Special Feature

SPECIAL FEATURE

Community Genetics: New Insights into Community Ecology by Integrating Population Genetics1

Community genetics is the study of the interaction between genes within a species and populations of other species in a community. This area of research was ®rst introduced by Janis Antonovics (1992 [Toward community genetics. Pages 426±429 in R. S. Fritz and E. L. Simms, editors. Plant resistance to herbivores and pathogens: ecology, evolution and genetics. University of Chicago Press, Chicago, Illinois, USA]) as a modern attempt to integrate community ecology and evolution. Research programs in community genetics span from understanding the mechanistic bases of the evolution of organisms in a community context to identifying the role of genetic variation in generating community patterns. Is community genetics an emerging subdiscipline that ®nally unites multispecies ecology with the genetics of evolutionary change? It is argued that a mature understanding of either ecological communities or the evolution of species will require community genetics. Others cry that this label is nothing more than funding-agency-style renaming of perfectly good, classic questions in evolutionary ecology. I have asked two of the leading groups in community genetics to lay down the de®nitions and set the agenda for future work: Neuhauser et al., who established a Center for Community Genetics at the University of Minnesota; and Whitham et al., who have been studying the consequences of genetic variation for diverse communities and community processes. The result is two fresh and controversial lead papers. Following these papers are commentaries from eight respondents. Some of these authors were pioneers in bridging the ®elds of community ecology and evolution; others are newcomers seeking to infuse ecology with novelty. The result is a mix of arguments that illustrate the vigorous debate within modern community ecology. The feature starts and ends with perspectives articulated by Janis Antonovics. Neuhauser et al. and Whitham et al. loosely follow Antonovics's (1992) reductionist and holist perspectives, respectively. The hallmark of the reductionist approach (Neuhauser et al.) is to document rapid evolutionary change. As nonequilibrial communities are affected by anthropogenic or other in- ¯uences, strong interactions catalyze ¯ux in abundances of individuals in multispecies communities, and these changes can rapidly alter the genetics of community members. It is argued that incorporation of spatial dynamics and population biology in interacting species is necessary to predict the eventual ecological and evolutionary outcome of perturbation. On the holistic side (Whitham et al.), the goal is to understand the consequences of particular genes (or gene combinations) in a community on higher levels of biological organization. Here, the in¯uence of particular genetic variants on communities and ecosystem properties is examined in more or less equilibrial communities. Variation generated by hybridization between species has been the pri- mary tool used to study the role of particular genes on communities; little is currently known about the role of variation within a single species. In addition, the role of genetic variance or diversity in a population in generating higher-level community properties such as productivity and species diversity is a central goal of the holistic approach to community genetics. Perhaps nobody could cogently argue that ecology and evolution can be disciplines without each other. Yet the respondents disagree in terms of the projected role for community genetics in ecology. Tension persists over the possibility of natural selection acting on levels higher than the individual. There is a persuasive call for examining and accounting for the in¯uence of interspeci®c (phylogenetic) relatedness within community studies. Add all this to the nagging age-old questions of how to de®ne a community, who the important dominant or keystone members are, and how to assess the relative importance of various factors in¯uencing community structure.

1 Reprints of this 59 page Special Feature are available for $9.75 each, either as pdf ®les or as hard copy. Prepayment is required. Order reprints from the Ecological Society of America, Attention: Reprint Department, 1707 H Street, N.W., Suite 400, Washington D.C. 20006 (e-mail: [email protected]).

543 Special Feature iaey h oli odvlpa nraigaiiyt ecieadpeitpten nnature. in patterns predict and describe to Ul- ability follow. increasing soon an will com- develop the Others to two of genetics. is benefactor or community of ®rst goal one to study study the the approaches than be the timately, the holistic more will Perhaps into and coevolution) meet? integrate genetics (diffuse reductionist ever to integrate communities bined approaches goal real to holistic worthy in and and coevolution a change reductionist of is genetic the will it population patterns, that of community agree study the to into able species be may we Although oeeiso omnte;pplto eeis pce richness. species genetics; population communities; of logenetics e od:ceouin iest;gopslcin nrseicvrain utlvlslcin phy- selection; multilevel variation; intraspeci®c selection; group diversity; coevolution; words: Key ᭧ 03b h clgclSceyo America of Society Ecological the by 2003 ÐA pca etrsEditor Features Special NURAG .A A. GRAWAL , Ecology, 84(3), 2003, pp. 545±558 ᭧ 2003 by the Ecological Society of America

COMMUNITY GENETICS: EXPANDING THE SYNTHESIS OF ECOLOGY AND GENETICS

CLAUDIA NEUHAUSER,1,4,5 D. A. ANDOW,2,4 GEORGE E. HEIMPEL,2,4 GEORGIANA MAY,1,4 RUTH G. SHAW,1,4 AND STUART WAGENIUS3,4 1Department of Ecology, Evolution and Behavior, University of Minnesota, 100 Ecology Building, 1987 Upper Buford Circle, St. Paul, Minnesota 55108 USA 2Department of Entomology, University of Minnesota, 1980 Folwell Ave, St. Paul, Minnesota 55108 USA 3Chicago Botanic Garden, 1000 Lake Cook Road, Glencoe, Illinois 60022 USA 4The Minnesota Center for Community Genetics, University of Minnesota, St. Paul, Minnesota 55108 USA

Abstract. Community genetics synthesizes community ecology and population genetics and yields fresh insights into the interplay between evolutionary and ecological processes. A community genetics framework proves especially valuable when strong selection on traits results from or impinges on interspeci®c interactions, an increasingly common phenomenon as more communities are subject to direct management or anthropogenic disturbances. We draw illustrations of this perspective from our ongoing studies of three representative communities, two managed and one natural, that have recently undergone large perturbations. The studied communities are: (1) insect pests of crop plants genetically engineered to produce insecticidal toxins; (2) insect-pollinated plants in habitats severely fragmented by agriculture and urbanization; and (3) a pathogen and its crop host now grown extensively outside their native ranges. We demonstrate the value of integrating genetic and ecological processes to gain a full understanding of community dynamics, particularly in nonequilibrium systems that are subject to strong selection. Feature Special Key words: anthropogenic disturbance; Bt maize; community genetics; Echinacea angustifolia; evolution of resistance; genetic engineering; habitat fragmentation; nonequilibrium dynamics; plant± insect interactions; plant±pathogen interactions; Ustilago maydis.

INTRODUCTION a given place. The strength of interaction between members of a community varies. Strong interactions Janis Antonovics (1992) articulated a vision for a can arise even for species that have associated only new ®eld of inquiry, ``community genetics'' (a term recently. For example, in the case of species invasions, suggested by Dr. J. P. Collins, Arizona State Univer- strong interactions may be apparent from the time that sity), to investigate the ``role of genetic variation in a species arrives at a given location (Pritchard and in¯uencing species interactions and determining com- Schluter 2001). Assembly of novel communities may munity structure.'' Community genetics is a synthesis have evolutionary, as well as ecological, consequences of community ecology and evolutionary genetics; it within few generations (Reznick et al. 1997, 2001, Da- directly assesses the interplay between genetic varia- vis and Shaw 2001). When, in addition, ecological in- tion and community dynamics to develop a mechanistic teractions strongly in¯uence the genetic composition understanding of the evolution of organisms in the con- of populations, the conceptual framework of commu- text of the communities that they occupy. nity genetics becomes valuable. Our community concept is that developed by Glea- Community genetics addresses questions about the son (1917, 1926, 1927), demonstrated by Whittaker evolution of interactions among organisms in a broader (1956), and supported by the work of Davis (1981) on context than that of the more stringent framework of community assembly. In this concept, species assemble coevolution where `àn evolutionary change in a trait in communities according to their individualistic attri- . . . in one population in response to . . . a second butes. We superimpose on this concept a contemporary population, [is] followed by an evolutionary response understanding of the ubiquity of genetic variation. by the second population to the change in the ®rst'' Thus, a community is the multispecies assemblage of (Janzen 1980). A situation in which the ecological suc- genetically variable populations that together occupy cess of one species depends on the genotypes of a second species would not necessarily be considered co- Manuscript received 8 April 2002; revised 10 June 2002; ac- evolution by Janzen's stringent criteria, but would ®t cepted 1 July 2002. Corresponding Editor: A. A. Agrawal. For well into the framework of community genetics, wheth- reprints of this special Feature, see footnote 1, p. 543. er or not the genetic composition of populations of the 5 Address correspondence to Department of Ecology, Evo- lution, and Behavior, University of Minnesota, 100 Ecology ®rst species is affected by the ecological interaction. Building, 1987 Upper Buford Circle, St. Paul, Minnesota For example, invasion of a novel pathogen could dra- 55108 USA. E-mail: [email protected] matically affect the population genetic structure of a 545 Special Feature nudrtnigo h neliggntcsrcueof without structure genetic consequence, underlying the a of As understanding the an selection. evo- from of of differ scale may that both spatial than that larger and or processes, ecological lutionary smaller of be scale spatial may the processes that species out of pointed (1980) Levin further studies and for Antonovics integrated diversity. ecology and abundance and an evolution advanced of gen- view (1976) and single a Antonovics scale in temporal detectable eration, is same change the evolutionary on that occur commonly es snail 1954). of Sheppard patterns and (van banding (Cain lepidoptera of shells in and 1958), mimicry Brower of Zandt 1971), An- al. 1952, et (Bradshaw soil tonovics metals of tolerance heavy of relation by 1955), in contamination (Kettlewell moths pollution of examples air forms to Classic melanic 1964). of Ford studies include 1960, Birch (Dobzhan- environment 1941, abiotic sky re- and biotic in the populations to among lation to differences sought genetics genetic ecological explain in studies genetic Early dynamics. and community and population conditions studying in composition ecological both considering the and organisms. distributions of on abundance of focused causes increasingly proximate studies identifying ecological as di- studies quite verged evolutionary pervasive, and of ecological ®eld was however, the soon, thinking of development evolutionary initial ecology, back the trace during also that points roots out Collins (1947). its Dobzhansky of and (1927) some Fisher to century; 20th beginning the the of to back genetics Collins ecological context. traces community (1986) a of inclusion the by netics com- a of members on munity. selection strong communi- impose com- and nonequilibrium of ties generate process historical assembly anthropo- (2) munity with or e.g., disturbances, structure, genic community abrupt an in (1) when change valuable particularly genetics community is a framework that populations argue We another. com- of to one from munity substantially composition differ may genetic species a within the his- population Thus, and context tory. community pop- the of may composition on collection genetic depend the a which as of each rather for (e.g., but ulations studies 1997), ecological al. purely et in Tilman as as species unit on a taxonomic view contingent not a does approach change It genetics composition. evolutionary community community rapid the for species, allows of for pair (1977) gen- Udovic few a and a Levin of by scale conceived As the erations. on operate can feedbacks that these acknowledges It and species. species within variability among genetic and within interactions ecological 1996). al. et (Alexander pathogen the population on effects genetic reciprocal without host, 546 bevn hteooia n vltoayprocess- evolutionary and ecological that Observing of importance the emphasizes genetics Ecological ge- ecological from emerged genetics Community between feedbacks on focuses genetics Community LUI EHUE TAL. ET NEUHAUSER CLAUDIA neprmn.Freape neprmnswith experiments of in outset example, the For at predictable experiment. in- readily an ecological be and not genetic may of terplay outcome demon- the have that natural strated approaches of theoretical and studies using populations, studies empirical coevolutionary experiments, of laboratory number A com- experimental eco- munities. simpli®ed, predict in to even patterns, impossible logical and be understand may to it histories, dif®cult evolutionary their and populations ouain;i atclr hywl esmlri some among in similar vary be will may they Traits particular, in (1) predic- populations; testable 1999). following (Thompson the yields tions it that is pothesis community the by affected composition. be and affect both var- Spatial can genetic iation intraspeci®c spots). to due (cold selection weakly in affect differences only organisms ®tness other's same each the different composition, a with community whereas strong spots), communities, (hot some occur in interactions that a such across range, qualitatively species' vary organisms Inter- between communities. actions different var- across spatial selection for of iation accounts the mosaic metapopulations, selection in geographic ecolog- organisms on of focusing interactions By ical 1980). (Janzen coevolution- de®nitions strict of ary constraints the evolutionary from of interactions study the frees further mosaic selection climate. in to change it predicted causing the response, behind lag evolutionary impede to are likely traits between correlations genetic antagonistic demonstrated that also study this traits However, in selection. variation under and genetic Kansas considerable in and grown Oklahoma, when population Minnesota a com- on selection strong the both over documented projected Etterson decades, warming ing rapid the for as proxy range climatic a and latitudinal extensive species' this n pedo eitn eoye oee,we sus- when However, ceptible genotype. resistant a origin of the spread following and extinction to driven was genotype 20)adEtro n hw(01 nasuyof study a legume, in annual an (2001) of Shaw populations and three Etterson Etterson and by change climate (2000) of case has the populations to among extended been differ may this how and change studies. pro- such genetics for framework Community out- a expected. alternative vides be which de- might under be comes can conditions models for predictive veloped included, are pa- genetic rameters when that illustrate approaches experimental dif- community such to importantly, More paths equilibria. different ferent along evolve may that communities demonstrated experiments and two (Lenski These might coexistence 1985). titers for Levin viral mechanism low the at been resistance have susceptible of the cost A with type. coexisted host that emerged type host fasciculata hracoli cheria natatv etr ftegorpi oachy- mosaic geographic the of feature attractive An geographic a of concept 1997) (1994, Thompson's osdrto fteptnilfrrpdevolutionary rapid for potential the of Consideration .coli E. rw rmarneo 00k.Treating km. 1000 of range a from drawn , netdb h iu 5 h susceptible the T5, virus the by infected eeifce ytevrsT,aresistant a T4, virus the by infected were clg,Vl 4 o 3 No. 84, Vol. Ecology, Chamaecrista Es- March 2003 COMMUNITY GENETICS 547 communities but not in others. (2) There is potential Both theoretical and empirical studies that simulta- for different outcomes of interspeci®c interactions due neously address ecological and evolutionary factors to genetic differences among populations and to dif- face challenges. Theoretical studies face the challenge ferences in the abiotic or biotic environment. (3) Where of multidimensionality. Even in the simplest frame- the range of a species extends over a mosaic community work of ordinary differential equations, two equations such that interactions and, consequently, selection are needed to model the ecological interactions of two vary, coevolved traits may not occur throughout the species. Adding genetic variation to one or both species range. Long-term studies, such as the Silene±Myco- quickly increases the number of equations beyond that botrium system (Antonovics 1992, Alexander et al. which is analytically tractable. However, multidimen- 1996), the Greya±Lithophragma system (Thompson sional models permit complex behavior (multiple stable 1999), and the Linum±Melampsora system (Burdon and equilibria, limit cycles, and chaos) and explicit con- Thrall 1999) provide evidence for the geographic se- sideration of spatial dynamics, such as in Thompson's lection mosaic. (1999) geographic selection mosaic. Empirical studies can be similarly challenged by multidimensionality. THE CONCEPTUAL FRAMEWORK Experiments to assess effects of multiple factors and their interactions require many treatment combinations The community genetics framework promotes new and, with replications, become very large. If the out- understanding when selection alters genetic composi- come of interactions varies in space and time, only tion on the same time scale as that on which numerical long-term studies over large spatial scales reveal the abundances change. Such concordant change is ex- full range of realized outcomes. Moreover, with vari- pected, with strong selection acting on traits mediating ation over space and time, comparison among data sets interspeci®c interactions. In these cases, ecological and is compromised and outcomes may appear unpredict- evolutionary processes can no longer be treated sep- able. Multidimensionality can also result in a prolif- arately. Although evolutionary genetics addresses eration of model parameters that cannot be adequately strong selection and the potential for rapid evolution estimated in empirical studies. Feature Special (Dobzhansky 1941), for instance, insecticide resistance We are optimistic that an integrated approach will (Georghiou 1986, Lenormand et al. 1999, Raymond et ultimately result in a general framework that can ac- al. 2001), drug resistance (Garrett 1994), and the evo- commodate the complexities that arise from consid- lution of competitors (Pimentel and Al-Ha®dh 1965, ering ecological and evolutionary processes simulta- Pimentel et al. 1965), the potential for evolutionary neously. We argue that the value of an integrated, com- change over a few generations remains underappreci- munity genetics approach is particularly great under ated. Darwin (1859), despite acknowledging rapid evo- three conditions: nonequilibrium, genetic variation lution in arti®cial selection, regarded natural selection within species, and strong selection. We claim that the as a weak force apparent only over the geological time co-occurrence of these three conditions is not rare. scale. Moreover, analysis of the consequences of strong We present empirical studies together with mathe- selection poses signi®cant theoretical challenges; al- matical models for which the framework of community though these have not been insurmountable (e.g., for genetics is particularly useful. These studies are char- quantitative genetics, Turelli and Barton [1994]), the- acterized by strong selection following a large pertur- oretical work has concentrated more heavily on the bation. The theoretical models are informed by empir- more tractable case of weak selection (e.g., Kimura ical studies and include both ecological and evolution- 1964, Kingman 1982, Neuhauser 1999). Thus, the fre- ary dynamics; leaving out one or the other aspect weak- quently mentioned distinction between evolutionary ens the model predictions. and ecological time scales is re¯ected in a modeling The ®rst two studies investigate the interplay of eco- framework in which theoretical analyses assume weak logical and evolutionary dynamics in the evolution of selection on the order of the inverse of the population resistance to transgenic crops (Bt maize). Evolution of size, and imply that the effects of selection must be resistance is often modeled neglecting much of the eco- manifested on a time scale on the order of the popu- logical context. We demonstrate that adding ecological lation size. In many population genetic models (e.g., interactions (such as population regulation or density- Kimura 1985), noticeable changes take thousands to dependent predation rates) can alter the predictions of millions of generations when the system is not in equi- simpler evolutionary models. The large-scale intro- librium. In contrast, in standard ecological models such duction of a transgenic crop, which represents a selec- as logistic growth or Lotka-Volterra competition mod- tion episode at an unprecedented scale, has motivated els, noticeable changes occur from one generation to research to develop management strategies informed the next when the system is not in equilibrium. Treated by the framework of community genetics. in this way, weak selection has only subtle effects on The ®rst study demonstrates that adding population population genetic dynamics over a few generations, regulation to a genetic model of spatially varying se- whereas ecological changes may be considerable, thus lection alters the prediction of the rate of evolution of resulting in a separation of time scales. resistance in the European corn borer (Ostrinia nubi- Special Feature Ͻ lalis 548 agtpsswl vlerssac oteCytoxins, Cry the to the resistance evolve rendering the will that is pests envi- plants target signi®cant these with the associated of risks One ronmental 2002). human (NRC and risks environmental investi- health potential scienti®c their to into attention gations drawing them, against ouain fln-ie,sl-noptbeplants self-incompatible long-lived, extensive ( previously of of persistence populations the on mentation which at rate the affect evolves. to resistance potential pred- the of has responses This ators. density-dependent by and females eitdb ifrne noioiinrtsin rates oviposition are in borers differences corn by of mediated genotypes on rates susceptible mortality and egg resistant Differential system. resistance same of evolution the the in on enemies natural of effect n eeta oe o rsaln rti (Cry protein from crystalline bacterium insects, some soil a to the toxic for selectively codes is which that toxin), gene a ing ol 1998). Gould cinbtenahs ln mie n t pathogen its and (maize) plant smut, (corn host a between action of persistence populations. promoting remnant to the fragmentation approaches of suggests consequences 2001) and the al. reveals et fully Glemen system more mating 2000, plant Kirkpatrick the and of consid- (Bataillon basis genetic explicit the that of Nee demonstrate eration We (e.g., 1992). framework May often ecological and purely is a fragmentation in Habitat modeled prairie. tallgrass ican ciae angustifolia Echinacea oia igorpi n otmoaygntcfac- genetic his- contemporary tors. by and caused be biogeographic might torical spot'' ``cold Such monocultures. evolutionary agricultural evo- an in accelerated virulence observed of resistance, commonly lution the durable to remarkably domestication. contrary exhibits following system species host This the of pansion fetv tkligtre et,i aycssallowing cases many in pests, target killing at effective ie oac,adtmt aebe rnfre to only transformed but been toxin, Cry have a tomato express maize, and soybean, cotton, tobacco, poplar, eggplant, potato, rice, canola, Presently, cabbage, crop. broccoli, a of genome atr fepeso.Mn fthese of and Many spectrum expression. toxicity of United unique pattern the own its in with commercialized each been used States, have been have that toxins crops At Cry in States. of types United different the nine in least grown commercially now are non- .%sria n®l odtos(no 2001). (Andow conditions ®eld in survival 0.1% rngnccoshv tre asosbt o and for both passions stirred have crops Transgenic Transgenic h hr td drse h fet fhbttfrag- habitat of effects the addresses study third The h orhsuycnen h vltoayinter- evolutionary the concerns study fourth The vlto frssac otransgenic to resistance of Evolution )to Bt az ed asdb etitdmvmn of movement restricted by caused ®elds maize Bt siaomaydis Ustilago az.Tescn td oue nthe on focuses study second The maize. N Bt ONEQUILIBRIUM Bt rp aebe eeoe yinsert- by developed been have crops rp nfetv Tbsnk1994, (Tabashnik ineffective crops ailsthuringiensis Bacillus ,a on nteNrhAmer- North the in found as ), fe asv ag ex- range massive a after ) Bt D YNAMICS otnand cotton Bt LUI EHUE TAL. ET NEUHAUSER CLAUDIA rp r very are crops Bt ( Bt nothe into ) crops Bt Bt maize and ewe eeain.Pplto euaini this in occurs regulation patches Population among Table generations. toxin; genes in the of given between to is mixing exposed genotypes Complete not the 1. is of 2 ®tness relative patch the and toxin exposed and the is with 1 to patch model Here (1977) regulation. population Comins without the use selection we directional problems, in dynamics generally population can of clarify role selection To the and polymorphism. migration genetic been how maintain has understand and dynamics to (1970:267). population Kimura used no and has Crow model in This described is model directional varying evolution spatially resistance of selection. of generally, understanding more our and, to adds mod- the el to if component what, dynamic population unclear a remains anything, popu- It treat explicitly. not dynamics does lation but model (1977), population Comins Levene's foreshadows niche-selection example, For two-patch, genetic. (1953) ge- quantitative population or whether netic model, kind some genetic have underlying to of necessary is it evolution, resistance dispersal. ge- interpatch the and dominance) on coef®cient, (selection largely parameters netic depend dynamics complex, evolutionary are the investigations counter- theoretical the (or Although these patches). of selection (or results patch no other and the in in resistance patches are selection) for more) models selecting (or these toxin one of a with All models, re- [2001]). patch concise Rausher a see un- (for our view, evolution to resistance added of have models derstanding 2002) Birch 1998, and Gould Hillier 1995, Andow and (e.g., Alstad simulation 1994, and Tabashnik Dobson submitted) and Andow and (May Ives 1986; mathematical additional Since in- time, dynamics. model that population mathematical and a selection developed corporating who Com- (1977) was it ins but (1914), Melander with began evolution ooyoeand homozygote and resistance dominant completely for 1 k 2) (patch ®tness Relative 1) (patch ®tness Relative refuge) 2, (patch Frequency toxic) 1, (patch Frequency blt ftegntpst uvv h oi,adgenerally, and toxin, the the survive to to related genotypes is the ®eld toxic of the ability in ®tness Relative population. tnadmgainslcinpplto genetic population migration±selection standard A understand To problem. simpler a on focus we Here resistance of dynamics evolutionary the of study The T ABLE eae otecs frssac,with resistance, is of refuge cost the the in to ®tness related Relative resistance. recessive pletely K Notes: ntetopthtypes. patch two the in eitn lee and allele, resistant L Here . Measurement .Feunisadrltv teso h genotypes the of ®tness relative and Frequencies 1. eoyefeunisaefrarnol mating randomly a for are frequencies Genotype h eemnsdmnneo eitne with resistance, of dominance determines ␩ h oiac ftecost. the of dominance the p stefeunyo the of frequency the is S (1 1 k (1 sassetbeallele, susceptible a is SR RR RS SS Ϫ Ϫ p p ) ) clg,Vl 4 o 3 No. 84, Vol. Ecology, 2 2 Genotype 2 ␬ ␩ϩ ␬␩ Lh 2 p p h ott the to cost the (1 (1 ϩ h Ϫ Ϫ k (1 ϭ (1 p p R Ϫ␩ ) ) o com- for 0 Ϫ allele. h ) ) R ␬ L p p sa is h RR 2 2 ϭ March 2003 COMMUNITY GENETICS 549 model is determined by simple density dependence (Hassell 1975). To reveal an effect of population regulation, we concentrate on a special case of resistance evolution, the high-dose plus refuge strategy, which now is used to delay resistance evolution in Bt crops in the United States (Alstad and Andow 1995). Under this model, the genetics of resistance are restricted to the case of recessive resistance and no cost of resistance. In addition, to simplify analysis, we assume that movement is random and that all individuals are redistributed at each time step. The genetic parameters for the model are described in Table 1. One can show that if resistance is not over- FIG. 1. Number of generations until control failure (i.e., dominant (0 Յ h Յ 1, where h is dominance), then the resistance allele frequency ϭ 50%) from the full Comins equilibria are the same for the model with and without (1977) model and the second-order Taylor approximation (Eq. population regulation. The resistant allele will go to 2) vs. percentage of area planted as refuge for levels of het- erozygote expression h 0, 0.001, 0.01, and 0.1, with random ®xation if the relative ®tness of the resistant homo- ϭ movement and high SS mortality (k ϭ 10Ϫ3). For each value zygote, L, exceeds the relative ®tness of the susceptible of h, simulations were performed for different levels of `èf- homozygote, k; the susceptible allele will go to ®xation fective'' reproduction in nontoxic ®elds; the lines are often when the reverse holds. In these simple directional se- coincident. Dashed lines give the predictions from the ap- lection models, there are no novel equilibria when pop- proximation. ulation regulation is added to a population genetic model. From an ecological perspective, however, the equilibria are only a part of the story. Indeed, for highly The Comins (1977) model with population regula- Feature Special forced systems (Palumbi 2001b), they may be only a tion yields different rates of evolution. The second- minor part of the story. It is through the nonequilibrium order Taylor approximation (p ϭ 0 and h ϭ 0) is derived dynamics that the evolutionary process will interact by Ives and Andow (submitted) and is most strongly with the surrounding ecological com- xЈ munity, and it is to these nonequilibrium dynamics that ⌬p ഠ (L Ϫ k) 1 p(p ϩ h) (2) xЈϩxЈ we turn. ΂΃12 We denote the frequency of resistant alleles in the where xЈ is evaluated at p ϭ 0; i.e., the various xЈ values population by p. For rare, recessive resistance (p and are the population sizes of SS genotypes in the toxic h very small), the evolutionary dynamic of the Comins and refuge patches. The second-order approximation (1977) model can be approximated by a Taylor expan- ®ts the full population genetic model until p Ͼ 0.2 and sion. For the Comins model, lacking population reg- h Ͼ 0.1 (Fig. 1). This is similar to Eq. 1, with the same ulation, a second-order Taylor expansion around p ϭ dependence on h and p. However, when population 0 and h ϭ 0 gives regulation is added to the model, the proportion of the SS population that occurs in the toxic ®eld also in¯u- (L Ϫ k)F (1 Ϫ Q) ⌬p ϭ 1 p(p ϩ h) (1) ences the rate of evolution. Compared to the model Fk12(1 Ϫ Q) ϩ FQ without population regulation, simple density-depen- where F1 and F2 are the fecundities in the Bt ®eld and dent population dynamics will alter the rate of evolu- non-Bt ®eld, respectively, and Q is the proportion of tion of resistance. We conclude that it often will be the refuge. This second-order approximation ®ts the necessary to consider both the population dynamics and full model until p Ͼ 0.2 and h Ͼ 0.1. However, when the genetic dynamics when investigating the rates of p Ͼ 0.2, resistance rapidly reaches ®xation in the pop- evolutionary change. ulation, so for all practical purposes, the approximation The directional selection model is one of the simplest is excellent. This approximation shows that resistance models of evolutionary biology. The results we present evolves at a faster rate with higher dominance (h) and here suggest that evolutionary rates will be different resistance allele frequency (p). Recessive resistance when simple density-dependent population dynamics (low h) evolves much more slowly than dominant re- are explicitly considered compared to a model without sistance (higher h). The fraction in Eq. 1 is the repro- an explicit population dynamic. Hence, nonequilibrium ductive advantage of the RR genotype (resistant ho- dynamics may persist for longer or shorter periods of mozygote) over the SS genotype (susceptible homo- times and perturbations from equilibria may occur more zygote) in the toxic patch, relative to the reproductive readily or less readily compared to predictions from rate of the SS genotype in both patches. Population pure population genetic models. The model that we density does not enter into the equation, and evolution discuss here is a simpli®ed directional selection model, does not depend on the population dynamics. with random movement and random mating and a rare Special Feature Bt change. evolutionary of pace the in- on considerable ¯uence have complex among may more communities interactions within to species ecological generalize models, do behav- results evolutionary richer these even direc- If show complex iors. will more models that selection likely tional is It allele. recessive 550 n ihdssof doses high ing 1994). Gould 1991, al. et resistant (Gould on genotypes attacks disproportionate by evolution reduced resistance be will of resistant rate is the on conversely, attack genotypes genotypes; the susceptible than on be greater attack disproportionately will the resistance inter- if herbivore enemy of accelerated herbivore±natural evolution the the The of in depending action. details decrease herbivores, the a in or evolution on enemies increase resistance natural an of that either rate suggested to lead (1991) could al. genotypes. mor- et herbivore the susceptible Gould biasing vs. in the resistant play of of can tality because enemies defenses natural plant that evo- role to resistance herbivores of rate of the lution affect may enemies how understand natural to us com- helps A perspective by genetics plants. munity evolution of defenses resistance inter- antiherbivore of to plant±insect herbivores rate of the outcome including the actions, path- affect and also parasitoids, ogens, predators, including Ei- tro- level, third and the phic (Boethel of members traits Reciprocally, 1986). these kenbary by indirectly affected be will directly) enemies (or natural herbivore's the follows that it behavior, and development, mortality, bivore ioe thg este Gud19;J White, J. 1994; her- (Gould than densities attack high enemy at natural of bivores risk at lower herbivores or low-density higher are becomes: ques- therefore The tion herbivores. resistant feeding-stage of among frequency genotypes the and density herbivore tween lo eitn eoye o`sae rmatcsby the attacks in from enemies would èscape' which natural to densities, genotypes low resistant at allow risk lower implies rate ffeigsaehrioe on herbivores density feeding-stage The 1994). of (Gould density herbivore pri- by mediated marily be will herbivores susceptible and sistant atclrylwpe este.Tu,teefcsof effects on herbivores the feeding-stage Thus, attacking enemies densities. of natural range prey a low attack over the dependent particularly density if inversely especially were resistance, rate of delay a evolution and the densities imply in herbivore would lower hand, at other risk the increased on rate, attack density- dependent inversely An evolution. resistance accelerated atidvdaswl omamc ihrfato of fraction higher the much on a feeding herbivores form the will individuals tant eodr fmgiuelwrta h est ffeed- of non- density on the herbivores than ing-stage lower magnitude of orders be ulse data published ntecs fsrn eitne(.. lnsexpress- plants (e.g., resistance strong of case the In her- on effects strong have traits resistance Because aua nme n h vlto fresistance of evolution the and enemies Natural lns hs eepc eaierltosi be- relationship negative a expect we Thus, plants. ?Apstvl est-eedn attack density-dependent positively A )? Bt oi) ifrnilatc fre- of attack differential toxin), Bt rp h eutwudbe would result The crop. Bt Bt lns oee,resis- However, plants. Bt lnsta ntenon- the on than plants lnswl initially will plants LUI EHUE TAL. ET NEUHAUSER CLAUDIA un- oee,i hs lnig r ag n ffemale if In and restricted. large likely, all are are at is plants plantings movement toxic these of if plantings ovipos- however, of in densities females Lower iting documented. been yet has not toxins insecticidal transgenic plants express of to avoidance engineered Female females female plants. ovipositing toxic (1) of of plantings densities plants: in lower toxic (2) her- and lower on choice, to rates lead could oviposition oviposition scenarios bivore equal two expect least we At rates? should non- But and plants. toxic when on toxic equal toxins re- are plant of rates rate oviposition to herbivore the herbivores on of effect evolution no sistance of have magnitude should or et mortality presence Gould egg the that Indeed, suggested genotype. (1991) plant al. se- upon precedes stage based egg lection the because transparent less are slowed, be versa. will vice evolution toxic resistance and on of greater rate are the rates attack plants, If plants. toxic pro- on nontoxic herbivores inversely vs. on be impact relative will their to evolution portional resistance of rate the rmntrleeyatc si h non- the in is refuge attack the enemy of natural Here, from rate 2). (Fig. the declines evolution mortality, resistance in- egg with Alternatively, density-dependent develop). larvae versely to susceptible, allowed not be but would resistant, more (i.e., ®elds Bt nqa vpsto ae in rates oviposition unequal ipootoaeymr avet eeto nthe in selection exposes to mortality egg larvae reduced more the disproportionately and attack enemy uino eitneto resistance of lution rmia] h uoencr oe.Because borer. corn European the Crambidae], ioeis bivore lososta h ee fegmraiyisl can itself mortality egg of level the that shows model The also bene®t. larvae susceptible and resistant both h fet fegmraiyo eitneevolution resistance on mortality egg of effects The otlt nnon- in higher disproportionately mortality density-depen- impose they positively rate, a attack pred- dent show egg parasitoids if that refuge or show and ators model toxic the both to of in Simulations assumed females ®elds. are with Males random at developed. mate they same the which in oviposit in to ®eld females refuge of nontoxic fraction a a includes allows in an and model toxin The of plant stage. high-dose eggs larval a the by the affected upon is that act herbivore to den- mortality inverse sity-dependent and density-dependent positive for lowed habitats. two the in eggs on rates attack the opportunity in¯uence an to have rates mortality egg sity-dependent ilb rsn tlwrdniisin densities lower at present be will fet the effect, eue,tedniyof density the refuges, n htti ifrnewl eicesnl ag as large increasingly be of will size difference the this that and n-eeain¯gtpro hnd non- do sec- than its period during species ¯ight this ond-generation of adults fewer far generate data az D .AdwadD .Alstad, N. D. and Andow A. (D. maize ecntutdapplto eei oe htal- that model genetic population a constructed We .I steeoelkl that likely therefore is It ). srnanubilalis Ostrinia Bt Bt lnig nrae.Udrcniin of conditions Under increases. plantings lnscnttt euefo natural from refuge a constitute plants Bt ed,wihaclrtsteevo- the accelerates which ®elds, Bt Ostrinia az in maize Bt (Hu s non- vs. g assi oe in lower is masses egg be)[Lepidoptera: Èbner) clg,Vl 4 o 3 No. 84, Vol. Ecology, Ostrinia Bt Ostrinia Bt az,akyher- key a maize, s non- vs. Bt Bt az,where maize, unpublished az,den- maize, Bt g masses egg Fg ) In 2). (Fig. Bt ed or ®elds Bt maize ®elds Bt March 2003 COMMUNITY GENETICS 551

