Functional Ecology

2007

21

, 793–804

793

© 2007 The Authors.Journal compilation© 2007 British Ecological Society

Blackwell Publishing Ltd

Functional diversity within a morphologically conservative genus of predators: implications for functional equivalence and redundancy in ecological communities

WILLIAM J. RESETARITS*† and DAVID R. CHALCRAFT‡

*

School of Biological Sciences, University of Southampton, Southampton SO16 7PX

,

UK

, ‡

Department of Biology, East Carolina University, Greenville, NC 27858-4353

,

USA

Summary

1.

The idea that sets of species may have similar effects on population, community orecosystem processes is a prevalent theme in many areas of ecology, especially in thecontext of biodiversity and ecosystem function. If indeed species are functionallyequivalent, limiting similarity suggests that it should be closely related, morphologicallysimilar species using similar resources in a similar manner.

2.

We assayed the functional equivalence of three congeneric, morphologically similarpredatory fish species (genus

Enneacanthus

). Functional equivalence was evaluatedusing aspects of both effects of fish on a variety of prey responses and the growthresponses of the fish themselves as a measure of energy consumption. Fish werematched by initial size to control for effects of body size. A strict definition of functionalequivalence based on niche theory was used to delineate it from the alternative offunctional diversity.

3.

Based on observed effects on larval anurans, only a single species pair could roughlybe judged functionally equivalent, but these two species showed the greatest differencesin growth rate and, hence, metabolic demand. Using the criterion of relative yield total,again, only a single pair could roughly be judged equivalent, however, members of thisalternative species pair were dramatically different in their effects on larval anurans.Thus, as previously shown for a more diverse set of species, grouping of species bysimilarity in effects depends upon the specific response variable.

4.

Overall range of effects produced on a variety of response variables was surprising,given the similarity in morphology and autecology, strong phylogenetic affinity, andthe fact that neither predator size nor growth explained significant variation. Each spe-cies appears to be interacting with the environment in a different manner, either as aconsequence of differences in metabolic demand or differences in preferences or effi-ciency with regard to prey types.

5.

Observed responses are consistent with the predictions of niche theory and supportan alternative explanation for observed relationships between diversity and ecosystemfunction. Our work suggests that functional equivalence may be uncommon, difficult topredict

a priori

, and that functional diversity, not functional equivalence, may underlieobserved diversity–ecosystem function relationships.

Key-words

: ecosystem function, functional redundancy, niche, species diversity, species turnover

Functional Ecology

(2007)

21

, 793–804doi: 10.1111/j.1365-2435.2007.01282.x

Functional Ecology

(2007)

xx

, 000–000

Introduction

Functional equivalence in community ecology encom-passes the idea that species in natural communities are

similar in effects on population, community or ecosystemprocesses, and therefore play the same ‘ecological role’(Lawton & Brown 1993; Kurzava & Morin 1998;Chalcraft & Resetarits 2003a,b; Loreau 2004), analogousto the concept that different structures may serve thesame function in different organisms (

sensu

Alfaro,Bolnick & Wainwright 2005). Ecology has a long history

†Author to whom correspondence should be addressed.E-mail: wjr@soton.ac.uk

794

W. J. Resetarits & D. R. Chalcraft

© 2007 The Authors.Journal compilation© 2007 British Ecological Society,

Functional Ecology

,

21

, 793–804

of describing similar species, particularly those withallopatric ranges, as ‘ecological equivalents,’ suggestingthe same functional role in their respective communities(e.g. Beauchamp & Ullyott 1932; Elton & Miller 1954;Ross 1957; Pianka 1973; Bazzaz & Pickett 1980). Thisconcept has been extended to species within com-munities and is now prevalent in both food web theory(e.g. Hairston, Smith & Slobodkin 1960; Fretwell 1977;Oksanen

et al.

1981; McQueen, Post & Mills 1986;Menge & Sutherland 1987), and theory concerning therelationship between diversity and ecosystem function(Lawton & Brown 1993; Aarssen 1997; Huston 1997;Allison 1999; Petchey & Gaston 2002a, 2006; Loreau2004). It also constitutes the fundamental assumptionof neutral theory (Hubbell 2005, 2006).

The idea of functional equivalence is specificallyembodied in the concept of ‘functional groups;’ (i.e.entities within which functional equivalence exists),(Lawton & Brown 1993; Stone, Dayan & Simberloff2000; Chalcraft & Resetarits 2003a; Loreau 2004;Petchey & Gaston 2006). Functional groups are typicallyidentified

a priori

based on similarity in morphology,life-history, and/or phylogeny (Schwartz

et al.

2000;see Chalcraft & Resetarits 2003b for review) and usedin the formulation of community assembly rules (Fox1987; Fox & Brown 1995). Though community assemblyrules based on functional groups remain controversial(e.g. Fox & Brown 1995; Wilson 1995; Stone, Dayan &Simberloff 1996; Brown, Fox & Kelt 2000; Stone

et al.

2000), as does the dependence of ecosystem function onfunctional redundancy (Chalcraft & Resetarits 2003a,b;Loreau 2004; Petchey & Gaston 2006), the idea that suchsimple criteria as membership in a genus, functionalgroup or guild can inform us about both communityassembly and ecosystem function is very compelling(Stone

et al.

2000).But does functional equivalence exist within com-

munities, and can we predict

a priori

which speciesare functionally equivalent and thus potentially inter-changeable or compensatory in effect? Functionalequivalence would suggest that ecologists can simplifysystems by focusing on functional groups (i.e. setsof such functionally equivalent species) rather thanindividual species (Leibold & McPeek 2006). Iffunctionally equivalent species co-occur within localcommunities; species can presumably be lost withoutsignificant negative effects on ecosystem function; ifthey co-occur at the regional level, regional loss ofspecies can be compensated by expansion of suchequivalent species (e.g. Fox 1987; Walker 1992; Naeem1998; Yachi & Loreau 1999; Jaksi 2003; Loreau 2004).The majority of data brought to bear on functionalredundancy is indirect, deriving from the observationthat experimental reductions in species diversity cancause a nonlinear change in ecosystem function(Naeem

et al.

1995; Naeem 1998; Wootton & Downing2003). Moreover, citing as evidence of redundancy thefact that different species may have very similar effectson a few very broad measures of ‘ecosystem function’

(Kurzava & Morin 1998; Chalcraft & Resetarits 2003a)ignores the vast complexity of ecological systems andthe variety of so-called ‘services’ they provide. Com-munities that dramatically differ in species compositionmay have the same levels of community respiration,productivity or similar measures, but provide dramatic-ally different levels of other important variables. Forexample, aquatic communities dominated by differenttaxa at a given trophic level (e.g. herbivorous tadpoles

vs

herbivorous zooplankton) may have similar totalherbivore production (one measure of ecosystemfunction), but the rate at which energy is transportedinto terrestrial communities as spatial subsidies (anothermeasure of ecosystem function) as a result of meta-morphosis is totally different (tadpoles metamorphoseinto terrestrial adults whereas zooplankton do nothave a terrestrial stage). Similarly, the extent of linkagebetween discrete aquatic communities and across theaquatic-terrestrial interface is markedly different if theherbivore trophic level is dominated by species withmarkedly different life-histories (Resetarits, Binckley& Chalcraft 2005).

