Systematic Position 0/ Luvarus imperiali.s. 281 



and anal unbranched, but not forming pungent spines. 

 Ventrals with one spine and four soft rays. 



The Luvaridai may be considered as ultra-specializ'id 

 Acanthuridai, and it is generally agreed that the Plectogiiathi 

 are descendants of the same fiiinilj, the origin ot" which is by 

 no means so settled a question. Jordan and Everniann place 

 them near the Ciiaetodonts, vvitli the Z mclidaj as a connecting 

 family; but this arrangement has recently been criticized by 

 Svvinnerton *, in his valuable paper on the morphology of 

 the Teleostean head-skeleton. He considers the mode of 

 attachment of the palatine arch to be a character of great 

 importance, and shows that most Acanthopterygii have the 

 palatine attached to the ethmoid region at two points— the 

 pre-ethmoid and par ethmoid comua — whilst the Plecto- 

 gnathi, with the allied Zanclidre and Acantliuridtfi, have only 

 the pre-ethmoid attachment. This leads him to the conclusion 

 that the origin of the Acanthuridas is not to be sought within 

 the Acanthopterygii, but lower dov/n, and that they may be 

 more nearly allied to tlie Scombresoces, which have a single 

 palatine attachment. This position would require strong 

 arguments to substantiate it, as the Acanthuridae are typically 

 Acanthopterygii in all other characters, and, judging by what 

 we know of the fossils, are not a family of great antiquity. 



I have not the least doubt that the Teutliididaj are closely 

 allied to the Acanthurida;, and on examination [ find that in 

 Teuthis the palatine is attached to pre-ethmoid and par-ethmoid 

 by two facets. Moreover, the author above mentioned, in 

 the same paper, describes the attachment of the palatine in 

 Chcetodon in these words : — " The palatine lies against the 

 par-ethmoid, it is united to that bone exclusively by ligaments, 

 and the articular surfaces have aborted." Here, then, we 

 have an intermediate condition between the mode of attach- 

 ment in the typical Acanthopterygii and that of the Acan- 

 thuridge, and it only needs the elongation of the ethmoid 

 region, thus removing the par-ethmoid from contact with the 

 })alatine, to produce the arrangement which characterizes 

 that family. It seems therefore that the position of the Acan- 

 thuridje near the Uhsetodonts is by no means weakened by the 

 corisideration of the ethmo-palatine attachment, and, as the 

 two families approach each other in many characters of 

 importance, that they have been rightly placed near together 

 in the system. 



* Quart. Journ. Micr, Sci. xiv. 1901, pp. 5o5, 582, & 583. 

 Ann. d& Mag. N. Hist. Ser. 7. Vol x. 21 



