Fisheries Research 114 (2012) 31–41 Contents lists available at ScienceDirect Fisheries Research jou rn al hom epage: www.elsevier.com/locate/fishres Examining cannibalism in relation to recruitment of silver hake Merluccius bilinearis in the U.S. northwest Atlantic Jason S. Link a,∗ , Sean M. Lucey a , Jessica H. Melgey b a National Marine Fisheries Service, Northeast Fisheries Science Center, 166 Water Street, Woods Hole, MA 02543 USA b School for Marine Science and Technology, University of Massachusetts-Dartmouth, 200 Mill Road, Suite 325, Fairhaven, MA 02719 USA a r t i c l e i n f o Article history: Received 16 November 2010 Received in revised form 11 April 2011 Accepted 30 April 2011 Keywords: Cannibalism Recruitment Spawning stock biomass Population dynamics a b s t r a c t Cannibalism can be an important feature for several fish species. From an ecological perspective, studying cannibalism in the northwest Atlantic fish community usually results in examining silver hake Mer- luccius bilinearis population dynamics. From a population dynamics (and hence fisheries assessment) perspective, silver hake pose some unique challenges and have defied easy, standardized approaches for evaluation of stock dynamics. Thus, here we propose to examine one facet of silver hake dynamics in light of their known cannibalistic propensity. We show for the two U.S. northwest Atlantic stocks that the per- cent diet composition of silver hake in silver hake is quite high and has been quite consistent over time. We also explore the size composition of cannibalized individuals as well, noting that the 0–10 cm group- ing is largely cannibalized by the 20–40 cm group, which is itself cannibalized by 40–60 cm group. From this, we fit stock-recruitment models (both Ricker and Beverton–Holt) that were typical and excluded cannibalism, that included cannibalism as an additional factor, and that changed the definition of the spawning stock biomass (SSB) from 20+ to those fish that are >40 cm. This latter shift resulted from the observed offset of recruitment production versus copious cannibalism as seen in the middle size group. Our results show that either the inclusion of cannibalism directly, or accounting for cannibalism indirectly by using a larger cutoff for the definition of SSB, are better model fits for both stocks than ignoring can- nibalism for this species. We discuss the broader implications of including cannibalism for this species, for other cannibalistic species, for changes in how we consider recruitment—particularly those species managed with recruitment-based reference points— and for how these might alter perceptions of stock dynamics. Published by Elsevier B.V. Introduction Cannibalism is one of the more intriguing aspects of population biology. As Fox (1975) and Polis (1981) both noted, it is cer- tainly a common phenomenon. Cannibalism has notable impacts; it can alter size and age structure, lead to unstable dynamics, alter growth rates, and in some cases ultimately regulate populations. In fact, despite a few limited benefits (Claessen et al., 2000; Polis, 1981), why cannibalistic populations persist (in light of all the well-documented ramifications and responses to cannibalism) has remained an important question in ecology (Claessen and de Roos, 2003; Magnusson, 1999). This feeding choice is particularly com- mon among piscivorous fishes (Smith and Reay, 1991) and has garnered periodic attention among fisheries scientists. Clearly, can- nibalism can have notable, nonlinear and counterintuitive impacts on population dynamics, all of which can add to the complexity of ∗ Corresponding author. Tel.: +1 508 495 2340; fax: +1 508 495 2258. E-mail address: Jason.Link@noaa.gov (J.S. Link). the population dynamics of commercially important fish popula- tions. When we have periodically examined food habits data to better elucidate cannibalism in the U.S. northwest Atlantic fish commu- nity (Table 1), the obvious results are: (1) silver hake Merluccius bilinearis is by far the most notable cannibalistic species and (2) there is an obvious need to further explore silver hake population dynamics in light of this cannibalistic behavior. This species has been well-noted as a cannibal in this and contiguous ecosystems from studies at several points in time, including the 1970s (Edwards and Bowman, 1979; Vinogradov, 1977), 1980s (Bowman, 1980, 1981, 1984; Bowman and Bowman, 1980; Cohen and Grosslein, 1981; Durbin et al., 1980; Koeller et al., 1989; Langton, 1982), 1990s (Helser et al., 1995; Sigaev, 1995; Waldron, 1989, 1992), and 2000s (Garrison and Link, 2000; Link and Almeida, 2000; Link and Garrison, 2002; Lock and Packer, 2004). Some authors (Bowman, 1980; Cohen and Grosslein, 1981; Edwards and Bowman, 1979; Waldron, 1989) were so impressed by the magnitude of the phe- nomenon that they were led to speculate about the potential controlling influence of cannibalism on silver hake stock dynamics, 0165-7836/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.fishres.2011.04.022