Stronger inducible defences enhance persistence of intraguild prey

Journal of Animal Ecology 2010, 79, 993–999

doi: 10.1111/j.1365-2656.2010.01705.x

Stronger inducible defences enhance persistence of intraguild prey Pavel Kratina1,2*, Edd Hammill1,3 and Bradley R. Anholt1,4 1

Department of Biology, University of Victoria, PO Box 3020, Victoria, British Columbia V8W 3N5, Canada; 2Biodiversity Research Centre and Zoology Department, University of British Columbia, 6270 University Boulevard, Vancouver, British Columbia V6T 1Z4, Canada; 3Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK; and 4Bamfield Marine Sciences Centre, Bamfield, British Columbia V0R 1B0, Canada

Summary 1. Intraguild predation is widespread in nature despite its potentially destabilizing effect on food web dynamics. 2. Anti-predator inducible defences affect both birth and death rates of populations and have the potential to substantially modify food web dynamics and possibly increase persistence of intaguild prey. 3. In a chemostat experiment, we investigated the long-term effects of inducible defences on the dynamics of aquatic microbial food webs consisting of an intraguild predator, intraguild prey, and a basal resource. We controlled environmental conditions and selected strains of intraguild prey that varied in the strength of expressed inducible defences. 4. We found that intraguild prey with a stronger tendency to induce an anti-predator morphology persist for significantly longer periods of time. In addition, model selection analysis implied that flexibility in defensive phenotype (inducibility itself) is most likely the factor responsible for the enhanced persistence. 5. As patterns at the community level often emerge as a result of the life-history traits of individuals, we propose that inducible defences increase the persistence of populations and may contribute to the widespread occurrence of theoretically unstable intraguild predation systems in nature. Key-words: extinction, food web dynamics, omnivory, protozoa, stability

Introduction Intraguild predation is a form of omnivory where a predator consumes an intermediate prey, as well as the resource of this prey (Polis, Myers & Holt 1989; Holt & Polis 1997). Although food webs with intraguild predation are considered inherently unstable, tending to rapid species extinctions (Pimm & Lawton 1977; Holt & Polis 1997; but see also Krivan 2000), intraguild predation is common in natural communities (Polis, Myers & Holt 1989; Arim & Marquet 2004), and most species above the herbivore trophic level are omnivorous (Thompson et al. 2007). Several factors have been proposed to promote stability and persistence of such systems. For example, persistence of food web structure is enhanced when an intraguild predator facilitates the growth of the intermediate prey, the intermediate prey is an essential food source for the intraguild predator, or refuges against predators are included (HilleRisLambers, van de Koppel & Herman 2006). Because anti-predator inducible defences also *Correspondence author. E-mail: [email protected]

protect prey and could be considered a form of prey refuge, they may be an important ecological mechanism promoting the persistence of food webs with intraguild predation (Holt & Polis 1997; Kimbrell, Holt & Lundberg 2007). Strong selection pressure applied by consumers on their resources has driven the evolution of a vast array of defence mechanisms, which can be broadly grouped into two core categories; constitutive and inducible. As opposed to permanently expressed constitutive defences, inducible defences are only expressed following cues from consumers, competitors or parasites (Harvell 1990; Harvell & Tollrian 1999). Such defences occur across a wide variety of taxa including plants (Karban & Baldwin 1997), unicellular organisms (Kuhlmann & Heckmann 1985), invertebrates (Gilbert 1966) and vertebrates (Bro¨nmark & Miner 1992; Werner & Anholt 1996), and induce shifts in the morphology, physiology, life-history, and ⁄ or behaviour of prey species. Inducible anti-predator defences ensure greater flexibility in biotic environments where the impact of natural consumers varies spatially and temporally and where constitutive defences are costly. The vast bulk of work on inducible defences has focused on short-

