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



crown of tentacles, and close to the genital aperture. I intend to point out those cases 

 in which several pores are to be seen. 



When examining the madreporic canal in Lcetmogone wyviUe-thomsoni it will be found 

 that, proceeding as usual from the dorsal part of the ambulacra! ring, it runs upwards and 

 backwards, being enveloped by the dorsal mesentery, and is attached at the medio-dorsal 

 line about 30 mm. behind the tentacular crown (PI. XLIII. fig. 4, a). The terminal 

 part is surrounded and entirely enveloped by a rather thick layer of connective tissue, 

 and gives off, four, five, or up to nine very fine branches, which pierce the body-wall, and 

 open at the tops of small papillae, 2 or 3 mm. long. These papillae vary in position ; 

 they are either closely crowded in front of, or at one side of, or behind the large genital 

 process, or they are situated in a semicircle along one side of the latter (PI. XXXVIII. 

 fig. 9, a). The fine canals, which pierce the perisoma, are rendered most distinct because 

 of the great abundance of red and violet pigment present in their walls. A transverse 

 section of the papillae proves that they are made up of a very thick layer of connective 

 tissue, containing a multitude of filaments, cells, and pigment, and that the canal which 

 penetrates them is very narrow, scarcely a fifth of the diameter of the papillae themselves. 



In Ilyodcemon maculatus (PI. XXXVIII. fig. 6) the terminal part of the madre- 

 poric canal divides into a greater or smaller number, sometimes up to fifty, branches, 

 which penetrate the body-wall immediately in front of the genital process, and about 

 20 mm. behind the anterior extremity of the body. The fine canals, which do not 

 run out in any papillae, are often slightly expanded and branched, and contain a great 

 quantity of pigment, but no deposits are present in them ; their walls are extremely thin, 

 and lined by an epithelium consisting of small, flat cells, which differ most strikingly from 

 the elongated, closely-crowded, cylindrical cells, which line the true rnadreporic canal. 

 In Benthochjtes dbyssicola the madreporic canal sometimes runs out by a single pore (PI. 

 XXXVIII. fig. 2, a), sometimes by four pores, which lie together at the top of a small 

 obtuse papilla (PI. XXXVIII. fig. 1). The canal is often more or less strengthened by 

 calcareous deposits, but it may also lack such, as, for instance, in Kolga hyalina, Dan. and 

 Kor. The fine canals or branches, which penetrate the body-wall seem to be always 

 without calcareous matters. 



The five main canals, which proceed from the circular vessel of the ambulacra! system, 

 are long and wide in some forms, and are always attached to the most anterior portion 

 of the alimentary canal by numerous threads or filaments (PL XLIII. fig. 1, a). Their 

 communication with the circular vessel is effected by means of a wide opening, while 

 their anterior, slightly distended ends, lying close to the calcareous ring, open into the 

 tentacles and the radial ambulacral vessels by a minute orifice. In Deima fastosum 

 especially I have had the opportunity of observing the manner in which the main 

 canals terminate in the tentacles and the ambulacral vessels. The anterior slightly 

 distended end of these canals (PL XLIII. fig. 7, a) lies close to the posterior portion of 



