3. XETEAGONtTEUS. 407 



of the head one-fifth ; the diameter of the eye equals the extent of 

 the snout (in mature specimens), and is nearly one-fourth oi the 

 length of the head. The first dorsal spine pungent, the others fili- 

 form ; the last dorsal rays produced, reaching to the hase of the 

 caudal. The ventrals end in a thread-like tip which overlaps the 

 commencement of the anal. The first dorsal and the anal commence 

 in nearly the same vertical. An even black stripe replaces the usual 

 silvery band, and is continued forwards over the giU-cover, upper 

 half of the eye, and sides of the snout ; it occupies the fifth series of 

 scales and the adjoining parts of the fourth and sixth. 

 Freshwater streams of Port Essington. 



a. Three inches long : ekin. King's River, near Victoria. — Type of 



the species. 



b. Adult : skin. Port Essington. From Mr. Gould's Collection. 

 c-e. Twenty-one lines long : skins. Severn River (New South 



Wales). — I am not quite sure whether these specimens really 

 belong to A. nigrans. The form is nearly the same, but they 

 have a very indistinct blackish band on the tail. 

 /. Adult: skin. Australia. 



Second Group. TETRAGONURINA. 



Teeth not very small, compressed ; scales keeled and striated. The 

 first dorsal continued on to the second. Air-bladder none. Pyloric 

 appendages numerous. 



3. TETRAGONURUS. 



Tetragonurus, Risso, Ichth. Nice, p. 347. 



Body subcylindrical, somewhat elongate ; snout of moderate length, 

 obtuse ; lower jaw elevated, with a convex dental margin. Teeth in 

 single series, compressed, triangular, rather small ; teeth on the 

 palate. Scales rather small, with deep stria), those of every single 

 transveree series coherent. Ventrals abdominal, but inserted below 

 the middle of the pectoral. Air-bladder none. Pyloric appendages 

 numerous. 



Mediterranean, and the neighbouring parts of the Atlantic. 



This genus has been annexed to very different families : first lo the 

 MugilidcB by Cuvier, then to the Scombridce by Lowe, and fiaally to 

 the NotacantJdni by Miiller. The attempt of the latter is certainly 

 the most unfortunate, and I doubt whether Miiller has known the 

 fish, except from Valenciennes' description. Even when the artificial 

 group called by Miiller Notacanthini is divided into natural families, 

 Tetragonurus has no affinity whatever either to Notacanthtts or to 

 Mastacembelus. ^Tien Mr. Lowe places this fish among the Scom- 

 beroids of Cuvier, he implies, in the words by which he modifies his 

 view, that he means that portion of them which I have called Tri- 

 chiuridce. But in Tetragonurus the ventral fins are truly abdominal, 