 2010 The Authors. Journal compilation  2010 British Ecological Society

994 P. Kratina, E. Hammill & B. R. Anholt term consequences for species survival and growth. Longterm predictions about intraguild predation are also often drawn from short-term studies due to logistical constraints (Briggs & Borer 2005). Although historically rare, multigenerational experiments are often necessary to avoid erroneous conclusions (Briggs & Borer 2005). Theoretical studies suggest that inducible defences have the potential to increase the long-term persistence of simple systems. Inducible defences have been shown to stabilize one predator–one prey models (Abrams 1982, 1984; Ives & Dobson 1987; Abrams & Walters 1996; Abrams & Matsuda 1997; Vos et al. 2004a). Increasing complexity slightly, predator-induced defences increased population persistence in a tritrophic food chain model parameterized to a rotifer-algae system (Vos et al. 2004a). More complex results were obtained from an alternative formulation of a tritrophic food chain discriminating between inducible defences affecting either attack rate or handling time of predators (RamosJiliberto et al. 2008). This model was also generally stabilized by inducible defences. Conversely, a destabilizing effect is often predicted by models which incorporate inducible defences but have considerable time-lags between the onset of predation risk and induction or relaxation of defences (Underwood 1999, Luttbeg & Schmitz 2000; Miner et al. 2005). Lagged time-scale responses are, however, not always destabilizing. Time-lags, for instance, did not affect stability of models that considered behavioral inducible defences (Abrams 1992). The importance of inducible defences has been shown empirically in linear food chains (Verschoor, Vos & van der Stap 2004; van der Stap, Vos & Mooij 2006; van der Stap et al. 2007, 2008). These experimental examples show how induced defensive spines in herbivorous rotifers reduced the strength of trophic cascades (van der Stap et al. 2007). Morphological inducible defences in the basal trophic level (algae) removed strong population fluctuations and increased species persistence (Verschoor, Vos & van der Stap 2004; van der Stap et al. 2008). However, defences in herbivores did not affect stability of food chains (Verschoor, Vos & van der Stap 2004). As opposed to above-cited studies comparing the effects of inducible defences to the effect of constitutive or no defences, we investigated how inducible defences of different strengths affect species persistence. Food webs containing both intraguild predation and different magnitudes of inducible defences have not been the subject of either theoretical or empirical studies even though both mechanisms have the potential to modulate species persistence. Although many studies focus on a single ecological mechanism, it is still unclear how these two important mechanisms combine to affect the persistence of natural communities. There is a clear need for more multi-causal approaches and multigenerational experiments in ecology (Vos et al. 2004b; Briggs & Borer 2005; Amarasekare 2007). In this study we investigate the dynamics of food webs that incorporate both intraguild predation and inducible defences. Our experimental food webs consist of the turbellarian flatworm, Stenostomum virginianum, (intraguild preda-

tor) feeding on hypotrich ciliates of the genus Euplotes spp. (intraguild prey) and the unicellular algae Rhodomonas minuta (basal resource). In the presence of predatory flatworms, Euplotes quickly change their morphology from an undefended morph to a circular, defended morph (Kuhlmann & Heckmann 1985). Defended morphs have a considerably greater body width than undefended individuals and are less likely to be swallowed by gape-limited predators (Kuhlmann & Heckmann 1994; Altwegg et al. 2006). Morphological changes are initiated within 2–4 h and maximum induction is usually reached within 18–24 h (Duquette, Altwegg & Anholt 2005). In the present study, we compared the population persistence of two highly inducible and two less inducible clones of Euplotes spp. and quantified the relationship between inducibility and persistence time. Our objective was to assess whether clones with a higher ability to induce defensive anti-predator morphology (more plastic clones) also have longer persistence times in food webs with intraguild predation.

Materials and methods VARIATION IN INDUCIBILITY

In the first experiment we measured variation in inducibility of four different Euplotes spp. clones. We exposed Euplotes to Stenostomum cue for 24 h and measured inducibility as an increase in the maximum body width – the difference between the body widths of individuals incubated with and without predator cue. Euplotes were originally obtained from K. Wiackowski (Jagiellonian University, Krakow, Poland) or isolated from a freshwater pond on the University of Victoria campus. First, we introduced 100 Euplotes in 400 lL of protozoan media to 24-well tissue culture plates (Costar, Corning, New York, USA). The media consisted of 1Æ5 crushed protozoan pellets (0Æ7 g each; Nr. 13-2360, Carolina Biological Supply Company, NC) dissolved in 2 L NAYA mineral water (Mirabel, Que´bec, Canada). We then added 400 lL of predator cue consisting of freezekilled Stenostomum at a density of 250 per 1 mL. We systematically scanned the bottom of each experimental well and photographed the first 12 Euplotes cells encountered using a Cohu CCD camera (San Diego, CA, USA) connected to an inverted microscope (Leica DM IRB; Leica Microsystems, Ontario, Canada). We measured the maximum cell width using Image Pro Plus 4.5 image analysis software (Media Cybernetics, Silver Spring, MD, USA). Twelve measurements were taken from each well and their median used as one replicate. Each clone was replicated three times, except clone AED33, which was replicated two times. Data were analyzed with ancova using original body size (without predator cue) as a covariate. As the covariate was not significant, we removed the term and repeated the analysis with anova and post hoc Tukey’s tests for multiple comparisons in R software, version 2.6.0 (R Development Core Team 2007).

