1904.] OSTEOLOGY OF THE ELOPID^ AND ALBULID^, 73 



the Teleostean series (excluding extinct forms such as the Pholi- 

 dophoridEe, in which " it is not yet quite clear that the mandible 

 was destitute of splenial and coronoid elements " (Smith Wood- 

 ward, Vert. Pal?eontol. 1898, p. 114), might more than any others 

 be expected to possess traces of this constituent. 



The size and shape of the dentary bone vary very considerably. 

 The bone is usually large ; but it is much reduced, especially in 

 its anterior parts, in Chanos. The coronoid pi'ocess is usually 

 formed entirely by the dentaiy, but the ectosteal ai'ticular occa- 

 sionally forms the hinder part of it. In Chatoessus the process is 

 situated unusually far forward, in a position recalling that of the 

 remarkable coronoid process of Gonorhynchus. 



In the extinct family Saurodontidfe the teeth are lodged in 

 sockets ; and Boulenger, in his recent synopsis of the families of 

 Teleostean fishes (Ann. & Mag. Nat. Hist. xiii. 1904, p. 164), 

 places the Chirocentridae next the Saurodontida?, and separates 

 them from the Clupeidiie because they have " teeth in sockets." 

 The teeth of the three specimens of Chirocentrus clorab examined 

 for the purposes of this investigation ai'e, however, cei'tainly not 

 lodged in sockets ; they are anchylosed to the edge of the bone, 

 and ai'e flanked by a slight ledge on the external side, exactly as 

 in Coilia, for instance. 



Hyopalatine Series. — The presence of one or two articular heads 

 for the front of the hyopalatine arch, a matter upon which Swin- 

 nerton (Quart. Journ. Micro. Sci. xlv. 4, 1902, pp. 551, 556-557, 

 ife 584) lays considerable stress, may possibly be an adaptive 

 feature related to the length of the ethmoid region of the skull, 

 for in Elops and Megaloj^s the ethmoid region is short, and the 

 front of the hyopalatine arch has but a single head, whereas in 

 Alhula the ethmoid region is long and there are two heads to the 

 palatine. The difference does not strike one at the outset as 

 likely to be of fundamental importance : extended investigations 

 upon dissected, ^. e. not dried, skulls alone can decide the point. 

 An unusual condition occurs in Coilia, in which the ethmoid 

 region of the cranium is short and the palatine has two heads ; 

 but these are right and left, and not anteiior and posterior. This 

 peculiarity is not shared by the allied genus EngrauUs. 



The palatine bone is fused with the ectopterygoid hi Arapaima, 

 Osteoglossum, and Rotopterus, and in these genera and in Hyoclon 

 there appears to be no endosteal part of the palatine. In the 

 Mormyroid fishes there is no sepaiute entopterygoid, and the 

 palatine bone is fused with the side of the vomer. 



The metapterygoid attains its maximum size relatively to the 

 adjacent bones in Coilia ; in Gonorhynchus the other extreme is 

 reached, for here it is reduced to a fine needle of bone. The 

 articulation between the metapterygoid and entopterygoid region 

 of the hyopalatine arch of Osteoglossum and the lateral process of 

 the parasphenoid described by Bridge in 1895 (Proc. Zool. Soc. 

 1895) is considered by Swinnerton (l. c. p. 572) on ernbryological . 



