﻿29 



67 



Ampliisile : 



Fig. 19. 



I.eft pectoral urcll. seen from 

 hin ; .supraclavicular removed. r( : clavicle ; 

 icige on clavicle; .'ic: scapula; co: coracoid ; 

 pel; postclavicle; bai: lowermost basal; 

 u; foramen. 



fig. 19) does not nearly contain its foramen, but is assisted above by the inner 

 plate of the clavicle"; and with this his ligure 6 also agrees. The true condition 

 I have shown in the figure of A. sciitala. We see here that the large, oval foramen 

 is i|uite surrounded by the scapular alone, as elsewhere in the bony fishes. Starks' 

 error has possibly arisen from investigation of a dried specimen, in which the 

 anterior, extremely brittle and almost unossified boundary of the opening has fallen 

 out. As in many other lishes the coraco-scapular cartilage has been |)reserved to 

 such an extent that there is but a small ossi- 

 fied portion, in parts exceedingly thin, outside 

 it. The whole of the inner part of the scapula 

 is cartilaginous and the cartilage on the front, 

 upper corner is quite uncovered by bone; from 

 the boundary between the scapula and coracoid 

 the cartilage extends into the latter as a broad 

 triangle, continuing from the lower angle as a 

 thin axis across the horizontal part and increa- 

 sing evenly in thickness towards the clavicle, 

 where finally the anterior end is quite cartila- 

 ginous. The horizontal part of the coracoid 

 unites so firmly with its fellow of the other side, that they give the impression of 

 having grown together; obviously a result of the extremely compressed condition 

 on the ventral side of the fish. The ridge / on the clavicle, to the hindmost part 

 of which the scapula is attached, continues under the scapula right to the posterior 

 end of the portion of the clavicle visible externally (c/, fig. 1, Tab. I). 



Fig. 19 shows likewise that the postclavicle {pctjhas essentially the same form 

 as in A. strigata, but is much shorter. In A. strigata it extends backward beyond 

 the pelvis as far as the posterior margin of the 8th ventral plate. 



In A. scuiata there is an oval opening o, bounded partly by the coracoid, 

 partly by the very considerable, lowermost (4th) basal (bai^). The part of the 

 clavicle visible externally is longer in scuiata than in the other two species, where 

 it also has a somewhat diflerent form (cf. PI. I, fig. 1 with text-figs. 2 and 3 p. 47 (9)). In 

 the other two species the shoulder girdle is on the whole not quite so elongated 

 as in sciitata and the pectoral lins are therefore not quite so far distant from the 

 gill-opening as in the latter; as Lütkkn (24a, p. 216) has correctly stated, the base 

 of the pectoral lies in sciitata (almost) above the middle of the 6th ventral plate, 

 whilst in the other two species it is (almost) over the boundary between the 5th 

 and 6th^. 



The number of rays in the pectoral fin I find to be 11 in A. strigata, 

 12 in A. piinctulata, 10 in A. scu/a/a (though 11 in one of 11 specimens); in addition, 

 there is in all 3 species a small, quite rudimentary ray at the upper border of the 

 base of the pectoral. In a single specimen of scutata, further, there was on the 

 left side 1 more rudimentary ray above and on the right side 1 rudimentary but 



