LV.—The Perciform Genera Gymnapogon and Nannatherina

Abstract

YMN.POGON japonicus, a new genus and species of fishes, was described by me in 1905 (Regan, Ann. Mag. Nat. Hist., (7) 15: 19) from a number of specimens measuring up to 45 mm. in total length, from the Inland Sea of Japan. I considered it to belong to the Chilodipteridae (=Apogonidae), and using Jordan and Snyder's review of the Japanese fishes of this family (Proc. U. S. Nat. Mus., 22, 1901: 891-913), I put it near Telescopias, which it appeared to resemble in having the spines feeble, canines anterior in the upper jaw, and lateral in the lower, preoperculum produced into a thin flap, etc. In 1913 (Classification of the Percoid Fishes, Ann. Mag. Nat. Hist., (8) 12: 111-145) Telescopias, as a synonym of Scombrops, was placed in the Pomatomidae, and Gymnapogon was mentioned as related to it. I have for some time thought it likely that this was a mistake, but it was not until I had read Herre's description of Henicichthys philippinus (CoPEIA (4), 1939: 200), evidently a Gymnapogon, that I decided to reexamine the types, and to study the skeleton. This I have done by examining an alizarin preparation and by dissecting another specimen. Henicichthys foraminosus Tanaka, 1915 (Tomiyama, Japanese J. Zool., 7 (1), 1936: 50, fig. 9) is a synonym of Gymnapogon japonicus. Tomiyama's figure is good, except that the pectoral fin is too high, and the caudal rays are too divergent. He considers this fish to be a gobioid, but it differs from all known members of this group in having the gill membranes separate and two anal spines, and I find that the skeleton is typically perciform. In most respects Gymnapogon japonicus resembles the Chilodipteridae, especially Chilodipterus, but it differs from them in the following characters: No scales; sides of head with a network of series of small pores; 20 or more vertical series of similar pores on each side of body; lateral line represented by a series ending below soft dorsal, and another on middle of side of tail. Fin spines very slender. Pectoral higher, the upper end of its base about equidistant from dorsal and ventral profiles. Operculum connected by membrane to pectoral arch, the gill opening ending a little above the pectoral. No supramaxillary (or supplemental maxillary). There appear to be no other differences from the Chilodipteridae. The dentition is similar to that of Dinolestes and Synagrops. The preoperculum, produced backwards below as a large thin flap, is also much as in these genera. The dorsal fin has VI, I 10-11 rays, the anal II 9-10. The caudal has 17 principal rays, 15 branched, and a series of about 10 procurrent simple rays above and below, a structure approached in Synagrops. The pelvics, with I, 5 rays, are close together, a little in advance of the pectorals, which have 13 or 14 rays. The pectoral fin is supported by four hour-glass shaped radials. The skull is essentially similar to that of Dinolestes, as described by Starks (Proc. U. S. Nat. Mus., 12, 1899: 113-120, pls. VIII-XI), but with parietal crests scarcely developed, and with a short ethmoid region, ending