Experiments have addressed whether specific speciespairs are functionally equivalent, however, there aredifficulties in using pairwise experiments conducted atdifferent times, by different investigators, under differ-ent conditions, to gain a general understanding of thepotential prevalence and distribution of functionalequivalence (Chalcraft & Resetarits 2003a). Paine (1992)found that a variety of consumers were functionallyequivalent largely as a consequence of being weakinteractors, while Harris (1995) found general equi-valence among a set of herbivores, but suggested thathers was not a fully adequate test. Our previous workidentified little functional equivalence among predatorsthat varied in higher level taxonomy, habitat and gape(Chalcraft & Resetarits 2003a,b). These studies dealtwith morphologically and taxonomically diverse speciesthat, nonetheless, could potentially be grouped undercommonly used criteria (trophic level, guild, taxon, etc.).

The competitive exclusion principle and the conceptof limiting similarity suggest that if equivalence exists,it is most likely among closely related, morphologi-cally similar species that use similar resources in asimilar manner (Fox 1987; Petchey & Gaston 2002a;Loreau 2004; Leibold & McPeek 2006). Functionalgroups, however, are typically identified

a priori

basedon similarity in morphology, phylogeny or life-historyusing a rather broad brush approach (Schwartz

et al.

2000; Chalcraft & Resetarits 2003b; Wright

et al

.2006). For example, levels of morphological similarityused to assign species to functional groups are rarelybased on empirical assessment of a species fundamentalniche or its functional morphology in relation to poten-tial niche axes. The basic question is; can we assumefunctional equivalence at any level in the absence ofmore detailed empirical assessment? The fundamentalutility of the functional group concept is that one canlump organisms by means of readily identifiable traits

795

Experimental test of functional equivalence

© 2007 The Authors.Journal compilation© 2007 British Ecological Society,

Functional Ecology

,

21

, 793–804

and then ignore differences below the level of thefunctional group; if we must empirically determinethe makeup of functional groups, or if functionalequivalence is rare, the utility is lost. Thus, experimentsare needed that examine functional equivalence at themost basic level, within sets of taxonomically andmorphologically similar species that would form thefundamental unit of any

a priori

classification of func-tional equivalence; ‘... our evolutionary considerationsdo suggest that ecologically similar species are morelikely to be closely related to one another ...’ (Leibold &McPeek 2006). Such species serve as a benchmark forexamination of functional equivalence and its poten-tial contribution to ecosystem function via functionalredundancy.

We tested functional equivalence among three speciesthat comprise the monophyletic centrarchid genus

Enneacanthus

(Sweeney 1972; Roe, Harris & Mayden2002). Following earlier work (Harris 1995; Chalcraft& Resetarits 2003a; Loreau 2004), we take a strict viewof functional equivalence; functional equivalence isthe lack of a significant difference in performancebetween monospecific cultures of species under thesame environmental conditions (Chalcraft & Resetarits2003a,b, Loreau 2004), or

sensu

Harris (1995) ‘in theabsence of further information it would seem wise toavoid grouping together any species that are knownto have statistically significant differences amongtheir effects ...’ Performance can be assayed in manyways, including growth and reproduction (e.g. relativeyield totals; Loreau 2004), and suppression of a givenresource(s) (e.g. Chalcraft & Resetarits 2003a,b). Anyless rigorous definition confuses functional equivalencewith functional diversity, which is an equally viablemechanism for the provision of ecological ‘insurance’(Petchey & Gaston 2002a,b, Chalcraft & Resetarits2003a,b; Loreau 2004; Petchey & Gaston 2006). In thiscontext, functional equivalence supports ecosystemfunction via functional redundancy, whereas functionaldiversity supports ecosystem function via functionalcomplementarity (Kurzava & Morin 1998). We discussthe importance of the distinction and the implicationsfor ecological and evolutionary theory below. In thepresent study, functional equivalence was not a featureof members of the genus

Enneacanthus

despite the sim-ilarity in morphology, life-history and phylogeny thatwould typically place them in the same functional group.

Materials and methods

Enneacanthus

is a distinctive genus of dwarf sunfish(Centrarchidae) confined primarily to the Atlanticslope of the eastern USA (Sweeney 1972).

Enneacan-thus chaetodon

(Baird),

E. gloriosus

(Holbrook) and

E. obesus

(Girard) (Fig. 1), are broadly sympatric in avariety of habitats (e.g. small ponds, river swamps,reservoirs).

Enneacanthus gloriosus

is the most commonand uniformly distributed, while

E. chaetodon

and

E.obesus

are typically found in more acidic, oligotrophic

habitats (Sweeney 1972; Gonzalez & Dunson 1989,1991). All pairs may be syntopic, but very rarely allthree.

Enneacanthus gloriosus

is the largest (adult size:61–78 mm SL), followed by

E. obesus

(51–70 mm SL)and

E. chaetodon

(29–66 mm SL). All three aremorphologically conservative, generalist predatorson small invertebrates and vertebrates. These specieswould clearly meet the

a priori

criteria for inclusionin the same functional group for most adherents, butdetailed studies of functional morphology and aute-cology for these, as well as most other species, couldlikely identify differences sufficient to support differ-ences in function (e.g. Graham 1986). Nonetheless,broad distributional overlap, similarity in morphologyand autecology, and importance in small lentic habitatsin the Atlantic drainage (e.g. Resetarits & Wilbur 1989;Bristow 1991; Binckley & Resetarits 2003; Chalcraft &Resetarits 2003a,b, 2004) make the genus an excellentcandidate for testing functional equivalence.

Fig. 1. Enneacanthus obesus (top), E. gloriosus (middle), E.chaetodon (bottom), illustrating morphological similarity.Drawings by R. Kuhler used with permission of the NorthCarolina State Museum. Reproduced with permission fromLee et al., 1980. Atlas of North American Freshwater Fishes.N.C. State Museum of Natural Sciences.

796

W. J. Resetarits & D. R. Chalcraft

© 2007 The Authors.Journal compilation© 2007 British Ecological Society,

Functional Ecology

,

21

, 793–804

We evaluated how a three-species ensemble of larvalanurans responded to variation in predator identity inreplicated artificial ponds. Our experiment consistedof four treatments, three species of

Enneacanthus

anda predator-free control, each replicated four timesin a randomized complete block design. We createda hexagonal array of 16 artificial ponds (1100 L cattletanks) in a clearing at Naval Security Group ActivityNorthwest (NSGANW) in SE Virginia, identifyingfour spatial clusters of four ponds each as statisticalblocks. We conducted all procedures (e.g. creation,manipulation and monitoring of ponds) by block tominimize variation within blocks not attributableto treatment.

We filled tanks on 5–6 April 2001 to 50 cm (volume:

c.

1000 L) with water from a nearby borrow-pit. A 2 mmmesh filter placed over the pump intake preventeduptake of larval anurans or their predators. Zooplankton,periphyton and phytoplankton could pass throughthe filter and served as alternative prey for predatorsand food for larval anurans. Immediately after filling,we covered tanks with tight-fitting fibreglass win-dowscreening to prevent unwanted colonization bypredatory insects and ovipositing treefrogs, and to retainexperimental organisms. On 7 April 2001 we added toeach tank 1 kg of forest leaf litter for structural com-plexity and nutrients, and a supplemental 500-mLmixture of zooplankton, periphyton, phytoplanktonand macroinvertebrates from ponds and ditches atNSGANW (potential predators were removed).

