Two classes of dead eggs and larvae not present 

 in November samples were found in February 

 samples: dead eggs with well-developed embryos 

 ready to hatch and dead larvae. Death of eggs 

 containing the well-developed embryos occurred 

 after November sampling and before hatching in 

 December. 



This prehatching mortality occurred in con- 

 junction with the December 3 to 12 freezeup. 

 The observed posthatching mortality may have 

 been caused, for the most part, by the subsequent 

 more severe periods of freezing in January and 

 February. 



I have already summarized in table 5 the 

 estimated fractions of 1956 brood year eggs and 

 larvae dying early, shortly before hatching, and 

 after hatching. Assuming that mortality at and 

 after hatching was caused by freezing, I estimated 

 mortality of the 1956 brood year from freezing to 

 be about 23 percent in Harris River riffle A, 65 

 percent in Indian Creek riffles B and C, and 14 

 percent in Twelvemile Creek riffles D and E. 



I visited the study streams during the February 

 22-25 freeze. Except for an occasional exposed 

 riffle, they were coated with ice several inches 

 thick. Freezing appeared to have a greater effect 

 on Indian Creek spawning beds than on Harris 

 River and Twelvemile Creek spawning beds. I 

 attempted to drive metal rods into the Indian 

 Creek streambed at a number of locations. The 

 gravel was often frozen, particularly where exposed 

 by drying of the stream. Anchor ice also had 

 formed at several points examined. 



One important conclusion to be drawn from 

 these data is that mortality associated with 

 freezing was highest in the stream having the 

 lowest minimum discharge (4 c.f.s. in Indian 

 Creek as compared with 12 c.f.s. in Twelvemile 

 Creek and 22 c.f.s. in Harris River). 



1958 Brood Year 



Freezing occurred after hatching in January 

 1959. Sampling was undertaken in March 1959 

 (2}i months after freezing), and the possibility 

 that larvae killed by freezing had decayed before 

 sampling could not be ignored. Therefore, the 

 best evidence of winter mortality possibly as- 

 sociated with freezing may he estimates of M, 

 which are summarized in table 14. 



In intertidal Harris River, mortality was high 

 before freezing and there was no evidence of 



518 



Table 14. — Estimated total mortality of 1958 brood year pink 

 and chum salmon eggs and larvae before and after freezing 

 in winter 1958-59 



1 This limit was set with log-transformed data. 



additional mortality over winter. Intertidal In- 

 dian Creek, on the other hand, experienced high 

 mortality over winter, thus supporting earlier 

 findings with 1956 brood year eggs and larvae. 

 There was no evidence of winter mortality in 

 intertidal Twelvemile Creek, but because of the 

 scarcity of larvae in the spawning bed, it is 

 highly unlikely that mortality would have been 

 detected here. 



Other Brood Years 



According to Hollis ah temperatures, there 

 were no prolonged periods of freezing that would 

 affect eggs and larvae of the 1957, 1959, and 1960 

 brood years. An examination of data on mortality 

 revealed little evidence of mortality possibly 

 caused by freezing. 



There was no evidence of winter mortality of the 

 1957 brood year. Prewinter and post winter 

 pooled estimates of M, for riffles B, C, E, and F 

 and then 90-percent confidence limits were: 



M r (pooled for November)=0.69±0.19 

 M r (pooled for March) =0.57 ±0.24 



With regard to the 1959 brood year, there was a 

 rather high incidence (IS percent) of dead larvae in 

 samples collected from intertidal Harris River on 

 February 26, 1960, but a low incidence (less than 1 

 percent) of dead larvae collected in samples from 

 intertidal Indian Creek. Factors other than 

 freezing probably were responsible for the mor- 

 tality observed in intertidal Harris River because 

 temperatures were mostly above freezing in winter 

 1959-60, and few dead larvae were found in inter- 

 tidal Indian Creek. Scarcity of eggs and larvae in 



U.S. FISH AND WILDLIFE SERVICK 