DYNAMICS OF EXPERIMENTAL FOOD WEBS

In the second experiment, we established 12 intraguild food webs using one of the four Euplotes clones, Stenostomum, and Rhodomonas. It has been previously observed that Stenostomum can feed on algae exclusively and survive for many generations without Euplotes prey (P. Kratina, unpublished data). We incubated the experimental food webs in continuous flow-through chemostats over 40 Euplotes

 2010 The Authors. Journal compilation  2010 British Ecological Society, Journal of Animal Ecology, 79, 993–999

Intraguild prey with inducible defences 995 or 20 Stenostomum generations (i.e. 58 days). Culture vessels held a constant volume of 200 mL Bold’s Basal Medium (BBM; Stein 1973) and dilution rate was set to d = 0Æ071 ± 0Æ001 SE. The chemostats were submerged in a water bath to maintain a constant temperature (21 ± 0Æ5C), and permanently illuminated to facilitate algal growth. Euplotes have been shown to persist under these conditions at equilibrium densities of about 300–500 individuals per 10 mL when Stenostomum were absent (P. Kratina and B. Anholt, unpublished results). Each of the four different Euplotes clones was replicated three times. Rhodomonas algae were introduced at density of 19Æ3 · 104 cells per chemostat vessel. The following day we added 4Æ5 · 103 Euplotes and 9 days after adding the algae we added 100 Stenostomum. We used sterile techniques throughout to prevent bacterial contamination. We sampled 10 mL (5%) of well-mixed culture directly from the culture vessels every second day at the same time (1400 h) and determined the densities of all three species. Algae were counted immediately using a particle size analyzer (Elzone II5390, Micromeritics, Norcross, GA, USA). Intraguild prey and predators were fixed with Lugol’s solution and counted under a dissecting microscope (Leica MZ8; Leica Microsystems, Ontario, Canada). Over the course of the experiment as well as in a similar studies performed in our laboratory, Euplotes induced anti-predator morphology and the magnitude of the induction varied over time (Altwegg et al. 2004; personal observations). At the end of the experiment the entire contents of each culture vessel was examined and all remaining Euplotes and Stenostomum were counted.

MEASURES OF POPULATION PERSISTENCE

We tested for significant differences in persistence time, measured as the day on which the Euplotes sample was first recorded as zero, using parametric survival analysis with log-normal error structure (Crawley 2007). In order to assess whether enhanced persistence is caused by larger inducibility, we chose highly reactive clones as well as clones with a smaller ability to induce, both represented by one large and one small clone. We constructed a series of survival analysis models incorporating inducibility, initial and final body sizes as explanatory variables (Table 1) and selected the best model using the sample-size adjusted Akaike’s Information Criterion and log-likelihood values (Burnham & Anderson 2002). Although some ecological factors other than defences (such as generation time or swimming speed) might lead to faster extinctions of less inducible types, such covariates are most likely related to Euplotes size. Survival analysis is recommended for studies of this type as it accounts for censored data

(microcosms where Euplotes persisted until the end of the experiment) and does not assume a constant variance (Crawley 2007). As variance in time to first recorded zero density is expected to increase with the mean (Fleming & Harrington 1991), this method is superior to using a simple anova. We used the functions surv() and psm() in the R programming language (R Development Core Team 2007) and followed the methods outlined in Crawley (2007). We also recorded the total time each Euplotes population density was below the detection limit (0 individuals in a 10 mL sample) during the course of the experiment. Initial analyses indicated nonhomogeneous error variances in regression models predicting both measures of persistence by inducibility. We employed nonparametric bootstrapping (1000 replications) to obtain robust estimates for linear regression model parameters. Similar measures of persistence were previously used in the analyses of other model consumerresource systems (van der Stap et al. 2008; van Veen, Brandon & Godfray 2009).

