TCHTHYOLOGICAL COLLECTIONS OF THE ALBATROSS. 453 



the eyes. Otherwise no pits or projection on head. A shallow triangular depression 

 on occiput. Gill-slit much loss obliciue than margin of preopercle, its length 1^ 

 times the distance between lower ends of gill-slits, the latter reaching the vertical 

 from middle of opercles. 



Dorsal fin of rather tiexible spines, not concealed in heavy fin membranes. The 

 origin of dorsal falls immediately behind axil of pectorals. Hinder margin of occi- 

 put midway between front of dorsal and middle of eye. Origin of anal well in 

 advance of middle of length, its distance from tip of suoat contained 1} times in its 

 distance from base of caudal. Pectoral short, rounded, its base separated by a wide 

 prepectoral area from gill-slit, the width of area three-fourths length of fin, the 

 latter equaling distance from tip of snout to middle of eye. No ventrals. Body 

 covered with lax naked skin, which also covers but does not obscure rays of dorsal 

 arid anal fins. No pores to lateral line. 



Color iu life : Reddish on head, body, and fins, due to the blood vessels in the skin. 



A single specimen, 180 mm. long, dredged north of Unalaska Island, at station 3312; 

 depth 45 fathoms. 



Family PTILICHTHYIDiE. 



124. Ptilichthys goodei Bean. 



A third specimen of this little-known fish ^as taken by dredging in shallow water 

 at the entrance to Unalaska Harbor, station 3311; depth, 85 fathoms. 



The genus Ptilichthys, of which this species is the sole representative, has been 

 doubtfully referred by Dr. Bean to the Mastacemhelidce, a family of fresh-water 

 fishes inhabiting the East Indies, characterized by having the shoulder girdle poste- 

 riorly placed and not articulating with the cranium (order Opisthomi Gill). The 

 necessity for preserving intact the unique type of the species prevented Dr. Bean 

 from making any anatomical examination of Ptilichthijs, and it was reserved for Dr. 

 Theodore Gill, in the Standard Natural History, 1885, p. 259, to express his disbelief 

 in the relationships which had been suggested and to make the fish the type of a 

 peculiar family, the Piilichthyidce, to be placed provisionally among the blennioid 

 series. His adherence to this view is again expressed in his list of Families and 

 Subfamilies of Fishes," appearing as the sixth memoir of volume vi, of the National 

 Academy of Sciences. He has doubtless indicated the jiroper position of this peculiar 

 fish as nearly as we are now able to determine it. An examination of its shoulder 

 girdle shows it to be entirely normal. The post-temporal is not furcate, but is a 

 very slender bony rod attaching to the epiotic region of the skull and giving loose 

 attachment posteriorly to the almost equally slender posterotemporal. The latter 

 overlaps the upper end of the clavicle in the usual manner. A postclavicle was not 

 detected. The coracoid i>ortion consists of a roundish oblong, perforated hypercora- 

 coid meeting the hypocoracoid directly, without intervening cartilage. The curved 

 line separating the two bones corresponds distally with the interspace between the 

 first (upper) and second actinosts. The hypocoracoid is broad and short. Its 

 mesially directed (i. e., inferior) process joins at its tip the clavicle, but is elsewhere 

 separated from the latter by the usual elongate membranous interspace. The acti- 

 nosts are four in number, of large size, hour-glass shaped. 



The jaws are normal, the premaxillary alone occupying the front and sides of 

 upper jaw and bearing the teeth, while the maxillary is a broad bone lying behind 

 it, overlapped proximally by the maxillary process of the palatines. Both vomer 

 and palatines seem to be toothless. The alimentary canal is almost perfectly 

 straight, with the anterior portion entirely enveloped in the long, narrow liver. At 

 the pylorus occurs a short and abrupt V-shaped flexure, scarcely noticeable on 

 account of the closeness with which the sides are joined and the fact that the width 

 of the flexure is no greater than the cross diameter of the tube. Pyloric cceca are not 

 evident. An air bladder is entirely wanting. The ovary is single, apparently without 

 (Oviduct, and contains in our specimen eggs which are comparatively very large. 



