1902.] 



PLECTOGNATHOUS FISHES. 285 



althouo-li there can be no question as to the close relationship of 

 the less speciahzed forms to the Acanthuridse, their differences 

 from that family are sufficiently great to admit of a definition 

 which separates them from the Acanthopterygii, and at the same 

 time includes the more aberrant and speciahzed forms. 



The feature of most importance in diagnosing the suborder 

 Plectognathi is the absence of ribs, although in some well-ossified 

 epipleurals are present which have been mistaken for ribs. Two 

 divisions are recognized, for which the names Sclerodermi and 

 Gymnodontes, originally proposed by Cuvier, are retained. The 

 Triodontidje, however, are removed from the latter division and 

 placed in the former, the structure of their pectoral arch and 

 vertebral column, as well as the presence of a pelvis and of well- 

 ossified epipleurals, indicating their close relationship to the 

 Triacanthidas and Balistidfe ; whilst the coalescence of the teeth 

 in the jaws is a feature of little importance, and has, as probably 

 as not, originated independently in these fishes and in the 

 Gymnodontes. 



The Ostraciontida?^ do not seem to me to differ sufficiently trom 

 the Sclerodermi to rank as another division— Ostracodermi. Their 

 very close relationship to the Balistida?. is apparent in their 

 physiognomy and in the structure of their skeleton ; whilst the 

 absence of epipleurals and of the pelvis is obviously due to 

 the development of the exoskeleton, which, however, is not very 

 different from that of JBcdistes, many species of which have 

 exoskeletal plates distinctly hexagonal in certain areas. I have 

 inserted notes after the diagnoses, explaining the omission of 

 characters before used or the addition of those now used for the 

 first time. 



Suborder PLECTOGNATHI. 



Similar to the Acanthopterygii, but without ribs ; with the post- 

 temporal short, simple, and completely united by suture to the 

 squamosal, and the pelvic bones, if present, more or less com- 

 pletely co-ossified. Branchial apertures very restricted. Pre- 

 maxillai-ies and maxillaries often firmly united, opercular bonts 

 more or less reduced, and scales usually osseous or spinate. 



Cuvier characterized the Plectognathi as having no ribs, but 

 other systematists have generally agreed that ribs are present in 

 Balistes, Triacanthus, Triodon, &c. I find that the so-called ribs oi 

 the Triacanthidse and Balistidse are epipleurals, which are attached 

 to the anterior caudal vertebrjs as well as to the prsecaudals 

 (text-tig. 56, p. 286), and are intermuscular bones, not bordering 

 the abdominal cavity. I have unfortunately had no opportunity 

 of examining the skeleton of Triodon, but have had to rely on 

 the memoirs of Hollard ^ and Dareste " ; but I think I am justified 

 in supposing tliat in Trlodon, as in Balistes and Trmcanthus, 



1 Ann. Sci. Nat. (3) xx. 1853, p. 71 ; (4) viii. 1857, p. 275, and (4) xiii. 1860, p. 1. 



2 Ann. Sci. Nat. (.3) xii. 1849, p. 68, and (3) xiv. 1850, p. lOo. 



