XXX IX 



(e. g. Synbranchus, Anguilla, Muraena), similar brain, so that in default 

 of sufficient opportunity to study the skeleton, 1 the author provisionally, 

 at least, retains them united, but admitting Cope's orders as suborders. 

 Their affinities through the more generalized forms of the order are pos- 

 sibly with the Gymnonoti, but the hints furnished by the elongated body 

 and increased number of vertebra?, etc., may be illusive. 



OPISTHOMI. 



The Notacanthidse and Mastacembelidae have recently been widely sepa- 

 rated, 2 and by Cope, an order (Opisthomi) has been established for the last, 3 

 but, as long ago shown by Johannes Miiller, both the forms in question 

 agree in the withdrawal of the shoulder girdle from the skull, and its con- 

 nection with the vertebral column, and this character seems sufficient, asso- 

 ciated as it is with general agreement in other respects between the two 

 families and great dissimilarity from other fishes, to isolate the forms thus 

 marked as a peculiar order ; 4 for this order, the name Opisthomi, pro- 

 posed by Cope for one of its members, will be very appropriate, and may 

 be adopted for the enlarged group. It is not obvious what better place 

 can at present be assigned to them than proximity to the Apodes, although 

 it will probably be eventually found to have closer relations with other 

 forms. 



HEMIBEANCHII. 



The order Hemibranchii, framed by Cope for the group here adopted, 

 seems to be also well worthy of recognition ; and, in addition to the char- 

 acters assigned by its founder, is distinguished [i. e., Gasterosteidse, Fis- 

 tulariidse) by the structure of the shoulder girdle and the skull, as shown 

 by Parker in the case of the Gasterosteidse {Shoulder Girdle, p. 39). 5 

 The nearest relations, according to Cope, are apparently with the Atheri- 

 nidse, but such are not obvious, nor are they more so with the Siphonog- 

 nathidae, with which tbey have also been in part compared. 



LOPHOBRANCHII. 



The order Lopiiobranchii, according to Prof. Cope, is most nearly related 

 to the Hemibranchii, and such appears to be probable ; some members of 

 the order Hemibranchii (Fistulariidae) had, indeed, been long previously 



1 I have only been able to study the osseous structure of Anguilla and Murtzna. 



2 See Gunther, Cat., v. 3, Syst. Synopsis, pp. viii. x. 



3 No reference is made by Prof. Cope to the Notacanthidae in any connection. 



* Of course, Tetragonurus, which Miiller, who was unacquainted with it, hinted 

 might belong here, has no relation with the group. 



5 Before I was aware of the peculiarities of the shoulder girdle, and only knowing 

 the characters assigned to the order by Cope, I retained it iu the order Teleocephali. . 



