Devonian Fisli-Funna of S)p}'tzher(jpn. 13 



whole leiii^th and forms a jjlaiie mpetin;>' tiiat of the onia- 

 iDcnted plate in an acute angle. Tlie lonj^ narrow clement 

 extending along the right side of the fossil [x) is exhibited 

 for the most part as an impression of the inner aspect ; hut 

 sufficient remains to prove that it was comparatively thin 

 except at the border that is now placed innermost, while it 

 tapers in front to an obtusely rounded extremity, on which 

 the superficial ornament is delicate and composed of longitu- 

 dinally directed nodose ridges. The bone consists of a single 

 nearly flat lamina, and seems to exhibit the characters of a 

 cheek-plate. 



Another problematical bone, of very large dimensions, is 

 preserved on the inferior aspect of the fossil, and is shown of 

 two thirds the natural size in PI. III. fig. 2. In general form 

 it is very suggestive of a clavicular element, and the face 

 exposed to view is marked only by structural fibres radiating 

 from the centre of ossification. At present, however, this 

 element cannot be assigned to a definite place in the skeleton. 



Form, and Loc. Ironstone, Mimers Valley. 



Subclass TELEOSTOMI. 



Order C R S S O P T E R Y G I I. 



Families Holoptychiidae and Rhizodontidse. 



Genera non det. 



It is somewhat remarkable that in the collection from the 

 Mimers Valley the Holoptychian fishes should be represented 

 exclusively by teeth, while the Rhizodonts are known only 

 by scales and one imperfect clavicle. A microscopical exam- 

 ination of the teeth has confirmed Lankester's surmise {op. 

 ci't. }). 6) that they are truly Dendrodont in structure, and 

 they are thus excluded from correlation with the numerous 

 Strejjsodus-Yike scales which are well figured in the memoir 

 already quoted. The teeth are not iniprobably referable to 

 two species of IJolopfi/chncSj and the scales are very sugges- 

 tive of those of Sauripterus ; but until the discovery of more 

 satisfactory specimens it seems unwise to attempt generic 

 and specific determinations. 



