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



conical or rovindcd elevations project at the proper position. Here and there structures 

 may be observed in wbich, instead of such rudiments of the four abortive rays, the 

 latter, or at least two of them are j^resent as well equijjped with many-ribbed umbels as 

 are the two rays usually alone developed. Carter has previously reported the occurrence 

 of this abnormality in Hyalonema sieboldii} I am inclined to compare the smooth 

 processes forming the bell-shaped umbel at each end of the stem with the terminal rays 

 of other sjaicules, and to regard them therefore as simple tooth or spine structures. I 

 have never observed any trace of an axial canal within them, and find, with Marshall and 

 others, that they are not formed in the first rudiment of the amphidiscs but develop 

 subsequently. In Hyalonema sieboldii only two different kinds of amphidisc occur in the 

 skin. On the one hand, there are large strongly developed amphidiscs, 0"42 mm. or more 

 in length, with an umbel 0'13 mm. or more in breadth, the axial beam of which is 0'035 

 mm. in thickness, and more or less rough, with irregularly scattered, inconspicuously arched 

 prominences. At each end there are about eight slightly arched umbel-rays, each of 

 which, with a longitudinally directed basal piece about 0'07 mm. long and O'OIS mm. 

 broad, springs from the side of the terminal portion of the axial beam, and passes 

 externally into the ploughshare-like lateral plates, which are bent towards the surface 

 of the umbel, and therefore at right angles to the basal plate (PI. XXVII. fig. 7). These 

 curved lateral plates do not usually grow together laterally, but form each a free sharp 

 lateral border. They are continued at their free end into a lancet-shaped terminal point, 

 while at the other extremity, near the main axis of the spicule, they end in a rounded 

 prominence, and thus enclose between them a central terminal pit. The free margin 

 of the basal plate exhibits a rounded edge which becomes gradually raised from the 

 axial beam of the amphidisc, and is continued with a hook-like bend to the end of the 

 lancet-shaped, curved, external lateral plate (PI. XXVII. fig. 7). 



The disposition of these large amphidiscs in the skin has not as yet been correctly 

 represented by investigators. They do not lie, as Marshall, 0. Schmidt, Bowerbank, 

 Carter and others have described, tangentially in the dermal membrane, but are radially 

 directed, so that one-half of the whole spicule projects freely from the external surface, 

 and the other extends internally into the subdermal space. I have found them 

 irregularly scattered, sometimes in comparative abundance, sometimes only sparsely, 

 l>ut could not distinguish any one region of the body as characterised by their 

 special frequency. They occur also on the external surface of the terminal sieve- 

 plate. 



Besides these large amphidiscs with broad shovel-like umbel rays, numerous smaller 

 forms, measuring only O'Ol to 0"016 mm. in length, occur, irregularly scattered in the 

 derma] membrane, in no special direction, but for the most part tangentially disposed. 

 Their umbels are terminally hemispherical! y curv'^ed, and consist of eight umbel-radii 



' Aim. and Mag. Nat. Hist., ser. 4, vol. sii. p. 372, anc\ pi. xiii. fig. 22. 