(T. ostriniae) suggests, however, that only a single re- lease is needed early in the season, making it more likely that releases may become economically viable (Wright et al. 2001, 2002). Do egg predators or parasitoids show density-dependent responses (either positive or negative) to Os- trinia egg density? Predators of European corn borer eggs are generalist feeders that typically also utilize aphids, other arthropods, and corn pollen in addition to European corn borer eggs (Andow 1996). Predation on Ostrinia eggs depends on the community composition of alternative prey, so that the response to egg density is indirect and complex (J. Harmon, unpublished data). Parasitism of Ostrinia eggs by T. ostriniae appears to be density independent at low host densities FIG. 2. Results of simulations illustrating the effects of different forms of egg mortality on the evolution of resistance and inversely density dependent at higher egg densities of herbivores to high-dose insecticidal toxins in plants. Sin- (Wang and Ferro 1998). Like other parasitoids, T. os- gle-locus recessive resistance is assumed with an initial fre- triniae presumably can become limited by the number quency of a resistance allele equal to 0.001. The area planted of eggs that they carry or by handling time if host to toxic ®elds is four times the area planted to nontoxic ®elds, resulting in a 20% nontoxic refuge. One-®fth of the females densities get high enough (Getz and Mills 1996, Ro- oviposit in the same ®eld where they developed as larvae, senheim 1996, Heimpel and Rosenheim 1998). If Os- and the remaining females are distributed evenly across toxic trinia egg densities are high enough to induce egg or and nontoxic ®elds. Egg mortality occurs prior to selection handling-time limitation in non-Bt maize and low in the model and is simulated in three ways: (1) no egg mor- enough for these factors not to come into play in Bt tality; (2) an egg mortality rate that increases with herbivore Feature Special density by N(i)/[k ϩ N(i)], where k is a constant and N(i)is maize, then egg parasitism could indeed be higher in the density of toxic or nontoxic plants (positive density-de- Bt maize, with a concomitant reduction in the rate of pendent egg mortality, PDD); and (3) an egg mortality rate evolution resistance in Ostrinia. that decreases with herbivore density by 1 Ϫ N(i)/[k ϩ N(i)] (inversely density-dependent egg mortality, IDD). The con- Habitat fragmentation stant k was set to equalize overall egg mortality rates in the PDD and IDD runs prior to ®xation of the resistance allele. Habitat destruction and resulting fragmentation are Herbivore fecundity was set at 100, and larvae that survive major causes of species extinctions. Investigators con- selection (or feed on nontoxic plants) are subject to density- cerned about the persistence of remnant plant popu- dependent intraspeci®c competition. lations have documented ecological and genetic effects of habitat fragmentation (Leach and Givnish 1996, in¯uence the rate of evolution resistance, with higher Young et al. 1996, Newman and Pilson 1997). To obtain rates of mortality slowing down the evolution of re- a community genetics perspective on the interplay of sistance. these effects, we have incorporated the empirical re- We now turn to the natural enemies of Ostrinia eggs sults of Wagenius (2000) on Echinacea angustifolia to evaluate if any are likely to cause density-dependent (Asteraceae), the narrow-leaved purple cone¯ower, into mortality. Ostrinia eggs are eaten by a number of egg an individual-based, spatially explicit, stochastic sim- predators that can collectively impose mortality rates ulation model. Echinacea angustifolia is native to the of up to (and in some cases exceeding) 50% (Andow North American tallgrass prairie, which has been re- and Risch 1985, Andow 1990, 1992). The most im- duced to isolated fragments in a matrix of agriculture portant Ostrinia egg predators are the native lady beetle during the past 150 yr. Contrary to the results of a Coleomegilla maculata, lacewing larvae, and various purely ecological model, we ®nd that the genetic prop- predatory bugs (Andow 1990, 1992). The recently in- erties of our study system exacerbate the risk of extir- troduced multicolored Asian lady beetle, Harmonia ax- pation. yridis, also feeds on Ostrinia eggs (Hoogendoorn and Echinacea angustifolia shares ®ve key features with Heimpel 2002). Ostrinia eggs are subject to parasitism, many of the plants that formerly dominated unbroken but naturally occurring egg parasitism impacts Ͻ1% of prairie: (1) long life (Echinacea is a long-lived peren- Ostrinia eggs over most of the species' North American nial and plants rarely ¯ower before their third year); range (Andow 1992). Augmentative releases of native (2) reproduction strictly by seed (Echinacea does not and introduced Trichogramma spp. can cause substan- spread vegetatively, so regeneration of populations de- tial egg mortality, but these releases, for the most part, pends exclusively on seed production); (3) self-incom- have not been considered economically feasible, in part patibility (seed set from each ¯oret depends on receipt because multiple releases have to be made each year of pollen from another plant; McGregor [1968], Leu- (Smith 1996, Andow 1997). Recent work with an egg szler et al. [1996], Franke et al. [1997]); (4) pollination parasitoid recently imported from northeastern China by generalist insects (service by nonspecialized pol- Special Feature fple n ed,a ela eeoyoiy sfunc- as local heterozygosity, as of well tions as dispersal seeds, and and regime, pollen ®re of the of function a re- as cruitment seedling mortality, and density-dependent growth persistence: population incorporates affect that and processes fragmentation additional previ- habitat on of builds models model ous sto- The explicit, model. simulation spatially chastic based, individually an veloped de- we persistence, population affects processes genetic and ecological both sur- factors. with genetic Not varies populations. vigor remnant plant prisingly, large be- and differs small composition tween genetic hypothesis, Under genetic remnants. of large ®re a and abundance small the in as differ as (such plants) exotic environment (such biotic environment hypothesis, or abiotic regime) ecological the an of under in- aspects might vigor: mating factors plant two small ¯uence Similarly, isolation; in fragments. assessed large with being and now are varies rates in- incompatibility of pollen receipt the that compatible evidence hypoth- is ecological there the However, support esis. not do results self- Pre- liminary plants. share non-isolated do that in- than plants alleles of incompatibility related proportion from greater pollen a compatible receive isolation, plants their isolated of but regardless pollen pollinator of amounts receive on plants similar hypothesis, restrictions genetic a to of According ¯ight. because conse- non- than plants and, insects pollinating isolated visits from fewer pollen less receive quently, plants isolated holds hypothesis that ecological An limitation. in¯u- pollen re- could ence and processes limitation non-exclusive Two pollen vigor. of duced causes the about potheses could ®ndings processes. these genetic of or ecological Each than from ones. result vigorous large less from are those populations seeds from small grown in plants that collected reduces found also limitation He pollen yield. that He seed and plants individuals. isolation individual the ¯owering of with increases limitation thousand pollen that from several found size in to varying populations one remnant 48 in plants specialized no (5) ( dispersal to [1992]); seed nectar of or Kunin mechanisms pollen for distances; ¯ights limit short to likely is linators 552 ad aeis(00 mapped (2000) Wagenius land, plants. perennial native many of conservation 16)mtpplto oe,wihdsrbsthe Levins' describes patches, on occupied which of based fraction model, is metapopulation framework (1969) The ecological system. our purely in fragmentation habitat con- of the predict sequences to insuf®cient is framework ecological oe ihgoa dispersal: global with model ieyt ooiedsathbttpths.Teefea- These patches). of habitat tures distant colonize to likely oass o h nepa fteeooia and ecological the of interplay the how assess To hy- genetic and ecological tests research Ongoing na60-asuyae fwsenMneoafarm- Minnesota western of area study 6400-ha a In h iuainmdldmntae htapurely a that demonstrates model simulation The Echinacea Echinacea aei utbemdlfrthe for model suitable a it make abundance. u Echinacea ( t ,i ni®iepatch in®nite an in ), Ͼ 2000 LUI EHUE TAL. ET NEUHAUSER CLAUDIA ed r un- are seeds Echinacea mligta h ouainde u ftefato of fraction the if habitat out destroyed dies population the that implying by equi- in given patches is occupied of librium fraction the that follows It D e parameter the Here, ilstefloig(e n a 1992): May and (Nee following the yields hw lp ls o0 niaigta ac otlt is mortality patch that triangles) indicating (open 0, time. to over plants close constant self-compatible slope a for reproduction shows The pollen. seed line compatible reduced of gression availability to reduced due by an indicating caused rate slope, self-incompatible positive extinction for a increasing line shows triangles) regression (solid The plants time. of function tnadeooia oesasm osatpatch constant a assume 3). models (Fig. ecological time over Standard increases extirpation of the probability that ®nd followed. we is alone, patches habitat self-incompatibility the Considering of the fate the of and fraction destroyed an then ®xed is until A fragmentation reached. habitat is equilibrium without of model fate initially the the we simulation, predict run the to In inbreeding, populations. remnant of the rate al- the self-incompatibility and our of leles For in number components the 1997). genetic varying Pilson model, include and we reasons, (Newman these remnants compro- small are inbreeding with in which declines ®tness of individual when both mised survival, output individual reproductive and on plants. depends (unrelated) persistence Likewise, compatible of to availability due fragmentation reduced with decreases reproduc- which on output, tive depends Neither colonization biogeography.) of holds: theory remain assumption the in patch made assump- a are (Similar of tions destruction. rate habitat extinction after the unchanged and rate zation eoe h xicinrt fec ac.I fraction a If patch. each of rate extinction the denotes F ftehbtti emnnl etoe,temodel the destroyed, permanently is habitat the of IG h oe nE.4asmsta ohtecoloni- the both that assumes 4 Eq. in model The .Rslso iuain npthmraiya a as mortality patch on simulations of Results 3. . du dt du dt ϭ D ϭ u cu Ã c xed 1 exceeds ϭ (1 eoe h ooiainrt and rate colonization the denotes cu 1 (1 Ϫ Ϫ Ϫ D D Ϫ u ) Ϫ Ϫ u Ϫ clg,Vl 4 o 3 No. 84, Vol. Ecology, ) c e e/c Ϫ eu. . eu. (3) (5) (4) March 2003 COMMUNITY GENETICS 553 extinction rate and thus cannot predict this trend. Therefore, standard ecological models overestimate the persistence of remnant populations. Ongoing work is exploring the further consequences of inbreeding in the context of habitat fragmentation. Habitat fragmentation poses threats from mechanisms as diverse as increased mating system incompatibility and reduced ®re frequency. The community genetics perspective considers the joint consequences of these aspects of fragmentation, promoting understanding of how ecological and evolutionary processes together affect population persistence.

