nificant. The data indicate that mortality was 

 high before hatching and low after hatching. 



Most dead eggs collected in November already 

 had decayed considerably and were classified as 

 fragments, suggesting that mortality occurred 

 shortly after spawning. This finding would not 

 be unexpected because dissolved oxygen levels 

 were unusually low during spawning. Further- 

 more, there was evidence of high biochemical 

 oxygen demand continuing well into autumn 1957. 

 Although oxygen levels had increased significantly 

 between August and November 1957, they were 

 still lower in November 1957 than in August 1958, 

 even though water temperatures were about 

 6° C. cooler in November 1957 than in August 

 1958 (McNeil, 1962b). 



The high percentage of fragments among dead 

 specimens collected from intertidal Indian and 

 Twelvemile Creeks in November 1957 remained 

 virtually unchanged through March 1958. The 

 values were: 



Indian Creek 



Riffle B: 92 percent in November and 96 



percent in March. 

 Riffle C: 70 percent in November and 70 



percent in March. 



Twelvemile Creek 



Riffle E: 66 percent in November and 76 

 percent in March. 



Riffle F: 97 percent in November and 95 

 percent in March. 

 This was nearly conclusive evidence that estimates 

 of M T obtained from egg fragments alone would 

 be little changed over late autumn and winter, 

 and estimates of M r based on presence of egg 

 fragments and made in early spring 1958 would 

 give essentially the same result as estimates made 

 the previous autumn. 



Intertidal Harris River was sampled about 

 April 1, 1958, when egg fragments or live larvae 

 were collected at 34 points. Only egg fragments 

 were found at 31 of the 34 points, giving minimal 

 estimates of 



M,>M r =31/34=0.91 



The available evidence suggested that condi- 

 tions in 1957 were unfavorable for egg survival, 

 and exceptionally high egg mortality occurred in 

 all study streams. This high mortality was 



associated with unusually low levels of dissolved 

 oxygen in intragravel water and low density of 

 spawners. Unexpectedly low levels of dissolved 

 oxygen prevailed into November 1957, suggesting 

 that the biochemical oxygen demand was unusually 

 high. There is a possibility that the density of 

 females spawning (five or less per 100 m. 2 ) was 

 too low to remove accumulated organic detritus 

 from spawning beds in quantities sufficient to 

 reduce materially the biochemical oxygen demand 

 (Ricker, 1962; McNeil and Ahnell, 1964). 



1958 Brood Year 



Adult escapements were moderately low in 

 1958, though considerably higher than in 1957. 

 Spawning densities ranged from 13 females per 

 100 m. 2 in intertidal Harris River and Indian 

 Creek to 1 and 2 females per 100 m. 2 in upstream 

 Twelvemile Creek and Harris River, respectively. 

 Density of spawning females in intertidal Twelve- 

 mile Creek was 11 per 100 m. 2 



Hydrological conditions during spawning fa- 

 vored a higher egg survival than in 1957. Indian 

 Creek maintained discharges of 10 c.f.s. or more 

 over the spawning period as opposed to a minimum 

 discharge of 4 c.f.s. during spawning in 1957, 

 when high early egg mortality occurred. 



Observations on mortality were made in inter- 

 tidal Harris River, Indian Creek, and Twelvemile 

 Creek. Spawning densities in upstream Harris 

 River and Twelvemile Creek were too light to 

 warrant investigations in these areas. Table 6 

 summarizes the results of the mortality estimates. 



For purposes of the present discussions, inter- 

 tidal Twelvemile Creek can be dismissed because 



A 



of extreme high early mortality (M ( =0.97 by 

 late November) apparently caused by scouring 

 during floods, which physically removed eggs 

 from the spawning bed. 6 As a consequence, 

 there were insufficient numbers of eggs remaining 

 to relate observations on mortality to observa- 

 tions on factors affecting oxygen supply. Workers 

 resampled the area in December primarily to 

 check the results of the November sampling. 

 The numbers of eggs collected approached 

 expectation in intertidal Harris River and Indian 

 Creek. Indian Creek was sampled on one 



s Potential egg deposition in intertidal Twelvemile Creek during September 

 1958 was estimated to be 175 per m. 2 By late November, mean density of 

 live plus dead eggs was estimated to be only four per m. 3 ; by late December 

 only one per m. 2 



SPAWNING BED ENVIRONMENT OF PINK AND CHUM SALMON 



511 



