Assessing Feeding Preferences of a Consumer Guild: Partitioning Variation Among Versus Within Species

vol. 192, no. 3 the american naturalist september 2018

Assessing Feeding Preferences of a Consumer Guild: Partitioning Variation Among versus Within Species

O. Kennedy Rhoades,1,* Rebecca J. Best,2 and John J. Stachowicz3

1. Bodega Marine Laboratory, University of California, Davis, Bodega Bay, California 94923; and Smithsonian National Museum of Natural History, Smithsonian Marine Station at Fort Pierce, Florida 34949; 2. School of Earth Sciences and Environmental Sustainability, Northern Arizona University, Flagstaff, Arizona 86011; 3. Department of Evolution and Ecology, University of California, Davis, California 95616 Submitted September 1, 2017; Accepted March 1, 2018; Electronically published July 20, 2018 Online enhancements: supplemental material. Dryad data: http://dx.doi.org/10.5061/dryad.5nq06fj. abstract: Interspecific variation in resource use is critical to un- predicts population persistence (Werner et al. 1983), species derstanding species diversity, coexistence, and ecosystem functioning. interactions (Sommer et al. 2001; Burkepile and Hay 2008; A growing body of research describes analogous intraspecific varia- Rasher et al. 2013), species coexistence (Chesson 2000), and tion and its potential importance for population dynamics and com- community assembly (Grant and Grant 2006). Similarly, a munity outcomes. However, the magnitude of intraspecific variation rel- growing body of work indicates that individuals within species ative to interspecific variation in key dimensions of consumer-resource consistently vary in their resource use (Bolnick et al. 2003; Sih interactions remains unknown, hampering our understanding of the im- et al. 2004a; Vellend and Geber 2005). Individual variation in portance of this variation for population and community processes. In fl this study, we examine feeding preference through repeated laboratory resource use in uences population persistence (Hughes and choice feeding assays of 444 wild-caught individuals of eight invertebrate Stachowicz 2011; Forsman and Wennersten 2016), ecological grazer species on rocky reefs in northern California. Between-species var- processes and community structure (Crutsinger et al. 2006; iation accounted for 25%–33% of the total variation in preference for the Hughes et al. 2008), community assembly (Jung et al. 2010; preferred resource, while between-individual variation accounted for 4%– Siefert 2012; Kraft et al. 2014), community stability (Agashe 5% of total variation. For two of the eight species, between-individual var- 2009; Pruitt et al. 2016; Wood et al. 2017), species coexistence fi iation was signi cantly different from zero and on average contributed 14% (Vellend 2006; Lichstein et al. 2007; Lankau 2009), and food and 17% of the total diet variation, even after accounting for differences due to size and sex. Therefore, even with clearly distinguishable between- web structure (Barbour et al. 2016).Yet itisoften assumedthat species differences in mean preference, diet variation between and within phenotypic variation among individuals is negligible com- individuals can contribute to the dietary niche width of species and guilds, pared with that among species (Lister 1976; Taper and Case which may be overlooked by focusing solely on species’ mean resource 1985; Schoener 1986; McGill et al. 2006; Petchey and Gaston use patterns. 2006). This assumption may be premature, given the limited fi fi number of empiricalassessments of intraspeci c diversity that Keywords: intraspeci c variation, feeding preference, dietary niche, fi herbivory, rocky shore, kelp forest. quantify the relative proportion of intraspeci c versus interspecific variation in resource use across an entire guild (Bol- nick et al. 2011; Siefert 2012; Kamath and Losos 2017). Also, Introduction existing assessments of individual specialization are likely biasedinfavorofspeciesknowntohaveconsiderableintraspe- Population dynamics and community structure are strongly cific variation, which may overestimate the general nature of influenced by the degree to which individuals and species spe- these effects (Bolnick et al. 2003). Although plant ecologists cialize on and partition available resources (Macarthur and have made significant progress in assessing the proportion of Levins 1967). Species vary broadly in their resource use within trait variation among and within species (Siefert et al. 2015), communities (Schoener 1974), and resource partitioning we have much less understanding of how these different levels of variation matter for species interactions, especially across trophic levels. * Corresponding author; email: [email protected]. In contrast with many other ecological traits (e.g., indirect ORCIDs: Rhoades, http://orcid.org/0000-0003-1456-1581; Best, http://orcid proxies of resource use, such as morphology and physiology), .org/0000-0003-2103-064X; Stachowicz, http://orcid.org/0000-0003-2735-0564. diet is a direct measure of consumer resource use for popula- Am. Nat. 2018. Vol. 192, pp. 000–000. q 2018 by The University of Chicago. 0003-0147/2018/19203-57924$15.00. All rights reserved. tions and species that also impacts the coexistence of compet- DOI: 10.1086/698325 itors and trophic interactions. Dietary niche width consists of

This content downloaded from 169.237.066.128 on August 01, 2018 09:36:51 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 000 The American Naturalist a between-individual component (BIC) and a within-individual largely through short-term laboratory experiments and field- component (WIC; Roughgarden 1972). In addition to predict- tracking surveys, which cannot assess consistent variation able between-individual variation due to size (Werner and among individuals in preference or variation in a single indi- Gilliam 1984), age (Polis 1984), and sex (Shine 1989), a broad vidual’s preference over time (Chapman 2000). range of taxa exhibit consistent between-individual differ- To address this knowledge gap, we repeatedly measured ences in diet that are not attributable to these characteristics the feeding preference of 444 individual herbivores of eight (Bolnick et al. 2003; Sih et al. 2004a). Indeed, some individuals species using a suite of generally available algal foods to es- may be specialists that each consume only a narrow range of timate the relative magnitude of interspecific and intraspe- the resources consumed by the population (Bolnick et al. cificdietvariation.Specifically, we asked two questions. First, 2003; Tinker et al. 2008; Woo et al. 2008; Kernaleguen et al. what is the relative magnitude of interspecific versus intraspe- 2015), whereas others are individual generalists that each cific variation in feeding preference? Second, is individual var- consume the full range of resources consumed by the popula- iation in preference within species primarily due to consistent tion (Costa et al. 2015). Individual specialization is fairly com- specialization among individuals (BIC) or generalist variation mon and BIC is large for some taxa, but the relative magni- within individuals (WIC)? We interpret our results in light of tude of BIC and WIC in contributing to dietary niche width how species that share similar feeding preference differ in the across entire consumer assemblages has rarely been quanti- magnitude and source of individual variation (within vs. be- fied (Bolnick et al. 2011; Maldonado et al. 2017). tween individuals), potentially exerting different structuring Individual dietary specialization can arise from differences influence on prey assemblages. in morphology, physiology or behavior, or genotype (Bolnick et al. 2003, 2011). Of these mechanisms, feeding behaviors are easily comparable across diverse taxa and trophic levels Methods and are the most proximate measures of diet. Moreover, variation in feeding preference (West 1988; Estes et al. 2003), We collected animals and algae from rocky reef habitats be- habitat use (Werner 1977; Matthews et al. 2010), and activity tween Bodega Bay (38.3332507N, 123.0480577W) and Point level (Pruitt et al. 2012, 2017) can produce interindividual var- Arena Cove (38.9140767N, 123.7089077W) in northern Cali- iation in the size, type, and traits of prey obtained by con- fornia. Among a diverse community of herbivorous mollusks, sumers. Furthermore, variation in these behaviors is closely crustaceans, and echinoderms on northern California rocky tied to and has clear implications for ecological processes and shores, we selected the most common, mobile invertebrate ecosystem dynamics; individual variation in feeding behavior macrograzers occurring in the intertidal and shallow subtidal directly impacts key ecological processes, such as top-down zones. These species included Mesocentrotus (formerly Stron- control (Belgrad and Griffen 2016; Michalko and Pekar 2017; gylocentrotus) franciscanus (red urchin, n p 27), Strongylo- Pruitt et al. 2017), interspecific competition, and coexistence centrotus purpuratus (purple urchin, n p 37), Haliotis ru- (Gibert and Brassil 2014; Hart et al. 2016). We therefore focus fescens (red abalone, n p 36), Tegula brunnea (brown on feeding behavior—specifically, feeding preference—as the turban snail, n p 79), Pugettia producta (northern kelp crab, proxy for the dietary niche and ecological function. n p 50), Pagurus samuelis (blueband hermit crab, n p 79), California kelp forests and adjacent rocky shores support Pachygrapsus crassipes (lined shore crab, n p 49), and Tegula complex food webs that host diverse assemblages of preda- funebralis (black turban snail, n p 87). These species live pri- tors, herbivores, and algae (Steneck et al. 2002; Blanchette et al. marily in the shallow subtidal and low intertidal zones, with 2016; Carr and Reed 2016). In these habitats, numerous herbi- the exception of P. crassipes and T. funebralis,whichoccupy vores influence community structure and dynamics via direct the mid- and high-intertidal zones but feed on many of the and indirect effects on algal abundance, diversity, and produc- same algae. We collected adult individuals of moderate size tivity (Paine and Vadas 1969; Sousa 1984; Aquilino and Sta- from a variety of habitats (surfgrass, kelp, bare rock, sand) chowicz 2012). Many of these herbivores occupy similar within one or two sites per species to account for potential habitats and forage on algae along a gradient from intertidal habitat-related variation in feeding preference (see table A1). to shallow subtidal zones. These grazers include crustaceans, Individuals of each species were brought back to the Bodega mollusks, and echinoderms that vary in size, morphology, life Marine Laboratory, initially housed together, fed the same algae history, and occurrence across these zones, and they exhibit subsequently used in experimental trials, and allowed to adjust well-documented between-species differences in foraging pat- to laboratory conditions for 1 week before feeding experiments terns (Lowry and Pearse 1973) and feeding preference for al- were conducted. We individually tagged, measured, weighed, gae (Leighton 1966; Thornber et al. 2008). Yet despite the rich and housed collected animals in separate flow-through contain- literature on the impacts of herbivory in these systems (Lub- ers or seawater tanks. Tagging and housing methods (and re- chenco and Gaines 1981; Menge 2000; Steneck et al. 2002; lated laboratory conditions, such as diurnal cycle) varied, de- Foster et al. 2006), herbivore preference has been documented pending on the size of the herbivore species (table A1).

