318 MEMOIRS OF THE CARNEGIE MUSEUM. 



with the Lophobranchs of Cuvier, and the family Pegasidse, whose position in the 

 system was admitted to be still somewhat doubtful. 



In 1903 ajDpeared an important paper by C. E. Starks^ on the Osteology of the 

 Hemibranchiate Fishes, in which he discussed the arrangement of families belonging 

 to this division in the sense originally proposed by Cope, and calls attention to the 

 fact that " Dr. Gill has pointed out how the tube-mouthed forms have descended 

 in an unbroken line from Gasterosteus through Spinachia and the family Aulorhyn- 

 chidae, these constituting the superfamily Gasterostoidea " (I. c, p. 622). The 

 Danish writer H. F. E. Jungersen published in 1908 a valuable memoir, in which it 

 was shown that the features of the Aulostomids, Centriscoids, and Lophobranchii 

 are such as to compel us to regard these divisions as constituting a natural group. 

 This view was supported by Regan in two papers published by him during the 

 years 1909 and 1910, the final arrangement of families advocated by him being as 

 follows : 



Order Thoracostei Swinnerton. 



" The order Thoracostei comprises the Gastrosteidse and Aulorhynchidae. 

 Swinnerton'' has shown that the dermal plate which appears as part of the coracoid 

 is in reality a distinct element. I cannot accept Jungersen's view that these fishes 

 belong to the Scorpsenoidea, although I readily admit that the Aulostomids are 

 more distinct from the Thoracostei than I recently considered them to be."^ 



Order Solenichthyes Regan. 



Under this caption are included by Regan the Aulostomids, Centriscoids, and 

 Lophobranchs, whose features show that they form a natural group. 



A few words may be said regarding the constitution of these orders, Thoracostei 

 and Solenichthyes. Under the first-named are now placed by Regan only the 

 Gasterosteidse and the Aulorhynchidae, in which procedure he follows the example 

 of Gill and Starks in their earlier arrangement of modern genera of sticklebacks. 

 As early as 1871 the former of these writers had associated the families Aulorhyn- 

 chidse and Gasterosteidse in a single division contrasting with the Aulostomids and 

 Centriscoids, and in a subsequent review of the forms of the order in 1884 he 

 remarks pointedly as follows :'' " Far from being able to see any close affinity between 



' The Shoulder-Girdle and Characteristic Osteology of the Hemibranchiate Fishes. Proc. U. S. Nat. 

 Mus. (1903), Vol. XXV, p. 619. 



* Quart. Journ. Micr. Sci., XLIX, 1905, p. 363. 



' Ann. Mag. Nat. Hist., 1909 (8), Vol. Ill, p. 84. 



' Gill, T. N., "On the mutual relations of the Hemibranch Fishes." Proc. Acad. Nat. Sci. Phila., 

 1884, p. 155. 



