Journal of Animal Ecology 2004 73 , 585–598 © 2004 British Ecological Society Blackwell Publishing, Ltd.Oxford, UKJAEJournal of Animal Ecology0021-8790British Ecological Society, 20045 2004733Original ArticleInteraction strengths in fod websE.L. Berlow et al. FORUM Interaction strengths in food webs: issues and opportunities ERIC L. BERLOW 1,2 , ANJE-MARGIET NEUTEL 3 , JOEL E. COHEN 4 , PETER C. DE RUITER 3 , BO EBENMAN 5 , MARK EMMERSON 6 , JEREMY W. FOX 7 , VINCENT A. A. JANSEN 8 , J. IWAN JONES 9 , GIORGOS D. KOKKORIS 10 , DMITRII O. LOGOFET 11 , ALAN J. MCKANE 12 , JOSE M. MONTOYA 13 and OWEN PETCHEY 14 1 University of California, White Mountain Research Station, Bishop, CA 93514, USA; 2 Department of Integrative Biology, University of California, Berkeley, CA 94720, USA; 3 Department of Environmental Sciences, Utrecht University, PO Box 80115, 3508 TC Utrecht, the Netherlands; 4 Rockefeller and Columbia Universities, 1230 York Avenue, Box 20, New York NY 10021–6399, USA; 5 Department of Biology, Linkoping University, S-581 83 Linköping, Sweden; 6 Department of Zoology, Ecology and Plant Sciences, University College Cork, Lee Maltings, Prospect Row, Cork, Eire; 7 NERC Centre for Population Biology, Imperial College, Silwood Park, Ascot, Berkshire SL5 7PY, UK; 8 School of Biological Sciences, Royal Holloway, University of London, Egham, Surrey TW20 0EX, UK; 9 School of Biological Sciences, Queen Mary, University of London, Mile End Road, London E1 4NS, UK, 10 Department of Marine Sciences, University of the Aegean, University Hill, 81100 Mytilene, Lesvos Island, Greece; 11 Laboratory of Math. Ecology, IFARAN, Pyzhevsky Pereulok 3, Moscow, 119017, Russia; 12 Department of Theoretical Physics, University of Manchester, Manchester M13 9 PL, UK; 13 Complex Systems Laboratory, IMIM -UPF (GRIB), Dr Aigvader 80, 08003 Barcelona, Spain; and 14 Department of Animal and Plant Sciences, University of Sheffield, Sheffield S10 2TN, UK Summary1. Recent efforts to understand how the patterning of interaction strength affects bothstructure and dynamics in food webs have highlighted several obstacles to productivesynthesis. Issues arise with respect to goals and driving questions, methods andapproaches, and placing results in the context of broader ecological theory. 2. Much confusion stems from lack of clarity about whether the questions posed relateto community-level patterns or to species dynamics, and to what authors actually meanby the term ‘interaction strength’. Here, we describe the various ways in which this termhas been applied and discuss the implications of loose terminology and definition forthe development of this field. 3. Of particular concern is the clear gap between theoretical and empirical investigationsof interaction strengths and food web dynamics. The ecological community urgentlyneeds to explore new ways to estimate biologically reasonable model coefficients fromempirical data, such as foraging rates, body size, metabolic rate, biomass distributionand other species traits. 4. Combining numerical and analytical modelling approaches should allow explorationof the conditions under which different interaction strengths metrics are interchange-able with regard to relative magnitude, system responses, and species identity. 5. Finally, the prime focus on predator–prey links in much of the research to date oninteraction strengths in food webs has meant that the potential significance of non-trophic interactions, such as competition, facilitation and biotic disturbance, has beenlargely ignored by the food web community. Such interactions may be importantdynamically and should be routinely included in future food web research programmes. Key-words : allometry, body size, ecological networks, interaction strength, keystonespecies, population dynamics, stability. Journal of Animal Ecology (2004) 73 , 585–598 Correspondence: Eric L. Berlow (E-mail: [email protected]); Joel E. Cohen (E-mail: [email protected]); Giorgos D. Kokkoris,(E-mail: [email protected]).586 E. L. Berlow et al. © 2004 British Ecological Society, Journal of Animal Ecology , 73 ,585–598 Introduction Food webs have long been a central concept in ecologyand are useful because they provide tractable abstrac-tions of the complexity and interconnectedness of nat-ural communities that potentially transcend system-specific detail. While potentially unifying, the food webconcept has proved scientifically divisive over the pastthree decades. Those interested in predicting speciespopulation dynamics have suggested that the structureof unweighted links alone tells little about the outcomeof a species perturbation (Levine 1976; Holt 1977;Paine 1980; Vandermeer 1980; Abrams 1987; Paine1988; Polis 1991), while others claim that generalitiesin unweighted link structure transcend the spatiallyand temporally variable details of individual speciesdynamics (Cohen 1978; Briand & Cohen 1987; Sugihara et al . 1989; Cohen et al . 1990; Martinez 1994).The discovery of keystone predation was among thefirst empirical examples that demonstrated dramat-ically how unweighted link structure by itself is nota good predictor of species and population dynamics(Paine 1969, 1974, 1980). Ironically, this example isalso one of the first to illustrate the critical dependence of dynamics on web structure. Keystone predationis not a single strong predator–prey interaction,but rather a particular configuration or structuralorganization of strong and weak links: strong preda-tion, relative to other predators of that guild, on acompetitively dominant prey species. It demonstrateshow a combined knowledge of both web structure andinteraction strengths is a key to understanding howecological communities function. Characterizing andabstracting this relationship between web structure,interaction strengths and population dynamics allowedothers to identify the presence (or predictable absence)of keystone effects in other sites and communities(Estes & Palmisano 1974; Lubchenco 1978; Castilla &Duran 1985; Paine et al . 1985; Carpenter & Kitchell1993).General patterns of food web structure also appearto be an emergent property of dynamical constraintson species interactions (e.g. Bastolla et al . 2001; Drossel et al . 2001; Fox & McGrady-Steed 2002; Montoya& Solé 2003). With interaction strengths assigned atrandom from a specified distribution, early theoreticalwork suggested that weak interactions are necessary fordiverse