The 90-percent confidence limits of k (k and k ) 

 are obtained from the expression 



(ito, k B )=kp ±l.U5[kp (l-po)] m 



(9) 



where k is the number of points sampled and p 

 is the fraction of points estimated to contain 

 fewer than 35 eggs and larvae per m. 2 



OBSERVATIONS ON ENVIRONMENT AND 

 MORTALITY 



Although numerous workers have postulated 

 factors causing mortality of eggs and larvae, few 

 have presented quantitative estimates of mortality 

 satisfying three essential criteria: (1) Estimates 

 free of bias, (2) estimates representative of natural 

 populations, and (3) estimates related directly to 

 causative factors. It is not surprising that these 

 criteria have not been met entirely in field studies, 

 for there are many difficult problems requiring 

 solution. In the present study, an effort was made 

 to reduce (or at least recognize) bias in mortality 

 estimates. Furthermore, because of the randomi- 

 zation techniques used, the samples were thought 

 to be representative of the populations studied. 

 However, the difficulties in associating observed 

 mortality levels with their causative factors are 

 formidable even with the first two criteria being 

 satisfied in part. The problem of relating observed 

 mortality levels to causative factors is complicated 

 in most instances because of interactions among 

 environmental factors. 



I attempted to account for interactions by classi- 

 fying environmental factors causing mortality into 

 generally inclusive groupings: (1) Oxygen supply 

 and related factors, (2) stability of the spawning 

 bed, and (3) freezing of intragravel water. I did 

 not consider one inclusive grouping — pathogenic 

 agents. 



OXYGEN SUPPLY AND MORTALITY 



Environmental requirements of salmon eggs and 

 larvae were briefly reviewed in an earlier section. 

 My purpose here will be (1) to describe the physical 

 characteristics of spawning beds where observa- 

 tions on mortality were made, and (2) to describe 

 the relation between environmental quality (as 

 related to oxygen supply) and the observed 

 mortality levels. 



Dissolved Oxygen Content of Intragravel Water 



The dissolved oxygen content of intragravel 

 water was consistently lower at the beginning than 

 at the end of the spawning period in all study 

 streams. There also were observed spatial differ- 

 ences in mean dissolved oxygen levels among and 

 within the study streams. 



In late August, at the beginning of spawning, 

 oxygen levels appeared to be lowest in the intertidal 

 Harris River and the upstream Twelvemile Creek 

 spawning areas and highest in the upstream Harris 

 River and the intertidal Indian Creek and Twelve- 

 mile Creek spawning areas. Near the end of 

 spawning, in late September, differences in mean 

 oxygen levels were no longer significant among 

 the spawning areas sampled. The data are sum- 

 marized in table 3, and mean values obtained in 

 1959 are shown in figure 5 to illustrate the kind of 

 relation observed. 



Differences among years in mean dissolved 

 oxygen levels were considerably greater than 

 differences among streams. Summer 1957 was 

 of particular interest in this regard because 

 unusually low levels of dissolved oxygen were 

 observed. Average values of all dissolved oxygen 

 determinations made in the study streams near 

 the beginning of spawning over the period 1956 



<s> 



o 

 o 



liJ 



>- 

 X 



o 



Q 



o 

 to 



CO 



o 



z 

 < 



LEGEND 



O INTERTIDAL HARRIS RIVER 

 • UPSTREAM HARRIS RIVER 

 a INTERTIDAL INDIAN CREEK 

 o INTERTIDAL TWELVEMILE CREEK 

  UPSTREAM TWELVEMILE CREEK 



SEPT I 



SEPT. 20 

 1959 



Figure 5. — Mean dissolved oxygen content of intragravel 

 water in the study streams over the period of spawning 

 in 1959. 



.IOC 



U.S. FISH AND WILDLIFE SERVICE 



