40 
BULLETIN OF THE BUREAU OF FISHERIES. 
fish in each of these districts, for although local spawning grounds for red salmon 
occur in the vicinity of these fisheries, it has been generally believed that they are 
wholly inadequate to account for the very extensive runs that occur. As regards 
Ikatan Bay there can be no question that the salmon are on their passage, circling 
the shores of the bay on their journey elsewhere, because the red salmon rims are 
at times of great magnitude in this bay, although no spawning grounds of any size 
occur there. In the case of Morzhovoi Bay there has seemed to be more ground 
for' divergence of opinion, for fresh-water lakes of large size, capable of producing 
very considerable runs of red salmon, are tributary to the head of the bay. On a 
preliminary examination of this district in 1918 it was considered not improbable 
that these lakes, together with the spawning grounds tributary to Thin Point, a 
short distance to the eastward, were the source of the Ikatan and the Morzhovoi 
runs. The investigations of 1922 failed to lend support to this hypothesis. 
The Unga Island fishery is likewise concerned with a red salmon run which 
may at times reach large proportions. Local spawning grounds for red salmon 
exist, like those of Red Cove, Acheredin Bay, and a number of smaller streams, but 
these seem obviously inadequate to account for the run. 
Another district in which the problem has arisen regarding the origin and 
destination of the run lies on the Bering Sea side of the Peninsula, immediately 
to the eastward of Port Moller. Here along a favorable stretch of coast, traps and 
purse seines have in certain years reaped an exceedingly rich harvest, whereas in 
other years the returns have been scanty. Two red salmon streams of importance — 
Bear River and Sandy River — have their mouths along this stretch of beach. Diver- 
gent views have been entertained concerning the destination of the Port Moller red 
salmon run, it being held by one faction that the run consists largely or wholly of 
local fish bound for Bear and Sandy Rivers, and by the other that, in the more 
prosperous years at least, the majority of the salmon are migrants, on their way to 
Bristol Bay. In our preliminary inspection of this field in 1918 all available evi- 
dence seemed to point to the local origin and destination of the run. The tagging 
experiments of 1922 gave no results in conflict with that theory. It must be 
recalled, however, that the runs in 1918 and 1922 were both small. Whether more 
prosperous years on the Port Moller grounds are in part or wholly due to Bristol 
Bay schools, which on those years more closely skirt the coast, is a problem still 
awaiting solution. 
The tagging experiments of 1922 were planned to throw light on as many of 
these problems as possible. Consecutively numbered aluminum tags (fig.- 1) were 
attached to the tails of 4,000 salmon, which were then released, and the time and 
place of recapture were recorded. Of these, 861 were attached at Unga Island, 200 
in Morzhovoi Bay, 2,300 in Ikatan Bay, and 639 in the vicinity of Port Moller. 
Of the 4,000 salmon tagged, 709, or 18 per cent, were reported recaptured, either in 
the vicinity where tagged or at more distant points. A detailed record of all recap- 
tures is presented in the tables given at the end of this paper. We here call atten- 
tion to some of the more striking results. 
1. Shumagin Islands . — The fish tagged at Unga Island, of the Shumagin Group, 
were obtained, through the highly appreciated cooperation of the Pacific American 
