4G0 



THE FUR SEALS OF THE PKIBILOF ISLANDS. 



dorsal. In these dark niarkiugs can be made out small roundish or polygonal spots of 

 of black, separated by reticulating lighter lines. Under side of head and body whitish, 

 the lower lip dusky; breast minutely black speckled; dorsal mottled light and dark, 

 without definite color pattern; anal whitish, with oblique series alternately of small, 

 roundish spots, and of still smaller dots; caudal with a dark crossbar at base, and two 

 or three more or less irregular ones on outer half; pectorals with a large black area at 

 base, the i)osterior i^ortion with elongate sjwts forming ill-defined cross series, the 

 lighter area with smaller dark markings of varying shapes; ventrals whitish, with 

 two or three faint dusky crossbars. 



Specimen 117 mm. agreeing well with types. The interorbital is narrow and very 

 deep. There is no cross ridge on occiput; the occipital ridges are very high, com- 

 pressed, knife-like, with serrulate edge. There are three strong hooks on one preoi)ercu 

 lar spine, two on the other. Each plate of the lateral line has a central, backwardly 

 directed spine. D., VII-14; A., 12; P., 18; iilates 35. Lines of i^lates converging at 

 interspace between dorsal, then again at end of second dorsal. 



91. Gymnocanthiis pistilliger (Pallas). (Plate LVIII.) 



Petropaulski ; station 3()4G, ofi' Kobben Island, 18 fathoms. Recorded by Dr. 

 Gilbert from many specimens from Bristol Bay, and by Dr. I>ean from Kyska, Point 

 Belcher, and Cape Tcluiplin, Siberia. Mr. Scofield found it at Port Clarence, and 

 Dr. Stejneger on Bering Island and at Petropaulski. 



No males are included in the number taken. The females differ from those we 

 have examined from Bristol Bay in having the top of head more extensively plated, 

 the rough plates extending onto middle of interorbital si>ace, or in one specimen onto 

 snout. Tlie preorbital ridges are less regular and have lower tubercles. The speci- 

 mens indicate an ajiproach therefore in this respect to G. f/aleatus. The fin rays are 

 as previously given. In eight specimens they are as follows: 



Numljer of ray 



Number of specimens. 



Dorsal 

 spines. 



IX X 



1 7 



Dorsal rays. 



14 15 16 

 16 1 



Anal rays. 



16 17 

 3 5 



92. Gymnocanthus galeatus Hean (Plato LIX). 



Seined in Cai)tains Harbor, recorded by Dr. Bean from the same waters, and by 

 Dr. (iilb.rt from Clieniofsky, all tliese localities being about the island of Unalaska. 

 Dr. Bean records it from Unalaska and from Cape Sabine iu the Arctic, and Mr. Sco- 

 field from Point Barrow. 



Many very young specimens, about 35 mm. long, were also collecrtedby Mr. Wil- 

 liam I'almer on St. Paul Island, in these, the preopercular spine is simply furcate 

 at tip, without trace of the upwardly directed processes characteristic of the adult. 

 The roughened plates on head are also undeveloped. Fin rays are: D., XI, 10; A., 19; 

 P., 20 or 21. 



93. Argyrocottus zanderi llerzenstein. (Plato LX.) 



Three specimens of this beautifully marked cottoid were taken in Shana Bay, 

 Iturup Island. Originally described from Sakhalin Island. 



The branchiostegal membranes are widely joined across the throat, narrowly 