We conducted feeding-choice assays every 1–2 weeks for Statistical Methods a total of four or five repeated measures of preference for every individual of each species between July 2013 and March We sought to (1) identify the important axes of variation in 2014 (table A1). During this time period, sea surface temper- feeding preference, (2) quantify how the variation in prefer- ature ranged from 10.57 to 157C, although short-term fluctua- ence is partitioned within the entire herbivore assemblage tions on the order of 27–47C within a single day or week were and within each herbivore species (between-species com- not uncommon. We conducted collections and laboratory ex- ponent, BIC, and WIC), and (3) explore the potential eco- periments simultaneously for two or three species at a time, logical factors (body size, sex, source locale/collection loca- using algae collected from a single location for all herbivores tion) contributing to significant BIC in preference. within each trial. Feeding-choice assays utilized four species We first performed a principal components analysis on of algae, including two kelps, Nereocystis luetkeana (bull kelp) algal consumption data. We used the proportion of total al- and Laminaria setchellii (stiff-stiped or split kelp), and two gae consumed, represented by each of the algal species as the species of foliose red algae, Mazzaella splendens (splendid ir- response variables, and treated each individual herbivore in idescent seaweed) and Dilsea californica (leathery strap sea- each trial as an individual sample. This allowed us to assess weed). These species of algae were selected in part because how variation in relative consumption of each algal species they are locally abundant, persist from late spring to midwin- contributed to differences in feeding preference among indi- ter, and are available as food (in an attached or detached viduals and species of herbivores. We found that principal form) for the focal herbivores. We chose these species of algae component (PC) 1 accounted for 80% of the total variation to represent a range of taxonomic and morphological char- and was strongly positively correlated with the proportion acteristics but also to minimize differences in algal morphol- of diet composed of bull kelp (N. luetkeana;see“Results”;ta- ogy; all four species are bladelike, and we selected fronds of ble S1.1; tables S1.1–S1.7, S2.1–S2.3 are available online). We comparable thickness, surface area, shape, and tissue weight found that PC2 accounted for 13.8% of the total variation and for our experiments to minimize the chance that area, shape, was strongly positively correlated with the proportion of diet or size would influence feeding rate by altering encounter composed of both species of red algae (M. splendens and D. frequency. californica). We therefore focused analyses on how the rela- We collected algae for each trial from a single rocky shore tive consumption of bull kelp varied among individuals and location and tidal height, using fronds that were ungrazed species of herbivores. We define this response variable as and nonreproductive. Blades of similar weight and length preference for bull kelp, calculated as the biomass of bull kelp were offered together to individual herbivores by clipping them consumed divided by the total biomass of algae consumed. to a looped cable tie secured within the container or tank. We This represented our main axis of variation in preference. offered the same amount of each alga to each given herbivore We define a second axis of variation as preference for red al- but varied the amount among herbivore species according to gae, calculated as the biomass of red algae consumed divided herbivore size. Controls of the same design were placed in iden- by the total biomass of algae consumed. This second axis of tical containers, either adjacent to experimental containers for variation represents a much smaller proportion of total vari- smaller herbivore species or in perforated, clear plastic con- ation in preference, and results concerning the partitioning of tainers inside experimental tanks for larger herbivore species. preference across the full herbivore assemblage and for each We terminated the experiment within 5 days, long enough for species are similar for both axes of variation. We therefore herbivores to consume a visible portion of the available algae present the results for preference for bull kelp in the main ar- but before any one choice had been completely consumed. We ticle, while the results of analyses of preference for red algae disregarded replicates in which herbivores either did not are presented in the supplemental material (“Supplemental consume any algae (55 instances) or consumed all biomass Information 2”). of any single algal species (157 instances). We constructed statistical models of individual prefer- To determine individuals’ relative feeding preferences ence for bull kelp and red algae to evaluate the relative mag- among the selected algae, algae from feeding-choice assays nitude of variation among individuals versus among species were blotted dry and weighed before and after feeding trials. in contributing to preference. The arcsine–square root trans- We measured change in algal mass as a result of herbivore con- formation was applied to preferences (for bull kelp and for red sumption for an experimental trial by subtracting the final algae) to normalize their distributions (see figs. S1.1–S1.3 wet mass from the initial wet mass after adjusting initial mass [figs. S1.1–S1.5, S2.1, S2.2 are available online] for histograms by the proportional growth observed in grazer-free controls of model residuals and tables S1.2 and S1.3 for a comparison (Stachowicz and Hay 1999). of estimates and significance of variance components for the Data from these experiments are deposited in the Dryad guild-level and species models using the arcsine–square root Digital Repository: https://dx.doi.org/10.5061/dryad.5nq06fj transformation of preference for bull kelp vs. PC1 of prefer- (Rhoades et al. 2018). ence). We used a mixed effects model with random intercepts

