4 E. KIRKPATRICK. 



and possibly also of a root-tuft. A few external (? velar) liypodermal oxypentactins are 

 present on the surface, and little tufts of broken-ofi' basalia project from the rounded 

 base of the specimen. There may be a well developed root-tuft in complete specimens ; 

 further doubts on this point were suggested by the condition of four specimens of a 

 species of Crnnielln from the Antarctic, three being almost smooth at the lower end, owing 

 to the root-tuft having been detached in dredging, whereas the fourth has a large root-tuft. 



The dimensions of the specimen are as follows : Height, 14cm. ; greatest breadth, 

 6cm. ; diameter of orifice, which is slightly torn, about 4cm. ; greatest thickness of 

 wall. 1 1cm. 



The lumen of the deep gastral cavity is occupied by numerous scattered internally 

 projecting plcuralia, which have been driven in from the outside. The circular 

 openings of the ostia, about 1 'Smm. in diameter, are clearly visible beneath the dermal 

 layer ; also the slightly larger postica are perceptible beneath the gastral layer, which 

 roofs them over with a fine lace-like reticulum. 



Skeleton. The skeletal framework is formed of bundles of diactins ; there are 

 no large hexactins. 



Spicilles. The principalia are oxydiactins, about 7000 x 80^ in size, with fine 

 tapering ends smooth or only very slightly spined. A much smaller kind, 1400 x 18/i, 

 separate, and not in bundles, with roughened ends and four central knobs, is common 

 beneath the dermal and gastral membranes. 



The autodermalia (IV. la) are spined hexactins, each ray being ISlju. in length 

 and 15'5/x broad at the base, slightly tapering to a blunt extremity; pentactins 

 (IV. Ib), with the odd ray proximal, occur, but very rarely. 



The hypodermalia (IV. Id, d 1 ) are oxypentactins with slender smooth tapering 

 paratangentials, each about 2700 x 40/x, the rays being either orthotropal or 

 anorthotropal. 



The autOgastralia (IV. Id 2 ) are spined hexactins with slightly longer and more 

 slender and sharply pointed rays than the autodermalia, each ray being 136 x 7'5/i. 



The intermedia. Holoxyhexasters* (IV. le) and hemioxyhexasters (IV. le 1 ), 

 varying in diameter from 100 to 120/i, are common. Monoxyhexasters* (IV. le 2 ), 108/i 

 in diameter, occur only rarely. Medium-sized holodiscohexasters (IV. If, f 1 ), from 

 80-100/u, in diameter, are also rare ; the short slender primary rays, about 7p- in length, 



base of the ray, and that the remaining portion of the ray is devoid of any trace of an axial canal, that it is, 

 in fact, more of the nature of a spine. Secondly, in many instances, all the transitions can be traced from 

 holohexasters, through several grades of heniihexasters to monohexasters. In a preparation of Kossella racomtaoe 

 Topsent, for example, these transitions can be traced in a crowd of discohexasters, the monodiscohexasters having 

 short thick primary rays with the axial canal extending only to the point where the thick portion (primary ray) 

 joins the more slender solid terminal portion? To call such spicules as these last discohexactins would be to 

 lose sight of the fact that they clearly have six primary and six single secondary rays, for the latter do not lose 

 the character of being secondary simply because they are single. A genuine hexactin would, by its definition, 

 have no secondary rays. The designation " hexactinose," used by Ijima, would, perhaps, be better written 

 " hexactinoid," or hexactin-like, but even this term is not without objection, since such a spicule with its primary 

 and secondary rays is seen, under a high power, not to resemble a true hexactin. Accordingly the prefixes 

 " holo-," " hemi-," and " mono-," added to " oxy- " or " disco- " hexaster, are suggested as indicating and 

 defining the form and relationship of these spicules. The following figures show the limitation of the axial 