Each tank received 300 newly hatched tadpoles eachof

Bufo terrestris

(Bonnaterre),

Rana sphenocephala

(Cope) and

Hyla squirella

(Bosc) on 11 June 2001 (initialdensity of 900 individuals/pond). Tadpoles repre-sented 12 clutches of

Bufo

, 8 of

Rana

and 37 of

Hyla

,differences reflecting clutch size. Each pond receiveda similar fraction from each clutch to ensure similargenetic diversity among ponds within species. Nextday we randomly assigned treatments to ponds withineach block with predator treatments each receivingone fish. We matched size ranges for our juvenile fish,ranked them by size within species, and assigned indi-viduals of the same rank size to blocks [initial mass(mean ± 1 SE): 2·828 ± 0·279 g (

E. chaetodon

), 2·910 ±0·327 g (

E. gloriosus

), 2·750 ± 0·272 g (

E. obesus

)]. Initialdensities fell within the range observed for adults innatural ponds (Bristow 1991; Chalcraft & Resetarits,personal observation). Using identical densities andmatched body sizes allows evaluation of per capitaeffects independent of differences in density and size.

We monitored the ponds daily, collected anurans asthey initiated metamorphosis (emergence of at leastone forelimb), and weighed metamorphs (wet mass)upon tail resorption. Between 24 and 25 September2001 (total duration 105–106 days) we drained allponds and counted and weighed survivors, producinga complete census of survivors. Most

Bufo

had com-pleted metamorphosis, while only a single

Rana

hadmetamorphosed. No

Hyla

survived in any tanks.

data analysis

We quantified predator effects on individual speciesresponses (survival and mass for

Rana

tadpoles; survival,metamorph mass and proportion metamorphs for

Bufo

)and larval ensemble responses (number and totalbiomass of anurans and final proportional abundanceof

Bufo

). Proportion metamorphs is a surrogate fordevelopment rate, measuring the fraction of thepopulation passing a particular developmental stage(metamorphosis). Total number and biomass measureimportance of larval anurans to pond food webs; foodweb models suggest ensembles with greater biomassor number have greater impacts on food resources(Hairston

et al.

1960; Fretwell 1977; Oksanen

et al.

1981;McQueen

et al.

1986; Menge & Sutherland 1987).Analysis of final proportional abundances detectedchanges in composition independent of changes in totalnumber (Morin 1983; Chalcraft & Resetarits 2003a).Proportional abundance of

Hyla

was not included as itwas invariant (0%), thus, proportional abundance wassolely dependent on the final proportional abundanceof either remaining species (we chose

Bufo

). Survivalwas angularly transformed (arcsine squareroot),proportion metamorphs squareroot transformed (thelatter because values were uniformly high), and bothmeasures of anuran importance were log

10

transformed.Three separate

manova

s were used to examine theeffects on: (i)

Rana

responses; (ii)

Bufo

responses;and (iii) larval ensemble responses. We also analyzedmean total biomass of

Bufo

metamorphs as a measureof energy export to the terrestrial environment using

anova

.Use of several

manova

s on logical data subsets ispreferable to a single

manova

with all response vari-ables, as a large number of variables dramatically reducethe power. Furthermore, interpretation of

manova

sconducted on subsets are more straightforward asresponse vectors describe a single effect of interest(Chalcraft & Resetarits 2003a). We also conductedunivariate

anova

s on each component (e.g. mass andsurvival) of the response vector, and pairwise com-parisons of means following only significant

anova

susing Fisher’s LSD. Such protected comparisonsusing LSD reduce experimentwise Type I errors whilealso minimizing the occurrence of Type II errors(Carmer & Swanson 1973; Bernhardson 1975; Carmer& Walker 1982), and are more consistent in detectingspecific treatment differences than other methods(Saville 1990, 2003). Block explained almost no varia-tion for any response variable (exclusion/inclusion didnot affect test outcomes), so block was rolled into theerror term for simplicity of presentation.

Results

Survival was 100% for all

Enneacanthus

species and allindividuals grew substantially (mean mass in grams[mean % change] ± 1 SE:

E. chaetodon

5·73 [203%] ±

797

Experimental test of functional equivalence

© 2007 The Authors.Journal compilation© 2007 British Ecological Society,

Functional Ecology

,

21

, 793–804

0·21,

E. gloriosus

9·58 [329%] ± 0·42;

E. obesus

7·79[283%] ± 0·88), attaining adult body size by terminationof the experiment. Growth differed significantly amongspecies (

F

2,8

= 9·99,

P

= 0·01), but initial size didnot significantly affect growth (

F

1,8

= 0·05,

P

= 0·83).Multiple comparisons revealed that

E. chaetodon

growthwas significantly different from that of

E. gloriosus

and

E. obesus

, but the latter two were not significantlydifferent.

Treatment had a highly significant effect on

Rana

survival and mass (

manova

, Wilk’s

λ

= 0·2507,

F

6,22

=3·66,

P

= 0·0114; survival

F

3,12

= 5·53,

P

= 0·0128;mean tadpole mass

F

3,12

= 8·66,

P

= 0·0025) (Fig. 2a).The

E. obesus

treatment had significantly lowersurvival of

Rana

than

E. chaetodon

and Control tanks,while the

E. gloriosus

treatment did not differ signifi-cantly from other treatments (Fig. 2a).

Rana

with

E. obesus

were significantly larger than with otherpredators or in Controls (Fig. 2a).

Treatment had a significant effect on

Bufo

survivaland proportion metamorphs but not on mass atmetamorphosis (

manova

Wilk’s λ = 0·2224, F9,24·5 =2·33, P = 0·0471; survival F3,12 = 3·54, P = 0·0483;proportion metamorphs F3,12 = 4·95, P = 0·0183; massF3,12 = 0·88, P = 0·4778) (Fig. 2b). Bufo with E. obesus

had significantly lower survival than with E. chaetodon;Control and E. gloriosus tanks were intermediate andnot significantly different from the other treatments(Fig. 2b). A greater proportion of Bufo metamor-phosed in E. obesus and E. gloriosus treatments than inControl and E. chaetodon treatments (Fig. 2b); therewas no significant difference in Bufo mean mass atmetamorphosis among treatments (Fig. 2b inset).

From a whole ensemble perspective, manova

revealed highly significant variation among treatmentsin the multivariate vector of total biomass, total numberof survivors, and relative abundance of Bufo (Wilk’sλ = 0·1359, F9,24·488 = 3·45, P = 0·007). In univariateanovas only total number was significant (F3,12 = 17·92,P < 0·0001, total biomass F3,12 = 1·92, P = 0·18, relativeabundance F3,12 = 2·22, P = 0·1389) (Fig. 2c). Multiplecomparisons revealed that the difference in totalnumber was driven by the E. obesus treatment.

