GREGORY: FISH SKULLS 



227 



Conceivably the fistulariids might well be derived from a short-bodied form like Cen- 

 triscus by the loss of the dorsal spine, the sudden elongation of the body and the fusion 

 of the already elongated first four vertebrae. Thus the general lines of ascent may be 



Solenostomus cyanopterus 



Fig. 107. Solenostomus. After Jungersen. 



visualized as first from Gasterosteus to Centriscus, thence on the one hand to the fistulariids, 

 and on the other, through Solenostomus, to the syngnathids. 



The osteology of the skull and skeleton of the hemibranchiate fishes has been fully 

 described by Starks (1902a) and by Jungersen (1908). Up to the time of the publication 



pJM 



Centriscus 



Fig. 108. Centriscus. After Jungersen. 



of Jungersen's monograph modern ichthyologists had generally accepted the view of Cope, 

 Gill, and Smith Woodward that the sticklebacks, aulostomids, tube-mouths, centriscids, 

 sea-horses and their collateral branches together constituted either a single order or at most 

 two orders, formerly called Hemibranchii and Lophobranchii. Jungersen, however, came 

 11 



