THE SHOULDER-GIRDLE IN FISHES. 23 



spindle-shaped, bi-polar bar is cupped in its broad middle part (fig. 2) ; below (fig. 3), it is sub- 

 carinate ; each ramus ends in a blunt point, and is sheathed below and behind, by the scooped 

 lower end of the clavicle. Here we stand, as it were, with one foot amongst the Plagiostomes 

 and the other planted within the Teleostean boundaries ; and the epicoracoids are cut off from the 

 coracoids, that Nature may be ready to throw them aside when she comes to the typical Fishes. 

 The Amphibia, which are also Dipnoi, will require these plates again, enlarged and expanded, 

 to form in them a double-breasted vestment of cartilage ; in them these cartilages will be found 

 to be continuous with the main part, a character not to be lost again until the observer has 

 risen above the Monotremes ; the parts themselves, however, are to be found even amongst 

 the Bats. Notwithstanding the Placoid character of the epicoracoids, there is nothing like the 

 Microscopic bony pavement of the Sharks and Rays, and the livid colour of the cartilage is in 

 strong contrast with the sea-green colour of the bones outlying it. In all the long rods of 

 cartilage the Lepidosiren agrees with the Sturgeons and Teleostei in having the ossification of 

 the ectosteal kind (fig. 1, p. b. 1), but, as in the Sturgeon and some of the Teleostei (e.g. 

 Gobius, Cyclopterus,) it is very feeble and arrested. 



The rays of cartilage (wholly unossified) in the long, fringed ( " crossopterygian ") fin are 

 rather compressed from side to side (fig. 1, c. 2), and are about three times longer than 

 they are broad. The convex proximal end of the first, or humerus (h.), articulates with the 

 glenoid cup (gl.) ; then comes the radio-ulnar (r. u.) ; first carpal (c. 1) ; second carpal (o. 2) ; and 

 phalanges (ph.) ; the smallest of those shown in the figure are the 20th and 21st ; but these were 

 not the terminal segments. 



PISCES TELEOSTEI (A. Siluroidei). 

 Example 1. Callichthys littoralis, Giinth. (the Round-headed Hassar). 



My figures (Plate I, figs. 9 13) of this exquisite fish are from a specimen five inches in 

 length, and are magnified two diameters. As this fish lies in the very channel which leads from 

 the Ganoids to the Teleosteans, and as the outer and inner bones of the shoulder-girdle 

 cannot be understood irrelatively, I shall make little apology for describing parts that lie beyond 

 my border. If Clarias be valuable as throwing light upon the fishes of the Old Red Sandstone, 

 a still more important link is found in Ccdlicltthys, a Siluroid in which the ganoid plates are 

 much more abundant and perfect. All the bones that can be seen in this fish immediately 

 beneath the extremely thin cuticle are dermal ; but the anterior and under parts of the face 

 are covered with thick skin, beneath which are to be found both subcutaneous dermals and 

 true endo-skeletals. All the bones that appear with their dense enamel-like coat at the 

 surface have unpolished parts hidden below, but I can find no separation into distinct ganoid 

 and subcutaneous bones in any region of the exo-skeleton. Such a separation does occur 

 in the Lophobranchii, as I shall soon show. In Plate I, fig. 9, there are no subcutaneous 

 bones shown ; and, indeed, the pra-maxillaries, maxillaries, dentaries, angulars, preoperculars 

 (squamosals), inter-operculars (the sub-operculars are wanting), branchiostegals, and the so-called 



