FISHERY BULLETIN: VOL. 79, NO 4 



as described by Hunter (1976). Eggs were spawned 

 from five groups of adults taken from laboratory 

 stocks on five occasions during spring 1980 (Table 

 1.) Eggs were transferred to noncirculated filtered 

 sea water in 400 1 black fiber glass tanks. Food for 

 larvae was the dinoflagellate Gymnodinium 

 splendens for 2- to 5-d-old larvae, and the rotifer 

 Brachionus plicatilis for older larvae. Water tem- 

 perature was maintained at 20° C. Larvae were 

 held under constant illumination of 2,000 Ix at the 

 water surface, provided by standard room fluores- 

 cent lights. 



Observations were made on predation of eggs 

 and 10 groups of larvae of different total lengths, 

 ranging from 0.29 to 1.2 cm (Table 1). Lengths were 

 based on measurements of a subsample from each 

 group at the time they were used for experiments. 

 Observations were concentrated on early larvae 

 because this is the period of greatest mor- 

 phogenesis (O'Connell 1981) and maturation of 

 response systems (Kimmel 1972; Webb and 

 Corolla 1981). 



Table l.— Total length of subsample (X±2 SE; N = 10) of 

 northern anchovy larvae populations used in predation experi- 

 ments with a model predator, the clown fish. Data are also given 

 on spawnings. 



'Eggs were 0.135±0. 005 cm long • 0.069 ±0.003 cm wide. 



Attempts to obtain natural predators of uniform 

 and suitable size were not successful. Therefore a 

 mimic for biting predators was used. This model 

 was the clown fish, Amphiprion percula, cultured 

 from eggs. The fish were held in 40 1 polypropylene 

 tanks at 20° C and under constant illumination. 

 The fish were fed daily on a diet of Tetramin^ and 

 northern anchovy eggs and larvae. The predator 

 was readily available at uniform sizes throughout 

 the duration of the experiment. The clown fish 

 were 4.4 ±0.2 cm total length (TL) and mass was 

 1.58 ± 0.17 g. The maximum depth was 1.8 ± 0.2 cm, 

 and maximum width 0.6 ±0.1 cm, both located 



■^Reference to trade names does not imply endorsement by the 

 National Marme Fisheries Service, NOAA. 



1.3 ± 0.1 cm from the nose (data are X ± 2 SE; N = 

 10). 



Interactions between the model predator and 

 northern anchovy eggs and larvae were observed 

 using schlieren optics (Holder and North 1963). 

 Briefly, a vertical collimated light beam was pro- 

 duced by a high intensity monochromatic point 

 source at the focus of a concave mirror (focal 

 length 140 cm) attached to the ceiling. A second 

 mirror, with the same focal length, was located at 

 the floor and focused the light on a black spot on a 

 glass plate. 



The observation arena was a cylindrical tank, 

 located in the light beam. It was 35 cm in diameter 

 and 10 cm deep and had a plate-glass bottom. The 

 tank contained a central cylinder, 26.5 cm in di- 

 ameter and 5 cm deep, which supported a circular 

 plate-glass lid. The lid and the bottom were paral- 

 lel to each other and normal to the light beam. 

 Discontinuities in refractive index (i.e., larvae in 

 the water) deflected the light from the focus spot 

 on the glass plate, and were seen as bright objects 

 against a dark background. Opaque objects (i.e., 

 the predator) were seen as black silhouettes. 



The clown fish were starved for 5 d to ensure 

 uniform high motivation. This was necessary be- 

 cause northern anchovy larvae grow rapidly and 

 loss of a day required starting new batches, par- 

 ticularly for young larvae. Individual clown fish 

 were placed in the observation arena and left for 

 24 h. After this period, up to 100 northern anchovy 

 eggs and larvae were introduced through a side 

 port in the central cylinder. After 10 to 30 min, the 

 predator began feeding on the prey distributed 

 through the tank. The reason for the delay in the 

 onset of feeding is not known. Behavior was re- 

 corded on video tape. Experiments were per- 

 formed at 20° C. 



The video tape was manually advanced to ana- 

 lyze images frame by frame (framing rate = 60 

 Hz). The following observations were made: the 

 number of attacks by clown fish, escape attempts 

 by prey, escape success of the prey, catch success 

 of the predator, and the number of predator er- 

 rors. The prey escape angle, distance traveled, and 

 mean speed during an escape attempt were mea- 

 sured. 



Animals from a wide range of taxa, including 

 fish, show avoidance responses to approaching, 

 presumably threatening objects (see Gibson 1980). 

 The image of such an object on the retina expands 

 rapidly and this magnification in time is called the 

 "looming effect." It is measured as the rate of 



728 



