44 THE VOYAGE OF H.M.S. CH^VLLENGER. 



median ambulacral and interambulacral spaces, leaving only here and there traces of its 

 former existence. The variation in the length of the spines is also very marked, so that 

 many specimens would at first sight readily pass as Cidaris 2}(i'2^ii^((ta (PL II. fig. 1). 

 It was one of these varieties of Goniocidaris canaliculata which Thomson described 

 and figured as Cidaris nutrix in his Voyage of the Challenger, vol. ii. pp. 226, 227 

 (fig. 42). There is nothing constant in the connection of these features. We find 

 specimens with long spines and deep median vertical furrows, and the reverse, 



Studer (Berlin Akad. Monatsb., 1876, p. 455) has noticed from Kerguelen Island 

 and the coast of Patagonia two species of Goniocidaris {Goniocidaris menibranipora 

 and Goniocidaris vivi^xira), both of which I am inclined to refer to Goniocidaris 

 canaliculata. The existence of large genital openings covered by a thin membrane 

 {Goniocidaris memhranipora) to facilitate the passage of the viviparous young, had not 

 been noticed in the older descriptions, and the differences upon which he separates 

 Goniocidaris vivipara from Goniocidaris memhranipora do not seem to be constant, 

 judging at least from the great variation in the size and position of the genital and 

 ocular plates in the sjoecimens collected by the Challenger, and from the great variation 

 in the length as well as ornamentation of the radioles. Thomson^ and Studer^ 

 published about the same time notices that Goniocidaris was viviparous, and that the 

 young were carried upon the abactinal system, protected by the upper spines of the test, 

 until their full development had taken place. 



On Plate II. are figured three of the most characteristic types of Goniocidaris 

 canaliculata, one (fig. 1) with long slender cylindrical spines, some of them twice the 

 diameter of the test. Fig. 2 represents from the abactinal side a specimen with compara- 

 tively short radioles, scarcely two-thirds the diameter of the test in length, but the 

 ornamentation of the spines is similar to tliat of the long cylindrical spines of fig. 1. 

 Both these specimens were covered with comparatively coarse papillae. In fig. 3 is 

 represented a specimen with much finer and more numerous papillas, and also slender 

 but short spines. Among the many specimens collected by the Challenger, all possible 

 combinations of fine and coarse papillae, with slender, long, or short, or with stout and 

 short radioles, were observed, showing, as in other species of the family, a most 

 extensive degree of variation, while in other Cidaridse with extremely variable spines the 

 characteristic features of the test are tolerably constant. In this species the variation is 

 not limited to the primaiy radioles and paj^iUae, but extends also to the ornamentation 

 of the test. This shows, as I have mentioned above, a median suture in all possible stages 

 intermediate between a broad suture (PI. III. fig. 4), or a deep sharply cut groove 

 (PI. III. fig. 6), and an almost indistinct bare space (PL III. fig. 5). 



An interesting account of the mode of carrying the young in this species is given by 

 Thomson (Voyage of the Challenger, vol. ii. p. 228). " The eggs, after escaping from 



' Joiirii. Linn. Soc, vol. xiii., June 1876. Berlin Akad. Monatsb,, July 1876. 



