FISHERY BULLETIN: VOL. 84, NO. 4 



mackerel is also a major fish stock in the region and 

 is both planktivorous and piscivorous, with juvenile 

 and adult northern anchovy being a dominant item 

 in the diet of larger individuals (Schaefer 1980; 

 Hunter and Lynn, unpubl. data, Southwest Fisheries 

 Center, La Jolla, CA). 



MATERIALS AND METHODS 



Experimental Fishes 



The northern anchovy larvae used in the experi- 

 ments were reared from the egg; 2,000-7,000 eggs, 

 from a laboratory brood stock (Leong 1971), were 

 stocked in 400 L black circular fiberglass tanks con- 

 taining about 150 L of filtered sea water. The culture 

 methods of Hunter (1976) were used to rear the lar- 

 vae, with extra additions of wild zooplankton when 

 the larvae were 10-25 mm SL. Temperature in the 

 rearing tanks was maintained at about 18 °C (range, 

 17.2°-19.2°C). 



Three groups of 5 adult northern anchovy (range 

 of mean standard length [SL] = 8.3-8.9 cm) and a 

 single group of 3 juvenile chub mackerel (mean SL 

 = 19.1 cm) were used as predators. They were fed 

 mainly adult brine shrimp and occasionally north- 

 ern anchovy larvae. The predators were not fed for 

 10 h before an experiment. 



Apparatus 



Predators were kept in two rectangular fiberglass 

 tanks (0.75 m x 2.15 m x 0.83 m = 1.35 m 3 ) with 

 a clear glass window on one side for observation. 

 Two 100 W tungsten household lamps produced 

 2,000-3,000 mc at the surface of each tank and a 

 black plastic tent enclosed the window, providing 

 a darkened compartment for an observer. Larvae 

 were released into the tank by gently submerging 

 a beaker at the water surface. Horizontal and ver- 

 tical metric scales on the tank window aided esti- 

 mation of predator attack distances. The tanks were 

 continuously supplied with ambient seawater rang- 

 ing from 20.5° to 23.8°C, except during an experi- 

 ment when the water was static. 



Experimental Procedure 



It was necessary to measure the feeding perfor- 

 mance of predators fed a standard prey because 1) 

 adult northern anchovy are easily frightened and 

 fright behavior reduces feeding motivation; 2) feed- 

 ing could be affected by satiation during an experi- 

 ment; and 3) feeding could be affected by the fish 



860 



learning and responding to cues associated with the 

 introduction of food. We used live adult brine shrimp 

 (Artemia sp., 6.4 mm mean total length, standard 

 deviation [SD] 1.2 mm, n = 25) as a standard prey. 

 Variation in feeding performance of the predators 

 could be more easily detected when Artemia were 

 used because unlike the larvae the Artemia did not 

 vary in size among experiments nor did they avoid 

 attack by the predators. 



Northern anchovy larvae and the adult Artemia 

 were added to the tank in groups of three. An addi- 

 tion of three of either prey constituted a trial. A trial 

 ended after 5 min or when all prey were taken. Dur- 

 ing a trial we used a computer compatible event 

 recorder to record observations of the interactions 

 between predator and prey. All the experiments 

 using northern anchovy as predators started with 

 5 consecutive trials in which 3 Artemia were offered 

 per trial. This was done to insure that northern an- 

 chovy predator groups had a similar level of feeding 

 motivation. Preliminary experiments indicated that 

 it normally took a few feeding trials before adult 

 northern anchovy fed consistently. After the 5 ini- 

 tial trials, predators were offered fish larvae and 

 adult Artemia alternately for 4-10 trials. Adult 

 Artemia were always used in the last trial to deter- 

 mine if satiation had occurred. A less rigorous 

 schedule was used for the chub mackerel predators 

 because their feeding behavior was less variable 

 than that of the northern anchovy. After 3 initial 

 Artemia trials, the chub mackerel were given 5 lar- 

 val trials followed by an Artemia trial. In most cases, 

 a second set of 5 larval trials were also given and 

 these were followed by a final Artemia trial to check 

 if satiation had occurred. 



The number of observations for each larval size 

 class was the total number of predator-prey inter- 

 actions observed among larvae in that size class. 

 This number exceeded the number of larvae tested 

 in many cases because, if a larva escaped the first 

 encounter with a predator, the subsequent en- 

 counter was also recorded as an event. The total 

 number of observations (predatory events) per lar- 

 val size class (mean SL), when northern anchovy 

 were the predators, was 5.9 mm, 24; 8.5 mm, 55; 

 11 mm, 48; 15 mm, 53; 21 mm, 82; and 33 mm, 62. 

 Those for the chub mackerel experiments were 6.7 

 mm, 19; 10 mm, 75; 16 mm, 54; 21 mm, 27; 31 mm, 

 47; and 50 mm, 39. 



Classification of Behavior 



Prey behavior was recorded only when the 

 predator attacked a prey. An attack was defined as 