Results MORPHOLOGY

We detected significant differences among Euplotes clones in their ability to increase maximum body width after 24-hr exposure to the predator cue (one-way anova, F(3, 7) = 9Æ96, P < 0Æ01, r2 = 0Æ81, Tukey’s HSD, P = 0Æ01, Fig. 1). Two clones with high inducibility, B8 and S5-1 increased their mean body widths from 49Æ93 lm (±1Æ66 SE) by 26Æ24 lm (±1Æ09 SE) and from 66Æ41 lm (±2Æ06 SE) by 23Æ67 lm (±1Æ67 SE), respectively. The two clones with low inducibility, SC8 and AED33 increased their mean body widths of 51Æ85 lm (±1Æ94 SE) by 18Æ12 lm (±0Æ74 SE) and 68Æ53 lm (±0Æ89 SE) by 17Æ77 lm (±1Æ77 SE), respectively (Fig. 1). Initial sizes and levels of induction in these four clones allowed us to measure the persistence of one large and one small clone with high inducibility and a large and a small clone with low inducibility.

PERSISTENCE TIME

Striking differences in population densities through time were detected among the four Euplotes clones (Fig. 2). The

Table 1. Model parameters and their significance for different survival analysis models of persistence time in the Euplotes clones. Initial body size, inducibility and final size all treated as linear variables. All models were fitted to the data by maximum likelihood and compared by the sample-size adjusted Akaike’s Information Criterion (AICc). The Akaike weights give the relative strength of support for one model over another within the model set; the best model is shown in italics. The model with initial size + inducibility + final size was not tested because the final size is a linear combination of the first two factors

Model

Parameter significance

z-score

Model v2

d.f.

Model P-value

AICc

Akaike weight

Inducibility only Initial size only Final size only Initial size + inducibility

Inducibility (P < 0Æ001) Initial size (P = 0Æ056) Final size (P = 0Æ497) Initial size (P = 0Æ141) Inducibility (P = 0Æ003) Final size (P = 0Æ141) Inducibility (P < 0Æ001)

3Æ34 )1Æ91 )0Æ68 )1Æ47 3Æ00 )1Æ47 3Æ77

7Æ96 3Æ23 0Æ45 9Æ97

1 1 1 2

0Æ0048 0Æ0720 0Æ5000 0Æ0069

77Æ93 82Æ73 85Æ53 79Æ60

0Æ505 0Æ046 0Æ011 0Æ220

9Æ97

2

0Æ0069

79Æ60

0Æ220

Final size + inducibility

 2010 The Authors. Journal compilation  2010 British Ecological Society, Journal of Animal Ecology, 79, 993–999

996 P. Kratina, E. Hammill & B. R. Anholt

Density (Ind/10 mL)

Fig. 1. Inducibility measured as an increase in the maximum body width of four Euplotes clones after 24-hr exposure to predator cue. Error bars are ±1 SE from the mean of three replicates (n = 2 for the clone AED33).

(a)

Low inducibility

(b)

Low inducibility

(c)

High inducibility

(d)

High inducibility

order in which the four clones dropped to the mean density of one individual per sample closely followed their inducibility. We predicted Euplotes persistence times using models containing various combinations of the following explanatory variables – inducibility, initial body size, and final body size (Table 1). In all survival analyses the response variable was the first day when zero Euplotes were recorded. Neither initial body size nor final body size had a significant effect on persistence time in any model (Table 1). The effect of inducibility was, however, significant, with clones expressing higher levels of inducibility surviving longer (Table 1, Fig. 3a, parametric survival analysis, c2 = 7Æ96, d.f. = 1, P = 0Æ0048). The best model predicting persistence, selected using the sample-size adjusted Akaike’s Information Criterion, included only inducibility. This model was supported 2Æ3 times more than the models that also included initial or final Euplotes body width (Table 1).

