loosened by gravel removal operation, bull- 

 dozing in streambeds, or fording by vehicles. 

 At Brooks Lake, however, the preference for 

 such areas was small. Perhaps under condi- 

 tions of severe gravel compaction and sparse 

 spawning populations, it would be greater. 



To ensure consistent methods and com- 

 parable results throughout an observation 

 period, personnel followed the procedures 

 outlined below. 



1. To avoid disturbing fish en route to the 

 tower from the field station, the observer 

 crossed the stream at least 100 yards above 

 the study area and then walked on the bank 

 until reaching the tower. 



2. To preclude recording behavior that 

 might be influenced by the observer as he 

 climbed the tower, 5 minutes elapsed before 

 observations were recorded. 



3. Observations were made of one row at 

 a time for 5-minute periods and were con- 

 fined to rows 7, 8, 9, and 10 after August 31. 

 These limitations were adopted when it became 

 apparent that a single observer could effec- 

 tively observe only a limited area where 

 visibility was best. 



4. Sex ratios on all redds were noted at 

 the start of each day's observations. 



5. Records of individual fish behavior by 

 sex were usually confined to defense activity 

 over redds with a 1:1 sex ratio. 



6. A code was devised to simplify and 

 expedite observations and recording. The code 

 was based on three categories of activity that 

 were encountered most frequently as follows: 



■ •{■■••{••I'll 



15 25 I 7 



SEPT. OCT. 



Figure 9. --Daily spawning activity of sockeye salmon, 

 as indicated by sex ratios in Brooks River observa- 

 tion area, August 16 to October 7. 1957. Three peaks 

 indicate periods of maximum spawningactivity. Prog- 

 ressive increase of 0:1 sex ratio signifies more 

 completed spawning, with females guarding redds 

 after death of males. 



a female; and c = one chasing or defense ac- 

 tivity by a male and female. Thus, b-4 would 

 signify that a female over a redd chased another 

 female from the redd area. 



Sex ratios of 1:1 and 0:1^ were most com- 

 mon during the spawning period and were 

 maintained vigorously by both sexes until the 

 male died (fig. 9). Male and female defense 

 activities (defined as overt actions against 

 other fish) for redds with a 1:1 sex ratio 

 were observed 2,333 times (table 4). Males 

 were almost twice as aggressive as females 

 in defending the redd against intruders (62 

 percent for males and 38 percent for females). 

 On only 1 day, September 2, were females 

 more active than males, and this day was 

 characterized by a severe wind and rainstorm, 

 which resulted in abnormal stream conditions. 



In the area under observation, three sepa- 

 rate and distinct waves of sockeye salmon 

 spawned. The first wave reached a peak of 

 activity August 26; the second, September 7 to 

 9; and the third, September 24. The last wave 

 was the largest and lasted longer than the 

 first two (fig. 9). 



The irregular breaking up of 1:1 sex ratios 

 toward the end of the initial wave (fig. 9) was 

 probably due to the occurrence of a severe 

 offshore wind and rainstorm on September 2, 

 which lowered the lake level at the outlet. 

 Concomitant with the decreased flow caused 



where a= one chasing or defense activity by 

 a male; b= one chasing or defense activity by 



'Males are designated first in all sex ratios in this 

 paper. 



13 



