Table 9. — Dates on which, rain 1 at Hollis, Alaska, exceeded 

 4.0 inches in 72 hours, 1956-60 



1 Data provided by Northern Forest Experiment Station, U.S. Forest 

 Service, Juneau, Alaska. 



continues over several days. To index periods of 

 heavy precipitation, dates on which total precipi- 

 tation exceeded 4 inches in 72 hours at Hollis 

 (fig. 4) are given in table 9. Stream discharge 

 records revealed that unusually high discharges 

 accompanied rainfall of this high intensity. The 

 analysis showed that rain was most intense in 

 1958 and 1959 and least intense in 1956 and 1957. 

 Although a comprehensive evaluation of mor- 

 tality of 1956 and 1957 brood year eggs and larvae 

 was not feasible because of limited sampling, 

 estimates of M r obtained for these brood years 

 (and discussed previously) gave no indication of 

 losses occurring from spawning beds. Observa- 



tions on subsequent brood years (1958 and 1959 

 particularly) have shown, however, that gravel 

 movement during high discharge is an important 

 cause of mortality in the study streams. 



1958 Brood Year 



Rainfall exceeded 4.0 inches in 72 hours three 

 times between October 19 and November 9, 1958 

 (table 9). Samples of eggs were collected before 

 and after the storms in intertidal Harris River 

 and after the storms in intertidal Indian and 

 Twelvemile Creeks (table 10). 



The number of eggs (P) estimated to be present 

 in intertidal Harris River near the end of spawning 

 agreed with the expected number (£') obtained 

 from the estimated density of females spawning 



(P=189 and £" = 201 per m. 2 ). After the three 

 periods of heavy rainfall, the abundance of eggs 

 had decreased significantly, and the fraction of 

 points containing fewer than 35 eggs per m. 2 had 

 increased significantly (X 2 =b.b, 1 d.f.). This was 

 good evidence that a significant mortality at- 

 tributable to gravel movement had occurred. 



In Indian Creek, the number of eggs estimated 

 to be present in November after the storms agreed 



with the expected number (P=222 and £" = 202 

 eggs per m. 2 ) ; hence, there was no evidence of 

 mortality from gravel movement in intertidal 

 Indian Creek. 



The number of eggs estimated in intertidal 

 Twelvemile Creek after heavy rainfall was con- 



Table 10. — Estimates of the population parameters E', P, p' , and M t used to evaluate mortality of 1958 brood year eggs 



possibly caused by movement of bed materials 



E i [mated number of live plus dead specimens in the spawning bed (P). 

 - These limits were set with log-transformed data. 

 ' Expected number of specimens based on potential epR deposition (/?')- 



;»i4 



U.S. FISH AND WILDLIFE SERVICE 