Days Fig. 2. Dynamics of twelve intraguild food webs incubated in continuous chemostats; filled circles (d) represent hypotrich ciliates of the genus Euplotes, triangles (D) algae Rhodomonas minuta, and open circles (s) denote predatory turbellarians Stenostomum virginianum. Algal densities are expressed in hundreds of individuals per 10 mL. Different Euplotes clones (a) AED33, (b) SC8, (c) S5-1 and (d) B8, vary in their ability to induce morphological defense in the presence of the predators. Samples were taken and all three species were counted every 2 days. Stenostomum were introduced 8 days after the introduction of Euplotes and the first predator counts started next day (absence of the symbols reflect 0 individuals detected in a 10 mL sample on that particular day).  2010 The Authors. Journal compilation  2010 British Ecological Society, Journal of Animal Ecology, 79, 993–999

Intraguild prey with inducible defences 997 (a)

(b)

Fig. 3. (a) Time since the start of the experiment when Euplotes first dropped below detection limit (regression slope estimated from nonparametric bootstrapping b = 1Æ69, empirical 95% CL are 0Æ37–3Æ02). and (b) total time that Euplotes spent below detection limit during the experiment (estimated regression slope b = )1Æ59, empirical 95% CL are from )2Æ89 to )0Æ09) as a function of their inducibility. Error bars are ±1 SE from the mean, n = 3 replicated food webs.

Inducibility had a strong effect on both measures of persistence. The first drop below detection limit in Euplotes clones increased with increasing inducibility (regression slope estimated from nonparametric bootstrapping b = 1Æ69, empirical 95% confidence limits are 0Æ37–3Æ02; Fig. 3a). In contrast, the amount of time Euplotes spent below detection limit decreased with increasing prey inducibility (regression slope b = )1Æ59, empirical 95% confidence limits are from )2Æ89 to )0Æ09; Fig. 3b). Although clone SC8 dropped very quickly to low densities, we were able to detect one individual per sample in one replicate for a long period of time, reducing the mean time spent below the detection limit. In contrast, clone S5-1 declined to low densities more slowly than clone SC8 but no individuals were detected after day 42 of the experiment.

Discussion Natural food webs are assemblages of species structured by complex interactions between predators and their prey. It is recognized theoretically that if these interactions are moderate or weak, systems are more likely to be stable (McCann, Hastings & Huxel 1998). The strength of species interactions can be reduced by including non-prey species (Vos 2001; Kratina, Vos & Anholt 2007) or by inducible defences rapidly expressed by prey (Vos et al. 2004a,b).In the present study, we investigated whether stronger inducible defences in Euplotes convey longer persistence of theoretically unstable food webs with intraguild predation (Pimm & Lawton 1977; Holt & Polis 1997). In our long-term experiment we detected a significant relationship between inducibility and persistence time of intraguild prey. Although extinction of the intraguild prey eventually did occur in most of the vessels (except for two replicates of the clone expressing the strongest inducibility), an increased ability to induce defensive morphology significantly prolonged prey persistence times. Interestingly, neither

the initial nor the final size of Euplotes had a significant effect on persistence, suggesting that the ability to modify body size has a greater effect on clone persistence than body size per se. This is further supported by the fact that the clone surviving the longest (B8), and showing the greatest inducibility, had a smaller post-induction size than post-induction AED33, the clone inducing the smallest change and persisting for the shortest time. Theory predicts that whether or not adaptively flexible defences are stabilizing largely depends on a time scale over which phenotype switching occurs relative to the time scale of species birth and death. Although adaptive plasticity can destabilize some models of population dynamics due to the time-lags (Abrams 1999), fast temporal responses to changing conditions often have stabilizing effects (Miner et al. 2005). Rapid inducible responses of Euplotes prey to Stenostomum predators (Duquette, Altwegg & Anholt 2005) are likely to increase system stability. Our results imply that species or clones with a larger ability to change their defensive phenotype are more likely to survive in a system with intraguild predation than less flexible species or clones. In such systems, intraguild prey face a risk of extinction not only through predation, but also competition. The clone with the greatest ability to be large and inedible in times of high predation risk, but also be small in times of strong competition should be the most likely to persist, as was observed in this study. Therefore, in systems with intraguild predation, we would expect the most plastic intraguild prey species to be the most successful. Another possible explanation for the patterns we observed is that more reactive clones able to produce the greatest range of potential sizes can achieve body widths closer to the optimal size for the current food web conditions than slower or less reactive clones. Multiple mechanisms delaying extinctions and allowing many species to coexist are likely to work simultaneously in natural settings. It is also important to notice that the differences in the mean population trajectories were detected especially at lower

