PETERSON ET AL.: FOOD HABITS OF JUVENILE SALMON 



then sinking passively through a portion of the 

 water column. 



Another explanation for the presence of large 

 numbers of hyperiids in salmonid guts is that 

 juvenile salmon may pick them from the surface 

 of medusae. The predominant hyperiid con- 

 sumed by chinook and chum salmon was Hyper- 

 oche medusarum, a species known to live on the 

 exumbrellar surface of medusae (Bowman et al. 

 1963; Harbison et al. 1977). The host may be easy 

 for salmon to locate, particularly the large Chrys- 

 aora fuscescens (bell diameter of several tens of 

 centimeters), which were very numerous in our 

 purse seine samples. 



Larval fishes were the other important prey 

 item. Information on their distribution and abun- 

 dance is limited to sampling done in 1971-72. 

 Data given in Table 9 are from Richardson and 

 Pearcy (1977) for larvae captured at stations 

 within 2-28 km from shore. Abundances were 

 200-400 larvae/10 m 2 , or 1-2 larvae/m 3 , assuming 

 they are all distributed only within the upper 20 

 m of the water column. 



To investigate the question of food limitation, 

 estimates are needed of salmonid feeding rates, 

 salmonid abundance, prey abundance, and prey 

 population growth rates. Feeding and digestive 

 rates can be inferred from field data, if there is 

 pronounced diel periodicity in stomach fullness 

 or state of digestion (Eggers 1977; Lane et al. 

 1979), but we have no evidence for this in our lim- 

 ited study. Thus, whereas estimates of stomach 

 fullness were obtained from this study, feeding 

 rates were estimated from other studies. The 

 average weight of food in full juvenile coho stom- 

 achs ( 1 .5 g wet weight) is equivalent to about 2.6% 

 of the 55 g body weight of an average juvenile 

 coho (160 mm long) (from Healey 1980). Walters 

 et al. (1978, fig. 6) showed that the maximum 

 ration of juvenile sockeye salmon weighing 55 g 

 is slightly <3% of body weight per day. On the 

 other hand, Brett (1971) found that the maxi- 

 mum daily intake of food was 7-8% of body weight 

 for a 50 g sockeye salmon. Therefore, we assume 

 that juvenile coho fill their stomachs between 1 

 and 3 times per day on the average. 



Averaged over the 2-wk period in June 1979, 

 the average 160 mm juvenile coho contained 37 

 euphausiids, 28 amphipods, and 4 fish (Table 1). 

 In order to locate this quantity of food, this salm- 

 on would have had to search a minimum of ap- 

 proximately 5.4 m 3 of water for the euphausiids, 

 7.8 m 3 for the amphipods, and at least 4.0 m 3 for 

 the larval fish. This assumes that all prey avail- 



able to plankton nets are also fully available to 

 juvenile salmon, and that annual differences are 

 minor. Considering the well-known problems of 

 zooplankton sampling variability and the fact 

 that samples from different years are being com- 

 pared, the agreement on water volume searched 

 by salmon to locate each prey item seems quite 

 good. 



The maximum abundance of juvenile salmo- 

 nids in any one purse seine was 123 fish, and the 

 average number of fish in sets in which at least 5 

 fish were caught was 26. The mean abundance in 

 these 16 sets was 17 fish/10 5 m 3 . Juvenile coho 

 abundances were about one-half as great, 8.7 

 fish/10 5 m 3 , or 1 fish/11,500 m 3 . If a juvenile coho 

 fills its gut once per day, it needs to eat all prey in 

 about 4-8 m 3 water/d. Thus, as a rough average, 

 one individual would consume at least 4/1 1,500- 

 8/1 1,500 (or 0.03-0.07%) of the available prey per 

 day. Should this individual coho fill its gut three 

 times each day, it would consume up to 0.1-0.2% 

 of the standing stock of prey per day. Coho and 

 chinook combined would consume aboutO.2-0.4% 

 of available prey per day. If growth rates of prey 

 population equal or exceed these loss rates, 

 predation by juvenile coho and chinook alone will 

 not reduce standing stocks of prey. Unfortu- 

 nately, estimates of these vital parameters are 

 lacking. 



Walters et al. (1978) examined the effect of 

 food limitation on juvenile salmon growth and 

 survival using a computer simulation model. In- 

 put variables included 1) zooplankton distribu- 

 tion, abundance, and production rates; 2) ration, 

 growth, and mortality of young salmon in rela- 

 tion to body size; and 3) timing of arrival of 

 smolts at sea and rate of migration along the 

 coast in relation to zooplankton production 

 cycles. They tentatively concluded that juvenile 

 salmonids were not food-limited, but rather 

 predator-limited. This conclusion rests on a cru- 

 cial assumption of the availability of zooplankton 

 prey, which may be in error. Their estimates of 

 zooplankton production and mortality and fish 

 consumption (their table 3, columns 5, 6, and 7) 

 were calculated using estimates ofthebiomass of 

 zooplankton within a 20-400 m water column, de- 

 spite their assumption that salmon forage only in 

 the upper 20 m of the water column. They as- 

 sumed that zooplankton prey removed by salmon 

 during the day will be replaced from deepwater 

 zooplankton populations at night. Since the sur- 

 face biomass is enhanced by diel vertical migra- 

 tions mainly at night and juvenile salmonids are 



849 



