Table 5. — Estimated mortality of 1956 brood year pink and chum salmon egqs and larvae based on ratios of dead to total 



specimens collected 



i All samples were collected from intertidal riffles. 



1957. Since spawning density was not estimated 

 in 1956, no estimates of M t are given. Also, 

 because of small sample size, it is not possible to 

 give meaningful estimates of p . Thus, for the 



1956 brood year, only estimates of M T are given to 

 indicate time and magnitude of mortality (table 5). 



In autumn 1956, egg mortality appeared to be 

 highest after embryos had eyed. The February 

 sample included three general categories of dead 

 specimens: (1) Embryos dying early (as evidenced 

 by the opacity and advanced decomposition of the 

 eggs), (2) embryos dying just before hatching (as 

 evidenced by development of body structures) , and 

 (3) larvae dying after hatching. Table 5 gives 

 estimates of the fractions of total deaths occurring 

 early, shortly before hatching, and after hatching. 



Two features of these data stand out. First, 

 early egg mortality was lowest in intertidal Indian 



A A 



Creek (M r =0.10 versus M r =0.27 in intertidal 

 Harris River and Twelvemile Creek). Second, 

 this relation had reversed by late February. Other 

 evidence, which will be considered later, strongly 

 suggests that freezing was the major cause of 

 mortality of the 1956 brood shortly before and 

 after hatching. 



With regard to early egg mortality possibly 

 associated with oxygen supply and related factors, 

 evidence from 1956 brood year embryos does not 

 contradict the possibility that intertidal Indian 

 Creek provides a more suitable environment than 

 either intertidal Harris River or Twelvemile 

 Creek. 



1957 Brood Year 



Estimated mortality of the 1957 brood year 

 provided the most striking evidence obtained in 

 the course of these studies on relation between 

 oxygen supply and egg mortality. As shown in 

 figure 6, mean dissolved oxygen levels during the 



spawning period in 1957 were less than 50 percent 

 of other years. These low levels of dissolved 

 oxygen occurred during a prolonged period of 

 low precipitation and discharge. For example, 

 between August 10 and September 25, discharge 

 of Indian Creek exceeded 20 c.f.s. only 30 percent 

 of the time. Also, over this period very low 

 discharge (4 to 10 c.f.s.) prevailed for 2 weeks 

 during and after spawning. Furthermore, clear 

 skies prevailing over the latter half of September 

 were thought to have contributed to an unusually 

 prolific growth of periphyton observed on the 

 surface of streambeds at the time. 



The escapement of adults was the lowest ob- 

 served during these studies. The density of adult 

 female pink and chum salmon spawning in inter- 

 tidal areas of the study streams was five or less 

 per 100 m. 2 . The period of spawning lasted only 

 from about September 7-17, the briefest period 

 observed. 



There was good evidence that egg mortality 

 was high after spawning in 1957. Eggs were 

 collected from riffles B and C in intertidal Indian 

 Creek and riffles E and F in intertidal Twelvemile 

 Creek during early November 1957 and late March 

 1958 and from intertidal Harris River in early 

 April 1958. 



Because differences in estimated values of M r 

 were not significant among the study riffles 

 sampled, samples collected from intertidal Indian 

 and Twelvemile Creeks were pooled by date to 

 give the following single estimates of M, and their 

 90-percent confidence limits: 



A 



November M r =0.69±0.1!t 



A 



March M,= 0.57±0.24 

 Difference between the two estimates i> not sig- 



.-,10 



U.S. FISH AND WILDLIFE SERVICE 



