Time and Magnitude of Mortality 



The amount of dissolved oxygen required by 

 salmon eggs and larvae for normal metabolism 

 approaches a maximum just before hatching (see 

 fig. 2). After hatching, levels of dissolved oxygen 

 limiting metabolism are greatly reduced, and 

 oxygen requirements are least likely to be satisfied 

 before hatching. 



Pink and chum salmon eggs begin to hatch in 

 the study streams in November, and most hatch 

 before mid-December. Figure 8 shows the per- 

 centages of live eggs and larvae collected from 

 intertidal Harris River and Indian Creek that had 

 hatched before the date of sampling. Twelvemile 

 Creek is thought to lag 1 or 2 weeks behind 

 Harris River and Indian Creek with regard to 

 time of hatching, because the peak of spawning- 

 occurs about 1 week later. 



After the spawning period, it is convenient to 

 consider two periods during which mortality 

 occurs — prehatching (autumn) and posthatching 

 (winter). The dissolved oxygen supply and re- 

 lated factors are thought to exert their greatest 

 influence on mortality before hatching, so the 



Id 

 o 



01 

 UJ 

 Q_ 



Q 



UJ 



I 



o 



1- 

 < 



I 



c/> 

 o 

 o 



UJ 



UJ 



> 



NOV. I DEC. I JAN. I 



DATE OF COLLECTION 



Figure 8. — Percentage of live eggs and larvae in collec- 

 tions from intertidal Harris River and Indian Creek 

 hatching before the date of sampling. (Brood years are 

 indicated by numerals.) 



discussions here will be limited primarily to 

 mortality of eggs. 



In this study, estimates of three population 

 parameters thought to provide evidence of the 

 effect of oxygen supply and related factors on egg 

 mortality were used: (1) Total mortality fraction 

 (M,), (2) mortality fraction from ratios of dead 

 eggs in samples (M T ), and (3) fraction of points 

 containing fewer than 35 dead eggs per m. 2 (p )- 



An estimate of M, includes mortality from all 

 causes, and this estimate includes eggs removed 

 from the spawning bed. Such removal can ob- 

 scure mortality from causes where there is no 

 direct removal of eggs; therefore, the estimated 



total mortality (M ,) was often of limited value 

 in measuring egg mortality from oxygen supply 

 and related factors. Furthermore, estimates of 

 M, were highly inefficient, and the resulting con- 

 fidence limits were often broad. 



The other two estimates (M, and p n ) also had 

 limitations imposed by the disappearance of eggs 

 from the spawning bed. Disappearance from 

 gravel shift alone would have little effect on 

 validity of estimates of ratios of dead to total eggs 



present in the streambed (M r ), provided live and 

 dead eggs disappear in numbers proportional to 

 their abundance. On the other hand, losses from 

 decomposition and scavenging affect only dead 

 and dying eggs and would cause mortality to be 



A 



underestimated by M r . 



Use of the number of points where eggs were 

 present or absent to index occurrence of mortality 

 also would be affected by disappearance because 

 evidence of mortality from factors not related to 

 the direct removal of dead eggs would have been 

 destroyed. Major losses due to gravel shift would 

 tend to invalidate the use of p to detect significant 

 mortality possibly caused by oxygen supply and 

 related factors. 



With these possible limitations in mind, the 

 estimates of M,, M T , and p are used to evaluate 

 prehatching mortality of the 1956-60 brood years 

 in the three study streams. Each brood year will 

 be considered separately. 



1956 Brood Year 



Sampling was confined to spawning riffles A 

 (Harris River), B and C (Indian Creek), and D 

 and E (Twelvemile Creek). Workers sampled the 

 riffles in late November 1956 and in late February 



SPAWNING BED ENVIRONMENT OF PINK AND CHUM SALMON 



509 



