1884.] NATURAL SCIENCES OP PHILADELPHIA. 155 



(H. Centrisciformes.) 



137. Centriscidm Centriscidm, Gthr., iii, 518-524. 



138. Amphisilidm Centriscidm, Gthr., iii, 518, 524-527." 



In the " Introduction to the Study of Fishes " (1880, p. 50T), 

 Dr. Giinther has referred the Aulorhynchoid fishes to the family 

 Fistulariidse. 



In the " Synopsis of theFishes of North America " (1 882, p. 387), 

 five families were recognized for American species by Messrs. 

 Jordan & Gilbert, and grouped as follows : — 



" * Bones of head produced into a long tube, which bears the 

 short jaws at its end. 

 a. Body short, compressed, scaly ; no teeth ; spinous dorsal 

 present. ..... Centriscidae, 60. 



aa. Body elongate ; teeth present. 

 6. Dorsal spines none ; a long caudal filament ; no scales. 



Fistulariidae , 61 

 bb. Dorsal spines present, disconnected ; no caudal filament, 

 c. Body covered with ctenoid scales. Aulostomatidae, 62. 

 cc. Body scaleless, with bony shields. 



Aulorhynchidae^ 63. 



** Bones of head moderately produced ; ventrals 1,1; dorsal 



preceded by free spines ; body scaleless, naked or mailed. 



Gasterosteidae, 64." 



On a recent review of the forms of the order, I am more than 

 ever convinced of the aptness of the classification proposed by 

 myself in 1872 and submit the following table and characters 

 which will, I think, amply justify that confidence. Far from 

 being able to see any close affinity between the Aulorhynchidae 

 and Aulostomidse, I am Unable to appreciate any very distinctive 

 differences from the Gasterosteidae, and the close affinity between 

 Aulorhynchus and Spinachia is such that I regard the family 

 Aulorhynchidae simply as a convenient one at the most, and as 

 expressing the culmination in one direction of the tendency 

 characteristic of the order. I should be scarcely disinclined to 

 dissent from any who should combine the Gasterosteidae and 

 Aulorhynchidae in one family. 



