300 TRANSACTIONS OF THE AMERICAN PHILOSOPHICAL SOCIETY 



skull, and the mouth appears at the end of a membranous pouch; the downward projection 

 would not catch the prey sighted by the forwardly directed eyes, and this is remedied by 

 the head being thrown up as the mouth is protruded." 



The cranium of Stylephorus, as figured by Starks (1908fl, Pis. 1, 2) bears a strange, 

 spurious resemblance to the skull of a bow-head whale. The size of the enormous orbits 

 determines the marked constriction and upward arching of the interorbital bridge and in 

 part also the narrowing of the ethmoid rostrum. The occiput is inclined forward (as in 

 whales), the supraoccipital being in broad contact with the frontals. The hyomandibular 

 is a small rod, sharply inclined backward and with a single articular head; the preopercular 

 is narrow and assists in the support of the simplified quadrate; the pterygoid arch is reduced 

 to a vestigial sliver. The opercular is reduced to a narrow triangular plate which bears 

 a tab-like subopercular. The subopercular is relatively large but has to be inclined back- 

 ward to reach the angular region of the relatively huge mandible, which is composed of 

 the usual elements. From the large size of the glossohyal and the ceratohyal it may be 

 inferred that the muscles of the floor of the mouth are strongly developed and that the 

 whole apparatus acts as a suction trap to drag in the prey. Notwithstanding these strange 

 specializations, both Starks and Tate Regan consider that this fish is distantly related to 

 the Tseniosomi (trachypterids and Regalecus), Tate Regan finally (1929) assigning it as a 

 section of the Allotriognathi. 



In conclusion, the skulls of the Allotriognathi are so very highly specialized that it 

 seemed best to defer treatment of them until after the more central percomorph types 

 had been described, even if the Allotriognathi an as order may have sprung off from the 

 berycoid stem independently of the percoid groups. 



SCOMBROIDEI (CrEVALLES, MaCKEREL, TuNNIES, ETC.) 



Pomatomus. — The blue-fishes {Pomatomus) are generally recognized as connecting the 

 Carangidse with the Serranidae (Tate Regan, 1909^, p. 68) or at least as affording inter- 

 mediate characters, not only in the skull (Fig. 177) but in general body-form and in the 

 caudal peduncle and caudal fin. However, it seems possible that while Pomatomus and 

 Scombrops may now be on the way to becoming carangoids, the real ancestors of the latter 

 may be some deep-bodied Cretaceous form, such as Jipichthys, which Smith Woodward 

 refers, along with other Cretaceous genera, to the Carangidse (1902, p. 3). 



Carangids. — It has been customary to bracket the Carangidse with the Scombridae 

 and related families in a single section of the percomorph order called by Jordan and 

 Evermann (1896, p. 863) the Scombroidei and by Boulenger (1910, p. 675) the Scombri- 

 formes. But those recent American authors who have abandoned all superfamily groups 

 simply list these families near each other, while Tate Regan (1909^, 1929, p. 321) refers 

 the carangoids to the percoid suborder and the true scombriforms to a separate suborder, 

 Scombroidei. Starks, however, as a result of his monographic studies (1909, 1910, 1911a) 

 on the osteology and relationships of the scombroid families, concludes that while the 

 carangids are closely related to the percoid fishes they are even more closely related to 

 the true scombroids. 



A comparison of the skulls of various carangids and scombroids with each other and 

 with those of Pomatomus and other percoids, checked by reference to vertebral, rib and 

 fin characters, inclines me to the opinion that the position of Boulenger and of Starks is 

 essentially correct. 



