350 



CLASSIFICATION OF FISHES. 



greatly flattened, and the ej'es small ; but the loAver jaw 

 is shortest, and the cirri are usually eight.* Here, then, 

 we have the true anguilliform or eel-shaped silures, dis- 

 tinguished in the most remarkable manner from all 

 those we have yet noticed ; and yet so closely allied are 

 they to the last genus, in every thing but their long 

 dorsal fin, that their affinity becomes apparent to every 

 one : the subordinate types or sub-genera, however, are 

 as yet very few, and therefore we cannot state how far 

 the contents of the circle represent that of the last, noi 

 can we trace the connection to the fifth genus, — a genus, 

 however, too remarkable to be confounded with any 

 other; we allude to Cetopsis of Agassiz ( f.g. 89.)- These 



silures present us, in their long anal fins, with the primary 

 character of the Silurincs ; but in all other respects show 

 a marked resemblance to Cyclopium in the circle of the 

 Pimelodince^ and to Astrohlepiis among the AspredincB : 

 hke these, and all other of their representatives, the eyes 

 are remarkably small — indeed,so minute, that they appear 

 as mere specks in the skin, — and their situation is nearly 

 on the top of the crown : the dorsal fin is single, trian- 

 gular, and placed near the crown, which is thick and 

 obtuse : the mouth is large, and the gape obliquely ver- 

 tical : there are no dorsal or pectoral spines : the tail is 

 hardly longer than the body ; and the four cirri are so 

 very short, as not to be so long as the head. Now, it 

 is to these two last characters that we must call the na- 

 turalist's attention. In commencing our survey of this 

 sub- family, it will be remembered that the first genus, 



* Batrachus, Bloch, pi. 370. fig. 1. ; M. Hamiltonn, Hamilt. pi. 26. fig. 45, 



