Jordan and Evermann. — Fishes of North America. 2267 



Couceruing the relations of this family, Dr. Gill has the following 

 pertinent remarks : 



" The family of Scombero'ides was constituted by Cuvier for certain forms 

 of known organization, among which were lishes evidently related to 

 Caranx, but which had free dorsal spines. In the absence of knowledge 

 of its structure, the genus EJacate was approximated to such because it 

 also had free dorsal spines. Dr. Giinther conceived the idea of disin- 

 tegrating this family, because, inter alias, the typical ScombcroUJcs (family 

 Svombrida) had more than 124 vertebr;e and others (family Carangidw) had 

 Just 24. The assumption of Cuvier as to the relation.ship of Elacatc 

 was repeated, but inasmuch as it has 'more than 24 vertebra'' (it has 

 25 = 12 + 13) it was severed from the free-spined Caraiu/idw and associated 

 with the Scombrida'. Elacatc has an elongated body, tlattish head, and a 

 colored longitudinal lateral band; Echencis has also an elongated body, 

 flattened head, and a longitudinal lateral band; therefore Echencis was 

 considered to be next allied to Elacaie and to belong to the same family. 

 The very numerous differences in structure between the two were entirely 

 ignored, and the reference of the Echencis to the Scombridw is simply due 

 to assumption piled on assumption. The collocation need not, therefore, 

 longer detain us. The possession by Echencis of the anterior oval cephalic 

 disk in i)lace of a spinous dorsal tin would alone necessitate the isolation 

 of the genus as a peculiar family. But that difference is associated with 

 almost innumerable other peculiarities of the skeleton and other parts, 

 and in a logical system it must lie removed far from the Scombridw, and 

 probably be endowed with subordinal distinction. In all essential respects 

 it departs greatly from the type of structure manifested in the Scombridw 

 and rather approximates — but very distantly — the Gobioidea and Blcn- 

 nioidca. In those types we have in some a tendency to flattening of the 

 head, of anterior development of the dorsal fin, a simple basis cranii, etc. 

 Nevertheless, there is no close affinity nor even any tendency to the 

 extreme modification of the spinous dorsal exhibited by Echeneis. In 

 view of all these facts Echeneis, with its subdivisions, uuxy be regarded as 

 constituting not only a family but a suborder. * * * Who can con- 

 sistently oljject to the i)roposition to segregate the Echeneididw as a sub- 

 order of teleocephaleous fishes ^ Not those who consider that the develop- 

 ment of 3 or 4 inarticulate rays (or even less) in the front of the dorsal fin 

 is sufficient to ordinarily differentiate a given form from another with only 

 1 or 2 such. Certainly the difierence between the constituents of a disk 

 and any rays or spines is much greater than the mere development or 

 atrophy of articulations. Not those who consider that the manner of 

 depression of spines, whether directly over the following, or to the right 

 or left alternately, are of ordinal importance; for such differences again 

 are manifestly of less morphological significance than the factors of a suc- 

 torial disk. Nevertheless, there are doubtless many who will jjassively 

 resist the proposition because of a conservative spirit, and who will 

 vaguely recur to the develoiiment of the disk as being a 'teleological 

 modification,' and as if it were not an actual fact and a development cor- 

 related with radical modifications of all parts of the skeleton at least. 

 But whatever may be the closest relations of Echencis, or the systematic 



