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



branches 1*5 to 2 mm. in diameter, and from 10 to 15 cm. long. These are from 3 to 6 

 cm. apart, and form a very wide angle with the stronger branches in the lower portion of 

 the corallum, but above they are closer together, and not so spreading. One of these 

 branches with the mass of branchlets and pinnules into which it becomes divided is re- 

 presented in PI. I. fig. 8. 



Such a branch becomes divided into a lame number of close-set subalternate branch- 

 lets, from 2 to 4 cm. long, which extend chiefly in one plane. There are usually about six 

 of these to a centimetre. The branches and branchlets are further clothed with in- 

 numerable short slender pinnules, which are not confined to one plane, but pass out in all 

 directions. These vary from 0'3 to 1 '5 cm. in length. Those up to about 0'4 cm. are usually 

 simple, but the others bear secondary pinnules about 0'5 cm. long, which have a subspiral 

 arrangement. A pinnule of 1 cm. in length may bear five to twelve secondary pinnules, 

 one or two of which may be again subdivided. The pinnules springing from the lateral 

 surfaces of a branch let usually become fused with those derived from adjoining 

 branchlets, but those on the anterior or posterior surfaces usually remain free. The 

 pinnules on the anterior surface of a branch, &c, are usually so crowded as to completely 

 hide the stronger portions under them. In the lower part of the corallum the smaller 

 branches are most irregular, and give off a number of short stiff branchlets, by means of 

 which a most irregular reticulum is formed (PI. I. fig. 10). In such portions the pinnules 

 are even more slender than those above. The longer ones are pinnate or bipinnate, and 

 form a confused mass of delicate hair-like twigs, which pass off in all directions and fill 

 up the meshes of the coarser reticulum. Such portions recall Morison's figure of his 

 zoophyte, No. 18 (6, pi. x.), a species which Pallas regarded as belonging to his Antipathes 

 clathrata. I have not seen a specimen which conforms to the definition of Antipathes 

 clathrata given by Pallas, and the species does not appear to have been studied by recent 

 observers. From the scanty information available, I am, however, inclined to think it 

 more nearly related to Arachnopathes ericoides (Pallas), and Arachnojxithes aculeata, 

 n. sp., than to the species under consideration. 



In Antipathella contorta the corallum is rendered still more complex from the fact 

 that on a number of the branches and stronger branchlets, many of the pinnules become 

 modified into a hollow cylindrical reticulum, which is inhabited by a parasitic Annelid. 

 The tubular reticulum has a structure similar to that of Tylopathes crispa, n. sp. 

 (c/. PI. III. fig. 2), but the meshes are closer and the arrangement more irregular. 



The polyps are very similar to those of Antipathella minor, but are frequently so 

 crowded that the outline of each is not well defined. The usual arrangement is consider- 

 ably closer than that shown in PI. I. fig. 9. The polyps on the reticulum, which serves as 

 shelter for an Annelid, are (in spirit) smaller and paler than the others. The tentacles 

 often project as small rounded processes, no larger than the median prominence of the 

 peristome on which the mouth opens. 