This content downloaded from 169.237.066.128 on August 01, 2018 09:36:51 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 000 The American Naturalist for individual, species, and trial and a fixed effect of seasonal including any variation among trials that affected multiple in- block to account for the fact that feeding trials were grouped dividuals in parallel. Total niche width (TNW) of a species for multiple species across seasons (table A1). The variance of consists of the sum of the variance between individuals the random effect of species represents the between-species (BIC) and the residual variance (WIC). The degree of indi- component, the variance of the random effect of individual vidual specialization is determined by the ratio (WIC/TNW) represents the BIC, and the residual variance represents the for each species, such that a value of 1 indicates that there is WIC in preference over time. WIC does not include the varia- no within-individual consistency in preference, while a value tion within individuals caused by algal or environmental condi- of 0 indicates high individual consistency (Bolnick et al. tions that varied across trials and seasons, which affected all in- 2002). To investigate potential mechanisms for specialization dividuals regardless of species; these are captured by the random in species with high BIC, we constructed additional single- effect of trial and the fixed effect of seasonal block. The signif- species models using the same structure described above but icance of random effects was tested using likelihood ratio tests, including the fixed effects of individual body size, source lo- whichcomparemodelswithandwithouttheserandomeffects. cale/collection location, and sex in species for which these To estimate BIC and WIC for each species, we also con- traits were able to be measured. structed separate models of individual preference for bull kelp To compare preference across herbivore species in the as- and red algae within each herbivore species. These models semblage, we constructed a ranked multinomial logit model contained random intercepts for individual and experimental of individual preference. This model contained two individual- trials. The random effect of individual represents the BIC, and specific variables, species and experimental trial. We plotted residual variance represents the WIC of preference, again not the predicted probabilities that each algal species ranked first

Figure 1: Results from the guild-level model of bull kelp preference. A, Estimates and 95% confidence intervals of variance. B, Proportion of total variance accounted for by the between-species component, the between-individual component, and the within-individual component extracted from the full model of individual preference for bull kelp across all herbivore species. Parameters estimated using maximum likelihood are shown in black fill, and those estimated using Bayesian inference are shown in white fill.

Figure 2: Results from species-level models of bull kelp preference. Shown are estimates and 95% confidence intervals of variance as well as the proportion of total variance accounted for by the between-individual component (shown in dark gray) and the within-individual component (shown in light gray) extracted from the models of individual preference for bull kelp for each herbivore species. Asterisks indicate the significance of the random effect of individual determined using likelihood ratio tests. Estimates were calculated using maximum likelihood methods (using linear mixed models). in order of preference for each herbivore species. Predicted the second largest source of variation in preference, repre- probabilities of rank preference coincided with estimates of senting 28%–42% of all variance in preference for bull kelp, strength of preference for bull kelp and red algae by herbivore although the estimates of between-species variance were species from the original linear mixed model. uncertain (wide confidence intervals in fig. 1A). Consistent Linear mixed models were fit to a normal distribution us- between-individual differences contributed substantially less ing maximum likelihood methods as well as Bayesian infer- (4%–5%) to total variance in preference, but BIC was still a ence (see arguments for use of Bayesian hierarchical models significant random effect in the model (table S1.4; P p :005). to estimate individual diet specialization [Coblentz et al. Seasonal block did not significantly impact preference for 2017]), and both produced similar predictor estimates (see bull kelp (P p :096) or red algae (P p :442), indicating that “Results”; figs. 1, 2, S1.4). We conducted statistical analyses species differences cannot be explained by seasonal variation in R version 3.4.0 (R Core Development Team 2017). Linear among trials for different groups of species. mixed models were fit using maximum likelihood methods The magnitude of BIC varied among grazer species (figs. 2, with the R package lme4 (Bates 2015) and using Bayesian in- S1.4). BIC accounted for 0%–25% of the total variation in pref- ference methods with the R package rethinking (McElreath erence for bull kelp, depending on the species (WIC/TNW 2016). We sampled from Bayesian inference models using ranged from 0.75 to 1.0; table 1). For Tegula funebralis and the Hamiltonian Monte Carlo method and noncentered pa- Pugettia producta,BICcomprised13%–14% and 25%, re- rameterization, with the R package RStan (Stan Development spectively, of the total variation in preference for bull kelp Team 2016). We used a ranked multinomial logit model to as- (WIC=TNW p 0:86, 0.75) and was a significant predictor sess rank preference with the package mlogit (Croissant 2013). of bull kelp preference (P ! :001, P p :014; table S1.5). Ac- counting for individual variation in body size, source locale/ collection location, and sex reduced the BIC from 25% to Results 17% of the total variation in P. producta preference for bull The full model of preference for bull kelp indicated that kelp but had minimal effects for T. funebralis (table S1.6) within-individual differences are the primary source of var- and did not affect the significance of BIC in determining iation in preference; WIC accounted for 55%–65% of all preference for bull kelp in either species (table S1.7). While variance in preference for bull kelp (fig. 1B). Species iden- BIC did comprise 11%–14% of the total variation in prefer- tity was a significant predictor of preference (P ! :001) and ence in M. franciscanus, the random effect of individual was

Table 1: Niche components for preferences for bull kelp, including estimates of the between-individual (BIC) and within-individual (WIC) components of variance, total niche width (TNW; TNW p BIC 1 WIC), and the degree of individual specialization (WIC/TNW) for each herbivore species, calculated from species-level models Repeated Species model N measures BIC WIC TNW WIC/TNW Preference Pagurus samuelis 79 4 .001 .031 .032 .983 .337–.5 Tegula funebralis 87 4 .011 .064 .075 .858 .376–.57 Pugettia producta 50 5 .012 .036 .048 .746 .511–.666 Pachygrapsus crassipes 49 4 .002 .088 .089 .981 .637–.765 Strongylocentrotus purpuratus 37 4 .012 .191 .203 .939 .829–.927 Tegula brunnea 79 4 .004 .156 .160 .977 .419–.585 Mesocentrotus franciscanus 27 4 .015 .098 .113 .863 .398–.55 Haliotis rufescens 36 5 .000 .079 .079 1.000 .644–.776 Note: N represents the number of individual herbivores tested repeatedly across trials. Repeated measures represents the number of repeated choice feeding assays conducted per individual animal. Preference gives the 95% confidence intervals for bull kelp preference for each herbivore species, measured as conditional predictions from the full model across all herbivore species. not distinguishable from zero because the majority of individ- dividuals prefer the same food or because individuals vary uals varied widely and inconsistently in preference across trials substantially in preference across trials (fig. 3). (fig. 3). For all other species (Haliotis rufescens, Tegula brunnea, Although all herbivore species exhibited the same rank Pachygrapsus crassipes,andPagurus samuelis), BIC contrib- preference (highest preference for bull kelp; fig. 4), species uted minimally (0%–2%) to the total variation in preference still varied in key aspects of their preference. Species var- for bull kelp (WIC=TNW p 0:98–1:0), either because all in- ied with respect to both the probability that bull kelp was

Figure 3: Raw preference for bull kelp separated by species, individual, and trial. Note that some species are invariant in their preference (Pagurus samuelis), some vary mostly within individuals across time (Strongylocentrotus purpuratus, Tegula brunnea), and others show some consistent variation among individuals (Mesocentrotus franciscanus, Pugettia producta, Tegula funebralis). See ﬁgure 2 for partitioning of within- and between-species components of variance.

