TRACHICHTHYS PRETIOSUS. 65 



ptyx. Indeed the whole resemblance is most marvellous ; and its complete 

 investigation may lead, perhaps, to more correct ideas of the affinities of 

 these singular little fishes than can be gathered from their present forced 

 and artificial place amongst the SalmonidfZ. They may prove merely 

 aberrant forms of Berycidce. 



Trachichthys australis, Shaw, appears, from the figure and descrip- 

 tion, to have the eye proportionately larger (its diameter equalling half, 

 instead of one-third the length of the head) ; the superscapulary spine is 

 stronger than the basal preopercular ; and the pointed scales forming the 

 abdominal keel are only eight in number, and are represented serrated 

 instead of entire. The fin formula, so far as can be ascertained, is 

 D. 4-HlO; A. 8-1-9; V. 1+6; M. B. about 8. In other points it 

 more resembles the Madeiran, than even the Mediterranean fish, figured 

 by MM. Cuvier and Valenciennes. 



The prompt kindness, however, of Mr. Gray, of the British Museum, 

 enables me to add the following decisive particulars : — " I have the plea- 

 sure to inform you," writes this zealous naturalist, in a letter dated 3rd 

 of September 1839, " that we have Dr. Shaw's specimen of Trachichthys 

 australis in the Museum Collection. It is not in a very perfect state, 

 the end of the vertical fins and one of the lobes of the tail having been 

 destroyed. It is very like Hoplostethus Mediterraneus of Cuvier, but 

 is quite distinct from it. It is much shorter compared with its height ; 

 and the scales of the whole body, the ridges of the head, and the whole of 

 the rays of the fins, spinous as Avell as soft, are covered with small spines ; 

 giving it a very rough appearance. It is the roughness of the keel of the 

 ventral plates Avhich gives them the appearance of being serrulated. There 

 are only three, or four at most, spines in front of the dorsal ; and the 

 dorsal and anal fins are shorter than in the Nice species. The (super) 

 scapulary spine is much longer, rough, and subdivided. Our specimen has 

 a broad band of velvet-like teeth on the intermaxillaries ; and a similar, 

 but not quite so broad a band parallel to these within them, which are 

 separated from each other by a small triangular group of teeth in the 

 centre (in front) between their ends." 



By a subjoined sketch, these inner bands of teeth are evidently ^^a/a- 

 izwa/, and the triangular group vomerine. This latter fact docs not 

 indeed correspond with the anticipations of MM. Cuvier and Valenciennes 

 (see p. 56 supra) ; but it cannot disturb the generic relation of the Au- 

 stralian, and Mediterranean or Madeiran fishes. It is only another in- 

 stance, like that of Therapon, and many others, of the little generic value 

 due to the character of an armed or unarmed vomer. 



It may be well to specify, that MM. Cuvier and Valenciennes' description 

 of the Mediterranean fish, is at variance with the Madeiran, mainly in the 

 following points: — In Hoplostethus Mediterraneus., the keel or ridge at the 



