44 THE VOYAGE OF H.M.S. CHALLENGER. 



are evidently very short and project little beyond the surface of the peristome. Usually 

 they appear in spirit specimens as round knob-like elevations arranged biradially around 

 the mouth. The members of each lateral pair of tentacles are separated from each 

 other by the stomodaeuni, but the two tentacles on each side are usually close together 

 and appear to form a pair as in Parantipathes (PI. III. fig. 9). 



AU the tentacles arise from the surface of the peristome in this genus, and the oral 

 cone forms a median prominence equalling the tentacles in height. The body-wall visually 

 passes almost immediately from the oral surface around the sclerobasic axis, so that the 

 outline of a zooid is rarely well defined. In specimens in which the sexual elements are 

 well developed the zooid becomes distended, and then its outline is more easily traced. 

 Spines project through the tissues of the zooid in all directions, and in numbers varying 

 with the species. This may be partly due to contraction in spirit specimens, but I have 

 satisfied myself that this feature is not altogether artificial, but is one of the peculiarities 

 of the genus, probably connected with the compressed type of zooid. In horizontal sections 

 spines may be seen projecting through the ccelenteron in many parts. These are usually 

 surrounded by a ring of mesogloea covered externally by a layer of entoderm. Such an 

 arrangement could not be brought about by shrinking, and I have regarded it as one of 

 the generic characters. The mouth is usually somewhat elongated in the sagittal axis, but 

 in some species (e.g., Aphanipathes canceUata) the aperture is usually wide and most 

 frequently circular. In some cases, though rarely, the oral aperture was observed to 

 consist of two terminal circular apertures united by a short median slit-like portion, thus 

 resembling a dumb-bell in shape. Whether in the living colony this may be the usual 

 shape of the mouth, as in many Actiniaria, I am unable to say. Should such prove to 

 be the case, the two terminal rounded openings would probably serve for the entrance 

 and exit of afferent and efferent currents. Unfortunately we as yet know little 

 concerning the living colonies of any of the Antipathidse. The mesenteries are ten in 

 number, and have the same arrangement as in Antipathes. Probably on account of the 

 compressed form of zooid, the secondary mesenteries are in this genus relatively more 

 important, and reach nearly to the base of the ccelenteron. They do not, however, appear 

 to bear convoluted filaments. 



Pteropathes. 



The single species at present included in this genus differs in several important 

 respects from any other species examined. The zooids form a regular linear series, 

 and are so closely pressed together that the line of demarcation between two adjoining 

 zooids never has a curved outline, but passes straight across the branchlet. Seen from 

 above (PI. IV. fig. 3) the zooids present a rectangular outline, and the elongation in the 

 transverse axis, when such exists, is not marked. The zooids are imperfectly separated 



