H^ PAPERS FROM TORTUGAS LABORATORY vol. xxxiv 



perfect rhythm as if dancing to music. In another entry he gave another descrip- 

 tion of the color, having at least 2 specimens before him, respectively 105 and 130 

 mm. long: "Color coppery, lighter below; peritoneum very slightly pigmented — 

 mere scattered points." 



The Tortugas collection contains 4 specimens, 44 to 101 mm. in standard 

 length, which have been compared with some specimens from Cuba in the U. S. 

 National Museum collection, identified as P. miilleri. The Tortugas material 

 does not agree entirely with the Cuban specimens. It is not possible, however, to 

 arrive at a definite conclusion concerning the relationship, as the Cuban material 

 especially is not in good condition, and not enough specimens are available. 

 Furthermore, the matter is complicated by some evident variation among the 

 Tortugas specimens. 



The specimens from Tortugas seem to be rather deeper and have a rather 

 larger eye than Cuban ones of about equal size. The principal difference, how- 

 ever, is in the scales; those covering a large area on the anterior part of the side, 

 especially the ones behind the pectoral fin, are prominently enlarged and very 

 distinctly cycloid and have concentric rings on their margins, whereas in the 

 Cuban specimens, though most of the scales are lost, enough remain to show that 

 they are not prominently enlarged on the anterior part of the side and that their 

 margins are definitely crenulate. 



In having cycloid scales on the sides the Tortugas specimens are related to 

 P. polio Breder (Bull. Bingham Oceanog. Coll., vol. 1, art. 1, 1927, p. 32, fig. 

 15). Prominently enlarged scales are not mentioned in the description of that 

 species, nor shown in the illustration. Furthermore, all the scales are not cycloid 

 in the Tortugas specimens, those in the lateral line and those on the back behind 

 the dorsal fin being distinctly ctenoid. In the 2 specimens of intermediate size the 

 scales in advance of the dorsal also are ctenoid, though not in the other 2, and all 

 except the smallest specimen have very strongly ctenoid scales on the chest and 

 on the dentary bones. In the smallest specimen scales are missing on these parts, 

 probably having been lost. 



The smallest Tortugas specimen agrees with the Cuban ones in having the 

 chest compressed to a more or less sharp edge, whereas in the larger specimens 

 from Tortugas the chest is round and covered with scales. 



It seems evident, then, that more specimens must be studied before the species 

 of this genus can be accurately defined. 



The following proportions and enumerations are based on the 4 specimens 

 from Tortugas: Head 3.1 to 3.3; depth 1.9 to 2.2; anal base 2.0 to 2.15. Eye in 

 head 2.0 to 2.2; snout 4.6 to 5.0; interorbital 3.5 to 3.9; maxillary 1.75 to 1.9; pec- 

 toral 1.1 to 1.2; ventral 2.25 to 2.5. D. Ill or IV,9 to 11; A. 111,3 1 to 345 P- l 5> 

 scales in lateral line to base of caudal 56 to 62, about 10 or 12 more on caudal fin; 

 gill rakers 17 to 19. 



The color of the preserved specimens is uniform brownish, with a darker 

 streak along the base of the anal. 



West Indies to Florida, and presumably to Brazil. Not reported previously 

 from United States waters. S. F. H. 