Figure 4: Mean estimates of rank preference among four algal species for eight herbivore species, based on a multinomial logit model of the individual rank order of the amount of each algal species consumed. See table 1 for full scientific names. their top-ranking food and their proportional consump- P. crassipes, T. brunnea,andS. purpuratus have the broadest tion of bull kelp over other species (42%–98%; fig. 4). Since TNW (0.089–0.203), but none of these species exhibit signif- species tested within the same seasonal block did not ex- icant BIC. Haliotis rufescens and T. funebralis have moderate hibit similar preferences (fig. S1.5), species preferences are TNW (0.079, 0.075), yet H. rufescens consists of many indi- not attributable to seasonal differences in temperature, light, viduals with similar preference and others that vary widely or other environmental conditions. Moreover, preference did in preference, while individuals of T. funebralis consistently not vary significantly among individuals collected from dif- vary in preference from one another. Pagurus samuelis and ferent locations in any species. However, for the two species P. producta have the narrowest TNW (0.032, 0.048), yet P. that showed significant between-individual variation in pref- samuelis consists of many individuals with similar preference erence for bull kelp, body size (for P. producta, P p :004; for that vary little across trials, while individuals of P. producta T. funebralis, P p :013), and sex (for P. producta, P ! :001) consistently vary in preference from one another. partly explained this preference variation, in which smaller individuals have a stronger preference for bull kelp. None Discussion of these predictors explained preference among individuals in any of the other six grazer species. Even among rocky reef herbivores that all generally prefer Although species also differed in their TNWs and the dis- bull kelp and exhibit the same rank preference for kelp over tribution of preference across individuals (table 1), TNW in red algae, we document significant variation among species this assemblage is not necessarily reflective of large BIC rela- and between individuals within species in feeding preference. tive to WIC, and in fact different combinations of BIC and Between-species differences account for a much greater pro- WIC produce the same TNW. Mesocentrotus franciscanus, portion of the total variation in feeding preference across the

This content downloaded from 169.237.066.128 on August 01, 2018 09:36:51 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 000 The American Naturalist guild than the BIC of preference, suggesting that focusing on for by all sources of BIC is comparable to the proportion of species-level differences in preference would approximate the between-species variation in preference for the entire herbi- feeding preferences of many of the species within this partic- vore guild. Thus, for some species between-individual differ- ular assemblage (Roughgarden 1974; Lister 1976; Taper and ences in phenotype may equal the magnitude of between- Case 1985; Schoener 1986). In contrast, a recent meta-analysis species differences (Kamath and Losos 2017; Des Roches suggests that the ecological effects of interspecific and intra- et al. 2018), while for others between-individual effects may specific variation are often of similar magnitude (Des Roches be quite small. This highlights the need for a better predictive et al. 2018). However, published studies available for meta- understanding of which species will have the strongest intra- analysis may be biased toward species with large intraspecific specific effects. variation and typically use only a single comparison species to Current hypotheses predict that the strength of between- estimate interspecific effects. Thus, meta-analyses may provide individual variation may be related to total species niche an upper limit of the relative strength of intraspecificvariation width, species interaction strength, or trophic level. For ex- rather than an estimate of its average importance. Here we show ample, the niche variation hypothesis (Van Valen 1965; Bol- that when systematically evaluating within- and between- nick et al. 2007; Costa et al. 2008; Maldonado et al. 2017) sug- species variation for all species in a single guild, interspecific gests that species with higher niche width should have higher effects were considerably stronger. Given that these species BIC. We found no evidence of this. However, our focus on have evolved combinations of between- and within-species comparable algal food sources may have limited TNW in niche variation that directly contribute to both their co- our study and reduced the strength of this test. For example, existence and their contribution to ecosystem function, this two species in this assemblage (Pagurus samuelis and Pa- across-guild evaluation is essential to understanding the rela- chygrapsus crassipes) are omnivorous and could exhibit con- tive ecological importance of these niche components. sistent diet variation with respect to additional axes of food Characterizing niche components in order to understand variation, as found for other omnivores (Svanbäck et al. their relative importance for community and ecosystem pro- 2015; Maldonado et al. 2017). We also did not find that spe- cesses requires several important methodological decisions. cies with large known interaction strengths (e.g., urchins, a For example, we deliberately tested feeding preferences of well-documented keystone species) were any more likely to individuals in isolation and so do not include individual var- have large BIC. This is similar to results from a comparison iation driven by intraspecific interactions (Svanbäck and of intraspecificandinterspecific effects on communities Bolnick 2007; Svanbäck et al. 2008; Tinker et al. 2008). This across many taxa (Des Roches et al. 2018). Finally, most may cause our estimates of individual variation to be conser- studies of individual specialization have focused on higher vative. However, interspecific competition or predation could trophic levels (Araujo et al. 2011), whereas ours focused on also suppress individual variation in preference (Eklöv and herbivores. This contrast is important because intermediate Svanbäck 2006; Bolnick et al. 2010; Araujo et al. 2014), so predators with greater environmental opportunity and vari- the net effect of this experimental design choice is not clear. ability in foraging-predation risk trade-offs are hypothesized We also provided only algae that could be consumed by all to exhibit a greater degree of BIC (Svanbäck et al. 2015). How- species, and we limited morphological variation among algae ever, recent work examined the magnitude of within-species to facilitate comparisons, potentially limiting BIC expressed variation across a range of trophic levels, from producers to in other diet axes (e.g., ephemeral algae and invertebrates con- secondary consumers, and did not find clear differences (Des sumed by some of the herbivores in this assemblage). We also Roches et al. 2018). presented herbivores with at least one highly palatable food Given that the species in this guild vary substantially in source, bull kelp, which was a priori known to be readily con- body size, mobility, and primary habitat type, we might also sumed by all species in this guild (Leighton 1966; Hines 1982; expect that these factors would explain some of the varia- Watanabe 1984; Foster et al. 2015) and which may have fur- tion in feeding preferences both between and within species. ther limited the expression of BIC amplified by limitation of For example, larger and more mobile invertebrates are more the preferred resource (Araujo et al. 2011; Tinker et al. 2012). likely to encounter a variety of algal food sources on a daily However, even with all these caveats, individuals within timescale (Best et al. 2014). In contrast, if individuals consis- some species displayed consistent variation in feeding pref- tently occupy home ranges with distinct algal or predator as- erence across trials. For the two species (Pugettia producta semblages, long-term habituation to local food availability may and Tegula funebralis)thatshowedsignificant diet varia- reinforce differences in feeding preference through physiolog- tion between individuals, BIC accounted for up to one- ical conditioning or morphological adjustment (Vesakoski quarter of the total variation in feeding preference within et al. 2009; Molis et al. 2015). Surprisingly, we found that source those species and for up to one-sixth of the total variation locale/collection location did not explain any of the BIC in in feeding preference after accounting for individual varia- the species for which it was significant. In addition, species’ tion due to size, source locale, and sex. The fraction accounted average body length, which varied from 22 to 200 mm, did

