320 MEMOIRS OF THE CARNEGIE MUSEUM. 



As for the constitution of the order Solenichthyes, this is made by Regan in 

 his later publications to comprise the Aulostomids, Centriscoids, and the old Cu- 

 vierian group of Lophobranchii, or specialized sticklebacks with " tufted " gills. 

 The family Pegasidse is admitted by Boulenger into the same association with the 

 foregoing, but is excluded from this order by Regan and placed in a group by them- 

 selves (order Hypostomides) . 



We have now to consider the position of two fossil forms, concerning which 



there is some difference of opinion. These are the genera Urosphen and Rham- 



phosus of Agassiz, both from the Eocene of Monte Bolca. They were both referred 



to the flute-mouths (Fistulariidae) by Dr. Giinther, but, as recognized by Dr. Gill, 



and following him A. Smith Woodward, " one of them is more nearly related to the 



Macrorhamphosidse and Gasterosteidse." For the one in question the new family 



Rhamphosidse was established by Gill to contain it, and Urosphen was also made 



the type of an independent family. The two new families proposed by Gill in 



1884 are thus defined bj^ him: 



Urosphenid^. 



Hemibranchs with the first four vertebrae much elongated, a moderately 

 elongated body, a long tubiform mouth (ventrals abdominal?, dorsal unknown), 

 and a very large cuneiform caudal. 



Rhamphosid^. 



Hemibranchs with the anterior vertebrae normal (not elongated) and separate, 

 about twenty-two (eight abdominal and fourteen caudal) vertebrae in all, plates 

 on the nape and shoulders only, with a tubiform mouth, subthoracic ventrals, a 

 dorsal spine behind the nuchal armature, and the second dorsal and anal far behind 

 and opposite. 



Regarding Urosphen, it may be recalled that Agassiz himself recognized its 

 intermediate position between the Aulostomids and Fistulariids. Unlike the 

 former, Urosphen is scaleless, and small teeth are present in the jaws. From 

 Fistularia it is distinguished chiefly by the form of the very large cuneiform caudal 

 fin, but in other respects approaches very closely to that genus. In grouping it 

 provisionally with recent flute-mouths, A. Smith Woodward gives the following 

 tabulation : 



^ V Synopsis of Genera. 



No free dorsal spines; caudal fin forked, with elongated median rays; no scales Fistularia. 



A series of free dorsal spines; caudal fin rhombic, without elongated ray; 



small ctenoid scales present Aulostonia. 



Imperfectly known, but all caudal fin-rays much elongated; no scales Urosphen. 