Domestication as invasion

FIG. 4. Correlation of genetic (as ⌽ST values) and geo- Beginning Ͻ150 yr ago, vast acreages of temperate graphic distances in pairwise comparisons of North American forests and grasslands in North America were con- populations of Ustilago maydis. Open symbols show pairwise verted to agricultural production, and in that process, comparisons between Ohio and other North American pop- novel crop plant genotypes were introduced (Smith ulations. The ⌽ST levels from pairwise comparisons show no correlation with distance, using a Mantel test (rY ϭϪ0.006, 1989). In North America, the conversion to modern P ϭ 0.35; Schneider et al. 1997). maize arguably represents one of the largest plant range expansions within human history. Although the impact of agricultural conversion on the landscape is readily Seven North American populations of U. maydis apparent, the impact on plant pathogens and the co- were sampled from ®eld corn and sweet corn, and ge- evolution of plant and pathogen is less apparent, but netic relatedness of the populations was assessed using Feature Special no less important. We focus on possible effects of the 11 genomic probes for restriction fragment length poly- rapid geographic expansion of maize and its associated morphisms (RFLPs). Using ⌽ST, a multilocus analog fungal pathogen, Ustilago maydis, corn smut. (Schneider et al. 1997) to Wright's inbreeding coef®- The maize±smut interaction is ideal for the study of cient, FST, genetic variation within and between North coevolutionary dynamics under nonequilibrium con- American populations was estimated. Values for ⌽ST ditions because genotypes of both species can be ma- varied from 0.07 to 0.26 across pairwise comparisons nipulated and the population history of maize is well of these populations. Overall levels of heterozygosity understood. The transition from small, genetically var- are quite high (ϳ0.4) and estimated levels of inbreeding iable teosinte populations of Central America to the within one population are low (FIS ϳ 0.06), despite the large, monotypic maize plantings of North American fact that sib matings could occur. Using an Index of agriculture is expected to accelerate the evolution of Association test (Agapow and Burt 1999), analysis of virulence in associated maize pathogens. U. maydis is linkage disequilibrium revealed that the alleles dem- a naturally occurring pathogen on both teosinte and onstrating signi®cant association between loci within maize (Duran 1987) and has tracked maize from do- populations were not the same in all sampled popu- mestication to present-day plantings. The history of lations. In contrast to the population genetic structure maize domestication and spread are well documented observed for many crop pathogen species (e.g., Boeger (Galinat 1992), as is the molecular genetic basis of its et al. 1993), we did not observe isolation by distance evolution from teosinte (e.g., Doebley 1992, Hilton and (Slatkin 1993), as one might expect with wind-borne Gaut 1998). Virtually every maize plant grown in North dispersal of spores and high migration rates (see Fig. America has a recorded pedigree, but much less is 4). Together, these data demonstrate that U. maydis known about U. maydis evolution. Historical records populations are sexually reproducing and maintain high show that maize breeding programs for smut resistance levels of heterozygosity, distributed unevenly among of the early 1900s were successful; surprisingly, the genetically variable populations across the major geo- pathogen has not evolved to overcome smut resistance graphic regions in which maize is grown (J. Garton and traits developed at that time (Christensen 1963), and C. Ramos, unpublished manuscript). smut resistance in maize has proven durable over 50 We developed two hypotheses to explain our obser- yr. Our long-term goal is to account for the difference vations. (1) The variation that we observe across smut between the observations of low levels of smut infec- populations re¯ects historical founder events as maize tion on widely planted corn varieties and an expectation was brought into North America about 1000 yr ago. for rapid evolution of virulence in agricultural mono- These populations have not come to equilibrium by cultures. More immediately, we ask how maize do- genetic drift and migration. (2) The genetic variation mestication and geographic range expansion have af- among smut populations is the result of host or regional fected the population genetic structure of U. maydis. environmental selection effects (e.g., Ahmed et al. Special Feature ⌽ fungus, the the for of data locus used pop- self-incompatibility We between migration occurred. little subsequently that introduc- ulations but the crop, with the in of brought to tion was comparable smut is that model assuming The migra- descendants. without among populations, tion ``an- descendant new an drawn found were in alleles to and placed America Central was of urn'' popula- variation cestral descendant observed All American tions. North establishing process in of the modeled alone, we Shaw, effects Frank with sampling collaboration historical for generate To expectations hypotheses. exclusive not are These 1996). 554 ae htw fe soit iharclua patho- agricultural with associate migration often high we of early that result of rates the ``footprints'' than the rather be introductions might America North in plantings. maize widespread the across evenly iation iomna atr ol aealreipc vra high over observe en- impact We large regional time. a other short have or could genotype factors host vironmental to that due suggests American selection collection Uruguay North one the with populations of comparison made. pairwise mid- were collections Further, at the which conducted from In was station research 4). resistance the smut Fig. for (see breeding America 1900s, North in parisons h ulhptei httedsrbto of distribution the that of rejection hypothesis allow null not the do results and simulation 20, 1000 Our repeated 10, times. was 5, parameters 3, of with at combination Each varied and 30. drawn populations alleles descendant of number 100 the and 20, out 5, carried with were simulations from The obtained model. those simulation with the 1997) al. et (Zambino lations ihmie n htisfcettm ic h intro- the since America time ( North insuf®cient duction that in di- and arrived maize, or smut with large of relatively populations that Al- yr. verse suggest 4000±5000 data of our period a together, by separated only been can populations have American North yet and species, Uruguay between the differences with associated cally evtosfrteocrec of ob- occurrence empirical the the for comparing on servations based test statistical constructed we a that model, this approximation so Using good sampling). enough a (binomial large was replacement was The with population drift. sampling source to assumed we due the and lost other that each be of to independent popula- were likely new draws less a an- be in the would arrival from tion, upon sampled and, approx- being population an of cestral have chance would allele equal each imately Thus, approximately 1997). (Zambino al. in selection et balancing occur to and due frequencies described equal are alleles 18 aafrteOi ouain hc ipaspairwise displays which population, Ohio the for data chi-square this in low test. is power statistical population. the ancestral However, an from by draws determined random were historical, populations current in alleles type ST eakby hn h urn ouain fsmut of populations current the then, Remarkably, h eodslcinhptei ssgetdb the by suggested is hypothesis selection second The aus( values Յ 00y)hseasdt itiueta var- that distribute to elapsed has yr) 1000 ϳ .4 wc hs o te ariecom- pairwise other for those twice 0.24) ⌽ ST b ( ϳ lee nsubpopu- in alleles .) austypi- values 0.4), LUI EHUE TAL. ET NEUHAUSER CLAUDIA b, b o which for mating- eeto netbihn h bevdpten fge- of patterns in observed variation netic the establishing rel- in the strong selection and resolve events founder now historical To of importance 1997). ative Silk and (Burdon gens nih o nyt oeulbimstain uhas such situations nonequilibrium to only not insight 1973). Valen (Van derivatives its Red and the model in abiotic Queen predicted as static community, among a a interactions within dynamic in organisms of result occur a also as evo- environment evolutionary can Rapid changes cycles. rapid seasonal lutionary with demonstrated coinciding Dob- changes (1943) and 1940s. (1940) the zhansky Timofeef-Ressovsky in traced others instance, be and can Dobzhansky For and by new work not to is back scales time short atively planning. urban and of genetics, management sustainable conservation in resources, valuable such be from will the derive studies on we that impact Principles human mounts. the landscape approaches as its compelling more offers, become in- genetics fundamental community the that to be addition sights In can disturbance. itself large management a and scarce increasingly are communities natural unmanaged managed impose and Indeed, natural communities. disturbances many on regimes are anthropogenic selection strong situations now; Nonequilibrium common variation. act genetic non- interactions strong by on which characterized in dynamics studies equilibrium community for a approach of value genetics com- the of demonstrated context We the munities. within evolution into yield insight to fresh populations interacting of dynamics numerical oiei ucm,t hmsns(99 evolutionary spots''? (1999) ``hot Thompson's to to op- outcome, but in conspire analogous, posite spot,'' ``cold interaction evolutionary maize±smut an genetic make the the and of we history smut, Could structure of factor. corn third interaction pathogen, a add common evolutionary might very the its and of re- maize view studies (Ber- additional our resistance If of inforce 1996). cost the Purrington to and and gelson 1997) Bur- al. 1991, Silk Burdon and et and don Jarosz (Burdon 1990, al. environment et Clarke and 1989, plants genetically host of distribution variable patchy previously the has to attributed structure been resistance host and virulence pathogen between the correspondence of to Lack pathogen's populations. this contribute in maize virulence may of evolution 1998), in slow strikingly al. resistance et quantitative Lubberstedt (e.g., the with combined system, reproductive pathogen clon- a Such or reproducing. inbreeding ally than rather primarily sexual, is and America outcrossing North in that population suggest espe- date pathogen to the results locations, Our America. diverse Central cially geographically in and sampling models, approaches, population experimental bine omnt eeisjitycniesgntcand genetic considers jointly genetics Community omnt eeisprpciecnbignovel bring can perspective genetics community A rel- on occur can change evolutionary that idea The .maydis U. D ISCUSSION ouain,w ilcom- will we populations, clg,Vl 4 o 3 No. 84, Vol. Ecology, March 2003 COMMUNITY GENETICS 555 we discuss here, but also to the case of equilibrium some others, the attraction of an herbivore's natural dynamics with strong balancing selection. For instance, enemies appears to be a component of the plant's de- Antonovics (1992) investigated a model of host±path- fensive reaction (DeMoraes et al. 1998, Bradburne and ogen interactions, in which he demonstrated that the Mithen 2000). In examples such as these, one can en- observed coexistence of susceptible and resistant plant vision evolution within entire multispecies complexes. genotypes was only consistent with a model that com- The importance of taking a combined ecological and bined ecological and genetic dynamics. We focused on evolutionary approach to understand the effects of nonequilibrium situations because they are more likely large-scale perturbations was recently emphasized by to produce pronounced effects of interactions. The four Palumbi (2001a)inaScience article entitled ``Humans studies in this paper demonstrate (1) that an ecological as the World's Greatest Evolutionary Force'' (see also or an evolutionary framework by itself is insuf®cient Palumbi 2001b). The rapid growth of the human pop- to understand or predict outcomes of organismal in- ulation has led to unprecedented alterations of natural teractions during the transient phase following pertur- ecosystems and widespread introduction of novel or- bation, and (2) that novel predictions about community ganisms. Consequences of human impact are felt in all change can emerge from mathematical models that in- areas, including epidemiology (emerging diseases, re- corporate both ecological and genetic processes. Only sistance to antibiotics), pest management (evolution of when we consider both factors in concert can we un- resistance to pesticides), species invasions (globaliza- derstand community dynamics following a large per- tion, homogenization), species extinctions (habitat turbation that imposes strong selection on the com- fragmentation, climate change), and expansion of ag- munity. In each of the studies that we discussed, strong ricultural land (to destroy natural habitats). selection occurs naturally; it is characterized by an The magnitude and spatial extent of disturbances are abrupt and large change that imposes strong selection staggering; no habitat seems to be fully protected. on the community. The transient phase to equilibrium Coastal marine environments around the world have may be long lasting and, with increased human impact, been massively perturbed by dredging, pollution, im- an equilibrium may never be reached. poundments, and an astounding number of invasive Feature Special We believe that our examples are illustrative of the species (Carlton 1999). Riparian habitats in remote common phenomenon of nonequilibrium conditions in mountains have been massively perturbed by removal contemporary communities. Agricultural and forested of beaver, damming of rivers, recovery of beaver, and areas, which cover over half of the earth's terrestrial biological species invasions. In highly perturbed sys- area, are subjected to massive ecological and evolu- tems, ecological and evolutionary forces are equally tionary disturbances, generating nonequilibrium dy- important. namics within those habitats, as illustrated in our ®rst A community genetics perspective relying on a the- example of resistance evolution. These same activities oretical framework not only leads to a more complete generate nonequilibrium dynamics in native habitats understanding of the consequences of human-induced by fragmenting and isolating these habitats, which we selection pressure, but also provides a sound basis for illustrated in our example of patch dynamics in prairies. the development of management strategies. Developing We recognize that many of our examples involve a uni®ed theoretical framework is of paramount im- only two-species ``communities.'' We have chosen to portance because our actions induce both ecological highlight these relatively simple interactions to lay bare and evolutionary change. Without such a comprehen- the structure of community genetics and illustrate some sive framework, our understanding of the complexity of the necessary conditions under which it can truly by which communities and ecosystems respond to our matter. We suggest, however, that this focus does not actions will be severely compromised, and we will nev- limit the generality of our analysis. Indirect interactions er be able to develop general management strategies in ecological communities appear to be common (Holt to address these responses. and Lawton [1994], Wooton [1994]; see also the com- panion paper by Whitham et al. [2003]). Hence, evo- ACKNOWLEDGMENTS lutionary change in two-species interactions has the We thank Don Alstad, James Garton, Jason Harmon, Frank potential to affect other members of the community Shaw, and Jennifer White for sharing unpublished information, and Eric Lonsdorf for preparing Fig. 3 and for com- through indirect species interactions. For instance, ge- menting on an earlier version of this manuscript. This work netic variation in resistance to herbivores in plants can was partially supported by a National Science Foundation affect the herbivores' natural enemies (Price et al. grant, DMS-083468. 1980). Plant defenses reduce the ®tness of predators and parasitoids of herbivores feeding on them (Camp- LITERATURE CITED bell and Duffey 1979, Duffey et al. 1986, Obrycki Agapow, P. M., and A. Burt. 1999. Multi-locus. Version 1.2b. 1986). Thaler (1999) also recently showed that the pro- Department of Biology, Imperial College, Silwood Park, UK. [Available online, URL: ͗www.bio.ic.ac.uk/evolve/ duction of plant defenses is correlated with the pro- software/multilocus͘.] duction of volatiles that attract parasitoids of the her- Ahmed, H. U., C. C. Mundt, M. E. Hoffer, and S. M. Coakley. bivores feeding on the defended plants. In this case and 1996. Selective in¯uence of wheat cultivars on pathoge- Special Feature 556 no,D .19.Caatrzto fpeaino egg on predation of Characterization 1990. A. D. evo- Andow, the Managing 1995. Andow. A. D. and N., D. Alstad, Jarosz, M. A. Antonovics, J. Thrall, H. P. M., H. Alexander, no,D .19.Agetn aua nme nmaize in enemies natural Augmenting 1996. A. D. Andow, no,D . n .J ic.18.Peaini diversi®ed in Predation 1985. Risch. J. S. and A., D. Andow, ogr .M,R .Ce,adB .MDnl.19.Gene 1993. McDonald. A. B. and Chen, S. R. ecology. M., population J. Boeger, in factor genetic The 1960. C. L. Birch, patterns Surveying 1996. Purrington. B. C. and J., Bergelson, depres- Inbreeding 2000. Kirkpatrick. M. and T., Bataillon, noois . n .A ei.18.Teeooia and ecological The 1980. Levin. A. D. and J., Antonovics, noois . .D rdhw n .C unr 1971. Turner. C. R. and Bradshaw, D. A. J., Antonovics, rdhw .D 92 ouain of Populations 1952. D. A. Bradshaw, noois .19.Twr omnt eeis Pages genetics. community Toward 1992. J. Antonovics, noois .17.Teiptfo ouaingenetics: population from input The 1976. J. Antonovics, no,D .20.Rssigrssac to resistance Resisting 2001. A. D. Andow, no,D .19.Ft feg f®rst-generation of eggs of Fate 1992. A. D. Andow, rdun,R . n .Mte.20.Guoioaege- Glucosinolate 2000. Mithen. R. and P., Inter- R. 1986. Bradburne, editors. Eikenbary, D. R. and J., D. Boethel, no,D .19.Itgaigbooia oto nIPM in control biological Integrating 1997. A. D. Andow, oemglamaculata Coleomegilla assof masses Science plants. 1894±1896. transgenic to resistance insect of lution disease. anther-smut of ge- study and case Ecology a dynamics disease: Population plant of 1996. netics Oudemans. V. P. and sn eeainldvriy ae 137±153 Pages diversity. vegetational using g ytm.EvrnetlEntomology Environmental systems. age gocsse:rltosbtenaccieldpredator coccinellid a between relations agroecosystem: Flor- Raton, Boca USA. Press, CRC ida, effects. health human and environmental assessing organisms: engineered Genetically o ewe egahcppltosof populations geographic between ¯ow Naturalist American Naturalist American plants. in 536±558. resistance of cost the in popula- Research ®nite Genetical in matter. mutations does size deleterious tions: mildly to due sion 452. lns nulRve fEooyadSystematics and Ecology in of Review regulation Annual dependent plants. density of consequences genetical ev ea oeac npat.Avne nEcological in Advances plants. in Research tolerance metal Heavy Illi- Chicago, Press, Chicago USA. of nois, University genetics. evolution, ecology, and pathogens: and herbivores to resistance odn eisB Series London, 426±449 245. `h e clgclgntc.'Sseai Botany Systematic genetics.'' ecological new ``The Ecology n .F yal dtr.Eooia neatosadbi- and USA. interactions Colorado, Boulder, Ecological Press, Westview editors. control. ological Nyvall, F. R. and aso h noooia oit fAmerica of Society Entomological the of nals eisadteatato fteahdparasitoid aphid the of attraction the and netics USA. York, New York, of New predators Wiley, and John parasitoids insects. and resistance plant of actions Phyto- markers. polymorphism pathology length fragment striction graminicola rapae 99±124 ytm.Pgs71±86 Pages systems. and (Food control FFTC Taiwan. Taipai, management. Biological Center), Technology pest editors. Fertilizer integrated Kiritani, in K. agents and Murakami, tritici nubilalis att edadzn osnn.Nature poisoning. zinc and lead to tant iiyof nicity .Phytopathology ). to in 77 Lpdpea yaia)i he osraintill- conservation three in Pyralidae) (Lepidoptera: yoparlagraminicola Mycosphaerella 22 in Brassica 7 srnanubilalis Ostrinia :1±85. :990±996. .K eoreuadB .Bros editors. Burrows, E. B. and Letourneau K. D. 83 :357±372. .S rt n .L im,eios Plant editors. Simms, L. E. and Frite S. R. (anamorph :1148±1154. 267 rceig fteRylSceyof Society Royal the of Proceedings . 94 :89±95. :5±24. in 86 n t od ora fApplied of Journal food. its and etratritici Septoria :454±458. .A no,D .Ragsdale, W. D. Andow, A. D. Lpdpea yaia) An- Pyralidae). (Lepidoptera: gotstenuis Agrostis (Anamorph eetdwt re- with detected ) 169 21 LUI EHUE TAL. ET NEUHAUSER CLAUDIA Mycosphaerella 75 :388±393. Bt :1098±1099. in 83 :75±81. Diaeretiella on Pages corn. .Ge,Y. Grey, G. :482±486. 11 Septoria Ostrinia 1 :411± :233± 268 148 resis- : : udn .J,A .Jrs,adG .Kry 99 Pattern 1989. Kirby. C. G. and Jarosz, M. A. J., J. Burdon, udn .J,adJ ik 97 iieiw ore and sources MiniReview: 1997. Silk. J. and J., J. Burdon, udn .J,adP .Trl.19.Sailadtemporal and Spatial 1999. Thrall. H. P. and J., J. Burdon, an .G,adP .Sepr.15.Ntrlslcinin selection Natural 1954. Sheppard. M. P. and G., A. Cain, lre .D,F .Cmbl,adJ .Bvn 90 Genetic 1990. Bevan. R. J. and Campbell, S. F. D., D. Clarke, hitne,J .16.Cr mtcue by caused smut Corn 1963. J. J. Christensen, alo,J .19.Tesaeadeooia consequences ecological and scale The 1999. T. J. Carlton, apel .C,adS .Dfe.17.Tmtn n par- and Tomatine 1979. Duffey. S. S. and C., B. Campbell, ozasy .14.Aatv hne nue ynatural by induced changes Adaptive 1947. T. Dobzhansky, ozasy .14.Gntc fntrlppltos IX. populations: natural of Genetics 1943. T. species. Dobzhansky, of origin the and Genetics 1941. T. Dobzhansky, eMre,C . .J ei,P .Pr,H .Abr,and Alborn, T. H. Pare, W. P. Lewis, J. W. M., C. Moraes, De adap- and shifts Range 2001. Shaw. G. R. and B., M. Davis, obe,J 92 apn h ee htmd az.Trends maize. made that genes the Mapping 1992. J. Doebley, ufy .S,K .Bom n .C apel 96 Con- 1986. Campbell. C. B. and Bloem, A. K. S., S. Duffey, osqecs nulRve fEooyadSystamatics and 20 Ecology of Review Annual and consequences. interactionsÐcauses plant±pathogen in patchiness and atrso iest npatptoei ug.Phytopa- fungi. plant±pathogenic in thology diversity of patterns atrsi ovligpatadptoe associations. pathogen and plant Naturalist American coevolving in patterns ua,R 97 siaiae fMxc.Wsigo State Washington Mexico. of Ustilaginales 1987. R. Duran, oln,J .18.Eouinr clg n h s fnat- of use the and ecology Evolutionary 1986. P. J. Collins, ai,M .18.Qaenr itr n h tblt of stability the and history Quaternary 1981. B. M. natural Davis, of means by species of origin the On 1859. C. pop- Darwin, to introduction An 1970. Kimura. M. and F., J. Crow, resis- insecticide of development The 1977. N. H. Comins, Cepaea rspiapseudoobscura Drosophila neatosbetween interactions 3±41. dmc odeh,TeNetherlands. The Dordrecht, Ac- ademic, Kluwer management. biodiversity and species Invasive fbooia nain ntewrdsoen.Pgs195± Pages oceans. world's the 212 in invasions biological of stcwss oeta noptblt fpatantibiosis Science plant control. of biological and incompatability potential wasps: asitic eeto nwl ouain of populations wild in selection eprlcagsi h opsto fppltosof populations of composition the in changes Temporal York, New Press, USA. University York, New Columbia edition. Second .H ulno.19.Hrioeifse lnsselec- plants Nature Herbivore-infested parasitoids. attract 1998. tively Tumlinson. H. J. Science change. climate 673±679. quaternary to responses tive New York, New Springer-Verlag, USA. York, application. con- succession: and Forest editors. cepts Botkin, B. D. and Shugart, idwfungus mildew dis nGenetics in 1±16. eune fsqetaino ln aua rdcsin products natural 31±60 plant Pages of interactions. plant±insect±parasitoid sequestration of sequences nvriy ula,Wsigo,USA. Washington, Pullman, University, John USA. insects. York, of New York, predators New and Wiley, plant parasitoids of Interactions and editors. resistance Eikenbary, D. R. and Boethel rlslcini clgclter.Junlo h History the of Journal Biology theory. of ecological in selection ural and pathogens UK. Pests, Oxford, Scienti®c, editors. Blackwell Leather, communities. R. plant S. and Burdon ln omnte.Pgs132±153 Pages communities. plant UK. London, Murray, selection. New York, New Row, USA. and York, Harper theory. genetics ulation Theoretical of Journal migration. of Biology presence the in tance :119±136. mrcnPyoahlgclSceyMonograph: Society Phytopathological American . in .T adud .J ce,adA and Schei, J. P. Sandlund, T. O. Genetics . 87 :664±669. 64 :177±197. 19 8 :302±307. :257±288. rspe®scheri Erysiphe 39 153, :89±116. cncovulgaris Scenecio upeet S15±S33. Supplement, Genetics . 205 ae 189±202 Pages . Drosophila 393 :700±702. clg,Vl 4 o 3 No. 84, Vol. Ecology, in :570±573. 28 .C et .H. H. West, C. D. Ê :162±186. ie,editors. Viken, . n h powdery the and siaomay- Ustilago Evolution . in in 292 .J. D. .J. J. 1 : : March 2003 COMMUNITY GENETICS 557

