258 Papers from the Marine Biological Laboratory at Tortugas. 



Gunther's excellent little outline drawing (text-fig. 2) has been previously 

 referred to and need detain us but a moment. One's attention is forcibly 

 called to the absence of spots on the anterior parts, to the elongated and 

 very rounded ventrals, and to the exaggerated spine. The dorsal fin and 

 head are well drawn, the eyes being almost invisible, but the spiracles are 

 placed rather too near the mid-dorsal line. 



Text-fig. 3. A. narinari, dorsal view, after Jordan and Evermann (1900), Florida. 



The only other figures of this fish (and the only drawing of the ventral 

 surface known prior to 191 3) are those previously referred to from Jordan 

 and Evermann's "Fishes of North and Middle America," and widely repro- 

 duced in various papers by one or the other of these ichthyologists, or by 

 other American writers, some of whose papers will be referred to later. 

 If text-figure 3 (their drawing) be compared with figure i, plate i (my photo- 

 graph), the following differences are observable: the pectorals are more 

 sharply pointed, the ventrals more widely spread, longer, and more rounded 

 than in my specimens. The spots are larger and fewer in number (especially 

 on the head and the edges of the pectorals) than in Beaufort and Key West 

 specimens. The most striking difference, however, is to be found in the 

 absence from the latter fish of the dark, roughly parallel lines, which extend 

 from the dorsum of the former out over the pectorals. 



I have elsewhere recorded (Gudger, 1910) the difficulties I had in 

 identifying my first specimen as A . narinari, but it may not be out of place 

 to repeat them here. The lines above referred to are sometimes present in 

 the living or at any rate in the freshly killed fish, but are so indistinct as 