This content downloaded from 169.237.066.128 on August 01, 2018 09:36:51 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). Dietary Niches: The Role of Individuals 000 not predict TNW or the degree of individual specialization. erences among individuals converge. However, if the win- Our results do offer some evidence that the presence of dow of plant susceptibility to grazing-induced mortality is between-individual variation in other traits, such as antipreda- small (e.g., size dependent), then high WIC could induce a tor behavior, could be associated with between-individual stochastic component to grazing preference that could main- variation in diet across this species assemblage. Individuals tain prey diversity at the landscape level. of P. producta exhibit phenotypically plastic color states that Although BIC is large for some of our species, within- they can alter when molting, depending on whether they individual differences represent the majority of total varia- are found on brown algae (kelp) versus red algae, and this tion in feeding preference both within and across this her- color matching affects their susceptibility to predation (Hult- bivore guild, supporting previous reviews of the magnitude gren and Stachowicz 2008). Thus, interactions between hab- of individual specialization within species (Bolnick et al. itat choice and feeding preference will jointly impact sur- 2003; Araujo et al. 2011). WIC may represent chance con- vival, and it is possible that individual variation could lead tact with different food sources (within-species noise), al- to preference-by-habitat segregation at smaller scales, such though housing and experimental conditions suggest that as at the level of algal blades rather than across tidal depths each herbivore contacted all algal choices during each assay. or among sites. Similarly, predator avoidance strategies in Alternatively, individuals might vary their diet over time as T. funebralis range from remaining stationary to rapid es- a feeding strategy to achieve a mixed diet (Aquilino et al. cape, leading to differential susceptibility to active versus 2012), possibly because such a diet prevents mouthpart wear sit-and-wait predators (Pruitt et al. 2012). These predator (Kitting 1980), provides complementary nutritional benefits avoidance strategies could influence grazer feeding behavior (Horn 1983; Aquilino et al. 2012), and/or dilutes the concen- by positive preference induction as a result of physiological tration of any particular defensive chemicals (Bernays et al. or morphological acclimation (Hultgren and Stachowicz 1994). Given seasonal and yearly fluctuations in the availabil- 2010) or by alteration of prey availability due to restricted ity of preferred foods (Cubit 1984; Edwards and Estes 2006), movement (West 1988; Coleman and Wilson 1998; Stacho- individuals with large WIC should be favored because they wicz and Hay 1999; Peacor and Werner 2001). Whatever the exploit a broader range of resources. While this may enhance cause, the existence of consistent individual variation in adult the stability and resilience of their populations over time habitat choice, antipredator traits, and feeding preference sug- (Wolf and Weissing 2012), greater within-species variation gest that BIC may be correlated across traits (Sih et al. 2004b) will also increase the strength of competition among species, and that individuals should be selected from a range of eco- decreasing the likelihood of coexistence among species (Hart logical contexts if the goal is to quantify the magnitude of et al. 2016). Where different herbivore species are clearly dif- BIC (Holbrook and Schmitt 1992; Kamath and Losos 2017). ferentiated along other niche axes (e.g., habitat, tidal height), Assuming that laboratory-measured patterns of individual high BIC could reduce intraspecific competition and increase preferences are reflective of herbivore feeding activity in the population size (by giving multiple individuals access to ex- field, we can speculate on the effects of this variation on prey clusive resources) without increasing interspecificcompeti- community composition as well as coexistence and competition in a way that might increase the likelihood of persistence. tion among herbivores. Specific impacts of species with greater In contrast with previous meta-analyses and reviews con- BIC on community structure may be evident (Post et al. 2008; sidering the prevalence and ecological impacts of individual Bassar et al. 2010) and will depend on the dominance hierar- specialization in natural assemblages, we find that consider- chy of prey. Selective herbivory (whatever the source) is widely ing longitudinal preference in all species from a single guild known to affect the diversity, biomass, and composition of yields distinct results; variation among individuals within a seaweeds on rocky shores (Lubchenco 1978; Sousa 1979; Un- species is substantially less, on average, than the variation derwood et al. 1983; Sousa 1984), with herbivores often pre- among all species. Yet even in this consumer guild dominated ferring early successional species that inhibit later successional by within-individual and between-species variation, between- perennials. Given that many of these herbivores are either rel- individual differences represent up to one-quarter of the total atively sedentary or site faithful, variation among individuals variation in feeding preference within particular species and in preferences could lead to spatial patchiness in intensity of one-sixth of the total variation if we account for body size, grazing on early successional species and thus to patchiness in source locale, and sex in these species. Moreover, such the rate of succession. Variation within individuals that is un- between-individual variation was not predicted by TNW or synchronized among individuals, in contrast, should create ecological importance across species, demonstrating that spatial variability in selectivity of grazing that can result in BIC represents a unique dimension of species resource use spatial patchiness in prey community composition, especially and ecological function. Finally, this study serves to highlight if the number of individual herbivores impacting a local as- that even in well-studied guilds of consumers (e.g., Hay and semblage is small. Over time, high WIC should eliminate any Steinberg 1992 and references therein) there are few empirical such variation as preferences vary temporally and mean pref- studies that repeatedly assess the preferences of the same

This content downloaded from 169.237.066.128 on August 01, 2018 09:36:51 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 000 The American Naturalist individual over time for several species. This limitation is oratory and ﬁeld will be critical to evaluating the effects of parallel to problems with characterizing species- and genotype- individual specialization on the structure, dynamics, and eco- level traits on the basis of a single time point in a single com- logical functioning of consumer assemblages. mon garden (Turcotte and Levine 2016) and poses similar constraints on our ability to predict coexistence under cur- Acknowledgments rent and changing conditions. For example, the relative importance of trait variation within and among species and the This work was supported by the Center for Population Biology plasticity of this variation in response to short-term (seasonal) at the University of California (UC), Davis; the UC Davis Nat- or long-term environmental change could dramatically alter ural Reserve; and grant 0841297 to S. L. Williams from the US the ecological consequences of local changes in the number National Science Foundation’s Graduate K–12 Program. Spe- versus abundance of species. For this reason, additional lon- cial thanks are due to Tom Schoener, who provided valuable gitudinal studies directly measuring resource use in the lab- feedback on the theory, design, and writing of the manuscript.

This content downloaded from 169.237.066.128 on August 01, 2018 09:36:51 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). APPENDIX All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).

Table A1: Collection locations, tagging, and housing methods of herbivore species tested

This content downloadedfrom 169.237.066.128 onAugust01, 2018 09:36:51AM Source Body Tempera- locale/collection GPS length Tagging ture range Dates of feeding Species location location Habitat type measure method Housing method Diurnal cycle (7C) assays Strongylocentrotus Point Arena 38.914307N, Subtidal reef Test Plastic Indoor/bucket 14-h day/10-h night 10.5–15 July 14–Aug. 10, purpuratus Cove 123.70737W diameter tubing (2 gallon) (6 a.m.–8p.m.) 2013 Pachygrapsus Mason’s Marina, 38.330387N, Midintertidal Carapace Nail polish Indoor/plastic 12-h day/12-h night 10.5–15 July 21–Aug. 10, crassipes Bodega Bay 123.05887W cobble width dots containers with (6 a.m.–6 p.m.) 2013 mesh sides (0.5 L) Tegula funebralis Shell Beach; 38.418237N, Multiple; Maximum Nail polish Indoor/plastic 12-h day/12-h night 10.5–15 July 22–Aug. 11, Schoolhouse 123.10377W; midinterti- shell dots containers with (6 a.m.–6 p.m.) 2013 Beach 38.376157N, dal reef and width mesh sides (0.5 L) 123.07867W boulders Pagurus samuelis Shell Beach 38.418273N, Low intertidal Maximum Nail polish Indoor/plastic 12-h day/12-h night 11–12.5 Oct. 7–Nov. 16, 123.10377W sand and shell dots containers with (6 a.m.–6 p.m.) 2013 tide pools width mesh sides (0.5 L) Tegula brunnea Fort Ross 38.524887N, Subtidal Maximum Nail polish Indoor/plastic 12-h day/12-h night 11–12.5 Oct. 8–Nov. 16, 123.23077W subcanopy shell dots containers with (6 a.m.–6 p.m.) 2013 kelp blades width mesh sides (0.5 L) Mesocentrotus Pebble Beach 38.696587N, Subtidal Test diam- Plastic Outdoor/cylindrical 10-h day/14-h night 11–12.5 Oct. 22–Dec. 2, franciscanus 123.43977W boulders eter (ex- tubing tank (1 m diame- (6:30 a.m.–5p.m.) 2013 cluding ter # 0.5 m deep, spines) 400 L) Pugettia producta Stillwater Cove; 38.548037N, Multiple; bull Carapace Nail polish Indoor/bucket 12-h day/12-h night 11–13 Jan. 15–Mar. 1, Shell Beach 123.29467W; kelp bulbs width dots (2 gallon) (6 a.m.–6 p.m.) 2014 38.418237N, and low 123.10377W intertidal reef with surfgrass Haliotis rufescens Stillwater Cove 38.548037N, Multiple; Maximum Stainless Outdoor/cylindrical 10-h day/14-h night 11–13 Jan. 17–Mar. 3, 123.29467W subtidal shell steel tank (1 m diame- (7 a.m.– 5:30 p.m.) 2014 open reef width washers ter # 0.5 m deep, and boul- and 400 L) der crevices wire 000 The American Naturalist

