﻿43 81 



means of cross-sections it is always possible to make sure thai the piece is inde- 

 pendent; externally it covers the base of the ray and extends forward under the 

 skin uncovered by scales as a striped, sculptured surface (like a "ganoin" bone 

 (cfr. PI. I, fig. 3 /)), resembling the spine itself. On the inner surface towards the 

 base of the spine it has some few, semicircular ridges, which fit into some corre- 

 sponding ridges on the base of the ray. The base of the dorsal spine is forked 

 and sits directly on the upper end of the interspinous bone; both lateral surfaces 

 of the latter are also provided with semicircular ridges, which fit in between corre- 

 sponding ridges on the inner surfaces of the fork of the dorsal spine. The latter 

 however form much longer curves than those on the interspinous bone; both sets 

 are very sharply cut with shining surfaces; they are more numerous than those 

 on the small bone / or those on the outer side of the ray. The whole of this 

 highly elaborate apparatus is obviously able to fix the dorsal spine in any definite 

 position (cf. W. Sørensen 32, pp. 65—66). In all details it agrees remarkably well 

 with the corresponding apparatus in Amphisile slrigata and punctulata; as in these 

 the elaborate articulation belongs to the 4th interspinous bone; on the sides of the 

 articulation there are the same paired bones /, and the unpaired bone t in the 

 Amphisile species quite corresponds in position with the process / in Cenlriscus; 

 this is probably an independent piece originally, which fuses later with the inter- 

 spinous bone. 



Taken as a whole, the 4 anterior interspinous bones and the corresponding 

 spinous processes in Centriscus form a vertical bony plate which builds, together 

 with the lateral pillars or buttresses formed by the dorsal armour, a kind of tripos- 

 stand for the support of the large dorsal spine. 



Of the interspinous bones of the second group the first (no. 5) lies behind 

 the point of the 6th spinous process, the following 5 (nos. 6—10) each in front of 

 its spinous process of the vertebræ 7 — 11. The distal portion of the first three 

 (nos. 5, 6, 7) is enclosed in the membrane of the anterior dorsal fin. The first (5) 

 bears a fin-ray, as a rule movable, the third spinous ray; but in many specimens 

 I find that the spinous ray is fused with the interspinous bone, but so that the 

 original articulation can easily be detected. The front surface of this interspinous 

 bone has a deep longitudinal furrow in which the point of the process / of the 

 4th interspinal is placed. 



The two following interspinous bones (6 and 7) are always, so far as I have 

 ■seen, fused with their spinous rays 4 and 5; the next two interspinals (8 and 9) 

 protrude through the skin as short, slightly irregularly shaped, shining points; 

 these also represent, very probably, rudimentary spinous rays (but may possibly 

 also be only modified scales?). The 10th interspinous bone usually has a very 

 small spine movably articulated to it (PI. IL fig. 2, r'); it is the first ray in 

 the second dorsal fin. I sometimes find however that this also is fused with the 

 interspinal. All the interspinous bones hitherto mentioned belong to the type 

 denoted by Bridge (4) as "unisegmented", as is the rule indeed for interspinous 



11* 