Etterson, J. R. 2000. Evolutionary potential of the annual Ives, A. R., and D. A. Andow. 2002. Evolution of resistance legume, Chamaecrista fasciculate, in relation to global to Bt crops: directional selection in structured environ- warming. Dissertation. University of Minnesota, Saint ments. Ecology Letters, in press. Paul, Minnesota, USA. Janzen, D. H. 1980. When is it coevolution? Evolution 34: Etterson, J. R., and R. G. Shaw. 2001. Constraint to adaptive 611±612. evolution in response to global warming. Science 294:151± Jarosz, A. M., and J. J. Burdon. 1991. Host±pathogen inter- 154. actions in natural populations of Linum marginale and Me- Fisher, R. A. 1927. On some objections to mimicry theory; lampsora lini: local and regional variation in patterns of statistical and genetic. Transactions of the Royal Ento- resistance and racial structure. Evolution 45:1618±1627. mological Society of London 75:269±278. Kettlewell, H. B. D. 1955. Selection experiments on indus- Ford, E. B. 1964. Ecological genetics. First edition. Broad- trial melanism in the Lepidoptera. Heredity 9:323±342. water Press, London, UK. Kimura, M. 1964. Diffusion models in population genetics. Franke, R., R. Schenk, J. Schaser, and A. Nagell. 1997. Ein- Journal of Applied Probability 1:177±232. ¯uss von Anbaumassnahmen auf Ertrag und Inhaltsstoffe Kimura, M. 1985. Diffusion models in population genetics von Echinacea pallida (Nutt.) Nutt. Zeitschrift fuÈr Arznei- with special reference to ®xation time of molecular mutants und GewuÈrzp¯anzen 2:173±185. under mutational pressure. Pages 19±39 in T. Ohta and K. Galinat, W. C. 1992. Evolution of corn. Advances in Agron- Aoki, editors. Population genetics and molecular evolution. omy 47:203±231. Japan Scienti®c Societies Press, Tokyo, Japan. Garrett, L. 1994. The coming plague: newly emerging dis- Kingman, J. F. C. 1982. The coalescent. Stochastic Processes eases in a world out of balance. Farrar, Straus, and Giroux, and Applications 13:235±248. New York, New York, USA. Kunin, W. E. 1992. Density and reproductive success in wild Georghiou, G. P. 1986. The magnitude of the resistance prob- populations of Diplotaxis erucoides (Brassicaceae). Oec- lem. Pages 14±43 in Committee on Strategies for the Man- ologia (Heidelberg) 91:129±133. agement of Pesticide Resistant Pest Populations Board on Leach, M. K., and T. J. Givnish. 1996. Ecological determi- Agriculture. Pesticide resistance: strategies and tactics for nants of species loss in remnant prairies. Science 273: management. National Academy Press, Washington, D.C., 1555±1558. USA. Lenormand, T., D. Bourguet, T. Guillemaud, and M. Ray- Getz, W. M., and N. J. Mills. 1996. Host±parasitoid coex- mond. 1999. Tracking the evolution of insecticide resis- istence and egg-limited encounter rates. American Natu- tance in the mosquito Culex pipiens. Nature (London) 400: pca Feature Special ralist 148:333±347. 861±864. Gleason, H. A. 1917. The structure and development of the Lenski, R. E., and B. R. Levin. 1985. Constraints on the plant association. Bulletin of the Torrey Botanical Club 44: coevolution of bacteria and virulent phage: a model, some 463±481. experiments, and predictions for natural communities. Gleason, H. A. 1926. The individualistic concept of plant American Naturalist 125:585±602. association. Bulletin of the Torrey Botanical Club 53:7± Leuszler, H. K., V. J. Tepedino, and D. G. Alston. 1996. 26. Reproductive biology of purple cone¯ower in southwestern Gleason, H. A. 1927. Further views on the succession con- North Dakota. Prairie Naturalist 28:91±102. cept. Ecology 8:299±326. Levene, H. 1953. Genetic equilibria when more than one ecological niche is available. American Naturalist 87:331± Glemen, S., T. Bataillon, J. Ronfort, A. Mignot, and I. Oli- 333. vieri. 2001. Inbreeding depression in small populations of Levin, S. A., and J. D. Udovic. 1977. A mathematical model self-incompatible plants. Genetics 159:1217±1229. of coevolving populations. American Naturalist 111:657± Gould, F. 1994. Potential and problems with high-dose strat- 675. egies for pesticidal engineered crops. Biocontrol Science Levins, R. 1969. Some demographic and genetic consequenc- and Technology 4:451±461. es of environmental heterogeneity for biological control. Gould, F. 1998. Sustainability of transgenic insecticidal cul- Bulletin of the Entomological Society of America 15:237± tivars: integrating pest genetics and ecology. Annual Re- 240. view of Entomology 43:701±726. Lubberstedt, T., D. Klein, and A. E. Melchinger. 1998. Com- Gould, F., G. C. Kennedy, and M. T. Johnson. 1991. Effects parative QTL mapping of resistance to Ustilago maydis of natural enemies on the rate of herbivore adaptation to across four populations of European ¯int-maize. Theoret- resistant host plants. Entomologia Experimentalis et Ap- ical and Applied Genetics 97:1321±1330. plicata 58:1±14. May, R. M., and A. P. Dobson. 1986. Population dynamics Hassell, M. P. 1975. Density dependence in single-species and the rate of evolution of pesticide resistance. Pages 170± populations. Journal of Animal Ecology 44:283±295. 193 in Pesticide resistance: strategies and tactics for man- Heimpel, G. E., and J. A. Rosenheim. 1998. Egg limitation agement. National Academy Press, Washington, D.C., in insect parasitoids: a review of the evidence and a case USA. study. Biological Control 11:160±168. McGregor, R. L. 1968. The taxonomy of the genus Echinacea Hillier, J. G., and A. N. E. Birch. 2002. A bi-trophic model (Compositae). University of Kansas Science Bulletin 48: for pest adaptation to a resistant crop. Journal of Theoret- 113±142. ical Biology 215:305±319. Melander, A. L. 1914. Can insects become resistant to sprays? Hilton, H., and B. S. Gaut. 1998. Speciation and domesti- Journal of Economic Entomology 7:167±173. cation in maize and its wild relatives: evidence from the Nee, S., and R. M. May. 1992. Dynamics of metapopulation Globulin-1 gene. Genetics 150:863±872. models. Pages 123±148 in I. A. Hanski and M. E. Gilpin, Holt, R. D., and J. H. Lawton. 1994. The ecological con- editors. Metapopulation biology: ecology, genetics, and sequences of shared natural enemies. Annual Review of evolution. Academic Press, San Diego, California, USA. Ecology and Systematics 25:495±520. Neuhauser, C. 1999. The ancestral selection graph and gene Hoogendoorn, M., and G. E. Heimpel. 2002. PCR-based gut genealogy under frequency-dependent selection. Theoret- content analysis of insect predators: a ®eld study. Pro- ical Population Biology 56:203±214. ceedings of the First International Symposium of Arthropod Newman, D., and D. Pilson. 1997. Increased probability of Biological Control, in press. extinction due to decreased genetic effective population Special Feature 558 enc,D . .H hw .H od n .G Shaw. G. R. and Rodd, H. F. Shaw, H. F. N., D. Reznick, enc,D,M .Bte,adH od 01 Life-history 2001. Rodd. H. and Butler, J. M. D., Reznick, amn,M,C etct .Wil .Pser n .Chev- C. and Pasteur, to N. Weill, M. resistance Berticat, C. pest M., and Raymond, Co-evolution 2001. D. M. Rausher, interspeci®c Declining 2001. Schluter. D. and R., J. Pritchard, N. J. McPheron, A. B. Gross, P. Bouton, E. C. W., P. Hayes. Price, T. J. and Wood, W. P. Feinberg, H. E. D., Pimentel, of control Ecological 1965. Al-Ha®dh. R. and D., Pimentel, aub,S .2001 P. S. Palumbi, aub,S .2001 P. S. Palumbi, mt,S 96 ilgclcnrlwith control North Biological eastern 1996. in S. agriculture Smith, of Origins 1989. D. B. Smith, lti,M 93 slto ydsac neulbimand equilibrium in distance by Isolation 1993. M. Slatkin, Excof®er. L. and Roessli, D. Keuffer, egg J.-M. for S., argument Schneider, evolutionary An 1996. A. J. Rosenheim, byk,J .18.Tei¯ec ffla uecneon pubescence foliar of in¯uence The 1986. J. J. Obrycki, ef- Environmental 2002. Council). Research (National NRC (Dordrecht) 97 vlaino h aeo vlto nntrlpop- natural ( in evolution guppies of of rate the ulations of Evaluation 1997. vlto ngpis I.Tecmaaieeooyof ecology Natural- comparative American ist The environments. low-predation VII. and high- guppies. in evolution lo.20.Isciierssac ntemosquito the in resistance Insecticide 2001. illon. Nature enemies. natural Research Ecology Evolutionary 3 past. competition of ghost the summoning displacement: character during competition Re- Systematics Annual and enemies. Ecology of natural among view be- and interactions Interactions herbivores on insect plants 1980. of tween in¯uence Weiss. levels: E. trophic three A. and Thompson, Naturalist coexistence American species. the ¯y and competing distribution, of spatial Selection, 1965. Amer- of Society parasite± ica Entomological the the in of Annals evolution system. genetic host by population parasite a New York, New USA. Norton, York, change. evolutionary rapid cause uinr oc.Science force. lutionary n aaiod n rdtr fiscs onWly New Wiley, USA. John York, resistance insects. New plant of York, predators of and Interactions parasitoids editors. and Eikenbary, D. R. mrc.Science America. o-qiiru ouain.Evolution populations. non-equilibrium Ge- of University Switzerland. Lab, Geneva, Biometry neva, and data genetic Genetics population for analysis. software a v1.1: Arlequin 1997. Evolution limitation. 1934±1937. pipiens noohgu pce.Pgs61±83 Pages species. entomophagous reg- of USA. D.C., adequacy Washington, and Press, Academy scope National the ulation. plants: transgenic of fects lution :209±220. acs ucse,adptnilo hi s.Ana Re- Annual use. their Entomology of of view potential and successes, vances, 157 ie xeietlppltosof populations experimental size: 58 :1±6. :126±140. hthv elandaotaatto?Genetica adaptation? about learned we have what : 51 :354±362. 112±113 b. a. 246 h vlto xlso:hwhumans how explosion: evolution The :287±296. uasa h ol' raetevo- greatest world's the as Humans 50 41 :1566±1571. oclareticulata Poecilia 411 293 :2089±2094. :375±406. :857±864. :1786±1790. 11 lri pulchella Clarkia :41±65. in Trichogramma 47 .J ote and Boethel J. D. .Science ). LUI EHUE TAL. ET NEUHAUSER CLAUDIA :264±279. 99 :97±109. Evo- . Culex 275 ad- : : aahi,B .19.Eouino eitneto resistance of Evolution 1994. E. B. Tabashnik, rgt .G,T .Khr .P ofan n .A Chenus. A. S. and Hoffmann, P. M. Kuhar, P. T. G., M. Wright, aeis .20.Promneo rii aigsystem mating prairie a of Performance 2000. S. Wagenius, hmsn .N 99 pc® yohsso h geographic the on hypotheses Speci®c 1999. N. J. Thompson, Chicago process. coevolutionary The 1994. cause N. J. defenses Thompson, plant Jasmonate-inducible 1999. J. Thaler, ioefRsosy .W 90 Zu 1940. W. N. Timofeef-Ressovsky, E. and Ritchie, M. Reich, P. Wedin, D. Knops, J. D., Tilman, htae,R .15.Vgtto fteGetSoyMoun- Smoky Great the of Vegetation 1956. H. R. Whittaker, Gehring, A. C. Martinsen, D. G. Young, W. G., T. Whitham, rgt .G,M .Hfmn,S .Ceu,adJ Gardner. J. and Chenus, A. S. Hoffmann, P. M. G., M. Wright, indirect of consequences and nature The 1994. T. J. Wootton, a ae,L .17.Anweouinr a.Evolution- law. evolutionary new A 1973. M. L. Valen, Van statistical and Genetic 1994. Barton. H. N. and M., Turelli, a ad rwr .15.Eprmna tde fmimicry of studies Experimental 1958. J. Brower, Zandt van ag . n .N er.19.Fntoa epne of responses Functional 1998. Ferro. N. D. and B., Wang, abn,P,J .Goh .Lkn,J .Gro,adG. and Garton, R. J. Lukens, L. Groth, V. J. P., Zambino, on,A,T ol,adT rw.19.Tepopulation The 1996. Brown. T. and Boyle, T. A., Young, aasplexippus Danaus rcormaostriniae Trichogramma 02 feto ncltv eessof releases inoculative of Effect 2002. 29±37. e ipra in dispersal pol- limited len and self-incompatibility habitat: fragmented in opimsbei morphismus Naturalist American coevolution. of mosaic USA. Illinois, Chicago, Press, University Nature herbivores. of parasitism increased opsto neoytmpoess Science and processes. 1302. diversity ecosystem functional of on in¯uence composition The 1997. Siemann. a)to dae) aoaoyad®l odtos niomna Entomol- Environmental conditions. ogy ®eld and laboratory an.Eooia Monographs Ecological R. tains. C. the and of Woolbright, consequence Ecology S. a phenotype. Lindroth, genetics: extended Community L. 2003. R. Kuske. Fischer, Bailey, G. D. Wimp, K. M. G. J. Shuster, M. S. Schweitzer, A. J. 01 ipra eairof behavior Dispersal 2001. Systematics Ecol- and of Review ogy Annual communities. ecological in effects chippus blatt. r Theory ary me what, traits: polygenic Genetics on normal? selection strong of analyses nsm ot mrcnbte¯e:Pr .Temonarch, The I. Part butter¯ies: American North some in est fMneoa inaoi,Mneoa USA. Minnesota, Minneapolis, Minnesota, of versity triniae a.19.Vraina the at Variation 1997. May. Control Biological corn. ®eld and corn sweet rnsi clg n Evolution plants. and for Ecology fragmentation in habitat Trends of consequences genetic bilalis maydis thuringiensis lctosfragettv eessagainst im- releases ®elds: augmentative corn for plications sweet in Trichogrammatidae) menoptera: 27 Phytopathology . 60 Lpdpea rmia) ilgclControl Biological Crambidae). (Lepidoptera: nppltosof populations on Evolution . :752±758. srnanubilalis Ostrinia :130±137. nulRve fEntomology of Review Annual . 1 :1±30. ciae angustifolia Echinacea dlabipunctata Adelia n viceroy, and , 138 12 :32±47. 25 :913±941. :443±466. 87 srnanubilalis Ostrinia Hmnpea Trichogrammati- (Hymenoptera: :1233±1239. Lpdpea yaia)under Pyralidae) (Lepidoptera: b rcormaostriniae Trichogramma aigtp ou of locus type mating 84 26 :559±573. :1±80. ieii rhpu ar- archippus Limenitis 11 clg,Vl 4 o 3 No. 84, Vol. Ecology, ilgshsZentral- Biologisches . rAayedsPoly- des Analyse Èr :413±418. israin Uni- Dissertation. . rcormaos- Trichogramma 399 153S n aaeto damage and srna nu- Ostriniae 39 23 :686±688. 277 :47±79. :149±155. :1±14. Ustilago :1300± Bacillus 22 (Hy- : Ecology, 84(3), 2003, pp. 559±573 ᭧ 2003 by the Ecological Society of America

COMMUNITY AND ECOSYSTEM GENETICS: A CONSEQUENCE OF THE EXTENDED PHENOTYPE

THOMAS G. WHITHAM,1,3,6 WILLIAM P. Y OUNG,1,3 GREGORY D. MARTINSEN,1,3 CATHERINE A. GEHRING,1,3 JENNIFER A. SCHWEITZER,1,3 STEPHEN M. SHUSTER,1,3 GINA M. WIMP,1,3 DYLAN G. FISCHER,2,3 JOSEPH K. BAILEY,1,3 RICHARD L. LINDROTH,4 SCOTT WOOLBRIGHT,1,3 AND CHERYL R. KUSKE5 1Department of Biological Sciences, Northern Arizona University, Flagstaff, Arizona 86011 USA 2Department of Forest Ecosystems, Northern Arizona University, Flagstaff, Arizona 86011 USA 3Merriam-Powell Center for Environmental Research, Northern Arizona University, Flagstaff, Arizona 86011 USA 4Department of Entomology, 1630 Linden Drive, University of Wisconsin, Madison, Wisconsin 53706 USA 5Biosciences Division, M888, Los Alamos National Laboratory, Los Alamos, New Mexico 87545 USA

Abstract. We present evidence that the heritable genetic variation within individual species, especially dominant and keystone species, has community and ecosystem consequences. These consequences represent extended phenotypes, i.e., the effects of genes at levels higher than the population. Using diverse examples from microbes to vertebrates, we demonstrate that the extended phenotype can be traced from the individuals possessing the trait, to the community, and to ecosystem processes such as leaf litter decomposition and N mineralization. In our development of a community genetics perspective, we focus on intraspeci®c genetic variation because the extended phenotypes of these genes can be passed from one generation to the next, which provides a mechanism for heritability. In support of this view, common-garden experiments using synthetic crosses of a dominant tree show that their progeny tend to support arthropod communities that resemble those of their parents. We also argue that the combined interactions of extended phenotypes con- Feature Special tribute to the among-community variance in the traits of individuals within communities. The genetic factors underlying this among-community variance in trait expression, particularly those involving genetic interactions among species, constitute community heritability. These ®ndings have diverse implications. (1) They provide a genetic framework for understanding community structure and ecosystem processes. The effects of extended phenotypes at these higher levels need not be diffuse; they may be direct or may act in relatively few steps, which enhances our ability to detect and predict their effects. (2) From a conservation perspective, we introduce the concept of the minimum viable interacting population (MVIP), which represents the size of a population needed to maintain genetic diversity at levels required by other interacting species in the community. (3) Genotype ϫ environment interactions in dominant and keystone species can shift extended phenotypes to have unexpected consequences at community and ecosystem levels, an issue that is especially important as it relates to global change. (4) Documenting community heritability justi®es a community genetics perspective and is an essential ®rst step in demonstrating community evolution. (5) Community genetics requires and promotes an integrative approach, from genes to ecosystems, that is necessary for the marriage of ecology and genetics. Few studies span from genes to ecosystems, but such integration is probably essential for understanding the natural world. Key words: community evolution; community genetics; community heritability; dominant species; ecosystems; extended phenotype; genetic variation; keystone species; minimum viable interacting population.

