Fishes of the Western North Atlantic 517 



It seems that the fish generally tend to continue upstream until they meet some 

 obstacle that bars their further progress. In the larger streams or rivers this is likely to 

 be the first waterfall; in many cases, however, the governing factor lies in the length 

 of the stream in which they spawn. In western Greenland, for example, the narrow- 

 ness of the ice-free belt limits the maximum distance of their upstream migrations to 

 25—30 miles. For more extreme cases it seems sufficient to refer the reader to 

 Dunbar and Hildebrand's report of spawning fish taken in the lowest half mile of a 

 stream tributary to the estuary of the George River (Ungava Bay, northern Quebec), 

 beyond which further progress is barred by falls {16: 99); and to Weed's account 

 of alfinus spawning about five miles up a brook tributary to Anatalek Bay, near Nain, 

 Atlantic coast of northern Labrador (j2 : 131, 132). Turner, on the other hand, reported 

 the presence of alpinus more than 1 10 miles up the Koksoak River, tributary to Ungava 

 Bay, and farther still up a tributary of the Koksoak (in 16: 99, 100, ftn.) — if indeed 

 the fish in question were alpinus, not Atlantic salmon, which run up the Koksoak 

 regularly in considerable numbers (p. 476). 



The descent to salt water usually takes place at about the time the river ice breaks. 

 At Southhampton Island, northwestern Hudson Bay, for example, the descending fish 

 have congregated within the mouth of Brooks River by the last part of June, and 

 catches have been made in salt water nearby as early as July 8 in some years (e. g., 

 1934) but not until after July 15 in other years (^7: 128, 129). In Frobisher Bay, 

 Baffin Island, in 1951 (the only year of record), the river ice broke on June 10 and 

 Arctic Charr were caught outside the mouth of the river in salt water shortly thereafter 

 {28: 355). I find no definite information as to the date of the downstream migration 

 for Labrador. Local ice conditions suggest, however, that it may fall at about the same 

 time there as in Frobisher Bay, though in western Greenland the outlets of the fresh- 

 water lakes remain frozen until July, when there is a rush of alpinus to the sea (McMil- 

 lan in J2: i). Similarly, in northeastern Greenland, where the river ice persists late 

 into the season, it was not until the first week of July that Johansen's nets, set in the 

 outlet of a lake tributary to Denmark Harbor, took this species (40: 669). Arctic Charr 

 in Novaya Zemlya are described as moving out into salt water during June and the 

 first half of July (yg: 69, 70). 



It is not yet known whether every adult returns to fresh water to spawn every 

 year, whether every fish that has spawned in a given autumn invariably returns to sea 

 the following summer, or whether any of them winter in the sea with any regularity 

 in any given locality. 



The few accounts that have appeared concerning the distribution of alpinus while 

 they are in salt water, based on their numerical distribution there, are unanimous to 

 the effect that they remain for the most part close to the mouth of their native rivers, 

 or at least within the bays into which these open. Perhaps the most definite evidence 

 in this regard is Manning's observation that, while they are regularly caught at Igloolik 

 Island, three miles off the mainland of Southampton Island, nortwestern Hudson 

 Bay, none was caught "at Walrus Island, 30 miles off Southhampton Island, although 



