444 THE FUE SEALS OP THE PRIBILOF ISLANDS. 



devise a scheme which shall present at the same time their relations to the parent 

 form and to the limitless number of other colonies. We have them along the entire 

 range of typical marine Q. cataplir actus, where they seem to present an accompanying 

 series of variant fresh-water groups, which have been in the main independently 

 derived from the marine stock. The amount of divergence which they exhibit from 

 typical cataphractus is not geographically progressive. Thus we have in the present 

 collection, from a lake on Medni Island, specimens in which all of the lateral plates are 

 invariably present, though narrow and perfectly smooth, the modification being evident 

 in the reduction of the spines, the pubic plate, and the pectoral fins. On the neigh- 

 boring island, the fresh-water form develops but 8 plates or less. (See Bean, Proc. 

 TJ. S. Nat. Mus., 1896, 250.) In San Prancisquito Creek at Palo Alto there is an 

 average of 6 plates; in San Gregorio Creek, but a few miles away, there is an average 

 of 20 plates, while other neighboring streams have fully plated specimens only. It 

 seems evident, therefore, that the partially plated form ("6. microcephalus'^) does not 

 possess the characteristics of a subspecies, its divergence from the parent form not being 

 geographically progressive. Still less can we recognize it as a species, the fully plated 

 groups into which it passes being indistinguishable from G. cataphractus. Further 

 evidence of complete intergradation of fully plated, partially plated, and wholly 

 naked forms is found in a carefully prepared table by C. Eutter (Proc. Cal. Acad. 

 Sci., 1896, 248). The naked forms are confined to a few streams in southern Cali- 

 fornia, and may be designated Gasterosteus cataphractus tcilliamsoni. A larger amount 

 of material and more detailed investigation may ultimately demonstrate the possibility 

 of recognizing among the fresh-water groups of sticklebacks subspecies coextensive 

 in range with the fresh- water faunal areas in which they occur, but this does not seem 

 probable. An apparently similar condition is presented by G. Mspinosus of the North 

 Atlantic coast of America and G. aculeatus of northern Europe. The characters alleged 

 to separate these species from each other and from G. cataphractus can not be con- 

 sidered satisfactory. It seems probable that all will be ultimately united under one 

 specific name. 



Specimens in the present collection are from Tareinsky, Kamchatka (collected by 

 Mr. Barrett-Hamilton); Mkolski, Bering Island; the Lagoon, St. Paul Island; Sum- 

 mer Harbor, Unalaska; Freshwater Lake, Medni Island. 



Among all the collections we note here as elsewhere the great preponderance of 

 females over males. The males can be distinguished at sight by the larger head and 

 longer pectorals. As already noted, the marine specimens exhibit little variation. 

 We note, however, that those from the Kamchatka coast, Bering and St. Paul islands, 

 as compared with those from southeastern Alaska and Puget Sound, exhibit slightly 

 longer heads, longer pectoral fins, deeper sculpturing on the head, and much rougher 

 spination of the plates, which are also deeper. In this they agree with specimens 

 which we have examined from Alaskan coast near Bering Straits. Unalaska speci- 

 mens agree, however, with the southern form. 

 52. Pygosteus pungitius (LinnaMis). 



Petropaulski Harbor, fresh-water lake near Tareinsky Bay. In the specimens 

 from Petropaulski the ventral spines average shorter than in those from the lake, 

 being contained 2| to 3 times in the head in the former, 2J to 2| in the latter. The 

 pubic bone varies greatly in length and in relative width. It is evident that neither 