Literature Cited Bolnick, D. I., L. H. Yang, J. A. Fordyce, J. M. Davis, and R. Svanbäck. 2002. Measuring individual-level resource specialization. Ecology Agashe, D. 2009. The stabilizing effect of intraspecific genetic varia- 83:2936–2941. tion on population dynamics in novel and ancestral habitats. Amer- Burkepile, D. E., and M. E. Hay. 2008. Herbivore species richness and ican Naturalist 174:255–267. feeding complementarity affect community structure and function Aquilino, K. M., M. E. Coulbourne, and J. J. Stachowicz. 2012. Mixed on a coral reef. Proceedings of the National Academy of Sciences of species diets enhance the growth of two rocky intertidal herbivores. the USA 105:16201–16206. Marine Ecology Progress Series 468:179–189. Carr, M. H., and D. C. Reed. 2016. Shallow rocky reefs and kelp forests. Aquilino, K. M., and J. J. Stachowicz. 2012. Seaweed richness and her- Pages 311–336 in H. Mooney and E. Zavaleta, eds. Ecosystems of bivory increase rate of community recovery from disturbance. Ecology California. University of California Press, Oakland. 93:879–890. Chapman, M. G. 2000. Poor design of behavioural experiments gets Araujo, M. S., D. I. Bolnick, and C. A. Layman. 2011. The ecological poor results: examples from intertidal habitats. Journal of Experi- causes of individual specialisation. Ecology Letters 14:948–958. mental Marine Biology and Ecology 250:77–95. Araujo, M. S., R. B. Langerhans, S. T. Giery, and C. A. Layman. 2014. Chesson, P. 2000. Mechanisms of maintenance of species diversity. Ecosystem fragmentation drives increased diet variation in an en- Annual Review of Ecology and Systematics 31:343–366. demic livebearing fish of the Bahamas. Ecology and Evolution Coblentz, K. E., A. E. Rosenblatt, and M. Novak. 2017. The applica- 4:3298–3308. tion of Bayesian hierarchical models to quantify individual diet Barbour, M. A., M. A. Fortuna, J. Bascompte, J. R. Nicholson, R. specialization. Ecology 98:1535–1547. Julkunen-Tiitto, E. S. Jules, and G. M. Crutsinger. 2016. Genetic Coleman, K., and D. S. Wilson. 1998. Shyness and boldness in pump- specificity of a plant-insect food web: implications for linking ge- kinseed sunfish: individual differences are context-specific. Animal netic variation to network complexity. Proceedings of the National Behaviour 56:927–936. Academy of Sciences of the USA 113:2128–2133. Costa, A., S. Salvidio, M. Posillico, G. Matteucci, B. De Cinti, and A. Bassar, R. D., M. C. Marshall, A. Lopez-Sepulcre, E. Zandona, S. K. Romano. 2015. Generalisation within specialization: inter-individual Auer, J. Travis, C. M. Pringle, et al. 2010. Local adaptation in diet variation in the only specialized salamander in the world. Scien- Trinidadian guppies alters ecosystem processes. Proceedings of tific Reports 5:13260. the National Academy of Sciences of the USA 107:3616–3621. Costa, G. C., D. O. Mesquita, G. R. Colli, and L. J. Vitt. 2008. Niche Bates, D., M. Mächler, B. M. Bolker, and S. C. Walker. 2015. Fitting expansion and the niche variation hypothesis: does the degree of linear mixed-effects models using lme4. Journal of Statistical Soft- individual variation increase in depauperate assemblages? Ameri- ware 67:1–48. can Naturalist 172:868–877. Belgrad, B. A., and B. D. Griffen. 2016. Predator-prey interactions me- Croissant, Y. 2013. mlogit: multinomial logit model, version 0.2-4. diated by prey personality and predator hunting mode. Proceedings Crutsinger, G. M., M. D. Collins, J. A. Fordyce, Z. Gompert, C. C. Nice, of the Royal Society B 283:20160408. and N. J. Sanders. 2006. Plant genotypic diversity predicts commu- Bernays, E. A., K. L. Bright, N. Gonzalez, and J. Angel. 1994. Dietary nity structure and governs an ecosystem process. Science 313:966– mixing in a generalist herbivore: tests of two hypotheses. Ecology 968. 75:1997–2006. Cubit, J. D. 1984. Herbivory and the seasonal abundance of algae on a Best, R. J., A. L. Chaudoin, M. E. Bracken, M. H. Graham, and J. J. high intertidal rocky shore. Ecology 65:1904–1917. Stachowicz. 2014. Plant-animal diversity relationships in a rocky Des Roches, S., D. M. Post, N. E. Turley, J. K. Bailey, A. P. Hendry, intertidal system depend on invertebrate body size and algal cover. M. T. Kinnison, J. A. Schweitzer, et al. 2018. The ecological impor- Ecology 95:1308–1322. tance of intraspecific variation. Nature Ecology and Evolution Blanchette, C. A., M. W. Denny, J. M. Engle, B. Helmuth, L. P. Miller, 2:57–64. K. J. Nielsen, and J. Smith. 2016. Intertidal ecosystems. Pages 337– Edwards, M. S., and J. A. Estes. 2006. Catastrophe, recovery and range 357 in H. Mooney and E. Zavaleta, eds. Ecosystems of California. limitation in NE Pacific kelp forests: a large-scale perspective. Ma- University of California Press, Oakland. rine Ecology Progress Series 320:79–87. Bolnick, D. I., P. Amarasekare, M. S. Araujo, R. Buerger, J. M. Levine, Eklöv, P., and R. Svanbäck. 2006. Predation risk influences adaptive M. Novak, V. H. W. Rudolf, et al. 2011. Why intraspecific trait var- morphological variation in fish populations. American Naturalist iation matters in community ecology. Trends in Ecology and Evo- 167:440–452. lution 26:183–192. Estes, J. A., M. L. Riedman, M. M. Staedler, M. T. Tinker, and B. E. Bolnick, D. I., T. Ingram, W. E. Stutz, L. K. Snowberg, O. L. Lau, and Lyon. 2003. Individual variation in prey selection by sea otters: J. S. Paull. 2010. Ecological release from interspecific competition patterns, causes and implications. Journal of Animal Ecology leads to decoupled changes in population and individual niche width. 72:144–155. Proceedings of the Royal Society B 277:1789–1797. Forsman, A., and L. Wennersten. 2016. Inter-individual variation pro- Bolnick, D. I., R. Svanbäck, M. S. Araujo, and L. Persson. 2007. motes ecological success of populations and species: evidence from ex- Comparative support for the niche variation hypothesis that more perimental and comparative studies. Ecography 39:630–648. generalized populations also are more heterogeneous. Proceed- Foster, M. C., J. E. K. Byrnes, and D. C. Reed. 2015. Effects of five south- ings of the National Academy of Sciences of the USA 104:10075– ern California macroalgal diets on consumption, growth, and gonad 10079. weight, in the purple sea urchin Strongylocentrotus purpuratus. PeerJ Bolnick, D. I., R. Svanbäck, J. A. Fordyce, L. H. Yang, J. M. Davis, C. D. 3:e719. Hulsey, and M. L. Forister. 2003. The ecology of individuals: inci- Foster, M. S., M. S. Edwards, D. C. Reed, D. R. Schiel, and R. C. dence and implications of individual specialization. American Nat- Zimmerman. 2006. Top-down vs. bottom-up effects in kelp forests. uralist 161:1–28. Science 313:1737–1738.

