230 TRANSACTIONS OF THE AMERICAN PHILOSOPHICAL SOCIETY 



"In the trunk region, however, though the pectorals are raised, the pelvics, abdominal, 

 and the arrangement of the other fins is the same as in Thoracostei; yet the complete ab- 

 sence of bony plates and infra-clavicles gives some excuse for not including the Scombere- 

 soces in the new sub-order." 



The peculiar connections of the nasal bones of Gasterosteus, as described by Starks 

 (1926(2, pp. 212-213), further separate this family from the mail-cheeked fishes. For in Gas- 

 terosteus the nasals are "closely attached to the frontals by dentate sutures and continue 

 the rugose surface of the frontals forward. They are widely separated, and the area be- 

 tween them drops abruptly to the level of the upper surface of the flat vomer, and forms a 

 deep fossa for the reception of the ascending premaxillary processes. From the lower sur- 

 face of each nasal a large process turns backward and is firmly attached against the side of 

 the parasphenoid, and against the lower, forward-projecting, plate of the prefrontal. The 

 nasals thus both roof, and (with the help of the prefrontal) floor the deep nasal fossa. 

 This is, as far as known, a unique condition." 



The backward extension of the third suborbital bone over the cheek might well be a 

 sign not of relationship with the scorpaenoids but of descent from primitive clupeoid iso- 

 spondyls. 



The evidence of the otoliths, as presented by Frost {\929b, p. 263), does not favor the 

 allocation of Gasterosteus to the scorpaenoids, since its sagitta "shows little resemblance to 

 those of the remainder of the order Scleroparei, but it resembles the Scopelid type." On 

 the other hand, Swinnerton's contention that Gasterosteus is related to Belone is not sup- 

 ported by a comparison of their otoliths, those of Belone and other synentognaths (figured 

 by Frost in 1926c, p. 471) conforming in general tothe"Elopine," "Clupeid" and "Biovate" 

 types, while that of Gasterosteus is rhomboidal, biconvex and of scopelid type. As to the 

 possible relationships of the sticklebacks to the pipe-fishes. Frost notes (1929^, p. 263) that: 

 "In the example examined of Spi^iachia spinachia, the otoliths were absent or microscopic, 

 as frequently occurs in the Pipe-fishes, which in some features it resembles." 



