150 Mr. G. A. Boulenger on the 



covered hy tlie ventral muscles. Amorif? the lower Teleo- 

 stomi the bones named clavicles proper by Ge^enbaur, the 

 morphological significance of which is not open to doubt, are 

 superficial dermal bones in Fulypterus and Sturgeons, whilst 

 in the Dipneusti they ossify round tlie primary cartilage of 

 the shoulder-girdle. Therefore, even if the " infraclavicle " 

 of Lampris should prove to ossify from the coraco-scapular 

 primary cartilage, this would be no insuperable obstacle to its 

 homology with the clavicle (infraclavicle), which, as we know 

 from tlie Mammalia, is not always a true membrane-bone*. 

 It would only show that in Lampris, in consequence of the 

 great development of the pectoral region, the bone has evolved 

 further than in any other fish. 



If this identification of the elements of tlie shoulder-girdle 

 of Lampris be conceded, all difficulties from the systematic 

 point of view disappear at once. Tlie Opah must be regarded 

 as more nearly allied to the Heraibranchii than to any other 

 group of fishes with which we are as yet acquainted. But 

 it possesses important features indicative of both greater 

 generalization and specialization which require the establish- 

 ment for its reception of a division of at least equal importance 

 with the Hemibranchii and Lophobranchii, and for this 

 division I propose the name Selenichthyes. The close rela- 

 tionship existing between the Hemibranchii (Sticklebacks and 

 Pipe-Fishes) and the Lophobranchii (Needle-fish and Sea- 

 horses), realized long ago by Kner and Steindachner f and 

 by Cope I, are now being admitted on all sides §; Dr. A. S. 

 Woodward || and Dr. Swinnerton ^ have independently 

 })roposed to unite them into one suborder, the former using 

 the term Hemibranchii in an extended sense, the latter pro- 

 posing the new nameThoracostei, " expressive of the presence 

 in all [Hemibranchii and Lophobranchii] of a more or less 

 complete bony armature, and especially of infraclavicles." 

 Unfortunately neither of these names would be appropriate 

 for the suborder after the addition of the Lamprididie, in 

 which the branchial apparatus is complete and the dermal 



* Cf. Gegenbaur, Yergl. Anat. i. pp. 20" & 496 



t Denkschr. Ak. AVien, xxi. 1862, p. 28. 



X Tr. Amer. Pbilos. Soc. (2) xiv. 1871, p. 457. 



§ On the difference in the structure of the gills, cf. the recent work of 

 A. Huot, Ann. des Sci. Nat. {6} xiv. 1902, p. 197, who showd that there 

 is no fundamental difference, only one of degree, between the so-civlled 

 tufted gill and the normal type, and that at a certain stage of develop- 

 ment the dispositim of the branchial lamellae is the same in a Synijnathns 

 and in an ordinary Teleostean. 



II Catal. of Fossil Fi.shes, iv. p. 369 (1901). 

 ^ Quart. Journ. Micr. Sci. xiv. 1902, p. 580. 