There was also significant variation among treat-ments in the extent of the spatial subsidies from theaquatic to the terrestrial environment, as measured bythe total biomass of Bufo metamorphs (univariateanova, F3,12 = 3·75, P = 0·041). In multiple compari-sons, the E. obesus treatment was significantly differentfrom both the E. chaetodon and E. gloriosus treatments,

Fig. 2. (a) Mean ± 1 SE for survival and mass of Rana, (b) mean ± 1 SE for survival, proportion metamorphs and mass (inset– no significant differences), (c) mean ± 1 SE for total number of survivors, relative abundance of Bufo, and total biomass (inset– no significant differences). Treatments sharing the same letter/case are not significantly different; capital letters represent thevariable on the y-axis, lower case the variable on the x-axis, (d) relationship between individual predator growth and totalnumber of survivors, illustrating lack of size/growth effects both among and within species. Legend same as in (a–c). Inset graphsin (b, c) represent variables with non-significant anovas, thus, multiple comparisons among the means were not performed.

798W. J. Resetarits & D. R. Chalcraft

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Functional Ecology, 21, 793–804

but not from the controls, while the Control andE. chaetodon treatments were also significantly different(Fig. 3).

Interestingly, very little variation in any responsevariable was explained by initial mass of predators orpredator growth (change in mass) (all r2 < 0·09 and allP > 0·35; e.g. Fig. 2d), which partly explains the lackof a block effect. Indeed, the rank order of effect size,which was consistent across all response variables(Fig. 2) (E. obesus > E. gloriosus > E. chaetodon) didnot mirror either mean initial mass (see Materials andmethods: E. gloriosus > E. chaetodon > E. obesus),which varied little among species, or growth (E. gloriosus >E. obesus > E. chaetodon), which varied considerablyamong species, as well as within E. obesus (Fig. 2d).

Discussion

functional equivalence in E N N E A C A N T H U S

We previously demonstrated that functional equi-valence is not characteristic of diverse sets of predatorsoften designated as functional groups (Chalcraft &Resetarits 2003a), and that species sharing broadlysimilar traits (e.g. taxonomic affinities, gape size ormicrohabitat use) are not necessarily more similar intheir functional effects (Chalcraft & Resetarits 2003b).The present study strengthens those conclusions bydemonstrating that, even within a small, morpho-logically conservative genus of predators, functionalequivalence is less likely than functional diversity. Ofthree species pairs, only one exhibits similarity remotelyconcordant with functional equivalence using thecriterion of effect on specific resources (larval anurans)(Chalcraft & Resetarits 2003a,b). Using similarity ofyield criterion (Loreau 2004) (similarity in ability to

extract energy from a community under similar con-ditions – similar to ‘importance’ in food webs), againonly one pair could be judged functionally equivalent.Yet, it is a different species pair and equivalence istenuous because of substantive differences in growthrate variation among individuals of the two species.

Enneacanthus obesus is distinct from both E. gloriosusand E. chaetodon in effects on almost all variables exam-ined, except growth, where it is closer to E. gloriosusbut shows considerable individual variation (Fig. 2d).Taxonomic and morphological similarities do notensure similarity of effects; in the case of E. obesus, noteven within a species. Differences in effects amongspecies occurred despite no differences in initial size,and size and growth rate explained virtually no vari-ation in effects on larval anurans. Further, similarityof effects could not be predicted based on similarityin morphology, as E. obesus is no less similar to E.chaetodon than is E. gloriosus (Sweeney 1972; Fig. 1),while molecular evidence indicates E. gloriosus andE. obesus are more closely related (Roe et al. 2002).

Figure 2d suggests an explanation for differences ineffect. Given similarity of morphology and physiology,differences in growth (i.e. final body size) (sans re-production) should indicate differences in metabolicdemand and functional non-equivalence (Chalcraft &Resetarits 2004; Loreau 2004). Enneacanthus chaeto-don exhibited relatively low growth and small effectson target prey. In contrast, E. gloriosus exhibited thehighest growth (> 60% higher than E. chaetodon), buthad similar effects on tadpole prey. Lastly, E. obesusdisplayed variable growth rates, overlapping with bothE. chaetodon and E. gloriosus, but had uniformlystrong effects on tadpoles. Enneacanthus chaetodon, asthe smallest member of the genus, likely had smalleffects on target prey because of relatively small meta-bolic demand tied to limited growth potential. Similareffects of E. gloriosus, despite high growth rate, suggestthat E. gloriosus focused on alternative invertebrateprey. Strong effects of E. obesus, irrespective of indi-vidual growth rate and metabolic demand, suggestthey are effective predators on tadpoles and may preferthem over the available invertebrate prey. Each speciesdoes something sufficiently different to generate differ-ent effects on prey assemblages. This is precisely whatniche theory would predict (Vandermeer 1972, Chase& Leibold 2003; Loreau 2004).

Even if productivity, community respiration orother metrics of ecosystem function remain the samewithin our aquatic habitats because of relative similarityin total anuran biomass, larval anurans are temporaryresidents of aquatic systems. Significant variation inthe output of metamorphs based on predator identity(Fig. 3) signals variation in export of energy to surround-ing terrestrial systems (spatial subsidies) and variationin connectivity between aquatic and terrestrial systems,as well as between discrete aquatic communities. Theseshifts in energy flow and connectivity with differentpredators further suggest that a single metric or limited

Fig. 3. Variation in energy transfer from the aquatic to theterrestrial environment as illustrated by the production ofBufo metamorphs in the four treatments (mean ± 1 SE).Shared letters indicate means that are not significantlydifferent. The E. obesus treatment produces significantlylower subsidies to the terrestrial environment than either E.chaetodon or E. gloriosus.

799Experimental test of functional equivalence

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Functional Ecology, 21, 793–804

set of metrics may lead to misleading and oversimplifiedresults.

What does it mean from the perspective of com-munity function if only one of three species pairs isfunctionally equivalent, at least with respect to larvalanurans? Or with respect to overall energy demandor export? At the local or regional level, only onepair may be substitutable with respect to any specificresponse variable. If E. gloriosus and E. chaetodoncoexist in a given community and one is lost, theoverall structure and function of the community maybe preserved from the perspective of larval anurans(Loreau 2004), but not necessarily for zooplankton.However, if either of the other two species pairs coexistand one is lost, the basic structure and function of thecommunity will necessarily change. On a regional scale,if we assume one species of Enneacanthus per com-munity (largely the case in nature), replacement ofone species by another in no way assures preservationof existing community structure and function. Inter-estingly, E. gloriosus and E. obesus are broadly sympatric,while E. chaetodon has a more restricted and sporadicdistribution. When all three are sympatric, E. obesusand E. chaetodon are more likely syntopic (Lee & Gilbert1980). Thus, across their distribution, co-occurring pairsare more likely complementary than equivalent ineffects on community structure. Again, given the basictenets of niche theory (Vandermeer 1972; Chase &Leibold 2003; Loreau 2004), this is hardly surprising.