INTRODUCTION ``community genetics,'' which emphasizes ``the analysis of evolutionary genetic processes that occur Population genetics is de®ned as ``the study of how among interacting populations in communities.'' This Mendel's laws and other genetic principles apply to entire populations'' (Hartl 1980). Community genetics de®nition allows us to examine complex genetic in- extends these same principles to the more complex are- teractions among diverse organisms and their potential na of communities and ecosystems. Jim Collins (cited ecosystem consequences (Loehle and Pechmann 1988), in Antonovics 1992) argued for a new discipline called but is not dependent upon the reciprocity of coevolution (Antonovics 1992). Our development of community genetics focuses on the role of intraspeci®c genetic Manuscript received 8 April 2002; revised 25 June 2002; ac- variation in dominant and keystone species, which in cepted 1 July 2002. Corresponding Editor: A. A. Agrawal. For reprints of this Special Feature, see footnote 1, p. 543. turn affects dependent species, community organiza- 6 E-mail: [email protected] tion, and ecosystem dynamics. Where population ge- 559 Special Feature otn edfrudrtnigorntrlwrd tis It world. natural im- our and understanding emerging for an ®eld is portant genetics community ar- that perspective complex Our gues environment. and the levels, ge- with trophic of interactions multiple consequences the variation, understand netic to world ``tangled natural Darwin's understand to are more include bank.'' we to if need the rever- factors emphasize Such 1996). outcomes Fritz in con- and sals basic Orians our also reverse (see matrix clusions potentially the juniper. can to of species interactions additional mutualist of one a just adding be three-way Thus, a can in mistletoe birds however, interaction, seed-dispersing If, their parasitic. two-way include is simple we a relationship as the junipers interaction, and relationship mistletoe the Whi- examine between and we if Ommeren that larger showed van (2002) the tham example, how es- For of works. is comprehension system matrix basic this our into to sential ®t not environment. species if physical how changing 100s, Understanding a encompass in species that of interactions 1000s, of within live matrix and live evolved a have not species do Most organisms vacuum. a that in truth messy, but recognizes simple, genetics the Community genetics? population used paper). de®nitions working this of consequences throughout summary a evolutionary for 1 Table and (see ecosystem, extended com- their these munity, explore we here Because heritable, be population. levels can phenotypes at the genes phenotype of than effects extended higher the as the 1982) de®ne Dawkins (sensu commu- we a from perspective, populations, nity of and expression individuals the in be to genes phenotype the considers netics community the of com- members affecting the in among particular variation interactions a genetic the intraspeci®c in `Ìf living of species role interacting The com- of the association at `Àn change phenotypic to leading selection Natural heritability Community genetics Community evolution Community Community 560 ABLE omnt eeisebae h opeiyo the of complexity the embraces genetics Community of ®eld complex already the to complexity add Why T xeddpeoyeTeefcso ee tlvl ihrta h ouainSnuDwis(1982) Dawkins Sensu population the than higher levels at genes of effects total The have and biomass community ``dominate that Species genetic Intraspeci®c phenotype Extended species Dominant esoeseisAseis`woeipc niscmuiyo csse is ecosystem or community its on impact ``whose species A Interacting Viable Minimum species Keystone variation ouain(MVIP) Population .Smayo okn entosue nti paper. this in used de®nitions working of Summary 1. emD®iinSource De®nition Term pig omnt,teitrcinwl ehrtbea the at heritable òff- be the level.'' will to community interaction community the `parent' community, the spring' from intact passed are dynamics ecosystem and organization munity level munity area'' eei aito on ihnaseiso yrdzn com- hybridizing or species a within found variation Genetic h ieo ouainnee omiti eei diversity genetic maintain to needed population a of size The mat htaelre u o ipootoaet their to disproportionate not but biomass'' large, are that impacts ag,addsrprintl ag eaiet t abun- its to relative dance'' large next) disproportionately the and to large, generation herita- one and from ¯ow traits gene of signi®cant transmission with ble unit largest (the plex tlvl eurdb te neatn pce ntecom- the in species interacting other munity by required levels at HMSG HTA TAL. ET WHITHAM G. THOMAS osqec fhrtbeetne phenotypes. extended ultimate heritable the of evolution, consequence con- community We exploring conservation. by for clude important are they environment, the how by and in¯uenced vertebrates, is to expression microbes their from how taxa diverse distrib- broadly across are uted or effects dominant their in how expressed species, be phe- keystone to extended likely most how are emphasizing notypes studies empirical series a of develop We we complex. individ- hybridizing an species, or within species ual among variation genetic not the to on species, concentrate generation a from within pass generation genes consequences Because community heritable. their are that phe- extended argue the and of notype ®rst mechanisms We the genetic levels. the how ecosystem develop conse- and examine important community to at have is quences genes paper of this phenotypes ecosys- of extended and goal communities The on tems. level genes higher the of understand to consequences up scaling of process the osqecs ecnetaeo eei variation there where genetic complexes hybridizing on and species evolutionary concentrate within their genetic we and the phenotypes consequences, understand extended to of reason, basis within this occurs ex- For notorious next ceptions). are (microbes the species to among not generation species, of one transmission The from im- (1) is traits reasons. species major within two variation for genetic portant mat- the do that species 2000, and that al. argue ter We et 2001)? Cabido Schwartz and (e.g., Diaz com- in ecosystems matter that and diversity agree munities species to or unable species are individual ecologists when perspective h ow eda nrseicgntcvariation genetic intraspeci®c an need we do Why I PRAC OF MPORTANCE G ENETIC S TUDYING V ARIATION odih (1990 Goodnight (1992); Antonovics (1997) Wilson sensu (1999) Molles hspaper this clg,Vl 4 o 3 No. 84, Vol. Ecology, I oe ta.(1996) al. et Power oe ta.(1996) al. et Power paper this hspaper this NTRASPECIFIC a ) March 2003 COMMUNITY GENETICS 561 is signi®cant gene ¯ow. (2) Only by partitioning total ni®cantly different response to selection than a large genetic variation into three classes (within populations, number of genes of small effect (Lynch and Walsh among populations within species, and among species) 1998), and can signi®cantly alter an extended pheno- can we determine the portion of total genetic variation type and the resulting interactions. Recent theory sug- that covaries among species. Thus, to understand in- gests that quantitative traits are determined by a com- teractions among species and communities, we must bination of a few loci of large effect and many loci of ®rst concentrate on the genetic variation within species. small effect, with a signi®cant portion of the variation In combination, these two points argue that an intra- being determined by the loci of large effect (reviewed speci®c perspective provides a mechanistic basis for in Mackay 2001; but see Wolf et al. 1998, Wade 2002). understanding the ecological and evolutionary conse- We will focus on genes of large effect because the quences of extended phenotypes. introduction of these genes through mutation or gene Community genetics integrates ecology and genetics ¯ow from other populations could signi®cantly alter by studying the genetic traits responsible for the species an extended phenotype, resulting in drastic changes in interactions that create communities. Species interac- community structure. Understanding genetic architec- tions are in¯uenced by extended phenotypes, and these ture can also reveal the presence of ϩ or Ϫ genetic interactions can be positive, neutral, or negative. Such correlations among traits (Hawthorne and Via 2001), interactions among species have been shown to con- which can cause rapid evolutionary responses in a spe- tribute to the among-community component of phe- cies (Widmer 2002). In the future, technological ad- notypic variance, a characteristic de®ned as community vances in bioinformatics and genomics may allow the heritability (Goodnight 1990a, b, Goodnight and Craig analysis of the actual genes or alleles that affect species 1996; see also Wade 1977). Laboratory studies on sim- interactions, greatly increasing our precision in map- ple, two-species communities demonstrate that an ping these effects. among-community component of variance can arise The potential for these molecular approaches to fa- within just ®ve generations (Goodnight 1990a, b, cilitate a community genetics perspective is illustrated Goodnight and Craig 1996). Selection on individuals by QTL analyses that have quanti®ed the genetic basis Feature Special within communities evidently favored particular ge- of ecologically important traits in plants (Alonso-Blan- netic interactions that, when community-level selection co et al. 1998, Kim and Rieseberg 1999), invertebrates was imposed, were passed intact from ``parent'' to (Page et al. 2000), and vertebrates (Robison et al. `òffspring'' communities. In more complex commu- 2001). Genes of large effect that could have community nities, similar genetic interactions are likely to arise and ecosystem consequences have been identi®ed in and contribute to community heritability (Swenson et QTL as being responsible for bud set and ¯ush (Frewen al. 2000). et al. 2000), tree growth and architecture (Bradshaw In our development of community genetics, it is im- and Stettler 1995), pathogen resistance (Newcombe and portant to demonstrate that genes affect traits that are Bradshaw 1996), and chemical defenses (Shepherd et likely to have community and ecosystem consequences. al. 1999). In addition, major qualitative phenotypic dif- Most of these traits are expected to be quantitative, ferences, such as changing a fungus from a pathogen meaning that they are determined by multiple genetic to a mutualist (Freeman and Rodriguez 1993), the num- and environmental factors (Lynch and Walsh 1998). ber of queen ants tolerated by workers (Krieger and The transmission of these factors from parents to off- Ross 2002), and trichome morphology (van Dam et al. spring provides the heritable, and thus selectable, var- 1999) are controlled by a single gene. iation for these traits in a population. Simple herita- Key points that emerge from this section include the bility estimates provide the ®rst step in linking a trait following: (1) a community genetics perspective is de- to species interactions within communities. More pre- pendent upon an understanding of intraspeci®c genetic cise estimates about the genetic factors responsible for variation, which is the source of heritable genetic var- these complex traits can be obtained with genetic map- iation; (2) laboratory experiments show that genetic ping techniques and Quantitative Trait Locus (QTL) interactions between species can be passed from ``parent'' to `òffspring'' communities (i.e., community her- analysis. QTL analysis detects a chromosomal region itability); and (3) QTL and other genetic analyses pro- containing one or more loci that affect a trait in a spe- vide powerful tools for quantifying and mapping the ci®c environment and can be used to estimate the num- extended phenotypes of genes that have community and ber of genes involved, magnitude and sign of their ecosystem consequences. The following three sections effect (ϩ or Ϫ), mode of gene action (additive, dom- emphasize studies of dominant and keystone species inant), and gene interactions (epistasis). A detailed un- in the wild because, as community drivers, their intra- derstanding of individual genes, including their gene speci®c genetic variation has especially important con- frequencies and the magnitude of their effects on the sequences for understanding community genetics. trait, is essential for understanding the genetic basis of quantitative variation (Falconer and Mackay 1996). GENETIC VARIATION IN DOMINANT SPECIES This is important because genetic variation resulting Many vegetation types are characterized by a few from a few genes of large effect will produce a sig- species that ``dominate community biomass and have Special Feature .fremontii P. genotype and environment, genetics, pro- to chemical due their the ®les in on variation extensive effects exhibit species Such often strong embedded. are particularly they which have in communities to likely re- species are phenotypes dominant these in extended variation genetic the from sulting effects, incom- genetic show all to munity Although likely are 1996). members al. community teracting et to (Power disproportionate not biomass'' but their large, are that impacts total 562 oia ris laeadWihm(95 lsie in- as classi®ed trees (1995) dividual morpho- Whitham and of Floate phenotype traits. as ``traditional'' logical predictable own as just plant's be the can supports it that organisms enemy. extended its were of aspen enemies in the to variation chemical genetic-based the of that indicating effects hosts, their in of neg- glycosides diet was phenolic the parasitoids of levels these with of correlated mass atively caterpil- adult the The which fed. upon lars genotypes aspen among fold o10.Tesm opud fettevaiiyof viability the affect pathogen compounds the same 0% The from 100%. ranged to rates survival genotypes, aspen ferent hoo omnte,teewsa9%lvlo agree- of level 98% ar- a their was or there traits communities, morphological thropod classi®ed own were the their trees upon from when based collected that found sets They data trees. same two upon based crosses, ysoso ae ( hares snowshoe by eesaeas fetd otn(01 on htsur- parasitoid that the found of growth (2001) and Holton vivorship affected. also are levels cinnata Lnrt n wn 96.We ys oh ( moths gypsy When mammals 1996). and Hwang pathogens, and (Lindroth insects, provide against glycosides resistance Phenolic levels. trophic higher and herbivores, aspen, between interactions in¯uence olites glyco- en- minimal phenolic variation. but vironmental of variation genetic levels marked that exhibit sides show studies contrast, In same 2001). Lindroth the and Osier 1997, 2000, al. Lindroth et and Osier (Hwang light, defoliation (e.g., availability and resource nutrients) genotypes, with among greatly variable vary con- and highly of are levels tannins that densed show studies Field common-garden 1996). Hwang and and Lindroth 1987, al. concentration et in (Lindroth 25-fold and vary tannins may glycosides condensed phenolic as such Major 1996). metabolites Grant secondary and (Lin- Mitton system 1996, Hwang defense and chemical droth its in extensive variation exhibits genetic America, North of much throughout heritable. be com- can community richness and that position and has consequences species dominant in community variation We genetic 1997). 1983, the Baldwin that and show McClure Karban 1992, and Simms and Fritz (Denno interactions vironment ati dispar mantria ln' xeddpeoyeo h omnt of community the of phenotype extended plant's A h xeddpeoye fteescnaymetab- secondary these of phenotypes extended The forests early-successional of tree dominant a Aspen, erdi oettn aeplas aidtwo- varied caterpillars, tent forest in reared , yoyo mammatum Hypoxylon and ,mjrdflaos eerae ndif- on reared were defoliators, major ), ouu fremontii Populus .angustifolia P. eu americanus Lepus n ope back- complex and , ,F swl sfeeding as well as , 1 opiuacon- Compsilura .Hge trophic Higher ). yrd between hybrids HMSG HTA TAL. ET WHITHAM G. THOMAS ϫ Ly- en- hme l [1999]). al. Whi- by et community-level review and tham (see (Aguilar variation demonstrate genetic oaks of also consequences and 1996), 1992) al. Boecklen et (Messina hybridizing sage- 1988), other brush Price and of (Fritz willows Studies including level. systems re¯ected community is the that at phenotype dominant extended a an of produces genotype taxo- plant underlying tree's the ®nd- a that This argues com- traits. of morphological ing predictor arthropod own its a the as good status study, nomic as ®eld just was this munity in Thus, ment. aetlseis h F Because the trees. species, both 953 parental of and communities upon insect family, the based accumulated genetic hybrids is the each analysis in at entire trees on the more found or arthropods three of least community point the Each multidi- represents techniques. nonmetric ordination (global scaling) results the mensional GNMDS shows using 1 com- Fig. obtained that heritable. evidence is powerful structure is munity next community the one to from generation phenotypes trans- extended Such 2000). of al. mission et (Dungey effects ge- additive have associations the plant±insect that these underlying suggesting factors netic parents, both of insect the communities accumulated crosses these of progeny the that quanti®ed. were taxa insect dis- the 30 planting, of after tributions years Three pedigree. known of aetlseis h igeF both single from The signi®cantly species. parental differed and space ordination re eutdi iia atrs .. h F the i.e., patterns, these similar of in chemistry resulted essential defensive 31 trees the the with of associated analyses oils ®ndings, these spe- agree- with common In ment patterns. few community-wide a represent by but driven equally,cies, not Because treated were were traits. patterns species these insect quantitative rare of and common inheritance the with fcmuiysrcuei eial.I h wild, the In heritable. is structure phenotype community extended the of that show experimentally and cnl rmteparental the from icantly rmthe from hi nme;()a oa cl,teetne phe- extended the scale, local a at (2) and enemies; consumers affects their which pro®les, (1) chemical variation their phenotype: genetic in extended signi®cant possess the species of dominant view our to damental and 2000). species al. parental et both (Dungey of intermediate were oils the all accumulated tde of Studies ed18) nlsso netadfna aao F and on taxa (Potts fungal swarm and insect hybrid of a Analyses 1985). form Reid to boundaries their on o hs atrs otoldcossof crosses ac- controlled might patterns, these that for count hypotheses separate To environmental 1994). and phenotypes al. genetic in et differed (Whitham parental signi®cantly communities types their pure cross these that and showed hybrids, backcross dalina amygdalina ao nigo hs omngre raswas trials common-garden these of ®nding major A hs tde eosrt he onsta r fun- are that points three demonstrate studies These and .amygdalina E. .risdonii E. aual yrdzswith hybridizes naturally Eucalyptus rdcdasnhtcpopulation synthetic a produced 1 aiiswr nemdaein intermediate were families aiis hc sconsistent is which families, nAsrlaobservationally Australia in .risdonii E. 2 aiydfee signif- differed family clg,Vl 4 o 3 No. 84, Vol. Ecology, aiis u not but families, .risdonii E. 1 .amyg- E. hybrids 1 E. at s, March 2003 COMMUNITY GENETICS 563

