 localities, which probably In-long to this species. The localities given on the labels are : 

 (1) off Erebus and Terror, 500 fathoms, January 22, 1902; (2) E. end of Barrier, 100 

 fathoms, January 29, 1902. In neither case is any statement made as to the nature of 

 the preservative used, but as in the first case the spicules are absent, and in the latter 

 small and apparently corroded, I cannot but suspect that they were treated witli picric 

 acid before being transferred to spirit in the same manner as some of the specimens of 

 Thviiiinllit antarctica (p. 9) were preserved. The specimens resemble C. antarctica 

 in the arrangement of the branches, in the size of the calicos, and in the size of the 

 iuternodes ; but the}- both differ from the other specimens of C. unturcticn in the 

 collection in having a well-defined main stem. The axis of this stem consists of nodes 

 about G mm. in length, with a diameter of from 1 to L'5 mm. In the specimen from 

 locality 2 the main stem is 160 mm. long, and the basal part, which is nearly devoid 

 of coeuenchym, supports colonies of Cephalodiscus hodgsoni. 



CERATOISIS SPICATA. 



(Plate II., figs. 16, 17, 18.) 



Locality : McMurdo Bay, February 8, 1902. 96 to 120 fathoms. 



The base is missing from the specimen, which is about 75 mm. in height. A main axis 

 may be traced through the whole height, from which secondary branches arise quite 

 irregularly on all sides. Notwithstanding this, however, by the bending of the secondary 

 branches the colony becomes almost fiabellate. The internodes both of the main axis and 

 the branches are 12 to 20 mm. in length. At the base the diameter of the internodes 

 is 1 mm. The terminal iuternodes are not very delicate, being about 5 mm. in 

 diameter where they join the last internodc, but they come to a sharp point distally. 

 The coenenchym is thin. The calices are irregularly scattered, cylindrical in shape, 

 about 2 mm. in height when full grown, and situated at distances of 1 to 2 mm. apart 

 on the lower part of the branches, but clustered at their distal extremities. 



The calices are covered with an armature of overlapping scales, some of which are 

 tri-radiate, others irregular in form (figs. 18b, c). Surrounding the crown there are two 

 or three circlets of spicules, of which one ray projects as a very prominent spine 

 (fig. 18a). The largest of these tri-radiate spicules are over 7 mm. in length, the 

 spine being as much as 0'45 mm., and the other two rays 0'35 mm. in length. The 

 other irregular spicules of the calices and the spicules that cover the coenenchvm are of 