Despite work dispelling the assumption that fre-quency and intensity of competition are always bestpredicted by morphological similarity (e.g. Brown &Davidson 1977; Polis & McCormick 1986; Morin,Lawler & Johnson 1988; Resetarits 1995a,b), limitingsimilarity remains a reasonable starting point forexamining functional equivalence, because for similarspecies both effects on a community and ecologicaltolerances should be reasonably concordant (Wellnitz& Poff 2001). Our data illustrate the difficulty ofpredicting functional equivalence a priori, as morpho-logically and taxonomically similar species are notnecessarily equivalent (Harris 1995; Chalcraft &Resetarits 2003a,b; Wright et al. 2006). Morphologicalsimilarity, similar phylogeny, guild membership, andeven the combination of the three seen in Enneacanthusdo not equate to functional equivalence. This leadsus back to consideration of the species itself as thefundamental unit for defining functional roles.

prospects for functional equivalence within and among communities

Recent work suggests that, in terms of both theeffects on ecological communities (functional roles)and responses to environmental variation (ecologicalresponses), species identity is critical. Spatial andtemporal turnover in species is associated with changesin the dynamics and functioning of the communities,metacommunities and ecosystems they comprise

(Leibold, et al. 1997; Chase & Leibold 2002, Chalcraft& Resetarits 2003a,b; Chalcraft et al. 2004; Resetarits2005). What does this mean for the idea of functionalequivalence? We suggest, based on our work on pre-dators and on fundamental principles of niche theory,that functional equivalence should be rare withincommunities and that the mechanisms underlyingdiversity–function relationships derive most commonlyfrom the diversity of ecological roles and the complexinteractions within species rich communities (Petchey& Gaston 2002b; Chalcraft & Resetarits 2003a; Loreau2004; Petchey & Gaston 2006), rather than from theprovision of insurance specifically in the form ofequivalence and redundancy (e.g. Elton 1958; Walker1992; Naeem 1998; Yachi & Loreau 1999; Hubbell2005, 2006).

Functional equivalence within communities nec-essarily implies that communities do not meet theequilibrium requirements for competitive exclusion(Gause 1934; Hardin 1960). Functional diversity impliesthat differences among species within communitiesexceed the ‘limiting similarity’ compatible withequilibrium coexistence (Hutchinson 1959; MacArthur& Levins 1967; Abrams 1983). We have known sinceGause (1934) that asymmetry in the competitioncoefficients is necessary for competitive exclusion,thus, it has always been an option that species musteither sufficiently differ in niche requirements, or mustbe sufficiently similar in both niche requirements andcompetitive abilities, in order to coexist (Aarssen 1983;Ågren & Fagerström 1984). Despite the possibility ofcoexistence by similarity, widespread functional equi-valence within communities remains largely antitheticalto ecological and evolutionary theory beginning withDarwin (Darwin 1859; Hutchinson 1959; Vandermeer1972; Chase & Leibold 2003; Loreau 2004; Leibold &McPeek 2006). While neutral theory itself and recentconsiderations of the potential relationship betweenniche and neutral processes have reinforced the idea ofthe evolution of ecological equivalence and its per-sistence in at the local scale within communities, noneof these considerations (with the obvious exception ofstrict neutral theory; Hubbell 2005, 2006) suggest thatco-occurring sets of functionally equivalent speciesare the norm within ecological communities (Leibold& McPeek 2006; Scheffer & van Ness 2006). Plausiblescenarios can be constructed that support local co-existence of ecologically equivalent species, but nicheprocesses are still considered the dominant force indetermining patterns of species coexistence (Aarssen1983; Ågren & Fagerström 1984; Leibold & McPeek2006; Scheffer & van Ness 2006).

While niche theory and the importance of com-petition have both fallen into and out of favour overthe years, the general ideas remain central to our viewof how ecological communities are assembled, as doesthe role of individual selection. And these conceptsare largely incompatible with the idea of functionalequivalence (Loreau 2004). Functional groups do not

800W. J. Resetarits & D. R. Chalcraft

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Functional Ecology, 21, 793–804

evolve; that is, the purview of populations and species.Scheffer & van Ness (2006) have demonstrated thatclusters consisting of essentially equivalent species canevolve and persist in theoretical communities, yet thisoccurs only in ‘saturated’ communities, and again isnot expected to be the norm. This reinforces the ideathat functional redundancy should not hold primacyas the mechanism driving relationships betweenbiodiversity and ecosystem function. Thus, althoughfunctional equivalence among species in ecosystemsexperiencing diversity loss might be beneficial toecosystem function, the ecological and evolutionaryimpetus for species to remain or become more similarin their resource use in sympatry remains problematic.The idea of character convergence, while not as oldas that of character displacement, has a considerablehistory (MacArthur & Levins 1967; Cody 1973). Yet,it has always been postulated as the exception ratherthan the rule in ecological communities (MacArthur &Levins 1967; Cody 1973; Slatkin 1980; Scheffer & vanNess 2006). Even its most ardent supporters neversuggested that ecological communities are dominatedby convergence of competitors. And while the Clem-entsian ideas of community selection and communityevolution have recently threatened to re-emerge (Morin2003; Wilson & Swenson 2003), conceivably providinga ‘mechanistic’ framework for functional equivalence,until we reconcile such notions with individual selec-tion and evolution we should be cautious in claimingthat effects of diversity on ecosystem function derivefrom functional equivalence and functional redun-dancy. This is especially true when an equally viableexplanation for ecological insurance, functionaldiversity, is supported by a vast corpus of ecologicaland evolutionary theory that suggests equivalenceshould be rare within communities and when present,selection should most often work to reduce such equi-valence (Fig. 4).

Lost in the discussion of functional equivalence aredifferences in expectation between species functioningin a redundant capacity at the local scale (within com-munities) and those redundant at the regional scale(among communities – within metacommunities).Niche theory tells us expectations for redundancy atthese two scales are not necessarily equivalent, becauselimiting similarity and competitive exclusion play amajor role in restricting coexistence at the local scale(Fig. 5), but may enhance coexistence at the regionalscale (Loreau & Mouquet 1999; Amarasekare & Nisbet2001; Mouquet & Loreau 2002). Under equilibriumconditions, local redundancy is constrained bylimiting similarity and competitive exclusion. If non-equilibrium mechanisms drive the local coexistence ofsimilar competitors, the consequences for local redun-dancy are modified or constrained by the dynamics ofmore complex interactions among multiple species (i.e.keystone predators) or the continuing effects of abioticfactors such as disturbance. Maintenance of redun-dancy is then dependent upon other species/processes

and is no longer a simple property of functionalgroups. Operationally, predicting functional equiva-lence at the local scale is highly problematic becausepredicting the limiting similarity and thus the optimumsimilarity for coexistence and functional equivalence isnot straightforward, and is likely an empirical question(Fig. 5).

Fig. 4. Projections of species functional roles withincommunities through time on the basis of niche theory.Functional equivalence implies that differences among spe-cies within a functional group are less than the minimum levelof similarity compatible with equilibrium coexistence(limiting similarity). In contrast, functional diversity impliesthat the differences are ≥ the minimum level of similaritycompatible with equilibrium coexistence, specifically theconditions expected in diverse, multi-species communities.The likelihood of finding functionally equivalent species in acommunity is dependent on: (i) the frequency with whichparticular ecological processes occur; (ii) the presence offunctionally equivalent species in the regional species pool;and (iii) the co-occurrence of these functionally equivalentspecies within the community in the past or the evolution offunctional similarity in situ. A large body of existing theoryand data suggest that very similar species are unlikely tocoexist in communities due to competitive exclusion (a).There is limited evidence suggesting that functional equi-valence may persist in communities if the intensity of competi-tive interactions are very strong (b) or if character convergenceoccurs (c). However, on the basis of niche theory and aconsiderable body of empirical work, the more likelyscenarios are that competition among functionally equivalentspecies forces niche divergence or character displacement(which promotes functional diversity, (d) or that the speciespresent in the community were never functionally equivalentand their coexistence derives from functional diversity (e).Consequently, the three most likely scenarios (a,d,e) supportthe idea of functional diversity within communities, whilethe remaining two scenarios (b,c) that support functionalequivalence have not been frequently documented, norare strongly supported by theory (see text for additionaldiscussion).