 2010 The Authors. Journal compilation  2010 British Ecological Society, Journal of Animal Ecology, 79, 993–999

998 P. Kratina, E. Hammill & B. R. Anholt Euplotes densities, where the dynamics may have been also affected by stochastic factors. Phenotypic plasticity in intraguild prey may not only change interactions with the intraguild predator, but may also influence interactions with the basal resource. Duquette, Altwegg & Anholt (2007) found the feeding rate was similar for induced and uninduced forms in two low reactive Euplotes clones. However, the clone with the largest morphological defense showed reduced foraging efficiency after induction of its anti-predator morphology. Despite this possible fitness cost, the two clones in this study with the largest ability to modify their defensive morphology after exposure to predators persisted significantly longer in our experiment than the two clones with weaker inducible defences. This implies that the relative advantage of inducible defences is possibly stronger than the disadvantage of reduced foraging in food webs with intraguild predators. The population dynamics of the top predator and the basal resource did not differ among treatments. Strong foraging efficiency of Stenosomum feeding on Euplotes has been documented (Kusch 1995; Altwegg et al. 2006) and is the most plausible explanation for the decline of Euplotes in this study. The intraguild predators, Stenostomum, were detected in all microcosms at a density of 35 ± 11 (mean ± SE, n = 12) predators per chemostat after the termination of the experiment, surviving exclusively on algae. Although the chemostats were well mixed by aeration, predator densities were probably underestimated due to Stenosomum’s affinity for glass or plastic walls (Kratina et al. 2009; personal observations). In a previous study of rotifers feeding on algae (Verschoor, Vos & van der Stap 2004), inducible defences at the basal trophic level of linear aquatic food chains increased the food web persistence by moving minimum population densities further away from zero. It has been suggested that inducible defences at the middle trophic level have substantially weaker stabilizing effects (Verschoor, Vos & van der Stap 2004; Vos et al. 2004a) than those at basal trophic levels. In this study, we bring some evidence that rapid inducible defences at the middle trophic level may enhance the persistence of intraguild prey, in particular when these defences are stronger. Our findings thus contribute to a more complete understanding of the interactive effects of intraguild predation and inducible defences in modules of real food webs. Recent studies show that increased persistence of intraguild predation in real food webs takes place by processes that either enhance the resources available to intraguild prey or reduce the strength of interaction between the predator and intraguild prey (Amarasekare 2007; Daugherty, Harmon & Briggs 2007; Janssen et al. 2007; Kondoh 2008). We propose that inducible defences with rapid reaction times are one of the ecological mechanisms able to promote persistence of intraguild predation systems by reducing the interaction strength between the predator and intraguild prey. Given that inducible defences are widespread, the results of our study may have broad implications across many taxa and ecosystems. As no previous ecological models considered

dynamical implications of inducibility per se, our study also opens an avenue for novel theoretical work.

Acknowledgements We thank Anita Narwani, Jonathan Shurin, and Matthijs Vos for valuable comments on previous drafts of this manuscript. We are also grateful to Jeremy Fox for helping to place our results in a broader theoretical concept. This research was funded by the Canada Research Chairs Program and an NSERC of Canada Discovery grant to B.R.A.