maturation date (Groot and Margolis 1991, Quinn and Unwin 1993), and reproductive energy allocation (Kin- nison et al. 2001). Large rivers have genetically dif- ferentiated salmon populations that migrate during most months of the year in different tributaries (Groot and Margolis 1991). Genetic variation in the timing of migration and energy allocation to reproduction is likely to cascade to affect the timing and ¯ux of nutrients from the ocean to aquatic and riparian ecosystems. The importance of salmon-derived nutrient in¯ux has been demonstrated in the riparian Sitka spruce forests of Alaska. Trees along reaches with spawning salmon exhibit three times more growth than trees along reaches without salmon (Hel®eld and Naiman 2001). In response to the temporal variation in migration and spawning, behavioral changes have occurred in bears, otters, mink, and eagles, which depend upon salmon as a major source of food (Cederholm et al. 1989, Ben- David 1997). Enhanced riparian plant growth derived from the transfer of nutrients to the terrestrial com- FIG. 1. In common-garden trials using crosses of known munity creates a positive feedback that increases the pedigree, F1 hybrids accumulated the arthropod communities of both parental species, suggesting a heritable component to survival of future salmon generations (Hel®eld and community structure. Each point represents the arthropod Naiman 2001). community of 30 insect taxa found on a family of trees, based In another example, the interaction between anthrax Feature Special upon a minimum of three trees and a total of 953 trees. Eu- (Bacillus anthracis) and browsing ungulates in South calyptus amygdalina (diamonds), E. risdonii (solid circles), Africa (K. L. Smith, D. T. Scholl, V.De Vos, H. Bryden, their F1 hybrids (open circles), and a single family of F2 hybrids (ϫ). Findings are based on the ®rst two dimensions M. E. Hugh-Jones, and P. Keim, unpublished manu- (Axis 1 and Axis 2), of a six-dimension global nonmetric script) shows how genetic factors underlying the ex- multidimensional scaling (GNMDS), and are adapted from tended phenotypes of pathogens may shift the balance Dungey et al. (2000). between woodlands and grasslands. Type B anthrax strain is associated with death rates 15 times higher notype of a plant (e.g., its dependent community) can than Type A. The virulence of these two strains dif- be just as predictable as the ``traditional'' phenotype ferentially affects 15 species of ungulates in Kruger (e.g., plant morphology) in distinguishing among plant National Park, which has a history of anthrax out- genotypes; and (3) synthetic crosses demonstrate a her- breaks. In addition to these differences in mortality itable component to community composition and spe- between strains, browsing ungulates kudu (Tragela- cies richness. phus strepsiceros) and nyala (T. angasii) suffer a death rate 10 times greater than ungulates that feed on grasses GENETIC VARIATION IN KEYSTONE SPECIES (K. L. Smith, D. T. Scholl, V. De Vos, H. Bryden, M. A keystone species is de®ned as `òne whose impact E. Hugh-Jones, and P. Keim, unpublished manuscript). on its community or ecosystem is large, and dispro- Because the expansion of woody shrubs (e.g., Acacia) portionately large relative to its abundance'' (Power et often follows outbreaks (Prins et al. 1993), it is likely al. 1996). Because of the disproportionate effects of that anthrax outbreaks promote woodland invasion of keystone species and their propensity to interact strong- grassland. Thus, anthrax outbreaks and the relative ly with a wide range of other species, genetic factors abundance of the two anthrax strains may ultimately underlying the extended phenotypes of keystone spe- cause a cycle between woodlands and grasslands. Sim- cies may have especially strong effects on communities ilar examples of keystone effects in other systems in- and ecosystems. Thus, those species with the strongest clude the bacterium that causes plague (Yersinia pestis; ecological effects are also likely to be those with the Biggins and Kosoy 2001), and ®g trees that are re- strongest community genetic effects. sources for vertebrate frugivores (Ficus spp.; Janzen Paci®c salmon are recognized as keystone predators 1979). in aquatic and marine communities (Power 1990), and These examples argue two points: (1) genetic dif- their decomposing bodies are a major source of nutri- ferences underlying the extended phenotypes of key- ents in both aquatic and terrestrial systems (Willson stone species have community and ecosystem conse- and Halupka 1995, Hel®eld and Naiman 2001). Eco- quences; and (2) these effects involve keystone plants, logically important traits that are heritable in salmon animals, and microbes from marine to terrestrial en- include the timing of juvenile and adult migrations, vironments. Special Feature ϳ 564 ) nteasneo oh,brshretagreater a harvest birds moths, of absence (Fig. the 1993) In Whitham 2). and (Christensen birds seed-dispersing mammals affects and turn, nearly in and which, 1985), shrubs Mop- per and (Whitham into production cone trees female eliminates moth upright the by normally trees susceptible turns of shoots terminal destruc- of The tion attacked. genetically chronically remaining are trees the susceptible whereas attack, little suffer motn oivsiae o xml,pno pine, pinyon example, For especially investigate. be to may combined important phenotypes the extended their species, of keystone effects and dominant acting ad,tetidms omnvgtto yei the in type vegetation common States. United most third the lands, h ai fteaudneo ee ulso eitn re eaiet ucpil re.Br ihvalues with Bars trees. susceptible to relative trees resistant on guilds seven of abundance the of ratio the ircraalbovittella Dioryctria ac oase-oigmoth, stem-boring a to tance esoeseis(rw ta.20) tsrsflsites, stressful cryptic At a 2001). al. becomes et (Brown insect and species keystone the moisture soil increase, its as stress on but nutrient impact genotype, little of has environ- regardless host, insect normal this Under com- conditions, 1991). environmental mental al. an et and (Mopper genetic ponent a both has which 0 ftesaegntclyrssatt h ohand moth the to resistant genetically are trees of 20% ( F u edulis nus eas aycmuiisaelkl ohv inter- have to likely are communities many Because re hno eitn re.A seikaoetebrdntsasaitclysgicn ifrneat difference signi®cant statistically a denotes bar the above asterisk An trees. resistant on than trees IG bnac fagido eitn re hno ucpil re;br ihvalues with bars trees; susceptible on than trees resistant on guild a of abundance dpe rmCrsesnadWihm(93,Bone l 20) n uk ta.(2003). al. et Kuske and (2001), al. et Brown (1993), Whitham and Christensen from adapted .Gntcvrainaogpno ie ( pines pinyon among variation Genetic 2. . I sadmnn reo iynjnprwood- pinyon±juniper of tree dominant a is , TRCIN OF NTERACTIONS iu edulis Pinus K EYSTONE tmbrn oh fet id aml n hzshr irb omnte.Tegahshows graph The communities. microbe rhizosphere and mammal, bird, affects moth) stem-boring a , xiisvraini resis- in variation exhibits D S MNN AND OMINANT PECIES ircraalbovittella Dioryctria HMSG HTA TAL. ET WHITHAM G. THOMAS iu edulis Pinus Pi- , loafc irba omnt of community microbial a affect also long- vs. local dispersal. on effects seed its distance important in an tree the in to results loop which feedback birds seeds, among for interactions mammals and competitive determines of herbivory outcome to the resistance (Vander- rodents, Thus, distances by 1997). shorter removal wall over seed seeds greater disperse Balda to which leads reduced and the crop moths, (Vanderwall of cone presence the distances in However, long 1981). the dispersing over potentially crop, seeds seed the of proportion omnt oiat nrssac oakytn herbivore keystone a to resistance in dominant) community a , hs eei ifrne nhrioesusceptibility herbivore in differences genetic These ta.20;Fg ) hs hfsaeipratbecause important are shifts These (Kuske 2). Fig. signi®cantly 2003; differ al. not et does bac- heterotrophic abundance only terial and pattern, opposite on the show abundant pseudomonads whereas trees, more susceptible than 30±200% de- resistant are four and of fungi Actinomycetes Three affected. heterotrophic 2001). also are al. guilds et composer Basidiomy- Brown by 2; (Fig. dominated cotina are Ascomycotina, trees subdivision resistant fungal whereas the of members dominated and by are (Gehring trees trees Moth-susceptible 1991). resistant Whitham on than trees moth- on ec- susceptible lower of 30% abundance is mutualists The fungal roots. tomycorrhizal pinyon with associated is Ͻ niaegetraudneo susceptible on abundance greater indicate 1 clg,Vl 4 o 3 No. 84, Vol. Ecology, P ϳ Ͼ 0 pce that species 600 Ͻ niaegreater indicate 1 .5 aaare Data 0.05. March 2003 COMMUNITY GENETICS 565 the structure of the microbial community can affect in their response to environmental change (Lynch and tree performance and ecosystem processes. Species of Walsh 1998). For example, Orians and Fritz (1996) ectomycorrhizal fungi vary in drought tolerance and found that under normal conditions, some willow ge- the ability to utilize organic nitrogen (Smith and Read notypes were two times more resistant to insect her- 1997), qualities that affect their positive feedback on bivores than were other genotypes. However, when fer- the tree. Among the decomposers, actinomycetes de- tilizer was added to simulate good environmental con- grade recalcitrant substances (McCarthy and Williams ditions, the formerly resistant genotypes became nearly 1992), whereas pseudomonads may promote plant three times more susceptible. Because willows domi- growth by competing with pathogens and acting as nate many riparian communities, such reversals in their helpers in mycorrhizal symbioses (DeÂfago and Haas resistance phenotypes due to an environmental inter- 1990, Garbaye 1994). Thus, the extended phenotypes action could result in a pronounced shift in the depen- of moth-resistant and moth-susceptible trees determine dent community of organisms. the community structure of hundreds of species from Genotype ϫ environment interactions are likely to microbes to vertebrates. take on additional signi®cance as humans continue to Genetic differences within a dominant plant species rapidly modify environments and the genotypes that can also affect the formation of keystone mutualisms. occur within them (Palumbi 2001). Humans have elim- For example, host plant genotype determines the pres- inated and fragmented habitats, introduced exotic spe- ence and strength of the mutualism between aphids and cies, and altered atmospheric chemistry, which can in- tending ants, which, in turn, affects an arthropod com- teract with genes of dominant and keystone species to munity of Ͼ90 species. Genetic differences among cot- alter communities. For example, in aspen, genes inter- tonwoods (Populus fremontii, P. angustifolia,F1 and act with environmental pollutants to affect multiple tro- backcross hybrids) affect the fecundity and distribution phic levels. Holton (2001) found that the performance of the aphid Chaitophorus populicola (Wimp and Whi- of the forest tent caterpillar (Malacosoma disstria) did tham 2001). When aphids were transferred onto trees not vary signi®cantly with aspen genotype when ex- pca Feature Special of varying genotype in a common garden, aphid fe- posed to elevated ozone (O3), but when exposed to the cundity across cottonwood genotypes differed approx- combination of elevated O3 and CO2, performance var- imately threefold in just 10 days, an effect that was ied 35% among aspen genotypes. This complex inter- mirrored in the distribution of aphids at ®eld sites. action has an extended phenotype in which a higher Given a suitable host genotype, the survival of the trophic level is affected; performance of the parasitoid aphid colony then depends upon the acquisition of ant (Compsilura concinnata) differed threefold among ge- mutualists such as Formica propinqua. Field obser- notypes under elevated O3 and CO2. vations and experiments showed that if an aphid colony These anthropogenic-caused environmental changes was Ͻ5 m from an ant mound, a mutualism would form, may lead to unpredictable genotype ϫ environment in- but if the distance was Ͼ5 m, it would fail and the teractions whose extended phenotypes dramatically al- colony would die out (Wimp and Whitham 2001). F. ter communities. The common reed, Phragmites aus- propinqua repels other herbivores, other species of tralis, was considered rare or uncommon in North ants, and generalist predators, yet, the mutualism at- America during the 1800s. However, the introduction tracts a unique group of predators and parasites with of an exotic genotype of this species from Europe, Af- adaptations for cryptic living among aphid±ant mutu- rica, or Asia (Saltonstall 2002), combined with human- alists. Because these specialists are found only in as- caused environmental disturbances (e.g., agricultural sociation with the ant±aphid mutualism, whereas others drainage, dikes, and urban expansion; Chambers et al. are found only in its absence, the greatest species di- 1999), has led to an expansion of P. australis, such versity is achieved in an environment that contains a that it is a dominant plant species in wetlands through- mosaic of tree genotypes in varying proximity to tend- out the mainland United States and southern Canada. ing ants. Its increased dominance has resulted in dramatic de- These examples illustrate two points: (1) genetic var- creases in the diversity of wetland plant and bird com- iation within dominants may be most important when munities (Chambers et al. 1999) and the apparent dis- it affects keystone species; and (2) these genetic dif- appearance of native P. australis genotypes from New ferences affect the composition and biodiversity of the England (Saltonstall 2002). The fact that an exotic ge- community (see the Conservation implications sec- notype of a native species has such large community tion). consequences emphasizes the importance of intraspeci®c genetic variation and the extended phenotypes that ENVIRONMENTAL INTERACTIONS come about through interactions with the environment. The environments in which the genes of keystone These studies demonstrate two points: genetic inter- and dominant species are embedded will greatly in¯u- actions with the environment affect dominant species, ence their extended phenotypes and subsequent effects whose extended phenotypes can cascade to affect mul- on communities and ecosystems. Genotype ϫ environ- tiple trophic levels; and human activities will probably ment interactions occur when different genotypes vary increase the importance of genotype ϫ environment Special Feature ouu fremontii Populus polymorpha M. on impacts human un- these to ecosystems. of important rami®cations the is derstand perspective genetics A environments. community introductions) climate) exotic (e.g., (e.g., biotic abiotic and in changes through interactions 566 n iosk(01 unie h osqecsof consequences in the variation quanti®ed genetic (2001) eco- Treseder Vitousek to addressed. been and framework rarely has evolutionary that studies and system pro- genetic This a cycling. nutrient vides and decomposition as processes such ecosystem-level on effects step their examine next to the is communities, in¯uence species keystone oeoytmfnto.Uigsnhtccossbetween crosses synthetic important Using probably function. are ecosystem that to ge- traits the speci®c establish of basis to netic analyses QTL ecosystem-level and of pedigree crosses to known experimental traced uses be approach This can processes. plant dominant a in un- are affect they that traits known. the and involved speci®c genes the a dynamics, of ecosystem with existence correlate the genetic demonstrate studies these processing. ecosystem Although of mosaic a var- into intraspeci®c translated by iation created mosaic genetic land- these the of scapes, both Across decomposition availability. litter ammonium to both and related affected positively that was chemistry matrix, litter distance genetic determined as a phenotype, by litter that found They ¯uxes. oyi iest mn iegntpsof genotypes nine among phe- diversity that found notypic in (2002) Hunter resulted and cycling. turn, Madritch nutrient root Likewise, in¯uence in as may which, that such feedbacks chemistry, traits positive litter plant di- leaf in alleleic differences and in to differences led small versity that differences concluded genetic They function. ecosystem in if differences nu- with determine reciprocal associated were with to garden and treatments common population trient a each in grown from were collected were populations. between Seedlings distance genetic the determine to sn nterpra oedmntae htteamount the that demonstrated zone riparian compar- the stand-level in Furthermore, isons 3C). (Fig. aquat- an system in ic decomposition explained litter in subsequently variation the con- which of 63% tannin 3B), in (Fig. differences cross centrations manyfold same found the from and leaves types used in (2000) Driebe differences Whitham levels. and higher genetic to of extended are effects production tannin The 3A). single (Fig. a by QTL for accounted tannins be pheno- condensed could leaves the of cottonwood production in of the portion in signi®cant variation typic a that found (2001) ytca)o csse ucini aai They in enzymes Hawaii. functional of in diversity function allozyme compared ecosystem on Myrtaceae) laevis eas h xeddpeoye fdmnn and dominant of phenotypes extended the Because eetsuisso htteefcso apdgenes mapped of effects the that show studies Recent a in®atipcso abnadnitrogen and carbon on impacts signi®cant had G ouain ln etlt gradient fertility a along populations NSTO ENES and ersdrspolymorpha Metrosideros .angustifolia P. E COSYSTEMS HMSG HTA TAL. ET WHITHAM G. THOMAS Woolbright , Quercus (òhià, n .G Whitham, 3D; G. Woolbright, (Fig. S. T. ratios Bailey, K. and N J. Hart, : C. S. lignin Schweitzer, A. or J. inputs biomass total than mineralization did N an- net through predicted stand better litterfall the nual entered that tannin condensed of motneo t eei iest o h eto the of rest the for the diversity diversity, genetic genetic its its of conserve importance to need we species hni o-aldlae rmtesm re J A. (J. trees same the from leaves Schweitzer, non-galled decomposition of in rates than slower tannins, 35±45% leaf in in increase susceptible result fold which 4±7 genetically a induce On aphids trees, 1989). (Whitham sistant csse rcse eg,teQLfrla tannins leaf mineralization). for N and QTL decomposition the affect affect to (e.g., steps few processes relatively in ecosystem act or direct be diffuse; be can not they need effects these (2) ecosystem and have consequences; for that QTL phenotypes the extended as have such tannins genes species, dominant a through extended traits. that plant suggest numerous 1996) for exist Vrieling Fritz phenotypes and and Dam re- van Orians pest 1988, 1994, and Price [2001]), and al. (Fritz and sistance et Osorio Benowicz [1991], [1994], en- Zwier Pereira changing and to Bassman resilience up- vironments; ef®ciency, carbon use (i.e., water processes take, physiological chemistry 2001), Lindroth plant 1995, al. al. et et Adler in 1993, al. and et differences (Nichols-Orians common Genetic are interactions important. complex and variation effects. direct the and that exceed effects tannins indirect may in condensed results resistance for aphid inter- for genes genes The the 1996). between Whitham action and and fungi, birds, (Dickson including arthropods taxa the other and 42 richness of species abundance affects also presence± aphids The of process. absence to ecosystem to production an tannin resistance affect for affect indirectly genes the that with genes interact al. the aphids et when Findlay 1988, result Choudhury 1996) also (see herbivory of eoye r ihyssetbet h gall-forming the tree to some susceptible aphid, example, highly of For are organism) QTL. genotypes or tannin step condensed third the a involve that interactions Karchesy deter and they 1989). (Hemingway and herbivores and 2000) systems pathogens Vitousek diverse and in cycling (Hattenschwiler nutrient com- these in¯uence because pounds are signi®cance general tannins of is condensed the nitrogen heritable that and and Knowing chemistry, decomposition mineralization. litter be- of leaf links processes trait, direct ecosystem the mapped demonstrate a system tween same the components terrestrial of and aquatic in studies combined lhuhi a ogbe eonzdta osv a save to that recognized been long has it Although hs tde mhsz w ons 1 hnacting when (1) points: two emphasize studies These genetic intraspeci®c that suggest also studies Other ti loipratt osdrteidrc ik (i.e., links indirect the consider to important also is It epiu betae Pemphigus nulse data unpublished C ONSERVATION nulse manuscript unpublished hra tesaehgl re- highly are others whereas , I .Sc `felf' effects `àfterlife'' Such ). MPLICATIONS clg,Vl 4 o 3 No. 84, Vol. Ecology, .These ). March 2003 COMMUNITY GENETICS 567 pca Feature Special

FIG. 3. Genetic variation in Populus hybrids results in differential expression of a condensed tannin QTL (Quantitative Trait Locus) that can be traced through ecosystem-level processes. (A) Composite interval mapping shows the location of a QTL for tannin concentration on a linkage group of a Populus fremontii ϫ P. augustifolia backcross family. One or more genes that in¯uence tannin production are located in the region where the signi®cance threshold level exceeds LOD 3.0 (log of the difference, equivalent to P Ͻ 0.001; Woolbright 2001). (B) In a common environment, these cross types express manyfold differences in the concentration of condensed tannins (Driebe and Whitham 2000). (C) The concentration of condensed tannins in different Populus cross types accounts for 63% of the variation in litter decomposition among cross types in a stream (Driebe and Whitham 2000). For panels (B) and (C), vertical lines represent 1 SE, and different letters indicate signi®cant differences among means (P Ͻ 0.05). (D) The concentration of condensed tannins also explains 57% of the variation in terrestrial net N mineralization rates between 12 stands that differ in cross type compositions (Schweitzer, unpublished manuscript). community has been much less appreciated. If indi- include another dimension that recognizes the impor- vidual species are dependent upon a subset of the ge- tance of genes in one species to other dependent spe- nome of another species, then their survival is more cies. closely tied to conserving the individuals possessing An example of one species being dependent upon those speci®c genes rather than all individuals in the the genetic makeup of another species is that of the other species' population. A consequence of the ex- gall mite, Aceria parapopuli, on cottonwoods. Mc- tended phenotype is that conservation genetics must Intyre and Whitham (2003) found that 99.9% of the Special Feature iespplto a ocnrtdo aual oc- naturally on F concentrated curring was population mite's 568 eas utbeF suitable Because ae fmts nwihteitiscrt fincrease of rate intrinsic the which ( in mites, of growth ``poten- rates different the vastly in in resulted population trees host tial'' among differences genetic trials, odadKee(90,Rie ta.(93,and between greater (1993), interactions even al. Gene-speci®c exhibit speci®city. et can (1988), Microbes Roinen al. (1994). (1990), et Thompson Keese Feder by and 1986). provided Wood (Scriber are speci®c host examples very Other be can it scale, local nitdadhgl pcaie ttelcllevel local the swallowtail, at tiger eastern specialized The 1994). highly (Thompson and limited. entiated host be still Even can species population. dependent for- riparian a in cottonwood ests, trees dominant larger are cottonwoods the though small a of is mites these subset for population host ``actual'' the et h ieo ouainnee omiti the maintain by to required needed levels population at diversity a genetic repre- of that size the (MVIP) sents size population interacting viable sin- against management. argument gle-species the furthers 1998). and support criticisms, to Simberloff these mechanism a 1991, provides Rohlf genetics Community (e.g., ecosystems communities or than con- rather researchers species with individual of associated Other servation weaknesses 1994). the recognized Thompson have (see oth- species with interactions er important support may to species small one too in be 1981) (Shaffer sizes (MVP) ulation 1993). al. et (Vijn and ®xation interaction nitrogen symbiotic the prevent disrupt will gene single a r lns( plants ocnev eei iest ndmnn n keystone and important dominant is in diversity it genetic phenotypes, conserve to extended their of because community. the of pheno- rest extended the their affect because types species, as keystone just dominant in be and variation may genetic the it conserve to species, important rare conser- target current strategies although vation Additionally, community species. the key- and stone consider dominant in not variation genetic do of consequences they if efforts ¯ounder conservation why may efforts for conservation mechanism future a for provide guide and points a end as These serve species. should dependent for important that end are interactions upper community conserve the to represents required con- is MVIP to that and required species, is the that end serve lower conserve the represents to MVP Thus, MVIP. required the is members, community what dependent other than smaller prob- are much species ably target population the viable conserving for sizes Minimum (MVP) species. interacting and agdfo o15o niiulte genotypes. tree individual on 1.5 to 0 from ranged ) imleguminosarum bium loglaucus ilio vngnrls pce a egntclydiffer- genetically be can species generalist Even eas fteecnen,w rps minimum a propose we concerns, these of Because pop- viable minimum that suggests speci®city Such h rcdn ruet mhsz orpit:(1) points: four emphasize arguments preceding The iu sativum Pisum 1 yrd.I v er fcommon-garden of years ®ve In hybrids. a ag ito otseis u ta at but species, host of list large a has , 1 yrdhssaerr ntewild, the in rare are hosts hybrid bv. euaesmiss h osof loss The symbiosis. regulate ) viciae n aieAgaipea Afghani native and other HMSG HTA TAL. ET WHITHAM G. THOMAS dependent Rhizo- Pap- MI)sz etrr¯csgast osreinterac- species. dependent conserve their to and goals tions re¯ects better size population (MVIP) interacting viable minimum among (4) interactions and of species; patterns single-spe- to also to but just survival, not cies linked is variation genetic be- goals cause conservation com- broaden of should understanding genetics munity an (3) dependent; are community members other which upon interactions crucial pre- to serve fail may management single-species (2) species; Gongt1990 (Goodnight oyi aine hti,to is, that variance, phenotypic of component among-community species the among to interactions contribute based populations genetically laboratory which on in pro- research this cited have of we evidence can cess, As evolution. phenotypes community to extended lead underlying on selection factors that genetic argued have We characters. heritable vlto,i sncsayt hwta,udrnatural under differences genetic that, on show acts to selection conditions, necessary is it evolution, trait individual among- in the expression. variance as of detectable component are community differences communi- distinct whose genetically interac- ties, to These lead ultimately heritability. tions community to contribute species associated their and phenotypes extended tween be- interactions genetic species, of keystone phenotypes and dominant extended on the focus underlying cite we factors that genetic examples the heritabil- Although exists. community ity that evidence direct them, duced ru n ihrlvlselection level of higher effects and the which group in histor- circumstances an on from emphasis arises ical selection multilevel of portance eietisc omnte Dne ta.2000). al. et on insects of (Dungey their communities ``offspring'' of Thus, communities richness insect and composition resident the affect dictably wno ta.20;seas ae17) ehave of We generations successive 1977). in trees of Wade genotypes also the see that shown 2000; al. et Swenson or ta.19,Arwle l 2001). al. et Agrawal 1997, al. 1992, al. et et re- Moore Goodnight statistically 1982, Craig been exis- 1978, (Wade has the solved selection over multilevel debate of philosophical tence The tractable precise. consequences and evolutionary its multilevel on of and effects analysis selection the level made higher have and ®tness individual group identifying for ods meth- Quantitative ®tness. relative signi®cantly selection, individual levels, in¯uence level these among higher interactions and as well group in as of circumstances effects ignores the It which approach case. extreme This an 1966). presents (Williams selection individual tus omnt level. community vltoaycag eut hnslcinat on acts selection when results change Evolutionary oudrtn h rae in®ac fcommunity of signi®cance broader the understand To re eebete`prn' omnte htpro- that communities ``parent'' the resemble trees Eucalyptus M ULTILEVEL C OMMUNITY uho h otoes vrteim- the over controversy the of Much a , ouain fkonpdge pre- pedigree known of populations b odih n ri 1996, Craig and Goodnight , S LCINAND ELECTION E VOLUTION omnt heritability community clg,Vl 4 o 3 No. 84, Vol. Ecology, supersede Eucalyp- hs of those tthe at March 2003 COMMUNITY GENETICS 569

