I04 PAPERS FROM TORTUGAS LABORATORY vol. xxxiv 



margin to end of dorsal base; ocular stripe most prominent, extending from 

 snout to caudal base, and there followed by a conspicuous spot. 



These fish may rest on the bottom, perched on their ventral fins. So resting, 

 they may display bands, which may appear less distinctly in those resting out of 

 contact with the bottom. 



Individuals may have hiding places they hold as their own. Some they dig by 

 lying down beside stones, vibrating their bodies rapidly, and throwing sand out 

 in clouds; some they find ready made. Once they possess a shelter, day after day 

 they may be seen beside it, darting in when approached by such fishes as the 

 swift carangids. 



Refuse from the Bird Key rookery examined included i specimen of no mm. 

 They themselves feed largely on crabs and shrimps. 



The species does not breed at Tortugas during the summer, as even in late 

 August the gonads are still small, but all of one type, having the appearance of 

 ovaries. In view of the normal hermaphroditism of Diplectrum bivittatum , this 

 is sufficient indication that the same condition prevails here. I may note that 

 D. formosum (Mus. Comp. Zool. no. 21767), 175 mm. long, has on the ventral 

 side of its gonad, between the ovarian lobes and behind the point of bifurcation, 

 just such a specialized part as yields spermatozoa in related species. 



Atlantic coast of tropical America, sometimes ranging northward on the south 

 Atlantic coast of the United States. W. H. L. 



Diplectrum bivittatum (Cuvier and Valenciennes) 



Serranus bivittatus Cuvier and Valenciennes, Hist. nat. poiss., vol. 2, 1828, p. 241 — Mar- 

 tinique. 



Diplectrum radiale Jordan and Evermann (part not of Quoy and Gaimard), Bull. U. S. 

 Nat. Mus., No. 47, pt. 1, 1896, p. 1204. Meek and Hildebrand, Field Mus. Nat. Hist., 

 Zool. Ser., vol. 15, pt. 2, 1925, p. 475, pi. 46, fig. 1. 



This species has been regarded recently as the same as Diplectrum radiale 

 (Quoy and Gaimard). Since it is so considered by Meek and Hildebrand, and is 

 manifestly different from the Pacific form, which has long been considered the 

 same as Quoy and Gaimard's Brazilian species, the authors mentioned have ele- 

 vated the west coast form to specific rank. It may be distinct, but if so the fact 

 must be established by comparison with Brazilian, not with West Indian material. 



Twenty-five specimens up to 150 mm. in length were taken between 40 and 55 

 fathoms. In one, 155 mm. in length (standard length 115 mm.), the proportionate 

 measurements are: Depth 3.8; head 2.9. Eye equal to snout, 4.0 in head; inter- 

 orbital width 6.7. Scales 6-52 to 54-17, about 8 in series between eye and angle of 

 preopercle, instead of 10 or n as in radiale. 



This species may be readily distinguished from formosum by the form of the 

 dorsal fin, the margin of which is very distinctly notched between the spinous 

 and soft parts as it is not in formosum; also the 3d dorsal spine in radiale is twice 

 as long as the 2d, whereas in bivittatum it is little longer than the 2d. The scales 

 on the cheek are larger and in only 7 oblique rows in radiale, in strong contrast 

 with the much smaller ones of formosum, which are in 1 1 or 12 rows. 



The usual color pattern consists of rather faint dark stripes on a light ground, 