References Abrams, P.A. (1982) Functional responses of optimal foragers. American Naturalist, 120, 382–390. Abrams, P.A. (1984) Foraging time optimization and interactions in food webs. American Naturalist, 124, 80–96. Abrams, P.A. (1992) Predators that benefit prey and prey that harm predators: unusual effects of interacting foraging adaptation. American Naturalist, 140, 573–600. Abrams, P.A. (1999) The adaptive dynamics of consumer choice. American Naturalist, 153, 83–97. Abrams, P.A. & Matsuda, H. (1997) Prey adaptation as a cause of predatorprey cycles. Evolution, 51, 1742–1750. Abrams, P.A. & Walters, C.J. (1996) Invulnerable prey and the paradox of enrichment. Ecology, 77, 1125–1133. Altwegg, R., Marchinko, K.B., Duquette, S.L. & Anholt, B.R. (2004) Dynamics of an inducible defence in the protist Euplotes. Archiv fu¨r Hydrobiologie, 160, 431–446. Altwegg, R., Eng, M., Caspersen, S. & Anholt, B.R. (2006) Functional response and prey defence level in an experimental predator-prey system. Evolutionary Ecology Research, 8, 115–128. Amarasekare, P. (2007) Trade-offs, temporal variation, and species coexistence in communities with intraguild predation. Ecology, 88, 2720–2728. Arim, M. & Marquet, P.A. (2004) Intraguild predation: a widespread interaction related to species biology. Ecology Letters, 7, 557–564. Briggs, C.J. & Borer, E.T. (2005) Why short-term experiments may not allow long-term predictions about intraguild predation. Ecological Applications, 15, 1111–1117. Bro¨nmark, C. & Miner, J.G. (1992) Predator-induced phenotypical change in body morphology in crucian carp. Science, 258, 1348–1350. Burnham, K.P. & Anderson, D.R. (2002) Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. Springer-Verlag, New York, New York, USA. Crawley, M.J. (2007) The R Book. John Wiley and Sons, Chichester, UK. Daugherty, M.P., Harmon, J.P. & Briggs, C.J. (2007) Trophic supplements to intraguild predation. Oikos, 116, 662–677. Duquette, S.L., Altwegg, R. & Anholt, B.R. (2005) Factors affecting the expression of inducible defences in Euplotes: genotype, predator density and experience. Functional Ecology, 19, 648–655. Duquette, S., Altwegg, R. & Anholt, B.R. (2007) Protozoan functional responses: effects of species, genotype, and anti-predator defences. Evolutionary Ecology Research, 9, 789–800. Fleming, T. & Harrington, D. (1991) Counting processes and survival analysis. John Wiley and Sons, New York, USA. Gilbert, J.J. (1966) Rotifer ecology and embryological induction. Science, 151, 1234–1237. Harvell, C.D. (1990) The ecology and evolution of inducible defences. Quarterly Review of Biology, 65, 323–340. Harvell, C.D. & Tollrian, R. (1999) Why inducible defences? The Ecology and Evolution of Inducible Defenses (eds R. Tollrian & C.D. Harvell), pp. 3–9. Princeton University Press, Princeton, New Jersey, USA. HilleRisLambers, R., van de Koppel, J. & Herman, P.M.J. (2006) Persistence despite omnivory: benthic communities and the discrepancy between theory and observation. Oikos, 113, 23–32. Holt, R.D. & Polis, G.A. (1997) A theoretical framework for intraguild predation. American Naturalist, 149, 745–764. Ives, A.R. & Dobson, A.P. (1987) Antipredator behavior and the population dynamics of simple predator-prey systems. American Naturalist, 130, 431– 447. Janssen, A., Sabelis, M.W., Magalhaes, S., Montserrat, M. & Van der Hammen, T. (2007) Habitat structure affects intraguild predation. Ecology, 88, 2713–2719.

