Classification of Tdeostean Fishes. 199 



clupea, and verified by the American authors in Beri/x, is well 

 known to be absent in the IMuUichu, and I have failed to find 

 it in Traclticlit}iys, Monoce/itris, nud Polyinixia ; it is there- 

 fore by no means distinctive of the group, and only shows 

 the nearer affinity which these Acanthopterygians bear to the 

 llaplomous Physostomes, from whiclw they are probably 

 directly descended. 



1 was very curious to see how Smith Woodward would 

 deal with the subji ct in his newly-issued fourth volume of 

 the ' Catalogue ot Fossil Fishes,' the Berycidaj being so 

 richly represented in Cretaceous deposits; but, to my dis- 

 appointment, nothing new appears in his definition of the 

 group, wliich is much in the style of his predecessors, and 

 consists merely of those few words: — "Division Berycitormes. 

 — Pelvic fins ihoracic, usually with more than five articulated 

 rays in addition to the spine. No bony stay between circuni- 

 orbital ring and preoperculum." The second character 

 appears merely for the purpose of contrast with the division 

 Scorpa-Miiibrmes. Three families are grouped under the 

 Beiycit'ormes : — Berycidfe (in CiiintliGr's sense), Aphredo- 

 derii'a?, and Cyttida^. As I shall explain presently, I have 

 every reason to think the author justified in thus placing the 

 Ajdiredoderidai near the l^erycidje, but I can see no reason 

 for the association with then) of the Cyttida?, which, in spite 

 of an extra lay to the ventral fins, are as little allied to them 

 as the so-called Scombritbrmes, a group in which families 

 with an increased number of ventral rays also occur (Grammi- 

 cole})ic;a3, Lam]jridida^) . 



1 have set myself the task of making a careful survey of 

 all the chaiacters available for defining the " Berycitormes," 

 but have absolutelj' failed to discover any single feature by 

 which they could be diagnosed from the '' Perciformes." An 

 examination of the skeleton has convinced me that PuJymixia 

 has been correctly placed near Benjx, and that it bears no 

 sj ecial affinity to the JMuilidie, which are themselves more 

 nearly related to the ISparicUe. The Anomalopidaj are 

 probaLly wrorgly })laced near the Berycidse, but 1 have not 

 been able to examine the skeleton, Aphredoderus, on the 

 other hand, has ail the esstntial characters of the Berycida?, 

 and may be regarded as the freshwater representative ot that 

 family. Its vertebral colun)n is of the same type, consisting 

 of 30 vertebrai (14: prafsacral and 16 caudal) *, the para- 



* 29 (14+15) according to Jordan and Everniann (Fisb. X. Am. i. 

 p. 785), •who raise Ap/<ndcclents to the rank of a suborder, Xeiiarcbi, with 

 the I'ollowiii^r dtliuition : — " AVu phice in a distinct suborder, next to the 

 Salnioperca- [I'frcop.^idjc, afaniily of Ilajilouii], the sin^ailar little faniilv 