801Experimental test of functional equivalence

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Functional Ecology, 21, 793–804

Similarly, the idea that entire communities areassembled based primarily on species similarity, ratherthan species dissimilarity and the resultant sortingamong a more diverse set of ecological roles, has littleempirical or theoretical support; ‘Consequently, itseems unlikely that entire assemblages of species willevolve to be equivalent, even though subsets of spe-cies may’ (Leibold & McPeek 2006). Leibold (1998)suggest that species similarities within ensemblesresult from general similarities in species responses tothe environment: Chesson’s (2000) ‘equalizing effects’(e.g. Emberton 1995), whereas species differences resultfrom or reflect species impacts on resources (whichdefines functional roles in our study; Chesson’s ‘stabiliz-ing effects’ (2000)).

Vandermeer et al. (2002) recently suggestedincreased levels of competition may promote speciescoexistence under certain conditions. While applica-tion of this idea (as well as previous ideas on coexist-ence through similarity) to functional equivalence lendssupport, it also points to a second major flaw. If similarspecies pack into communities based on similar responsesto the environment, or based on increased resourceoverlap and intensity of competition, then environ-mental change causing species loss should do so notvia uniform loss across functional groups, but by elim-inating similar sets of species, that is, members of afunctional group or entire functional groups. Speciesmore similar in effects on ecological communities are

also likely to be more similar in responses to environ-mental change (Wellnitz & Poff 2001). Thus, extinctionprobabilities should be correlated among members ofa functional group, discounting the value of functionalredundancy for providing ecological insurance. This isespecially true if they share a common phylogenetichistory. Counteracting this, of course, is co-occurringspecies with similar requirements diverging in resourceuse, which is again what niche theory would predict.Such divergence in resource use should, however,generate divergence in functional roles or, underVandermeer et al.’s (2002) scenario, a reduction incompetition and loss of species richness. Neither ofthese processes is compatible with functional equiva-lence underwriting ecological insurance. The broaderconcept of functional groups used to examine patternsof community assembly remains a potentially validconstruct (see Fox 1987; Fox & Brown 1995; Wilson1995; Stone et al. 1996; Brown et al. 2000; Stone et al.2000), but whether functional groups, rather thanniches themselves, form the basis of communityassembly processes is an unresolved question.

Any attempts to simplify ecological systems have acost in lost information. Our recent work and that ofothers suggests that when we simplify the world byignoring species, that cost is often too great (Leiboldet al. 1997; Chalcraft & Resetarits 2003a,b; Loreau2004). While sets of species may have equivalent effectson natural communities, identification of those species

Fig. 5. Graphical model of the expected relationship between functional redundancy and phylogenetic/morphological similarityat the local (community; dash curve) and regional (metacommunity; solid curve) scales based on niche theory (see text). Highβ diversity allows functional redundancy to be higher at the regional scale than at the local scale. Although limiting similarityprevents redundant species from co-occurring at the local scale under equilibrium conditions (except in cases of near identity –see text), processes that generate spatial variation in species composition could allow functionally redundant species to co-occurat the regional scale. If there is little or no species turnover, the maximal level of redundancy at the regional scale is equal to thatat the local scale.

802W. J. Resetarits & D. R. Chalcraft

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Functional Ecology, 21, 793–804

and assessment of their ability to coexist in ecologicalcommunities remain empirical questions. If, as Petchey& Gaston (2006) suggest, ‘For functional classifica-tions ... the correct number of traits is the numberthat are functionally important’, the quantification offunctional diversity (or functional equivalence) withincommunities is a major endeavour in and of itself,comparable in methodology and scope to phylogeneticsor population genetics. This endeavour is justified interms of its potential importance in understandingecosystem function, community structure, and theevolution of diversity within communities, but simul-taneously obviates the idea that simplistic notions offunctional groups and ecological insurance based onfunctional redundancy allow us to ignore, withoutsuch detailed understanding, the basic differencesin form and function implied by the classification oforganisms into species. Patterns may exist in the dis-tribution of functional equivalence among species,however, no simple a priori criterion is likely to providea ‘breakthrough’ (sensu Stone et al. 2000) in simplifyingcommunities and applying assembly rules (Fox 1987;Fox & Brown 1995). This clearly has consequences forhow we continue to approach the study of ecologicalcommunities.

We have demonstrated that even morphologicallysimilar congeners of the same initial size are not char-acterized by functional equivalence. Combined withrecent theoretical (Loreau 2004) and empirical(Chalcraft & Resetarits 2003a,b) evidence, this suggeststhat functional equivalence is likely rare, at least athigher trophic levels, and supports the accompanyingproposition that the positive effects of species diversityon ecosystem function derives, rather, from functionaldiversity. This should not disillusion ecologists in thesearch for pattern and generality. On the contrary, thisproposition derives directly and logically from 150 yearsof ecological and evolutionary theory and accompany-ing empirical research. Rather, it is the prospect thatfunctional equivalence is the norm within communitiesthat should shake us to our foundational roots and forceus back to the drawing board of ecological theory.

Acknowledgements

We thank C.A. Binckley and J.F. Rieger for fieldassistance, CAB, H. Vance-Chalcraft, M. Rosenshield,and several anonymous reviewers for valuable com-ments, and the Center for Aquatic Ecology, IllinoisNatural History Survey, and R. Turner and Com-manding Officers, NSGANW, for additional logisticalsupport. Thanks to the North Carolina State Museumfor allowing us to reproduce the original drawingsof Enneacanthus by R. Kuhler for the Atlas of NorthAmerican Freshwater Fish. Funding was provided byINHS and the Program in Ecology and EvolutionaryBiology, University of Illinois, an NSF-DIG (DEB-99-02188), and NSF (DEB-0096051) and EPA-STAR(R825795-01-0) grants.

References

Aarssen, L.W. (1997) High productivity in grassland ecosystems:effected by species diversity or productive species? Oikos80, 183–184.

Aarssen, L.W. (1983). Ecological combining ability andcompetitive combining ability in plants: toward a generalevolutionary theory of coexistence in systems of competition.American Naturalist 122, 707–731.

Abrams, P. (1983) The theory of limiting similarity. AnnualReview of Ecology and Systematics 14, 359–376.

Ågren, G.I. & Fagerström, T. (1984). Limiting dissimilarity inplants: randomness prevents exclusion of species withsimilar competitive abilities. Oikos 43, 369–375.

Alfaro, M.E., Bolnick, D.I. & Wainwright, P.C. (2005)Evolutionary consequences of many-to-one mapping ofjaw morphology to mechanics in labrid fishes. AmericanNaturalist 165, E140–E154.

Allison, G.W. (1999) The implications of experimental designfor biodiversity manipulations. American Naturalist, 153,26–45.

Amarasekare, P. & Nisbet, R.M. (2001) Spatial heterogeneity,source-sink dynamics, and the local coexistence of com-peting species. American Naturalist 158, 572–584.

