PACIFIC SALMON 



77 



not be a complete barrier to the west drift 

 on the Pacific Ocean side of the Peninsula. 

 However, it is a barrier sufficient to divert 

 a large part of this water to the south and 

 to allow a drift from the Ikatan and Morz- 

 hovoi Bays area. Thus, east of the Shu- 

 magin Islands the fresh-salt water gradient 

 would be from the east, of necessity, since 

 the current or drift is from this region. 

 The Shumagin Islands region is the parting 

 of the ways of the west and east fresh-salt 

 water gradients and is also the parting of 

 the ways of the red salmon's migratory 

 movements. No doubt red salmon pass 

 through Unimak Pass and even westward 

 on their way to Bristol Bay. Tagging ob- 

 servations and the westward drift of the 

 ocean current would lead one to conclude 

 that the red salmon which pass through 

 Unimak Pass and west do not come from 

 the east but from the west feeding grounds, 

 and would lose no time in reaching spawn- 

 ing ground areas. That is, at no place 

 would they be delayed as in Morzhovoi and 

 Ikatan Ba3-s and form a good fishing 

 ground. 



What is the gradient to which the red 

 salmon is reacting? Let us now consider 

 the observations made by Gilbert and Rich 

 (1927b) on Karluk River, Kodiak Island, 

 one of the two best red salmon streams in 

 Alaska. The other is the Chignik River 

 which is west and on the Pacific Ocean side 

 of the Peninsula. Even though the red 

 salmon may respond to a fresh-salt sea 

 water gradient out in the open ocean, this 

 cannot be true after it has entered a river 

 proper. Still the red salmon is quite selec- 

 tive of the streams flowing into Karluk 

 Lake. The best spawning streams are : 

 Salmon Creek "a tributary of some size 

 comes brawling in ... is a favorite spawn- 

 ing ground for red salmon. Far more 

 spawn in this stream than in Thumb Lake 

 or all of the other tributaries ' ' ; Fall Creek 

 ". . . the majority went into Fall Creek 

 . . . the stream grows boisterous"; ''Can- 

 yon Creek of real value ... it enters hilly 

 country, has a swift current and rocky floor 

 . . . " ; and ' ' two short creeks, Grassy Point 

 Creek and Halfway Creek . . . are brawl- 



ing streams, with coarse boulder beds and 

 only occasional very limited patches of 

 gravel . . . but in spite of their small size 

 and unfavorable character they carry more 

 red salmon in proportion to their size than 

 do the more suitable spawning streams of 

 the Thumb and head of lake." What are 

 the common characteristics of the better 

 spawning streams? The first and perhaps 

 most important characteristic is that all are 

 boisterous streams. This assures a carbon 

 dioxide tension of the water approximating 

 the carbon dioxide partial pressure of the 

 air. From these observations one is led to 

 suspect that the red salmon, at least after 

 having entered a stream, are responding to 

 the carbon dioxide tension of the water. 

 Further evidence will be presented that 

 will show that this supposition is most 

 probably correct. 



Other observations on the inshore tagging 

 of red salmon dominated by gradients of 

 fresh water from streams near at hand 

 (Rich and Suomela, 1927) show that red 

 salmon as well as other species scatter, i.e., 

 move to different spawning grounds from 

 one and the same tagging station. This 

 same tendency is shown to be true (David- 

 son and Christley, 1938) in the case of the 

 pink salmon (0. gorhuscha). The most 

 interesting fact is that none of the salmon 

 tagged in close proximity (marked gradi- 

 ents to streams) ever strayed to streams 

 beyond this proximity — outside and beyond 

 their gradients. 



When the tagging experiments carried 

 on in Canadian waters (Pritchard, 1930; 

 Clemens, 1930 ; Williamson, 1927 ; William- 

 son and Clemens, 1929 and 1932 ; Clemens, 

 1932; Pritchard, 1932, 1934a and 1934b) 

 are examined one is struck by the pre- 

 dominant movements toward the south 

 of all species of salmon tagged. When 

 the tagged salmon moved north in the 

 waters east of Vancouver Island, or even 

 on the west side for that matter, and 

 in the water at and near to Queen Char- 

 lotte Islands, it would be interesting to 

 know just what were the effects of the tides 

 and other factors in changing gradients of 

 fresh-salt water mixtures. We find the fol- 



