172 Mr. G. A. Boulenger on the Suborders and 



illaries ; jaws toothed ; no suborbitals ; prae- 

 operculum well developed ; dorsal and anal very 

 long, formed of soft rays ; ventrals absent ; vent 

 immediately behind the gill-opening 5. Fierasferidce. 



Suborder VII. Catosteomi. 



Air-bladder, if present, without open duct. Parietal bones, 

 if present, separated by the supraoccipital. Pectoral arch 

 suspended from the skull ; no mesocoracoid arch ; coracoid 

 usually very large, or produced posteriorly. Ventral fins, if 

 present, abdominal, or pelvis attached to the coracoid bones. 



The mouth is bordered by the prgemaxillaries, or by the 

 prgemaxillaries and a small portion of the maxillaries. Air- 

 bladder present, except in the Solenostomidre and Pegasida3. 



Following the suggestions of Kner and Steindachner and 

 Cope to their logical conclusion, A. S. Woodward has united 

 the Lophobranchs of Cuvier with the Hemibranchs of Cope, 

 a course which seems fully justified, and has received further 

 support from the recent investigations of Swinnerton *, who 

 has proposed to unite the two groups under the new name of 

 Thoracostei. The structure of the Lophobranchs (Soleno- 

 stomidse and Syngnathidse) shows that these fishes are only 

 extremely specialized forms of the group of which the 

 Sticklebacks are the well-known type, and the character of 

 the "tufted" gills alone is surely not of sufficiently great 

 importance to warrant the retention of the Lophobranchii as 

 a division equivalent to the suborders adopted in the present 

 classification. Besides, as recently pointed out by A. Huotf, 

 there is no fundamental difference, only one of degree, 

 between the so-called tufted gill and the normal type ; each 

 " tuft " corresponds to one branchial lamella, and at a certain 

 stage of development the disposition of the branchial lamella? 

 is the same in a Syngnathus and an ordinary Teleostean. I 

 have recently attempted to show \ that the Lamprididee are 

 related to the Hemibranchii, although sufficiently distinct to 

 warrant the establishment of adivision, named Selenichthyes§. 



The affinities of the Lamprididae are very doubtful. Lam- 

 pris has usually been placed with the Acanthopterygians, a 



* Quart. Journ. Micr. Sci. xlv. 1902, p. 503. 



t Ann. Sci. Nat. (8) xiv. 1902, p. 197. 



X Ann. & Ma«-. Nat. Hist. (7) x. 1902, p. 147. 



§ E. C. Starks, in an important paper (P. U.S. Nat. Mus. xxv. 1902, 

 p. b'19), has shown that the so-called " infraclavicle " of sticklebacks and 

 allies does not exist as a distinct element. The definition of the Cato- 

 steomi as I had originally drawn it up has accordingly been modified. 



