278 



Fishery Bulletin 94(2). 1996 



present study, Atlantic mackerel larvae 3-14 mm 

 long preyed primarily on newly hatched fish larvae 

 (conspecifics and others). This suggests that in the 

 days following hatching, fish larvae quickly acquire 

 enough motility to escape attack from Atlantic mack- 

 erel larvae in this size range. Thus, the predation 

 rates estimated in the present study ( 7% to 57% /d of 

 the standing stock of larvae of suitable prey size) 

 would apply only over the relatively short period 

 before the development of an efficient escape re- 

 sponse. Growth during this period would determine 

 cumulative predation mortality from co-occurring 

 Atlantic mackerel larvae 3-14 mm in length. In the 

 laboratory, survival to metamorphosis in Pacific 

 mackerel larvae was positively correlated with 

 growth over the period during which sibling canni- 

 balism occurred (Hunter and Kimbrell, 1980). 



Effect of prey density on predation by 

 Atlantic mackerel larvae 



Surprisingly few studies have related the feeding 

 success offish larvae in the sea to prey availability 

 (Heath, 1992; Fortier et al., 1995). In the majority of 

 these studies, there are indications that feeding was 

 limited below some threshold density of prey. For 

 example, the feeding ratio of larval Atlantic cod, 

 Gadus morhua, was adequately described by an Ivlev 

 equation, with feeding limited at nauplii concentra- 

 tions <10/L (Ellertsen et al., 1989; Sundby and Fossum, 

 1990). The same model was used to describe stomach 

 fullness and feeding incidence ( proportion of predators 

 with prey in the gut) in larvae of southern bluefin tuna, 

 Thunnus maccoyii, as a function of zooplankton settled 

 volume (Young and Davis, 1990; Heath, 1992). In this 

 case, the feeding of larval tuna appeared limited at 

 densities of approximately <0.5 prey/L (Heath, 1992). 

 In ice-covered Hudson Bay, the feeding incidence and 

 feeding ratio of Arctic cod, Boreogadus saida, and lar- 

 vae of sand lance, Ammodytes sp. in relation to nauplii 

 density were also adequately modeled by an Ivlev equa- 

 tion (Fortier et al., in press). Feeding was limited at 

 densities <40,000 nauplii/m 2 , corresponding approxi- 

 mately to <1.6 nauplii/L. In large enclosures, the growth 

 of larvae of capelin, Mallotus villosus, was related to 

 prey biomass by a function resembling an Ivlev func- 

 tion (Frank and Leggett, 1982). 



In the present study, there was no indication that 

 predation on copepod nauplii and copepods by At- 

 lantic mackerel larvae was dependent on the den- 

 sity of these prey in the environment. This may indi- 

 cate that naupliar (261 to 4,134/m 3 ) and copepod 

 (1,170 to 7,190/m 3 ) densities were at or above the 

 level at which feeding Atlantic mackerel larvae satu- 

 rate. However, the relationship between the incidence 



of predation on fish and the density of fish larvae 

 was adequately described by an Ivlev equation. Limi- 

 tation began at prey densities <0.1 fish larva of suit- 

 able prey size per m 3 . This threshold was only slightly 

 less than the mean density offish larvae of suitable 

 prey size (0.14 larva/m 3 ), and predation on fish ap- 

 peared limited in about 40% of our collections. Alter- 

 natively, Atlantic mackerel larvae preying on fish 

 larvae appeared swamped by their prey in approxi- 

 mately 60% of the collections. Thus, the contracted 

 production of large numbers of larvae in species such 

 as yellowtail flounder and silver hake may reduce 

 total cumulative predation by Atlantic mackerel lar- 

 vae. Frank and Leggett (1982) suggested that the 

 synchronized emergence of beach-spawned capelin 

 larvae during onshore wind events in eastern New- 

 foundland resulted in the saturation of their preda- 

 tors and a reduction of percent predation mortality. 



Effect of alternative prey on cannibalism and 

 piscivory by Atlantic mackerel larvae 



Controlled studies have invariably reported that the 

 presence of alternative prey provides fish larvae with 

 some protection from larval, juvenile, or small-bod- 

 ied adult fish predators. In the laboratory, the rate 

 of cannibalism on larvae by juvenile Cape anchovy, 

 Engraulis capensis, decreased in the presence of al- 

 ternative copepod prey ( Brownell, 1985 ). Bay anchovy 

 juveniles preferred newly-hatched (2-d-old) sea 

 bream, Archosargus rhomboidalis, but shifted to cope- 

 pods when the density of the latter increased (Dowd, 

 1986). The presence of alternative copepod prey strongly 

 protected larval Atlantic mackerel (Kean-Howie et al., 

 1988) and Atlantic cod (Gotceitas and Brown, 1993) 

 from predation by the three-spine stickleback, 

 Gasterosteus aculeatus. The addition of Daphnia ma- 

 gna reduced by five to tenfold the predation rate of ju- 

 venile bluegill, Lepomis macrochirus, on larval white 

 perch, Morone americana (Margulies, 1990). 



In the present study, we found some evidence that 

 the presence of alternative prey reduced the preda- 

 tion by Atlantic mackerel larvae on fish larvae (Fig. 

 8). The daily predation rate on fish larvae of suit- 

 able prey size (assuming a digestion time of 24 h) 

 declined with increasing density of copepod nauplii 

 (the only other prey, along with fish larvae, selected 

 for by Atlantic mackerel larvae), but not with an in- 

 creasing density of copepods. The evidence for a re- 

 duction of predation was stronger when the actual 

 consumption of alternative prey rather than their 

 density was considered (Fig. 8). On average, the daily 

 predation rate on fish larvae of suitable prey size 

 was reduced from 47% to 12%/d when the average 

 number of copepod prey in the gut of Atlantic mack- 