 2010 The Authors. Journal compilation  2010 British Ecological Society, Journal of Animal Ecology, 79, 993–999

Intraguild prey with inducible defences 999 Karban, R. & Baldwin, I.T. (1997) Induced Responses to Herbivory. The University of Chicago Press, Chicago, IL, USA. Kimbrell, T., Holt, R.D. & Lundberg, P. (2007) The influence of vigilance on intraguild predation. Journal of Theoretical Biology, 249, 218–234. Kondoh, M. (2008) Building trophic modules into a persistent food web. Proceedings of the National Academy of Sciences of the United States of America, 105, 16631–16635. Kratina, P., Vos, M. & Anholt, B.R. (2007) Species diversity modulates predation. Ecology, 88, 1917–1923. Kratina, P., Vos, M., Bateman, A. & Anholt, B.R. (2009) Functional responses modified by predator density. Oecologia, 159, 425–433. Krivan, V. (2000) Optimal intraguild foraging and population stability. Theoretical Population Biology, 58, 79–94. Kuhlmann, H.W. & Heckmann, K. (1985) Interspecific morphogens regulating prey-predator relationships in protozoa. Science, 227, 1347–1349. Kuhlmann, H.W. & Heckmann, K. (1994) Predation risk of typical ovoid and winged morphs of Euplotes (Protozoa, Ciliophora). Hydrobiologia, 284, 219–227. Kusch, J. (1995) Addaptation of inducible defence in Euplotes daidaleos (Ciliophora) to predation risk by various predators. Microbial Ecology, 30, 79–88. Luttbeg, B. & Schmitz, O.J. (2000) Predator and prey models with flexible individual behavior and imperfect information. American Naturalist, 155, 669– 683. McCann, K., Hastings, A. & Huxel, G.R. (1998) Weak trophic interactions and the balance of nature. Nature, 395, 794–798. Miner, B.G., Sultan, S.E., Morgan, S.G., Padilla, D.K. & Relyea, R.A. (2005) Ecological consequences of phenotypic plasticity. Trends in Ecology and Evolution, 20, 685–692. Pimm, S.L. & Lawton, J.H. (1977) Number of trophic levels in ecological communities. Nature, 268, 329–331. Polis, G.A., Myers, C.A. & Holt, R.D. (1989) The ecology and evolution of intraguild predation: potential competitors that eat each other. Annual Review of Ecology and Systematics, 20, 297–330. R Development Core Team (2007) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL: http://www.R-project.org. Ramos-Jiliberto, R., Mena-Lorca, J., Flores, J.D. & Morales-Alvarez, W. (2008) Role of inducible defences in the stability of a tritrophic system. Ecological Complexity, 5, 183–192.

van der Stap, I., Vos, M. & Mooij, W.M. (2006) Linking herbivore-induced defences to population dynamics. Freshwater Biology, 51, 424–434. van der Stap, I., Vos, M., Verschoor, A.M., Helmsing, N.R. & Mooij, W.M. (2007) Induced defences in herbivores and plants differentially modulate a trophic cascade. Ecology, 88, 2474–2481. van der Stap, I., Vos, M., Tollrian, R. & Mooij, W.M. (2008) Inducible defences, competition and shared predation in planktonic food chains. Oecologia, 157, 697–705. Stein, J. (1973) Handbook of Phycological Methods. Culture Methods and Growth Measurements. Cambridge University Press, Cambridge, UK. Thompson, R.M., Hemberg, M., Starzomski, B.M. & Shurin, J.B. (2007) Trophic levels and trophic tangles: the prevalence of omnivory in real food webs. Ecology, 88, 612–617. Underwood, N. (1999) The influence of plant and herbivore characteristics on the interaction between induced resistance and herbivore population dynamics. American Naturalist, 152, 282–294. van Veen, F.J.F., Brandon, C.E. & Godfray, H.C.J. (2009) A positive trait-mediated indirect effect involving the natural enemies of competing herbivores. Oecologia, 160, 195–205. Verschoor, A.M., Vos, M. & van der Stap, I. (2004) Inducible defences prevent strong population fluctuations in bi- and tritrophic food chains. Ecology Letters, 7, 1143–1148. Vos, M. (2001) Plant-mediated indirect effects and the persistence of parasitoid-herbivore communities (vol 4, pg 38, 2001). Ecology Letters, 4, 38– 45. Vos, M., Kooi, B.W., DeAngelis, D.L. & Mooij, W.M. (2004a) Inducible defences and the paradox of enrichment. Oikos, 105, 471–480. Vos, M., Verschoor, A.M., Kooi, B.W., Wackers, F.L., DeAngelis, D.L. & Mooij, W.M. (2004b) Inducible defences and trophic structure. Ecology, 85, 2783–2794. Werner, E.E. & Anholt, B.R. (1996) Predator-induced behavioral indirect effects: consequences to competitive interactions in anuran larvae. Ecology, 77, 157–169. Received 25 September 2009; accepted 16 April 2010 Handling Editor: Kevin McCann

 2010 The Authors. Journal compilation  2010 British Ecological Society, Journal of Animal Ecology, 79, 993–999