Bazzaz, F.A. & Pickett, S.T.A. (1980) Physiological ecology oftropical succession: a comparative review. Annual Review ofEcology and Systematics 11, 287–310.

Beauchamp, R.S.A. & Ullyott, P. (1932) Competitive relation-ships between certain species of fresh-water triclads. Journalof Ecology 20, 200–208.

Bernhardson, C.S. (1975) Type I error rates when multiplecomparison procedures follow a significant F test of anova.Biometrics 31, 229–232.

Binckley, C.A. & Resetarits, W.J. Jr (2003) Functionalequivalence of non-lethal effects: generalized fish avoidancedetermines distribution of gray treefrog, Hyla chrysoscelis,larvae. Oikos 102, 623–629.

Bristow, C.E. (1991) Interactions between phylogeneticallydistant predators: Notophthalmus viridescens and Ennea-canthus obesus. Copeia 1991, 1–8.

Brown, J.H. & Davidson, D.W. (1977) Competition betweenseed-eating rodents and ants in desert ecosystems. Science196, 880–882.

Brown, J.H., Fox, B.J. & Kelt, D.A. (2000) Assembly rules:desert rodent communities are structured at scales fromlocal to continental. American Naturalist 156, 314–321.

Carmer S.G. & Swanson M.R. (1973) An evaluation of tenpairwise multiple comparison procedures by Monte Carlomethods. Journal of the American Statistical Association68, 66–74.

Carmer, S.G. & Walker W.M. (1982) Baby bear’s dilemma: astatistical tale. Agronomy Journal 74, 122–124.

Chalcraft, D.R. & Resetarits, W.J. Jr (2003a) Predatoridentity and ecological impacts: functional redundancy orfunctional diversity? Ecology 84, 2407–2418.

Chalcraft, D.R. & Resetarits, W.J. Jr (2003b) Mapping func-tional similarity on the basis of trait similarities. AmericanNaturalist 162, 390–402.

Chalcraft, D.R. & Resetarits, W.J. Jr (2004) Metabolic ratemodels and the substitutability of predator populations.Journal of Animal Ecology 73, 323–332.

Chalcraft, D.R., Williams, J., Smith, M.D. & Willig, M.R.(2004) Scale dependence in the species richness-productivityrelationship: the role of species turnover. Ecology 85, 2701–2708.

Chase, J.M. & Leibold, M.A. (2002) Spatial scale dictates theproductivity–biodiversity relationship. Nature 416, 427–430.

Chase, J.M. & Leibold, M.A. (2003) Ecological Niches: LinkingClassical and Contemporary Approaches. University ofChicago Press, Chicago.

803Experimental test of functional equivalence

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Functional Ecology, 21, 793–804

Chesson, P.L. (2000) Mechanisms of maintenance of speciesdiversity. Annual Review of Ecology and Systematics 31,343–366.

Cody, M.L. (1973) Character convergence. Annual Review ofEcology and Systematics 4, 189–211.

Darwin, C. (1859) On the Origin of Species by Natural Selectionor the Preservation of Favoured Races in the Struggle for Life.John Murray, London.

Elton, C.S. (1958) The Ecology of Invasions by Animals andPlants. Methuen, London.

Elton, C.S. & Miller, R.S. (1954) The ecological survey ofanimal communities: with a practical system of classifyinghabitats by structural characters. Journal of Ecology 42,460–496.

Emberton, K.C. (1995) Sympatric convergence and environ-mental correlation between two land-snail species. Evolution49, 469–475.

Fox, B.J. (1987) Species assembly and the evolution of com-munity structure. Evolutionary Ecology 1, 201–213.

Fox, B.J. & Brown, J.H. (1995) Reaffirming the validity of theassembly rule for functional groups or guilds. Oikos 73,125–132.

Fretwell, S.D. (1977) The regulation of plant communitiesby food chains exploiting them. Perspectives in Biology andMedicine 20, 169–185.

Gause, G.F. (1934) The Struggle for Existence. Williams &Wilkins, Baltimore.

Gonzalez, R.J. & Dunson, W.A. (1989) Differences in low pHtolerance among closely related sunfish of the genus Ennea-canthus. Environmental Biology of Fishes 26, 303–310.

Gonzalez, R.J. & Dunson, W.A. (1991) Does water pH controlhabitat segregation of sibling species of sunfish (Enneacan-thus)? Wetlands 11, 313–323.

Graham, J.H. (1986) Niche ontogeny and progressive deviationin two congeneric sunfishes, Enneacanthus obesus and E.gloriosus (Centrarchidae). PhD thesis, Rutgers University,New Brunswick, 66 pp.

Hairston, N.G., Smith, F.E. & Slobodkin, L.B. (1960) Com-munity structure, population control, and competition.American Naturalist 94, 421–425.

Hardin, G. (1960) The competitive exclusion principle. Science131, 1292–1297.

Harris, P.M. (1995) Are autecologically similar species alsofunctionally similar? A test in pond communities. Ecology76, 544–552.

Hubbell, S.P. (2005) Neutral theory in community ecologyand the hypothesis of functional equivalence. FunctionalEcology 19, 166–172.

Hubbell, S.P. (2006) Neutral theory and the evolution of eco-logical equivalence. Ecology 87, 1387–1398.

Huston, M.A. (1997) Hidden treatments in ecologicalexperiments: re-evaluating the ecosystem function of bio-diversity. Oecologia 110, 449–460.

Hutchinson, G.E. (1959) Homage to Santa Rosalia or whyare there so many kinds of animals? American Naturalist93, 145–159.

Jaksic, F.M. (2003) How much functional redundancy is outthere, or, are we willing to do away with potential backupspecies? How Landscapes Change (eds G.A. Bradshaw &P.A. Marquet), pp. 255–262. Springer-Verlag, Berlin.

Kurzava, L.M. & Morin, P.J. (1998) Tests of functionalequivalence: complementary roles of salamanders and fishin community organization. Ecology 79, 477–489.

Lawton, J.H. & Brown, V.K. (1993) Redundancy in eco-systems. Biodiversity and Ecosystem Function. (eds E.D.Shultz & H.A. Mooney), pp 255–270. Springer Verlag,New York.

Lee, D.L. & Gilbert, C.R. (1980) Enneacanthus gloriosus;Enneacanthus chaetodon; Enneacanthus obesus. Atlasof North American Freshwater Fishes (eds D.S. Lee,C.R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister

& J.R. Stauffer Jr). North Carolina State Museum of NaturalHistory, Raleigh.

Leibold, M.A. (1998) Similarity and local co-existence ofspecies in regional biotas. Evolutionary Ecology 12, 95–110.

Leibold, M.A., Chase, J.M., Shurin, J.B. & Downing, A.L.(1997) Species turnover and the regulation of trophicstructure. Annual Review of Ecology and Systematics 28,467–494.

Leibold, M.A. & McPeek, M.A. (2006) Coexistence of theniche and neutral perspectives in community ecology.Ecology 87, 1399–1410.

Loreau, M. (2004) Does functional redundancy exist? Oikos104, 606–611.

Loreau, M. & Mouquet, N. (1999) Immigration and themaintenance of local species diversity. American Naturalist154, 427–440.

