466 



PROCEEDINGS OF THE NATIONAL MUSEUM. 



VOL. 37. 



ray 2.1 in head; longest anal ray 2.3 in head; second anal spine 3.3 

 in head; caudal convex; ventrals inserted in advance of upper axil of 

 pectorals their length 1.7 in head, nearly reaching vent; pectoral long, 

 subsymmetrical, its lower rays reaching to middle of anal. 



Color bright red in life. In spirits yellowish white, with about 

 three very indistinct whitish longitudinal streaks on side; fins pale, 

 unmarked. 



Known from the type, a specimen 6.25 inches long, taken at 

 Yamagawa, near Kagoshima, above described. A second specimen, 

 apparently of the same species, from the same locality has been named 

 Sayonara mitsukurii. The differences seem to us to be due in part 

 to the small size of the specimen t)f Smith and Pope and in part to 

 errors in the figure of the typical species. 



The nominal species, Sayonara mitsukurii, is said l)y the authors 

 to differ from Sayonara satsunise, as figured and described, in its 



Fig. 15. — Sayonara satsum.k (type of sayonaka mit.sukukii i. 



larger eye, longer tubules in lateral line, long and slender gill-rakers, 

 higher dorsal fins, unbranched pectoral rays and color. We note 

 that the youth and small size (2 1 inches) of Smith and Pope's speci- 

 men may explain its slightly larger eye, higher dorsal fins, unbranched 

 pectoral rays, and traces of dark cross l)ands. The type of Sayonara 

 satsumx has some of the pectoral rays not at all branched and the 

 others branched only very far out toward their ends. The drawing 

 of Sayonara satsumse incorrectly represents the length of the tubules 

 of the scales in the posterior half of the lateral line ; shows the oper- 

 cular spines rather too much concealed ; and represents somewhat too 

 small the scales on suborbitals and maxillaries. The statement in the 

 description of S. satsumse that the gill-rakers are "short" is hardly 

 accurate, the longest being nearly equal to the corresponding fila- 

 ments. Only a careful comparison of the types of the two species 

 will show whether they are really distinct, but the probabilities are 

 in favor of their identity. 