Gibert, J. P., and C. E. Brassil. 2014. Individual phenotypic variation Lowry, L. F., and J. S. Pearse. 1973. Abalones and sea urchins in an reduces interaction strengths in a consumer-resource system. Ecology area inhabited by sea otters. Marine Biology 23:213–219. and Evolution 4:3703–3713. Lubchenco, J. 1978. Plant species diversity in a marine intertidal Grant, P. R., and B. R. Grant. 2006. Evolution of character displace- community: importance of herbivore food preference and algal ment in Darwin’s finches. Science 313:224–226. competitive abilities. American Naturalist 112:23–39. Hart, S. P., S. J. Schreiber, and J. M. Levine. 2016. How variation between Lubchenco, J., and S. D. Gaines. 1981. A unified approach to marine individuals affects species coexistence. Ecology Letters 19:825–838. plant-herbivore interactions. 1. Populations and communities. Hay, M. E., and P. D. Steinberg. 1992. The chemical ecology of plant- Annual Review of Ecology and Systematics 12:405–437. herbivore interactions in marine versus terrestrial communities. Macarthur, R., and R. Levins. 1967. Limiting similarity convergence Pages 371–413 in G. A. Rosenthal and M. R. Berenbaum, eds. Her- and divergence of coexisting species. American Naturalist 101:377– bivores: their interactions with secondary plant metabolites. 2nd ed. 385. Vol. 2. Ecological and evolutionary processes. Academic Press, San Maldonado, K., F. Bozinovic, S. D. Newsome, and P. Sabat. 2017. Test- Diego, CA. ing the niche variation hypothesis in a community of passerine Hines, A. H. 1982. Coexistence in a kelp forest: size, population- birds. Ecology 98:903–908. dynamics, and resource partitioning in a guild of spider crabs Matthews, B., K. B. Marchinko, D. I. Bolnick, and A. Mazumder. 2010. (Brachyura, Majidae). Ecological Monographs 52:179–198. Specialization of trophic position and habitat use by sticklebacks in Holbrook, S. J., and R. J. Schmitt. 1992. Causes and consequences of an adaptive radiation. Ecology 91:1025–1034. dietary specialization in surfperches: patch choice and intraspecific McElreath, R. 2016. Rethinking: statistical rethinking book package, competition. Ecology 73:402–412. version 1.59. Horn, M. H. 1983. Optimal diets in complex environments: feeding McGill, B. J., B. J. Enquist, E. Weiher, and M. Westoby. 2006. Rebuilding strategies of two herbivorous fishes from a temperate rocky inter- community ecology from functional traits. Trends in Ecology and tidal zone. Oecologia 58:345–350. Evolution 21:178–185. Hughes,A.R.,B.D.Inouye,M.T.J.Johnson,N.Underwood,and Menge, B. A. 2000. Top-down and bottom-up community regulation M. Vellend. 2008. Ecological consequences of genetic diversity. in marine rocky intertidal habitats. Journal of Experimental Ma- Ecology Letters 11:609–623. rine Biology and Ecology 250:257–289. Hughes, A. R., and J. J. Stachowicz. 2011. Seagrass genotypic diversity Michalko, R., and S. Pekar. 2017. The behavioral type of a top pred- increases disturbance response via complementarity and domi- ator drives the short-term dynamic of intraguild predation. Amer- nance. Journal of Ecology 99:445–453. ican Naturalist 189:242–253. Hultgren, K. M., and J. J. Stachowicz. 2008. Alternative camouflage strat- Molis, M., R. A. Scrosati, E. F. El-Belely, T. J. Lesniowski, and M. egies mediate predation risk among closely related co-occurring kelp Wahl. 2015. Wave-induced changes in seaweed toughness entail crabs. Oecologia 155:519–528. plastic modifications in snail traits maintaining consumption effi- ———. 2010. Size-related habitat shifts facilitated by positive preference cacy. Journal of Ecology 103:851–859. induction in a marine kelp crab. Behavioral Ecology 21:329–336. Paine, R. T., and R. L. Vadas. 1969. Effects of grazing by sea urchins, Jung, V., C. Violle, C. Mondy, L. Hoffmann, and S. Muller. 2010. In- Strongylocentrotus spp., on benthic algal populations. Limnology traspecific variability and trait-based community assembly. Jour- and Oceanography 14:710–719. nal of Ecology 98:1134–1140. Peacor, S. D., and E. E. Werner. 2001. The contribution of trait- Kamath, A., and J. B. Losos. 2017. Does ecological specialization mediated indirect effects to the net effects of a predator. Proceed- transcend scale? habitat partitioning among individuals and spe- ings of the National Academy of Sciences of the USA 98:3904– cies of Anolis lizards. Evolution 71:541–549. 3908. Kernaleguen, L., J. P. Y. Arnould, C. Guinet, and Y. Cherel. 2015. Petchey, O. L., and K. J. Gaston. 2006. Functional diversity: back to Determinants of individual foraging specialization in large marine basics and looking forward. Ecology Letters 9:741–758. vertebrates, the Antarctic and subantarctic fur seals. Journal of Polis, G. A. 1984. Age structure component of niche width and in- Animal Ecology 84:1081–1091. traspecific resource partitioning: can age groups function as eco- Kitting, C. L. 1980. Herbivore-plant interactions of individual limpets logical species? American Naturalist 123:541–564. maintaining a mixed diet of intertidal marine algae. Ecological Mono- Post, D. M., E. P. Palkovacs, E. G. Schielke, and S. I. Dodson. 2008. graphs 50:527–550. Intraspecific variation in a predator affects community structure Kraft, N. J. B., G. M. Crutsinger, E. J. Forrestel, and N. C. Emery. 2014. and cascading trophic interactions. Ecology 89:2019–2032. Functional trait differences and the outcome of community assem- Pruitt, J. N., D. I. Bolnick, A. Sih, N. DiRienzo, and N. Pinter- bly: an experimental test with vernal pool annual plants. Oikos 123: Wollman. 2016. Behavioural hypervolumes of spider communities 1391–1399. predict community performance and disbandment. Proceedings Lankau, R. A. 2009. Genetic variation promotes long-term coexis- of the Royal Society B 283:20161409. tence of Brassica nigra and its competitors. American Naturalist Pruitt, J. N., K. A. Howell, S. J. Gladney, Y. Yang, J. L. L. Lichtenstein, 174:E40–E53. M. E. Spicer, S. A. Echeverri, et al. 2017. Behavioral hypervolumes Leighton, D. L. 1966. Studies of food preference in algivorous inver- of predator groups and predator-predator interactions shape prey tebrates of southern California kelp beds. Pacific Science 20:104–113. survival rates and selection on prey behavior. American Naturalist Lichstein, J. W., J. Dushoff, S. A. Levin, and S. W. Pacala. 2007. In- 189:254–266. traspecific variation and species coexistence. American Naturalist Pruitt, J. N., J. J. Stachowicz, and A. Sih. 2012. Behavioral types of 170:807–818. predator and prey jointly determine prey survival: potential impli- Lister, B. C. 1976. The nature of niche expansion in West-Indian cations for the maintenance of within-species behavioral variation. Anolis lizards II: evolutionary components. Evolution 30:677–692. American Naturalist 179:217–227.