MacArthur, R.D. & Levins, R. (1967) The limiting similarity,convergence, and divergence of coexisting species. AmericanNaturalist 101, 377–385.

McQueen, D.J., Post, J.R. & Mills, E.L. (1986) Trophicrelationships in freshwater pelagic ecosystems. Journal ofthe Fisheries Research Board of Canada 43, 1571–1581.

Menge, B.A. & Sutherland, J.P. (1987) Community regulation:variation in disturbance, competition, and predation inrelation to environmental stress and recruitment. AmericanNaturalist 130, 730–757.

Morin, P.J. (1983) Predation, competition, and the com-position of larval anuran guilds. Ecological Monographs53, 119–138.

Morin, P.J. (2003) Community ecology and the genetics ofinteracting species. Ecology 84, 577–580.

Morin, P.J., Lawler, S.P. & Johnson, E.A. (1988) Competitionbetween aquatic insects and vertebrates: experimentalmeasures of interaction strength and higher order inter-actions. Ecology 69, 1401–1409.

Mouquet, N. & Loreau, M. (2002) Coexistence in metacom-munities: the regional similarity hypothesis. AmericanNaturalist 159, 420–426.

Naeem, S. (1998) Species redundancy and ecosystem reliability.Conservation Biology 12, 39–45.

Naeem, S., Thompson, L.J., Lawler, S.P., Lawton, J.H. &Woodfin, R.M. (1995) Biodiversity and ecosystem function-ing: empirical evidence from experimental microcosms.Philosophical Transactions of the Royal Society, London, B.347, 249–262.

Oksanen, L., Fretwell, S.D., Arruda, J. & Niemela, P. (1981)Exploitation ecosystems in gradients of primary productivity.American Naturalist 118, 240–261.

Paine, R.T. (1992) Food web analysis: field measurements ofper capita interaction strength. Nature 355, 73–75.

Petchey, O.L. & Gaston, K.J. (2002a) Functional diversity(FD), species richness, and community composition.Ecology Letters 5, 402–411.

Petchey, O.L. & Gaston, K.J. (2002b) Extinction and the lossof functional diversity. Proceedings of the Royal Society ofLondon Series B, Biological Sciences 269, 1721–1727.

Petchey, O.L. & Gaston, K.J. (2006) Functional diversity:back to basics and looking forward. Ecology Letters 9,741–758.

Pianka, E. R. (1973) The structure of lizard communities.Annual Review of Ecology and Systematics 4, 53–74.

Polis, G.A. & McCormick, S.J. (1986) Scorpions, spiders, andsolpugids: predation and competition among distantlyrelated taxa. Oecologia 71, 111–116.

Resetarits, W.J. Jr (1995a) Competitive asymmetry and co-existence in size-structured populations of brook troutand spring salamanders. Oikos 73, 188–198.

Resetarits, W.J. Jr (1995b) Limiting similarity and theintensity of competitive effects on the mottled sculpin, Cottusbairdi, in experimental stream communities. Oecologia104, 31–38.

804W. J. Resetarits & D. R. Chalcraft

© 2007 The Authors.Journal compilation© 2007 British Ecological Society, Functional Ecology, 21, 793–804

Resetarits, W.J. Jr (2005) Habitat selection links local andregional scales in aquatic systems. Ecology Letters 8, 480–486.

Resetarits, W.J. Jr, Binckley, C.A. & Chalcraft, D.R. (2005)Habitat selection, species interactions, and processes ofcommunity assembly in complex landscapes: a metacom-munity perspective. Metacommunities: Spatial Dynamicsand Ecological Communities (eds M. Holyoak, M.A. Leibold& R.D. Holt), pp. 374–398. University of Chicago Press,Chicago.

Resetarits, W.J. Jr & Wilbur, H.M. (1989) Choice of ovipositionsite by Hyla chrysoscelis: role of predators and competitors.Ecology 70, 220–228.

Roe, K.J., Harris, P.M. & Mayden, R.L. (2002) Phylogeneticrelationships of the genera of North American sunfishesand basses (Percoidei: Centrarchidae) as evidenced by themitochondrial cytochrome b gene. Copeia, 2002, 897–905.

Ross, H.H. (1957) Principles of natural coexistence indicatedby leafhopper populations. Evolution 11, 113–129.

Saville, D.J. (1990) Multiple comparison procedures: thepractical solution. American Statistician 44, 174–180.

Saville, D.J. (2003) Basic statistics and the inconsistencyof multiple comparison procedures. Canadian Journal ofExperimental Psychology 57, 167–175.

Scheffer, M. & van Nes, E.H. (2006) Self-organized similarity,the evolutionary emergence of groups of similar species.Proceedings of the National Academy of Sciences (USA)103, 6230–6235.

Schwartz, M.W., Brigham, C.A., Hoeksema, J.D., Lyons,K.G., Mills, M.H. & van Mantgem, P.J. (2000) Linkingbiodiversity to ecosystem function: implications for con-servation ecology. Oecologia 122, 297–305.

Slatkin, M. (1980) Ecological character displacement.Ecology 61, 163–177.

Stone, L., Dayan, T. & Simberloff, D. (1996) Community-wideassembly patterns unmasked: the importance of species’differing geographic ranges. American Naturalist 148, 997–1015.

Stone, L., Dayan, T. & Simberloff, D. (2000) On desert

rodents, favored states, and unresolved issues: scaling upand down regional assemblages and local communities.American Naturalist 156, 322–328.

Sweeney, E.J. (1972) The systematics and distribution of theCentrarchid fish tribe Enneacanthini. Dissertation. BostonUniversity, Boston.

Vandermeer, J.H. (1972) Niche theory. Annual Review ofEcology and Systematics 3, 107–132.

Vandermeer, J.H., Evans, M.A., Foster, P., Höök, T.,Reiskind, M. & Wund, M. (2002) Increased competitionmay promote species coexistence. Proceedings of theNational Academy of Sciences (USA) 99, 8731–8736.

Walker, B.H. (1992) Biodiversity and ecological redundancy.Conservation Biology 6, 18–23.

Wellnitz, T. & Poff, N.L. (2001) Functional redundancy inheterogeneous environments: implications for conservation.Ecology Letters 4, 177–179.

Wilson, D.S. & Swenson, W. (2003) Community genetics andcommunity selection. Ecology 84, 586–588.

Wilson, J.B. (1995) Null models for assembly rules: the JackHorner effect is more insidious than the Narcissus effect.Oikos, 72, 139–144.

Wootton, J.T. & Downing, A.L. (2003) Understanding theeffects of reduced biodiversity: a comparison of twoapproaches. The Importance of Species: Perspectives onExpendability and Triage (eds P. Kareiva & S.A. Levin),pp. 85–104. Princeton University Press, Princeton.

Wright, J.P., Naeem, S., Hector, A., Lehman, C., Reich, P.B.,Schmid, B. & Tilman, D. (2006) Conventional functionalclassification schemes underestimate the relationship withecosystem functioning. Ecology Letters 9, 111–120.

Yachi, S. & Loreau, M. (1999) Biodiversity and ecosystemproductivity in a fluctuating environment: the insurancehypothesis. Proceedings of the National Academy of Sciences(USA) 96, 1463–1468.

Received 22 January 2007; revised 28 February 2007; accepted15 March 2007Editor: Duncan Irschick