We suggest that community-level selection is wide- sures the intensity of selection on individual pheno- spread, and that multilevel selection can be detected types as a result of their membership within particular using contextual analysis, a conventional, statistical communities. framework based on multiple regression (Sokal and How can these parameters be used to detect multi- Rohlf 1995). Contextual analysis makes use of the stan- level selection in nature? As an example, consider the dard evolutionary genetic de®nition of selection as the gall-forming aphid, Pemphigus betae, and its cotton- covariance between phenotype, z, and relative ®tness, wood host trees. Here, aphid survival depends upon at w (Cov[w, z]; Lande and Arnold 1983, Lynch and least three factors: the genotype of the aphid (individual Walsh 1998). Here, relative ®tness, w, equals an in- effects); the genotype of the tree, which in¯uences local dividual's absolute ®tness divided by the average ®t- aphid density (group effects); and the genotypes and ness of all individuals. With contextual analysis, w is numbers of other organisms associated with the tree partitioned into within- and among-group components. and its aphids (community effects). Aphid genotype Thus, it identi®es individual and group effects on in- in¯uences its ability to form a gall and reproduce, and dividual relative ®tness (Goodnight et al. 1992). It also tree susceptibility affects the distribution and density shows that even when selection acts only on individuals of aphids, which in turn affect many other species. (e.g., soft selection; Wade [1985]), indirect effects on These species (e.g., arthropod and avian predators) af- ®tness arising from group membership contribute sig- fect aphids and the host tree. Considering the genotype ni®cantly to the total variance in relative ®tness, i.e., to total selection (Crow 1958). Although this approach of the aphid (k), the genotype of the tree ( j), and the is limited in its ability to identify the source of genetic aggregate genotypes of the community of other organ- interactions (Agrawal et al. 2001), it does show when isms associated with each tree (i) as independent var- and how strongly multilevel selection acts. Moreover, iables in multiple regression, we can quantify how each its use removes the need to consider only situations in level of selection contributes to the relationship be- which the intensity of community-level selection ex- tween the aphid's ability to form a functional gall (i.e., ceeds that of individual selection, or those situations the phenotype, z) and its subsequent fecundity relative Feature Special in which direct competition among communities occurs to other aphids (i.e., relative ®tness, w). (the conditions of Johnson and Boerlijst 2002). To understand community evolution, we must un- We can rewrite the standard equation of Goodnight derstand four phenomena: (1) the nature of genetic var- et al. (1992) to include a term that accounts for com- iation underlying traits within species; (2) how trait munity effects on individual relative ®tness, w, as interactions within and among species contribute to the among-community component of phenotypic variance w ϭ bzϩ bzϩ bz. wz´zij.´zi.. ijk wzj.´z´zi.. ij. wzi..´z´zi. i.. (e.g., community heritability); (3) how these direct and Here, the effects of selection acting at individual, indirect genetic effects in¯uence the relative ®tness of group, and community levels are identi®ed by the three individuals and communities; and (4) how selection terms on the right side of the equation. Each term has acts at hierarchical levels within communities. Iden- two parts, a partial regression coef®cient and an in- tifying the levels at which selection is most powerful dividual, or average, phenotype. In the ®rst term, will reveal whether individual-, group-, and commu- bwz´zij.´zi.. describes the partial regression of relative ®t- nity-level selection have signi®cant effects on relative ness on individual phenotype, with the effects of the ®tness. This will allow researchers to focus their efforts average group and community phenotypes held con- on the causes of ®tness differences. Once traits that stant. It is multiplied by z , the phenotype of the kth ijk represent the extended phenotypes of dominant and individual in the jth group, within the ith community. keystone species are identi®ed, their in¯uence on other Thus, the ®rst term measures the intensity of individual species can be rigorously explored with factorial ex- selection acting on individual phenotypes. clusion experiments. Controlled crosses, QTL, and se- In the second term, the coef®cient b describes wzj.´z´zi.. quence analyses can then be used to explore the genetic the partial regression of relative ®tness on group phenotype, with the effects of individual and average com- architectures underlying these traits. Four major points are raised in this section. (1) The munity phenotype held constant. It is multiplied by zij, the average phenotype in the jth group. The subscript issue of group selection vs. individual selection is out- ``.'' represents the average across all individuals within moded; selection can operate simultaneously at mul- each group. Thus, the second term measures the inten- tiple levels. (2) Due to the heritability of extended phe- sity of selection on individual phenotypes as a result notypes and multiple levels of selection, community of their membership within particular groups. The third evolution is likely. (3) Realistic statistical methods and coef®cient, bwzl..´z´zi., describes the partial regression of experiments allow us to measure the relative impor- relative ®tness on community phenotype, with the ef- tance of selection acting at different levels. (4) To the fects of individual and average group phenotype held extent that species interactions affect genetic covari- constant. It is multiplied by zi.., the average phenotype ances, species evolution must be placed in a community within the ith community. Thus, the third term mea- context. Special Feature de,L . .Shit n .D oes 95 Genetic 1995. Bowers. D. M. and Schmitt, J. S., L. Adler, 570 o eserfrpeaigilsrtos n .Arwl P. Agrawal, comments. reviewers constructive anonymous A. and their Wilson, and S. for D. Rehill, illustrations, B. insights, Hart, preparing S. and Keim, for discussions Ari- Redsteer stimulating Northern their Ron at We for group University seminar DE-FG03-94ER61849. zona genetics community grant the DOE thank 97-35302-4241, and grants 95-37302-1810, USDA DEB-9306981, DEB-0074427, DEB-9615313, DEB-9726504, DEB-0075563, DEB-0087017, (G expression Because gene MVIP). in¯uences of min- population, environment genome the interacting the of viable concept of the imum subset (i.e., a drivers on community these depen- species the other to of due dence important especially be may keystone of species and extensive. dominant effects in be variation the to genetic expect Conserving phenotypes we extended 2), combined they Fig. their and (e.g., both, interact have often communities spe- most keystone Because or cies. dominant conse- in are ecosystem expressed and when phenotypes Our community quences have extended diffuse. to be these likely most not that need suggest and studies in steps exerted extended be few these can remarkably of processes effects ecosystem on The miner- phenotypes 3). N as Fig. such (e.g., processes alization com- ecosystem the to to and levels, trophic munity, other to individuals trait, the the from possessing traced be The can phenotype. phenotype re- extended extended perspective an our of consequences examination, the of veals level every at ics, lxitrcin,adteefcso anthropogenic of effects the change. and perspective com- interactions, its world, genetics natural plex the understand community to us help a should the of interactions ecosystems, and development species communities de®ne Because and critically structure hypotheses. to these exist issue, analyses test controversial and this experiments on tractable position one's of Re- evolution. gardless community of likelihood the combi- enhance heritability In community and wild. phenotypes the extended in nation, evolution com- and understanding structure toward step munity important an is (e.g., 1) heritable Fig. be can richness species and composition understand. the to on important impacts especially species are keystone human and the dominant and of reason, phenotypes change extended this global For of consequences. effects unex- large with levels and trophic pected change, multiple en- affect environments one to as in cascading expressed expressed be not can are vironment that phenotypes tended aito ndfnieceityin chemistry defensive in variation Patgnca)adisefc nteseils herbivore specialist the coenia on Junonia effect its and (Plantaginaceae) hsrsac a upre yNFgat DEB-0078280, grants NSF by supported was research This hogotordvlpeto omnt genet- community of development our Throughout h xeietldmntainta community that demonstration experimental The Nmhlde.Oecologia (Nymphalidae). A L CKNOWLEDGMENTS ITERATURE C ONCLUSIONS C ITED lnaolanceolata Plantago HMSG HTA TAL. ET WHITHAM G. THOMAS 101 :75±85. ϫ ) ex- E), gia,J . n .J ocln 92 atrso her- of Patterns 1992. Boecklen. J. W. and M., J. Aguilar, gaa,A . .D rdeII n .J ae 01 On 2001. Wade. J. M. and III, Brodie D. E. F., A. Agrawal, ioyi the in bivory nietgntcefcsi tutrdppltos American Naturalist populations. structured in effects genetic indirect asa,J . n .C we.19.Gsecag char- exchange Gas 1991. Zwier. C. J. and H., J. Bassman, ign,D . n .Y oo.20.I¯ecso intro- of In¯uences 2001. Kosoy. Y. M. and E., D. Biggins, rdhw .D,J. n .F ttlr 95 oeua ge- Molecular 1995. Stettler. F. R. and Jr., D., H. Bradshaw, eoiz . .R alo,M .E-asb,adY A. Y. and El-Kassaby, R. M. Carlson, R. G. A., Benowicz, mink: wild in reproduction of Timing 1997. M. Ben-David, huhr,D 98 ebvr nue hne nleaf-litter in changes induced Herbivore 1988. D. Choudhury, akn,R 92 h xeddpeoye xodUni- Oxford phenotype. De extended The 1982. R. selection Dawkins, measuring for possibilities Some 1958. F. J. selec- individual Crow, versus selection Group 1982. M. D. Herbivore Craig, 1993. Whitham. G. T. and M., K. Christensen, iko,L . n .G hta.19.Genetically-based 1996. Whitham. G. T. plant and L., difference: L. Dickson, la Vive 2001. Cabido. M. and S., Diaz, and plants Variable 1983. McClure. S. M. and F., R. Denno, noois .19.Twr omnt eeis Pages genetics. community Toward 1992. J. Antonovics, eehl,C . .B oso,D .Cl,adW .Scar- J. W. and Cole, L. D. Houston, B. D. J., C. C. Cederholm, and Ernest, Morgan K. S. Whitham, G. T. H., J. Brown, hmes .M,L .Myro,adK atntl.1999. Saltonstall. K. and Meyerson, A. L. M., R. Chambers, lnoBac,C,S .E-sa,G opad n M. and Coupland, G. El-Assal, E. S. C., Alonso-Blanco, ouu trichocarpa Populus ceitc of acteristics USA. Illinois, Chi- Chicago, of Press, University pathogens. cago and herbivores to resistance rs adns,gsecag n rwh iveGenetica Silvae growth. and 50 exchange gas hardiness, frost rmeooyo lgei sa ora fMammalogy of Journal Asia. in plague 82 implications of mammals: ecology American from North on plague duced eiso rwhaddvlpetin development and growth of netics euapapyrifera Betula fg,G,adD as 90 suooasa antagonists as Pseudomonads 1990. Haas. D. and G., Âfago, lKsay 01 eei aito mn ae birch paper among variation ( Genetic 2001. El-Kassaby. Zoology Journal Canadian of salmon. Paci®c spawning of in¯uence the 8 ot mrc.AutcBotany Aquatic America. North eoreqaiy elce seto ebvr neco- in herbivory Oikos of dynamics. aspect nutrient neglected system a quality: resource fsibrepatptoes oe fato n genetic and action Biochemistry of Soil modes analysis. pathogens: plant soilborne of USA. York, New York, New Press, versity Biology Human man. in intensities Evolution analysis. experimental an tion: Ecology seeds. pin- pine for mammals yon and birds between competition on impact ln eitnetat fetatrpd,fni n birds. and fungi, arthropods, affect Oecologia traits resistance plant Evolution and Ecology Trends in processes. ecosystem to matters diversity functional Academic systems. USA. York, managed New York, and New Press, natural in herbivores 426±449 749±764. sad ctpsof ecotypes Islands et 99 aeo ooslo acse nspawning Sci- Aquatic in and carcasses Fisheries salmon of ences coho Journal of Canadian Fate streams. 1989. lett. Science dynamics systems. the and ecological interactions of Complex 2001. Gehring. A. phenology Genetics and tree. form, forest growth, a on in traits effects large with QTLs xaso of Expansion orne.19.Aayi fntrlallcvrainat variation allelic the in natural loci of time ¯owering Analysis 1998. Koornneef. Oikos plex. :145±159. :7±3. :906±916. 46 158 in :1347±1355. 106 :308±324. 75 .S rt n .L im,eios Plant editors. Simms, L. E. and Fritz S. R. 64 ouu trichocarpa Populus uru grisea Quercus hamtsaustralis Phragmites :376±382. :400±406. :498±504. as. ouain ngermination, in populations Marsh.) ϫ rbdpi thaliana Arabidopsis .deltoides P. adbr erecta Landsberg 74 16 :2270±2278. 6 :646±655. :249±291. ϫ 139 293 .gambelii Q. 64 51 , clg,Vl 4 o 3 No. 84, Vol. Ecology, lns rePhysiology Tree clones. :261±273. :963±973. ouu deltoides Populus :643±649. 30 :389±393. notdlwtad of wetlands tidal into Populus :1±13. 36 Genetics . n aeVerde Cape and :271±282. V Mapping IV. . pce com- species and , 149 : March 2003 COMMUNITY GENETICS 571

Driebe, E., and T. G. Whitham. 2000. Cottonwood hybrid- Hwang, S.-Y., and R. L. Lindroth. 1997. Clonal variation in ization affects tannin and nitrogen content of leaf litter and foliar chemistry of aspen: effects on gypsy moths and forest alters decomposition. Oecologia 123:99±107. tent caterpillars. Oecologia 111:99±108. Dungey, H. S., B. M. Potts, T. G. Whitham, and H.-F. Li. Janzen, D. H. 1979. How to be a ®g. Annual Review of 2000. Plant genetics affects arthropod community richness Ecology and Systematics 10:13±51. and composition: evidence from a synthetic eucalypt hybrid Johnson, C. R., and M. C. Boerlijst. 2002. Selection at the population. Evolution 54:1938±1946. level of the community: the importance of spatial structure. Falconer, D. S., and T. F. C. Mackay. 1996. Introduction to Trends in Ecology and Evolution 17:83±90. quantitative genetics. Fourth edition. Longman Science and Karban, R., and I. T. Baldwin. 1997. Induced responses to Technology, Harlow, UK. herbivory. University of Chicago Press, Chicago, Illinois, Feder, J. L., C. A. Chilcote, and G. L. Bush. 1988. Genetic USA. differentiation between sympatric host races of the apple Kim, S., and L. H. Rieseberg. 1999. Genetic architecture of maggot ¯y Rhagoletis pomonella. Nature 336:61±64. species difference in annual sun¯owers: implications for Findlay, S., M. Carreiro, V. Karischik, and C. G. Jones. 1996. adaptive trait introgression. Genetics 153:965±977. Effects of damage to living plants on leaf litter quality. Kinnison, M. T., M. J. Unwin, A. P. Hendry, and T. P. Quinn. Ecological Applications 6:269±275. 2001. Migratory costs and the evolution of egg size and Floate, K. D., and T. G. Whitham. 1995. Insects as traits in number in introduced and indigenous salmon populations. plant systematics: their use in discriminating between hy- Evolution 55:1656±1667. brid cottonwoods. Canadian Journal of Botany 73:1±13. Krieger, M. J. B., and K. G. Ross. 2002. Identi®cation of a Freeman, S., and R. J. Rodriguez. 1993. Genetic conversion major gene regulating complex social behavior. Science of a fungal plant pathogen to a nonpathogenic, endophytic 295:328±332. mutualist. Science 260:75±78. Kuske, C. R., L. O. Ticknor, J. D. Busch, C. A. Gehring, and Frewen, B. E., T. H. H. Chen, G. T. Howe, J. Davis, A. Rohde, T. G. Whitham. 2003. The pinyon rhizosphere, plant stress W. Boerjan, and H. D. Bradshaw, Jr. 2000. Quantitative and herbivory affect the abundance of microbial decom- trait loci and candidate gene mapping of bud set and bud posers in soils. Microbial Ecology Journal, in press. ¯ush in Populus. Genetics 154:837±845. Fritz, R. S., and P. W. Price. 1988. Genetic variation among Lande, R., and S. J. Arnold. 1983. The measurement of se- plants and insect community structure: willows and saw- lection on correlated characters. Evolution 37:1210±1226. ¯ies. Ecology 69:845±856. Lindroth, R. L., M. T. S. Hsia, and J. M. Scriber. 1987. Sea- sonal patterns in the phytochemistry of three Populus spe- Fritz, R. S., and E. L. Simms. 1992. Plant resistance to her- Feature Special bivores and pathogens: ecology, evolution, and genetics. cies. Biochemical Systematics and Ecology 15:681±686. University of Chicago Press, Chicago, Illinois, USA. Lindroth, R. L., and S.-Y. Hwang. 1996. Diversity, redun- Garbaye, J. 1994. Tansely Review Number 76. Helper bac- dancy and multiplicity in chemical defense systems of as- teria: a new dimension to the mycorrhizal symbiosis. New pen. Recent Advances in Phytochemistry 30:25±56. Phytologist 128:197±210. Lindroth, R. L., T. L. Osier, S. A. Wood, and H. R. A. Barnhill. Gehring, C. A., and T. G. Whitham. 1991. Herbivore-driven 2001. Effects of genotype and nutrient availability on phy- mycorrhizal mutualism in insect-susceptible pinyon pine. tochemistry of trembling aspen (Populus tremuloides Nature 353:556±557. Michx.) during leaf senescence. Biochemical Systematics Goodnight, C. J. 1990a. Experimental studies of community in Ecology 30:297±307. evolution I: the response to selection at the community Loehle, C., and J. H. K. Pechmann. 1988. Evolution: the level. Evolution 44:1614±1624. missing ingredient in systems ecology. American Naturalist Goodnight, C. J. 1990b. Experimental studies of community 132:884±899. evolution II: the ecological basis of the response to com- Lynch, M., and B. Walsh. 1998. Genetics and analysis of munity selection. Evolution 44:1625±1636. quantitative traits. Sinauer, Sunderland, Massachusetts, Goodnight, C. J., and D. M. Craig. 1996. The effect of co- USA. existence on competitive outcome in Tribolium castaneum Mackay, T. F. C. 2001. The genetic architecture of quanti- and Tribolium confusum. Evolution 50:1241±1250. tative traits. Annual Review of Genetics 35:303±339. Goodnight, C. J., J. M. Schwartz, and L. Stevens. 1992. Con- Madritch, M. D., and M. D. Hunter. 2002. Phenotypic di- textual analysis of models of group selection, soft selection, versity in¯uences ecosystem functioning in an oak san- hard selection and the evolution of altruism. American Nat- dhills community. Ecology 83:2084±2090. uralist 140:743±761. McCarthy, A. J., and S. T. Williams. 1992. Actinomycetes Groot, C., and L. Margolis. 1991. Paci®c salmon life his- as agents of biodegradation in the environmentÐa review. tories. University of British Columbia Press, Vancouver, Gene 113:189±192. British Columbia, Canada. McIntyre, P. J., and T. G. Whitham. 2003. Plant genotype Hartl, D. L. 1980. Principles of population genetics. Sinauer, affects long-term herbivore population dynamics and ex- Sunderland, Massachusetts, USA. tinction: conservation implications. Ecology 84:311±322. Hattenschwiler, S., and P. M. Vitousek. 2000. The role of Messina, F. J., J. H. Richards, and E. D. McArthur. 1996. polyphenols in terrestrial ecosystem nutrient cycling. Variable responses of insects to hybrid versus parental Trends in Ecology and Evolution 15:238±243. sagebrush in common gardens. Oecologia 107:513±521. Hawthorne, D. J., and S. Via. 2001. Genetic linkage of ecological specialization and reproductive isolation in pea Mitton, J. B., and M. C. Grant. 1996. Genetic variation and aphids. Nature 412:904±907. the natural history of quaking aspen. BioScience 46:25± Hel®eld, J. M., and R. J. Naiman. 2001. Effects of salmon- 31. derived nitrogen on riparian forest growth and implications Molles, M. C., Jr. 1999. Ecology. McGraw-Hill, Boston, for stream productivity. Ecology 82:2403±2409. Massachusetts, USA. Hemingway, R. W., and J. J. Karchesy. 1989. Chemistry and Moore, A. J., E. D. Brodie III, and J. B. Wolf. 1997. Inter- signi®cance of condensed tannins. Plenum Press, New acting phenotypes and the evolutionary process. I. Direct York, New York, USA. and indirect effects of social interactions. Evolution 51: Holton, M. K. 2001. Effects of elevated carbon dioxide and 1352±1362. ozone on tree±insect±parasitoid interactions. Thesis. Uni- Mopper, S., J. B. Mitton, T. G. Whitham, N. S. Cobb, and K. versity of Wisconsin, Madison, Wisconsin, USA. M. Christensen. 1991. Genetic differentiation and hetero- Special Feature 572 oio,B . .A hee,K udn .Ska n G. and Sikka, P. Sundin, K. Wheeler, A. P. D., B. Robison, history life in Variation 1993. Unwin. J. Herbivore M. and P., 1993. T. Quinn, Jeugd. der van P. H. and T., H. H. J. Prins, W. Menge, A. B. Estes, A. J. Tilman, D. E., M. Science Power, webs. food river in ®sh of Effects 1990. E. M. Power, ae .E,J. .K odk .J ut .Guzman-Novoa, E. Hunt, J. G. Fondrk, K. M. Jr., E., R. Page, ihl-ras .M,R .Fiz n .P lue.1993. Clausen. P. T. and Fritz, S. R. M., C. Nichols-Orians, se,T . .Y wn,adR .Lnrt.20.Within- 2000. Lindroth. L. R. and Hwang, S.-Y. L., T. Osier, soil-nutrient and Genetic 1996. Fritz. S. R. and M., C. Orians, atntl,K 02 rpi nainb o-aiege- non-native a by invasion Cryptic 2002. K. Saltonstall, onn . .Voie,J avnie,adR Julkunen- R. and Tahvanainen, J. Vuorinen, J. H., Roinen, Endangered the why reasons biological Six 1991. J. D. Rohlf, sro . n .S eer.19.Gntpcdfeecsin differences Genotypic 1994. Pereira. S. J. and J., Osorio, genotype, of Effects 2001. Lindroth. L. R. and L., T. Osier, ot,B . n .B ed 95 nlsso yrdswarm hybrid a of Analysis 1985. evo- Reid. greatest B. J. world's and M., the B. Potts, as Humans 2001. R. S. Palumbi, cie,J .18.Oiiso h einlfeigabilities feeding regional the of Origins 1986. M. J. Scriber, G. K. Hoeksema, D. J. Brigham, A. C. W., M. Schwartz, ao ou nunigebyncdvlpetrt in rate development ( trout embryonic rainbow reveals in¯uencing mapping locus interval major Composite a 2001. Thorgaard. H. Sciences Aquatic and Fisheries 1414±1421. of populations. Journal salmon Canadian chinook Zealand New among patterns Africa. East Ecology in of structure Journal woodland key- and for crashes population quest the in Lubchenco, Challenges J. Castilla, 1996. BioScience C. stones. Paine. J. T. Daily, R. G. and Mills, S. L. Bond, 250 Botany of Journal Australian Labill. .A upre,K gyn n .S ree 2000. Greene. S. A. and ( honeybee Nguyen, of dissection K. Genetic Humphries, A. M. h eei ai o aito ntecnetaino phe- of concentration in the glycosides in nolic variation for basis genetic The Research est fqaigapn( aspen quaking phytochemistry season of early in variation between-year and Oecol- willow. on herbivores of ogia abundance the on effects ae ifrne.BohmclSseaisadEcology and Systematics 21 Biochemical differences. based oyeo h omnreed, common the of notype Oecologia ito 93 otpeeec n loyedifferentiation saw¯y, allozyme galling shoot and in preference Host 1993. Tiitto. Con- it. about do to Biology what servation workÐand doesn't Act Species 92 ecme . n .D rdhw r 96 Quantitative 1996. Jr. Bradshaw, D. H. and G., Newcombe, ae s fcec and ef®ciency use water Ecology Chemical of Journal pests. 1289±1313. insect phytochem- and aspen on istry defoliation and availability, nutrient Ecology and Systematics Biochemical between Science force. lutionary Heredity of Journal behavior. globulus populicola ntetgrsalwalbte¯:eooia monophagy ecological butter¯y: the swallowtail and tiger the in con- for Oecologia implications ecology. Linking function: servation 2000. ecosystem Mantgem. van to J. P. biodiversity and Mills, H. M. Lyons, of Academy National the (USA) Sciences of Proceedings America. North :535±542. ri oicnern eitnei yrdppa to poplar hybrid in resistance conferring loci trait her- Evolution to stress. resistance environmental with and associated bivory pine pinyon in zygosity :811±814. :16±22. 105 rePhysiology Tree . aii luu australis glaucus Papilio uaytsrisdonii Eucalyptus :388±396. h as fla pt aainJunlo For- of Journal Canadian spot. leaf of cause the , 71 26 :94±103. nohnhsmykiss Oncorhynchus :1943±1950. 99 46 81 5 ai sericea Salix ouu tremuloides Populus :2445±2449. :273±282. :609±620. :305±314. ur atra Euura 13 293 14 iciiainin discrimination C :871±882. :1786±1790. okf and Hook.f. 91 122 hamtsaustralis Phragmites psmellifera Apis lnlvrainadsex- and variation clonal : :474±479. Evolution . :297±305. useisi Florida. in subspecies 33 .Junlo Heredity of Journal ). :543±562. 28 :197±208. HMSG HTA TAL. ET WHITHAM G. THOMAS ih. clones. Michx.) .amygadalina E. 47 45 . foraging L.) Eucalyptus :300±308. :989±999. Septoria into , 50 27 : : hfe,M .18.Mnmmpplto ie o species for sizes population Minimum 1981. L. M. Shaffer, rsdr .K,adP .Vtue.20.Ptnilecosys- Potential 2001. Vitousek. M. P. and K., K. Treseder, Univer- process. coevolutionary The 1994. N. Arti®cial J. Thompson, 2000. Elias. R. and Wilson, S. D. edition. W., Third Biometry. Swenson, 1995. Rohlf. F. J. and symbiosis. R., Mycorrhizal R. 1997. Sokal, Read. J. D. and E., S. Smith, is keystones: and umbrellas, Flagships, 1998. D. Simberloff, Genetic 1999. Teasdale. R. and Chaparro, X. J. M., Shepherd, hta,T . .A orw n .M ot.19.Plant 1994. Potts. M. B. and Morrow, A. P. G., T. Whitham, herbivory: Chronic 1985. Mopper. S. and G., T. Whitham, Dungey, S. H. Floate, D. K. Martinsen, pests. D. for G. G., sinks T. Whitham, as zones hybrid Plant 1989. G. T. Whitham, imr .20.Atgtln ewe pcaiainand specialization between link tight A 2002. A. Widmer, adral .B,adR .Bla 91 clg n evo- and Ecology 1981. Balda. P. R. and B., S. Vanderwall, a mee,R . n .G hta.20.Cagsin Changes 2002. Whitham. G. T. and J., R. Ommeren, van in . .dsNvs .vnKme,H rnsn n T. and Franssen, H. Kammen, van A. Neves, das L. I., Vijn, ae .J 02 eeseeve feitss selection, epistasis, of view eye gene's A selec- 2002. kin J. M. selection, Wade, hard selection, Soft group of 1985. models J. the M. of Wade, review critical A 1978. J. M. selection. Wade, group of study experimental An 1977. J. M. Wade, ilo,M . n .C auk.19.Aarmu ®sh Anadromous 1995. Halupka. C. K. and F., M. a Willson, selection: natural and Adaptation 1966. C. G. Williams, adral .B 97 ipra fsnlla io pine pinon singleleaf of Dispersal 1997. B. S. Vanderwall, a a,N . .D ae n .El.19.Inheritance 1999. Elle. E. and Hare, D. J. M., N. Dam, van osrain BioScience conservation. e-ee fet fgntcvrainaogpopulations among variation genetic of of effects tem-level USA. Illinois, Chicago, Press, Chicago of sity (USA) Sciences Academy National of the of Proceedings selection. ecosystem USA. California, Francisco, San Freeman, H. W. UK. London, Press, Academic edition. Second Conservation ological passe management single-species Genetics Applied and interspeci®c 1207±1215. Theoretical an hybrid. in composition eucalypt monoterpene of mapping a a,N . n .Viln.19.Gntcvariation Genetic 1994. Vrieling. K. and M., N. Dam, van uiyo w nei amna uayt.Oecologia eucalypts. Tasmanian endemic 97 two com- herbivore of the biodiversity: munity of centers as zones hybrid pine. pinyon of expression Science sex and architecture on impacts of understanding genetic-based affect Ecology zones a structure. hybrid community Plant for 1999. tools Keim. P. biodiversity: and Potts, M. B. Science 346. pcain rnsi clg n Evolution and Ecology in Trends speciation. uino odsoaebhvo nconifer-seed-caching in behavior fu storage Zeitschrift corvids. food of lution Mammalogy neatosbtenjnpradmsltemdae by mutualism. mediated potential Oecologia to mistletoe parasitism frugivores: and avian shared juniper between interactions isln.19.Ndfcosadndlto npat.Sci- plants. in nodulation and ence factors Nod 1993. Bisseling. n pcain ora fEouinr Biology Evolutionary of Journal speciation. and Naturalist American selection. group and tion, Biology of Review Quarterly selection. Evolution nvriyPes rneo,NwJre,USA. Princeton Jersey, thought. New Princeton, evolutionary Press, current University some of critique iu monophylla Pinus ( n itiuino rcoepeoye in phenotypes trichome of distribution and Oecologia Hawaii. in ncntttv n nuil yrlzdn laodlevels alkaloid pyrrolizidine in inducible and constitutive in Heredity of Journal ygosmof®cinale Cynglossum ersdrspolymorpha Metrosideros :481±490. 260 228 244 :1764±1765. 31 130 :1089±1091. :1490±1493. :134±153. 78 :281±288. :181±191. 97 yse-ahn oet.Junlof Journal rodents. seed-caching by ) :9110±9114. rTierpsychologie Èr 83 90 126 :247±257. :220±227. 31 .Oecologia L. :266±275. :131±134. rmasi etlt gradient fertility soil a from ntelnsaeea Bi- era? landscape the in Â 80 :416±428. clg,Vl 4 o 3 No. 84, Vol. Ecology, 99 56 53 :374±378. :217±242. :101±114. auawrightii Datura 17 125 :161. 15 :61±73. :337± 99 : . March 2003 COMMUNITY GENETICS 573

