FISHERY BULLETIN: VOL. 79, NO. 1 



0.02-0.11 g C/m^ per d. Primary production in the 

 Bight averages 0.5-1.0 g C/m^ per d (Eppley et al. 

 1979). The northern anchovy feeds primarily 

 upon zooplankton. Assuming a 20% conversion of 

 primary to secondary production, the spawning 

 population ofE. mordax consumes 10 to >100% of 

 secondary production in the Southern California 

 Bight. This figure appears to give the right order 

 of magnitude, since >80% of acoustic targets 

 from pelagic surveys in this area are estimated to 

 be northern anchovy schools (Mais 1974). It 

 should be noted that the Southern California 

 Bight appears to be an area of net zooplankton 

 consumption, since zooplankton densities are 

 typically greater in the California Current north 

 of the Bight than in the Bight itself (Reid 1962). 

 Presumably the zooplankton are being consumed 

 as they are carried into the area. The northern 

 anchovy also supplements its diet by filter feed- 

 ing on ph3rtoplankton. 



If the feeding of the northern anchovy is size 

 selective, its impact on the community of zoo- 

 plankton could be considerable. Furthermore, 

 the northern anchovy population is, relatively 

 speaking, not all that large. The population of 

 the Peruvian anchovetta, which is predominant- 

 ly phytophagous but is a zooplankton feeder in 

 certain areas and at certain stages of its life 

 history (Rojas de Mendiola 1971), is estimated to 

 have been an order of magnitude more densely 

 concentrated (Walsh et al. 1980). 



Most quantitative studies of the feeding selec- 

 tivity of marine fish have been conducted in the 

 laboratory with small groups of fish (Leong and 

 O'Connell 1969; O'Connell 1972; Durbin and 

 Durbin 1975). This permits only a crude approx- 

 imation of their impact upon marine systems, 

 where these fish populations are predominantly 

 found in massive shoals. For example, approxi- 

 mately 90% of the biomass of the northern 

 anchovy population is found in schools >25 t 

 (calculated from Hewdtt et al. 1976). While Eggers 

 (1976) modelled the energetics of planktivorous 

 fish schools using the extensive literature on the 

 feeding of individual fish, there is no experimen- 

 tal data on the feeding of fish schools to test such 

 models. Without better data on the feeding of 

 schooling fish on the zooplankton, contemporary 

 models of marine zooplankton community d5rnam- 

 ics have perforce concentrated upon interactions 

 among the lower trophic levels (Steele 1974; 

 Steele and Frost 1977). 



I report here the results of in situ measure- 



ments of the feeding selectivity of schools of E. 

 mordax in the nearshore waters of the Southern 

 California Bight. These represent the first direct 

 quantitative field measurements of the feeding of 

 schools of planktivorous fish. 



METHODS 



A vessel with side-scanning sonar and echo 

 sounder was used to track and determine the 

 dimensions of large (25-200 t), near-surface 

 schools of northern anchovy. A school was consid- 

 ered appropriate for study when 1) the school was 

 near the surface and of sufficient size (>50 m 

 along the axis perpendicular to the school's 

 movement) that plankton samples could unequiv- 

 ocally be taken in its wake, 2) the school did not 

 show signs of being disturbed by the ship's pres- 

 ence, and 3) the school was either directly 

 observed to be feeding (October 1976) or its gen- 

 eral configuration and movement were consistent 

 with feeding behavior. It was assumed that when 

 feeding, a school would either form an amorphous 

 "ball" (Radakov 1973) or that its long axis would 

 be normal to its axis of motion (Weihs 1973), and 

 that the school's velocity would not exceed sever- 

 al body lengths per second. 



When a school was selected for study, a cruci- 

 form drogue with surface buoy was dropped into 

 its center. The school's movements in relation to 

 the drogue were monitored for 10-25 min, during 

 which time the school usually moved 100 to 

 several hundred meters from the drogue. A 

 weighted buoy was then placed over the school. 

 Thus a transect was established, over which the 

 school had passed while presvunably feeding. The 

 school's physical dimensions and swimming 

 speed relative to the water could be determined 

 using the ship's sonar, echo sounder, and by 

 timing the school's movement between the two 

 buoys. 



In general, the sampling regimen consisted of 

 taking two replicate samples with zooplankton 

 nets first in the wake of the school between the 

 buoys and then in "control" areas either in front 

 of or several hundred meters to the side of the 

 school. The nets were lowered obliquely from the 

 surface to the average depth of the school (as 

 determined by echo sounder), towed at that depth 

 for 2 min, and then hauled to the surface (total 

 length of tow about 100 m). A 0.5 m diameter 

 plankton net (102 /um mesh) with a TSK^ flow- 

 meter was towed in a harness with a 1.0 m 



132 



