Fortier and Villeneuve: Cannibalism and predation by Scomber scombrus larvae 



277 



RMT 



5 to < 6 mm 



/P = 31.3(1-8 - 2 ' ) 



r* =0 79. P<0 005 



RMT + Multi-net 



1 2 0.3 4 0.5 16 0.0 1 2 3 4 5 16 



Density of fish larvae of suitable prey size (no./m 3 ) 



Figure 7 



Incidence of piscivory (IP) in larval Atlantic mackerel, Scomber 

 scombrus, as a function of the density offish larvae of suit- 

 able prey size (D), by standard-length classes of mackerel lar- 

 vae. (A-Cl RMT collections only; (D-F) RMT and multi-net 

 collections. The percentage of the variance (r 2 ) in the inci- 

 dence of piscivory explained by the Ivlev function (Ivlev, 1961 ) 

 and the significance level of the regression (P) are given. Note 

 the broken scale on the abscissa. 



identified ( 74 out of 710), Atlantic mackerel larvae exhib- 

 ited little preference for any of the four dominant species 

 offish larvae available in July in the area studied. 



Despite their relatively low frequency in the diet ( 29c 

 to 16% by number), fish larvae represented an impor- 

 tant fraction of the energy intake of Atlantic mackerel 

 larvae. For example, most of the conspecifics ingested 

 ranged from 3.0 to 5.5 mm in length, which translates 

 into 20 to 200 fjg in dry weight (Table 3 in Peterson 

 and Ausubel, 1984). By comparison, the approximate 

 dry weight of copepod prey (0.1-0.6 mm in cephalotho- 

 rax length) ranged from 0.006 to 2 /Jg (Cohen and 

 Lough, 1981). Thus, assuming that the nutritional 

 value of copepods is four times that of fish larvae per 

 unit biomass (Gotceitas and Brown, 1993), the energy 

 intake represented by one fish larva corresponded to 

 that of 2.5 to 25 of the largest copepods ingested. Pa- 

 cific mackerel, Scomber japonicus, and Atlantic mack- 

 erel larvae are characterized by fast growth, high me- 

 tabolism, and high food requirements (Hunter and 

 Kimbrell, 1980; Peterson and Ausubel, 1984; Ware and 



Naupln density (thousands/m 3 ) 



Feeding ratio on copepods 

 (mean number per gut) 



Figure 8 



Estimated daily predation rate (PR, see text) of larval 

 Atlantic mackerel. Scomber scombrus, on fish larvae of 

 suitable prey size as a function of (A) the density of 

 copepod nauplii iDN) in the environment (PR = 61.6 - 

 0.0189LW [ra=8, r 2 =0.505, F=6.1, P=0.048]), and (B) as 

 a function of the feeding ratio of Atlantic mackerel lar- 

 vae for copepods (FP, all developmental stages), (PP = 

 1/(2.1 + 0.612FR) [n=31, r-=0.462, F=24.9, P<0.001]). 

 In (B), each data point in the regression was weighted 

 by the number of Atlantic mackerel larvae in the sample 

 ( given near each point ). The outlier marked by the open 

 symbol was excluded from the analysis. 



Lambert, 1985). The width of ingested fish larvae 

 closely matches the mouth width of Atlantic mackerel 

 larvae (Ware and Lambert, 1985), and fish larvae are 

 the largest prey ingested ( Grave, 1981; Ware and Lam- 

 bert, 1985; this study). Hence, fish larvae probably rep- 

 resent the most profitable prey available in the plank- 

 ton to satisfy the high food requirements of Atlantic 

 mackerel larvae. 



Piscivores generally select small prey in labora- 

 tory experiments and consistently ingest smaller 

 than average prey in field situations, a potential re- 

 flection of an inability to capture larger prey ( Juanes 

 and Conover, 1994, and references therein). In the 