R Core Development Team. 2017. R: a language and environment for sta- Svanbäck, R., M. Quevedo, J. Olsson, and P. Eklöv. 2015. Individuals tistical computing. R Foundation for Statistical Computing, Vienna. in food webs: the relationships between trophic position, omnivory Rasher, D. B., A. S. Hoey, and M. E. Hay. 2013. Consumer diversity and among-individual diet variation. Oecologia 178:103–114. interacts with prey defenses to drive ecosystem function. Ecology Taper, M. L., and T. J. Case. 1985. Quantitative genetic models for 94:1347–1358. the coevolution of character displacement. Ecology 66:355–371. Rhoades, O. K., R. Best, and J. Stachowicz. 2018. Data from: Assessing Thornber, C. S., E. Jones, and J. J. Stachowicz. 2008. Differences in feeding preferences of a consumer guild: partitioning variation herbivore feeding preferences across a vertical rocky intertidal gra- among versus within species. American Naturalist, Dryad Digital dient. Marine Ecology Progress Series 363:51–62. Repository, http://dx.doi.org/10.5061/dryad.5nq06fj. Tinker, M. T., G. Bentall, and J. A. Estes. 2008. Food limitation leads to Roughgarden, J. 1972. Evolution of niche width. American Natural- behavioral diversification and dietary specialization in sea otters. ist 106:683–718. Proceedings of the National Academy of Sciences of the USA ———. 1974. Niche width: biogeographic patterns among Anolis 105:560–565. lizard populations. American Naturalist 108:429–442. Tinker, M. T., P. R. Guimarães Jr., M. Novak, F. M. Darcie Marquitti, Schoener, T. W. 1974. Resource partitioning in ecological communi- J. L. Bodkin, M. Staedler, G. Bentall, et al. 2012. Structure and ties. Science 185:27–39. mechanism of diet specialisation: testing models of individual var- ———. 1986. Resource partitioning. Pages 91–126 in J. Kikkawa iation in resource use with sea otters. Ecology Letters 15:475–483. and D. J. Anderson, eds. Community ecology: pattern and process. Turcotte, M. M., and J. M. Levine. 2016. Phenotypic plasticity and Blackwell Scientific, Palo Alto, CA. species coexistence. Trends in Ecology and Evolution 31:803–813. Shine, R. 1989. Ecological causes for the evolution of sexual dimor- Underwood, A. J., E. J. Denley, and M. J. Moran. 1983. Experimental phism: a review of the evidence. Quarterly Review of Biology 64: analyses of the structure and dynamics of mid-shore rocky inter- 419–461. tidal communities in New South Wales. Oecologia 56:202–219. Siefert, A. 2012. Incorporating intraspecific variation in tests of trait- Van Valen, L. 1965. Morphological variation and width of ecological based community assembly. Oecologia 170:767–775. niche. American Naturalist 99:377–390. Siefert, A., C. Violle, L. Chalmandrier, C. H. Albert, A. Taudiere, A. Vellend, M. 2006. The consequences of genetic diversity in compet- Fajardo, L. W. Aarssen, et al. 2015. A global meta-analysis of the itive communities. Ecology 87:304–311. relative extent of intraspecific trait variation in plant communities. Vellend, M., and M. A. Geber. 2005. Connections between species di- Ecology Letters 18:1406–1419. versity and genetic diversity. Ecology Letters 8:767–781. Sih, A., A. M. Bell, and J. C. Johnson. 2004a. Behavioral syndromes: Vesakoski, O., J. Rautanen, V. Jormalainen, and T. Ramsay. 2009. Di- an ecological and evolutionary overview. Trends in Ecology and vergence in host use ability of a marine herbivore from two habitat Evolution 19:372–378. types. Journal of Evolutionary Biology 22:1545–1555. Sih, A., A. M. Bell, J. C. Johnson, and R. E. Ziemba. 2004b. Behavioral Watanabe, J. M. 1984. Food preference, food quality, and diets of three syndromes: an integrative overview. Quarterly Review of Biology herbivorous gastropods (Trochidae: Tegula) in a temperate kelp 79:241–277. forest habitat. Oecologia 62:47–52. Sommer, U., F. Sommer, B. Santer, C. Jamieson, M. Boersma, C. Werner, E. E. 1977. Species packing and niche complementarity in Becker, and T. Hansen. 2001. Complementary impact of copepods three sunfishes. American Naturalist 111:553–578. and cladocerans on phytoplankton. Ecology Letters 4:545–550. Werner, E. E., and J. F. Gilliam. 1984. The ontogenetic niche and spe- Sousa, W. P. 1979. Experimental investigations of disturbance and cies interactions in size-structured populations. Annual Review of ecological succession in a rocky intertidal algal community. Eco- Ecology and Systematics 15:393–425. logical Monographs 49:227–254. Werner, E. E., J. F. Gilliam, D. J. Hall, and G. G. Mittelbach. 1983. ———. 1984. Intertidal mosaics: patch size, propagule availability, An experimental test of the effects of predation risk on habitat and spatially variable patterns of succession. Ecology 65:1918–1935. use in fish. Ecology 64:1540–1548. Stachowicz, J. J., and M. E. Hay. 1999. Reduced mobility is associated West, L. 1988. Prey selection by the tropical snail Thais melones:a with compensatory feeding and increased diet breadth of marine study of interindividual variation. Ecology 69:1839–1854. crabs. Marine Ecology Progress Series 188:169–178. Wolf, M., and F. J. Weissing. 2012. Animal personalities: conse- Stan Development Team. 2016. Rstan: the R interface to Stan, ver- quences for ecology and evolution. Trends in Ecology and Evolu- sion 2.14.1. tion 27:452–461. Steneck, R. S., M. H. Graham, B. J. Bourque, D. Corbett, J. M. Erlandson, Woo, K. J., K. H. Elliott, M. Davidson, A. J. Gaston, and G. K. Davoren. J. A. Estes, and M. J. Tegner. 2002. Kelp forest ecosystems: biodiversity, 2008. Individual specialization in diet by a generalist marine pred- stability, resilience and future. Environmental Conservation 29:436– ator reflects specialization in foraging behaviour. Journal of Animal 459. Ecology 77:1082–1091. Svanbäck,R.,andD.I.Bolnick.2007.Intraspecific competition drives Wood, C. M., S. T. McKinney, and C. S. Loftin. 2017. Intraspecific increased resource use diversity within a natural population. Pro- functional diversity of common species enhances community sta- ceedings of the Royal Society B 274:839–844. bility. Ecology and Evolution 7:1553–1560. Svanbäck, R., P. Eklöv, R. Fransson, and K. Holmgren. 2008. Intraspe- cific competition drives multiple species resource polymorphism in Associate Editor: Richard Svanbäck fish communities. Oikos 117:114–124. Editor: Daniel I. Bolnick

This content downloaded from 169.237.066.128 on August 01, 2018 09:36:51 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).