THE SHOULDER-GIRDLE IN FISHES. 31 



fibrous, and very different from what we have seen in the Fishes already described, and also from 

 the typical Teleostei. In S. acus there are two dorsal-line plates behind the occiput, and then 

 the back is concave, being finished above by the supero-lateral series. There is then a lateral 

 series ; the first of these is shown in the Woodcut, fig. 3 A, 1. 1. 1 : then an infero-lateral, figs. 

 3 A and 3 B, i. 1. 1, i. 1. 2 ; whilst the trunk is walled-in below by the abdominal-line series, a. 1. 1, 

 a. 1. 2. These plates vary as to the number of rows, even in the British species of Syngnathm 

 (see Yarrell's 'British Fishes,' vol. ii, p. 325 341.) 



There is a post-temporal scale (fig. 3 A, p. t.) below the first post-occipital ; but there is no 

 distinct supra-clavicle ; hence the large size of the clavicle (fig. 3 A, cl.) This bone is T-shaped, 

 the descending crus being almost vertical ; it leans a little forwards in its descent, and there is a 

 large backward development of the bone, strengthened by an internal beam, below. The lower 

 third of the anterior plate is similarly fortified, and these two bars swell into a rough boss at 

 their junction. Above, the bone is smooth, but thick ; and it is full of the elegant radiate 

 markings common to this and the ordinary scales. The bone all along the top is composed of two 

 distinct plates ; the outer being the inner half of the skin ossified and answering to the ganoid plates 

 generally, and the inner lamina being a layer of ossified subcutaneous fibres. The post-temporal, 

 (fig. 3 A, p. t.), is a triangular ganoid plate, notched in front ; the supra-clavicle has no separate exist- 

 ence ; and here, as in many other Teleostei, the upper parts of the dermal Shoulder-bones are, to 

 some degree, ganoid as well as subcutaneous ; and thus the space between the Ganoids and Siluroids, 

 on one hand, and the typical Teleostei, on the other, is partly bridged over. The main partof the 

 descending crus of the clavicle appears within, as a rounded rod, which bifurcates below ; the 

 outer fork being the shortest, and having the first infero-lateral scale strongly clamping it in front 

 (see fig. 3 A, cl., i. 1. 1) ; whilst the longer, inner spur rests upon the submesial part of the first 

 abdominal-line plate (fig. 3 A, cl., a. 1. 1). In some Fishes (e.g. Cottoidei) there is a large 

 development of the clavicle internally ; the strong aponeurotic band which lies directly outside 

 the "peritoneum" being ossified, and lying inside the true Shoulder-girdle (see Plate II, 

 7 n, fig. 12, cL cr.). 



There is no such plate growing from the clavicle within and below in Syngnathus ; but this is 

 replaced by two separate bones not seen, as a rule, in the Teleostei generally. The descending rod 

 of the clavicle, however, has an outer wing to it, the greater part of which is seen in fig. 3 A, cl., 

 overlying and partly protected by the first lateral plate (1. 1. 1) : this infero-posterior wing is, in 

 some degree, ganoid ; there is evidently a scale connate with it, as in the upper part of the clavicle. 



Referring again to the clavicle of Coitus bubalis (Plate II, fig. 12, cl.), it will be seen that the 

 clavicular plate which passes within the coracoid (cr.) has two additional ossicles (a. i. cl., p. i. cl.) : 

 these two bones are represented in Syngnatlius by much larger bones, which take the place of the 

 whole inner plate of Cottus (Woodcut, fig. 3 A, a. i. cl., p. i. cl.). Both these bones are behind the 

 inner fork of the base of the clavicle, and, therefore, neither of them can represent the " ischio- 

 pubis," which, in the Cottus, as in other Thoracic Fishes, passes within the lower part of the 

 clavicle, and is tilted up in front, so as to leave some space between the pubic extremity and the 

 base of the clavicle, where it rests on the skin, below. But the Syngnatlim has no pelvic members ; 

 and if it had a " pelvis " without the fins, yet, as the posterior arch is always morphologically 

 inferior to the anterior or scapular, it could not be bony in this instance, seeing that the Shoulder- 

 girdle itself is entirely unossified. 



Moreover, these " interclavicular " bones are entirely fibrous (see fig. 3 A, a. i. cl., p. i. cl.) ; 



