HARGREAVES: PREY PREFERENCE OF PREDATORS 



prey, can often be directly verified. For the two ex- 

 periments reported here, prey size (fork length) was 

 the characteristic that was most important. The 

 observed difference between the average length of 

 the tagged prey that were released and the tagged 

 prey subsequently recovered from the stomachs of 

 predators was 2.5 mm (6.6%) in the first experiment, 

 and 3.7 mm (8.3%) in the second experiment. These 

 are small differences, and the possibility that they 

 might be due to experimental error rather than 

 predator selectivity must be considered. All length 

 measurements were made to the nearest millimeter 

 and numerous remeasurements indicated that mea- 

 surement errors were negligible at this level of ac- 

 curacy. To eliminate the possibility that the length 

 of the tagged fish might decrease if they were pre- 

 served (Parker 1963), live fish were used to deter- 

 mine the length-frequency distribution of the prey 

 prior to release and all prey recovered from the 

 stomachs of predators were immediately measured. 

 Burgner (1962) reported that the length of sockeye 

 salmon smolts decreased by 2-3% because of rigor 

 mortis alone. However, experiments conducted in 

 Masset Inlet in 1984 indicated much smaller changes 

 occur after death in juvenile pink salmon. At tem- 

 peratures of 9°-10°C, the average length of 26 

 juvenile pink salmon of known length, fed to and 

 subsequently recovered from the stomachs of 19 

 yearling coho salmon, decreased less than 1% for 

 periods of up to four hours after ingestion (Har- 

 greaves unpubl. data). In the two experiments re- 

 ported here, numerous beach seine sets were made 

 to capture potential predators, but all of the tagged 

 pink salmon found in their stomachs were recovered 

 within four hours of the releases of tagged prey. 

 Shrinkage of the prey after ingestion therefore can 

 account for only a small portion of the observed 

 differences in size between the prey that were 

 released and those that were found in the stomachs 

 of predators. 



The fourth assumption, that all tagged or marked 

 prey remain equally "available" to predators 

 throughout the experiment, will usually prove to be 

 the most difficult to assess and verify. The avail- 

 ability of prey to predators frequently depends on 

 characteristics of the predators (hunger level, visual 

 acuity, mobility, body or gape size, individual or 

 group behavior, etc.), the prey (abundance, colora- 

 tion, size, speed, endurance, behavior, etc.), and the 

 environment (habitat complexity, light conditions, 

 etc.). These parameters can interact in a complex 

 manner, so that it is typically only in the simplest 

 situations that all factors that affect the availabil- 



ity of prey to a predator can be thoroughly inves- 

 tigated and understood (Curio 1976; Zaret 1980). 



Predators consumed significantly more of the 

 smaller prey in both experiments reported here, 

 despite substantial differences in the physical char- 

 acteristics of the two release sites, time of year, and 

 various characteristics of the predators (abundance, 

 species composition, size, feeding history, etc.). This 

 suggests that the availability of prey to the preda- 

 tors was not substantially affected by variations in 

 the characteristics of either the predators or the en- 

 vironment. It also appears reasonable to assume 

 that prey of all sizes remained equally available to 

 predators during both experiments. All of the 

 tagged prey were one species and received identical 

 treatment prior to release. There is no reason to 

 think there were any substantial differences in the 

 physical characteristics among the prey at the time 

 of release, aside from the desired variation in size. 



It is conceivable, however, that differences in prey 

 behavior or size might have indirectly influenced the 

 availability of prey to the predators. For example, 

 extensive sampling of juvenile salmon in Masset 

 Inlet has indicated a tendency for larger pink salmon 

 to be concentrated further offshore than smaller 

 pink salmon during the early sea-life period (Har- 

 greaves et al. 1987a, b). Swimming speeds of salmon 

 are also known to increase rapidly with increasing 

 body size (Brett 1965). Thus, if there was any 

 tendency for tagged salmon to rapidly swim away 

 from the release sites, larger fish may have left 

 quicker than smaller fish. The result could be a 

 decrease in the average size of tagged salmon found 

 in the immediate vicinity of the release site and the 

 incorrect conclusion that predators were selective- 

 ly feeding on the smaller prey. 



In fact, however, this possibility appears unlike- 

 ly. In both experiments the size of the live, tagged 

 fish recovered along with the predators was not 

 significantly different than the size of the fish that 

 had been released as much as nine hours earlier 

 (Table 1). There is also no indication that the mean 

 size of these fish changed in a consistent manner 

 over the course of the second experiment. These 

 results suggest that, if there was any segregation 

 of tagged prey after release, it was probably minor 

 and did not appreciably affect the availability of prey 

 to the predators. 



In general, complications arising from variations 

 in the availability of prey to the predators may be 

 reduced or eliminated by limiting the duration of the 

 experiment. If all of the prey are released at one 

 time and location, it is reasonable, and in most cases 



769 



