THE SHOULDER-GIRDLE IN FISHES. 35 



pointed behind and below, where it clamps itself on to the outside of the next piece. This 

 second post-clavicle (1. p. cl) is narrow and pointed above, and then soon spreads into a broad 

 trowel of bone, which stretches half way from the oblique pectoral opening to the end of the 

 curiously tesselated ganoid box. The lower and hinder part of the fore-and-aft down-growth of the 

 clavicle seems to answer to the distinct " anterior inter-clavicle'' of Syngnathm ; but the posterior 

 inter-clavicle of that Fish is represented in Ostracion by a very large bone (figs. 4 A, 4 B, i. cl.), 

 having three wings, bent on each other at a very obtuse angle. 



The broad, slightly arcuate base of the interclavicle rests on the abdominal line, close behind 

 the clavicle ; it grows upwards as a high but regular triangle, with a rounded apex. The posterior 

 margin of the clavicle leans forwards, and the anterior wing of the inter-clavicle widens upwards to 

 be attached to it ; this thin wing, free from the growth-markings of the middle part, ends almost 

 horizontally above. The posterior wing is separated from the main part by a deepish sulcus ; it is 

 lanceolate, sinuous behind, and runs upwards above the middle part into a point, which wedges in 

 between the prse-coracoid and lowest "brachial" (fig. 4 B, i. cl., p. cr., b. 5). This pointed plate 

 and the rounded end of the main triangle are anchylosed to the posterior margin of the prae- 

 coracoid, reminding one of the Siluroids ; but the gently in-turned front wing is permanently free 

 from the base of the coracoid, their mode of junction being by " harmony," and there is a very 

 narrow but clear chink between them. 



The scapula of Ostracion auritus (fig. 4 B, sc., p. sc.) forms nearly half of the Shoulder- 

 girdle moiety; the other half being the prse-coracoid (p. cr.). The whole plate is large, 

 oblong, and bent forward in such a manner as to leave a very open angle in the front margin. 

 Along the front margin there is a selvedge of clear cartilage, and passing through between the 

 two bony plates there is a considerable band, largest at its ends, quite uncovered with bony sub- 

 stance. The shape of the scapula is roughly pentagonal ; it is divided into a scapula proper (sc.) 

 and a pra3-scapula (p. sc.), the latter being the narrowest ; this division, which is arrested, is made 

 by a transversely egg-shaped fenestra (sc. f.), the apex of which is directed backwards. The 

 " glenoid " margin is thick, and it is here that ossification commences in an " ectosteal " plate, 

 clamping the cartilage by an outer and an inner extension ; these plates are arrested in front and 

 below. The prae-coracoid (p. cr.) is an elegant rhomboid, with the posterior margin notched to 

 receive a peg from the inter-clavicle ; at this point, and a little above and below it, there has been 

 anchylosis of the great underpropping splint with the ectosteal selvedge of the back of the 

 praa-coracoid. Like the scapula, this prae-coracoidal element is only feebly ossified, except at the 

 hinder margin ; the bony-plate has grown but little into the substance of the cartilage. 



If I had surveyed these parts fresh from the typical Fishes they would have caused me 

 much trouble ; but the way has been made much smoother by my coming to them through by- 

 paths. We have seen this coalescence of parts of the outer and inner skeleton, and shall see it 

 again, and it is noteworthy that, although the splint-bones are ready enough to coalesce with endo- 

 skeletal bones, yet, retaining their old nature as scales, they are very slow to combine with each 

 other. The affinity of the dermal system for the endo-skeleton becomes stronger and stronger as 

 we ascend the vertebrate scale ; and in the warm-blooded classes the most remarkable metamorphic 

 combinations take place. The most careful microscopical observation of the inter-clavicle (i. cl.) 

 of the young Ostracion reveals no trace of cartilage-cells within it although the scapula and 

 coracoid have their cells, for the most part, unchanged between the ectosteal plates. 



I think, therefore, that the coracoid proper has been completely aborted by this large and 



