574 H. H. SWlNNElil'ON. 



Systematic Position and Affinities of 

 Gasterosteus. 



The present views of the relationships of the Gasterosteidas 

 can be best stated by entering briefly into the history of their 

 growth. 



On account of the presence of spines in the fins, and of the 

 union between the suborbital bone and the pre-opercukTm, 

 Cuvier and Valenciennes (29) classified this family with the 

 mail-cheeked Acanthopterygii, which included the Triglidas 

 and Cottidge. Thirty years later Giinther (61), whilst still 

 retaining it among the Acanthopterygii, separated it from 

 the mail-cheeked forms ; and because of its spinous anterior 

 dorsal fin and the abdominal position of the pelvics, 

 put it with the Fistularida? in a separate division, the 

 Acanthopterygii Gastrostei formes, and indicated his ideas 

 of the affinities of this by placing the Mugiliformes and 

 the Centrisciformes on either side. In a paper, which 

 has since formed a basis for much recent systematic 

 work. Cope, because of the position of the pelvic fins 

 (70, p. 456), separated these three divisions from the 

 Acanthopterygii, creating for the Mugiliformes the distinct 

 order Percesoces, and for the other two, the Hemi- 

 branchii. The latter he regarded as annectant between the 

 former and the Lophobranchii ; looking upon the weak 

 branchial apparatus, the presence of interclavicles, the simple 

 post-temporals, the prolongation of the muzzle, and the 

 presence of ganoid plates on the body, as approximations to 

 the Lophobranch type; and upon the character of the dorsal 

 spines as an indication of relationship to NematocenLris, one 

 of the Atherinidas. Later systematists have followed Cope 

 more or less closely. Gill (93) placed the Hemibranchii next 

 the Lophobranchii, and considered it as equivalent not 

 merely to the Acanthopterygii, but to an order — the Telo- 

 cephali — containing these and the Haplomi, Scomberesocidse, 

 and Percesoces. Jordan and Evermann (96) acknowledge 

 that it is closely allied to the Lophobranchii, and that though 



