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



surround the gullet, but is attached to it by a great many filaments (PI. XLIII. fig. 1), 

 an interval being thus constituted, which connects the oesophageal sinus with the 

 peritoneal cavity ; when this communication exists as, for instance, in Oneirophanta, 

 Deima, &c, the openings between the five main canals never appear closed, consequently, 

 even here the sinus communicates with the peritoneal cavity. In Benthodytes san- 

 guinolenta, Lcetmogone, and many other forms where such a circular interval is present 

 between the gullet and the circular water-vessel, the oesophageal sinus is closed by 

 a very thin membrane which not only unites the five main vessels with one another, but 

 is continued as a circular band lying close behind the circular vessel connecting this 

 with the oesophagus (PI. XL. fig. 5, m). Sometimes as in llyodcemon maculatus, &c, 

 the circular water-vessel carries a great many very small csecal sacs, the function of 

 which is unknown. Most of the Elasipoda are provided with a single Poban vesicle, 

 but in Parelpidia elongata, Parelpidia cylindrica, Elpidia incerta, Elpidia willemo'esi, 

 Peniagone horrifer and Peniagone vitrea, a couple of such csecal prolongations of the 

 circular water- vessel may be observed. These vesicles are always ventral in position, 

 and when only one is present it lies to the left. In the family Elpidiidse the Polian 

 vesicle has a rounded form, and does not attain any considerable size, while in the two 

 other families it is more elongated, almost cylindrical and often remarkably large. 

 Its size varies most strikingly in different members of the same species, this being 

 most apparent in Oneirophanta ; one individual of this form dredged at Station 244 

 is remarkable for possessing a Polian vesicle, which measures about 134 mm. and is 

 almost as long as the animal itself. Even Orphnurgus asper is distinguished by a Polian 

 vesicle of unusual size. As a matter of fact, the number of Polian vesicles varies in a 

 remarkable manner in the Pedata, but above all in the Apoda, though it must be 

 remembered that even in these orders many forms have only a single one. 



The tubular prolongations of the circular water-vessel, which are termed the madre- 

 poric canal, always single in the Elasipoda, run in the medio-dorsal line, and are intimately 

 attached to the interradial dorsal mesentery (PL XLIII. fig. 4, a, and fig. 6, a). It is very 

 well known that in the other Holothurioidea, if, even as a rule, a single dorsal madreporic 

 canal with a single terminal madreporic tubercle is present, many exceptions are to 

 be found, which having been fully described in the splendid memoirs of Semper and 

 Selenka do not require repetition here. For a long time it was considered as charac- 

 teristic of all the Holothurioidea that the madreporic tubercle was always internally 

 placed, consequently the interior of the ambulacral system could not communicate 

 with the exterior, but only with the peritoneal cavity. As will be presently pointed 

 out, many of the deep-sea Holothurids form exceptions in having the stone canal pierc- 

 ing the perisoma in the medio-dorsal line at greater or smaller distances from the crown 

 of tentacles, thus bringing the water- vascular system in communication with the exterior 

 as is the case in other Echinoderms. In Elasipoda two kinds of madreporic canals are 