as keystone species for invertebrate communities. Conser- direct genetic effects. Trends in Ecology and Evolution 13: vation Biology 9:489±497. 64±69. Wilson, D. S. 1997. Biological communities as functionally Wood, T. K., and M. C. Keese. 1990. Host plant-induced organized units. Ecology 78:2018±2024. assortative mating in Enchenopa tree hoppers. Evolution Wimp, G. M., and T. G. Whitham. 2001. Biodiversity con- 44:619±628. sequences of predation and host plant hybridization on an Woolbright, S. 2001. Genetic analyses of a synthetic popu- aphid±ant mutualism. Ecology 82:440±452. lation of hybrid cottonwoods with implications for com- Wolf, J. B., E. D. Brodie III, J. M. Cheverud, A. J. Moore, munity-level processes. Thesis. Northern Arizona Univer- and M. J. Wade. 1998. Evolutionary consequences of in- sity, Flagstaff, Arizona, USA. pca Feature Special Ecology, 84(3), 2003, pp. 574±577 ᭧ 2003 by the Ecological Society of America

WHAT CAN WE LEARN FROM COMMUNITY GENETICS?

JAMES P. C OLLINS1 Department of Biology, Arizona State University, Tempe, Arizona 85287-1501 USA

INTRODUCTION biology will prove immensely useful for exploring the intersection of genetics, ecology, and evolution. Sec- Throughout the 20th century, investigators argued ond, advances in genomics will hasten the day when that genetics should be incorporated into ecological we can document the genes in each individual that are explanations (Collins 1986). C. C. Adams (1915) sug- responding to other organisms. In a manner analogous gested very early in the century that emerging concepts to studies, especially in the 1950s, that delimited eco- in Mendelian genetics could help ecologists to explain systems by tracing the paths of radioisotopes, a map the distribution of land snails in the genus Io. Genecol- of the genetic bases of ecological interactions will de- ogy developed from 1920 to 1950, with research fo- ®ne a community. We are closing in on this possibility. cused on intraspeci®c variation that anticipated eco- ``Community genetics'' is a neologism, and although logical genetics, which developed in the 1950s and the papers in this Special Feature present new advanc- 1960s. Evolutionary ecology emerged in the 1960s, es, they also address classic questions in ecology. driven by empirical results in three areas (Collins When, how, and why should genetics and evolution be 1986): ecologically signi®cant traits like competitive incorporated into ecological explanations? Neuhauser ability had a genetic basis; some kinds of evolutionary et al. (2003) say a great deal about this question. Whi- change progressed within the time required for many tham et al. (2003) raise again the old question, ``What ecological process to reach completion; and, natural is a community?'' They also raise the more recent ques- selection operated over spatial scales suf®ciently small tion, ``Should we expect selection to act often at levels such that microevolution partially explained the dis- above the individual, including the community?'' Both tribution and abundance of populations over relatively papers led me to ask: ``What can these studies in com- short distances. By the late 1960s, ecologists were also munity genetics tell us about how we do ecology?'' becoming increasingly sensitive to the level of analysis at which natural selection was expected to operate. Fu- WHY COMMUNITY GENETICS? tuyma (1986:307) integrated these ideas in de®ning evolutionary ecology as ``the analysis of the evolu- Neuhauser et al. (2003) focus on non-equilibrial sys- tionary origin of ecological phenomena with an explicit tems and understanding population and community dy- Special Feature recognition of the distinction among, and the conse- namics over short time scales. For them, a community quences of, selection at various levels (gene, organism, is a set of interacting species that may or may not have kin group, population, or higher).'' been together for very long. Their cases have the fol- While on sabbatical at Duke University in 1982, I lowing important quality: a prediction about the out- discussed population genetics and ecology with Janis come of interactions might be false unless the analysis Antonovics as I worked on a study of the history of assumes that the interactions may lead to gene fre- the integration of ecology and evolutionary theory quency changes, hence evolution, in one or more of leading to the emergence of evolutionary ecology (Col- the species involved. Conceptually, then, community lins 1986). My efforts to understand the intellectual genetics has an important place within ecology. Neu- issues that drove the integration led to the question: To hauser et al. care most about what is happening `ìn what extent is the genetic composition of populations practice.'' Their four leading examples are from hu- in a community a function of the other species com- man-dominated systems: evolution of resistance to prising the community? Antonovics (1992) outlined a transgenic Bt crops; natural enemies and the evolution research program in community genetics that began to of resistance; population persistence and the interplay address this question. of habitat fragmentation with genetics; and domesti- The papers for this Special Feature are the most re- cation as invasion. These are important examples in cent use of genetics in ecology, but community genetics light of human-accelerated evolution (Palumbi 2001), prompts a certain optimism for two reasons. First, as especially in human-dominated urban environments Neuhauser et al. (2003) show, our ability to model these (Collins et al. 2000). Their models show nicely that interactions is improving. Advances in computational without population regulation, simple density-dependent population dynamics will alter the rate of disease resistance; i.e., predictions about population dynamics Manuscript received 8 July 2002; accepted 9 August 2002; Corresponding Editor: A. A. Agrawal. For reprints of this Special differ when genes are included or excluded. They gen- Feature, see footnote 1, p. 543. eralize this result and conclude that ecological inter- 1 E-mail: [email protected] actions among species in communities may accelerate 574 March 2003 COMMUNITY GENETICS: COMMENTARY 575 the pace of evolution. The four cases illustrate how Whitham et al. outline a more provocative program ecological theory related to communities is incomplete than Neuhauser et al., and it is one with more pitfalls. if it does not account for the fact that ecological and Whitham's team is interested in multilevel selection evolutionary processes jointly affect community dy- and community evolution. For them, a community is namics. an equilibrial assemblage of organisms whose structure Whitham et al. (2003) focus on equilibrial systems is heritable. They propose analyzing the genetic mech- composed of species where interactions have evolved anisms at the root of what they envision as the com- over a long time. The interactions have a genetic basis munity's extended phenotype, and they argue that the at the individual level, and the authors also argue (p. ``transmission of extended phenotypes from one com- 568) that, ``These interactions ultimately lead to ge- munity generation to the next is powerful evidence that netically distinct communities, whose differences are community structure is heritable.'' This is an important detectable as the among-community component of var- claim because, for them, the expectation that selection iance in individual trait expression.'' The claims that acts above the individual level means that community ``selection acts on genetic differences at the commu- evolution is likely. If true, their argument would sup- nity-level'' (italics theirs), and ``community-level se- port the now rarely held view that ecological com- lection is widespread'' are provocative, and if sup- munities are analogous to superorganisms (Odum ported, have important implications for how we con- 1969), a position that also runs counter to the expec- ceive of communities. tation of the neutral argument (Bell 2001, Hubbell Neuhauser et al. and Whitham et al. also discuss the 2001) that communities are `òpen and easily invaded'' usefulness of community genetics for developing con- (Whit®eld 2002:480). servation strategies in a rapidly changing world. Sev- At the heart of Whitham et al. is the assumption that eral recent reports add to the mounting evidence of organisms matter, natural history matters, and individ- global warming. Fitter and Fitter (2002:1689) have ual species matter. For this team, the theory on which concluded that, ``. . . large interspeci®c differences in our understanding of communities as organismal as- this response [to increasing temperature] will affect semblages rests must incorporate genetics and evolu- Feature Special both the structure of plant communities and gene ¯ow tionary biology. Many of us would agree to this point. between species as climate warms.'' As we move from But they go on to argue that communities are a complex a focus on conserving individual species to conserving network of co-evolved relationships that support se- communities and ecosystems, it will be important to lection above the individual level. Many of us would understand what we must do to retain interactions disagree here. Their view raises issues related to levels among organisms, interactions expected to have a ge- of selection that are addressed by many including Wade netic basis. (1978), Wilson (1980), and Williams (1992), as well as philosophers of science like Hull (1980), Sober WHAT ISACOMMUNITY? (1984), and Brandon (1990). Whitham et al. must iden- For Neuhauser et al. (2003), studying interspeci®c tify a community-level trait that is under selection to interactions must include genetics and the possibility distinguish selection of genes at the individual level of evolutionary change in order to predict a system's from selection for a trait at the community level. Gene future state. This raises the question, ``What is a com- frequencies can change by virtue of the life or death munity?'' Relevant here is the issue of how long a of groups, but that is not necessarily the same as se- group of species must associate if genetics and evo- lection for a group or community trait (Sober 1984). lution are to matter. Neuhauser et al. claim that the association of a group of species need only be brief, WHAT CAN THESE STUDIES IN COMMUNITY placing them in a community ecology tradition that GENETICS TELL US ABOUT HOW WE DO ECOLOGY? originates with Gleason (1917) and that found further The papers in this Special Feature are end points. expression in the 1960s when ecologists studied Dro- For Neuhauser et al., communities can be loose amal- sophila communities, diatom communities, and bird gams of species that can evolve quickly, whereas Whi- communities. At that time, `ècologists departed from tham et al. see communities as co-evolved networks of the functional de®nition of the community to a rather species that take time to develop. Throughout the 20th arbitrary concept that de®nes the community as the century, ecologists struggled to answer the question, group of organisms being studied.'' (Wilbur 1972:3). ``What is a community?'' Among other things, com- This differs from a view in which the long-term prox- munity genetics provides a basis for investigating how imity of species leads to many coevolved interactions the interactions among species might be more than just and a network of species that, in an extreme, might a series of encounters among organisms with similar express one or more traits at the community level that physiological requirements. If the interactions among can serve as a basis for selection. Whitham et al. (2003) organisms living in the same habitat are evolved re- subscribe to this latter view, which places them at the sponses to other species in that habitat, then this in- other end of a continuum relative to Neuhauser et al. terspeci®c genetic network can be the basis for de®ning (2003). a community in a manner analogous to the intraspeci®c Special Feature lnr n nedsilnr praht on science doing to approach interdisciplinary and plinary variation. genetic as of well products are as them resilience, ecology on and confer inter- stability that the like and which properties webs, to food de®ne accelerates degree that the communities actions understand many to in need species our of changing are mix actions human the multispecies which at of rate evolution The frame- systems. the a understanding offers for way genetics that work Community remain time.'' to food likely some is possible for It their uncertain. is restricts shapes further web and multispe- which systems of functions cies at the concluded affects rate evolution (1982:193) com- the ``How Pimm that web than evolve. food faster can in much populations change be and predator may versa; from position vice roles or shift prey might to species point species one the to complex; which even vary are at can but webs ®xed, not within are change interactions evolutionary such of consequenc- of Pimm the effects because es structure variation. the web food consider on genetic change not necessary did popula- the (1982) no because have evolve not needed tions qualities will a the stability in with confer present species or,to already more assemblage; organisms or stable one of habitat, kinds a the the in of on result more based or could one be- that to reasons: Pimm inaccessible species two is for (sensu habitat nature, stable the in cause be occur not should con®gura- might predict web 1982) we food The that might. tions commu- They for for genetics? matter nity explanations evolves diversity alternative community these grass- how no Do are there snakes. example, for eating them; we spe- ®ll and no to ready with ®nite, niches'' cies is `ùnoccupied variation imagine easily genetic can that clear (1978) Lewontin it However, makes niche. novel a the of ®ll evolution to subsequent phenotype support the and to species, resources new suf®cient a with niche'' ``vacant a opportunity. ESA, the sets the while variation evolution for genetic of context the setting Meeting ecology as vari- different views Annual genetic very these of recent characterized sort (2002) right Roughgarden the the At just is ation. but non qua phenotypes, sine novel the produce gene±development±environ- that of interactions product ment biologists, a evolutionary is For diversity species?'' ®lled be another could resource yet that niche by on empty ``Based an there resources. becomes, is availability, by question controlled leading the as The community envision a often of Ecologists diversity evolution. hence, and, interactions. of bases using genetic community a the it of de®nition make this a soon employ may to by possible technology, meth- microarray united genomic as species such evolving ods, same Rapidly pool. the gene of common collec- individuals a as of population tion a delimits that network genetic 576 ial,ec fteeporm ssamultidisci- a uses programs these of each Finally, imagining by views these integrate to possible is It genetics and ecology integrates genetics Community AE .COLLINS P. JAMES em fivsiaoswt klsars ag of range a across subdisciplines. skills and by disciplines with scienti®c fruitfully investigators in addressed of questions be teams larger can to ecology al. et answers evolutionary Whitham how and illustrate (2003) al. (2003) et Neuhauser The by 2003). com- al. papers larger, et down (Collins break problems to environmental plex way is a research as increasingly Collaborative seen ecologists. biologists, community population commonly and less geneticists, found research population doing of among style a in- is teams, individual it in but just work often not ecologists collaborators, Ecosystem ef- vestigators. of collective teams the of on rely fort that ecosystems) com- or to munities employ (genes strategies programs research integrated Both vertically studying for 2002). tactic a (Collins as itself communities of and in interesting is that ehue,C,D .Adw .E eme,G a,R G. R. May, G. Heimpel, E. G. Andow, A. D. C., Neuhauser, dm,C .11.Tevrain n clgcldistribu- ecological and variations The 1915. C. C. Adams, ul .L 90 niiult n eeto.Ana Review Annual selection. and Individuality 1980. L. D. Hull, biodiversity of view neutral uni®ed The 2001. P. S. Hubbell, the of development and structure The 1917. A. H. Gleason, in and changes ecology re¯ections: Rapid on Re¯ections 2002. 1986. Fitter. J. R. D. Futuyma, S. R. and D. H., Fagan, A. F. Fitter, W. Grimm, B. N. Kinzig, P. A. P., J. Collins, eotn .C 98 dpain cet® American Scienti®c Adaptation. 1978. C. R. Lewontin, lg rn B 976 upre rprto fteman- the of preparation supported uscript. 9977063 IBN grant ology oln,J . .Chn .W aisn .E ogoe and Longcore, E. J. Davidson, W. E. Cohen, N. P., J. Collins, using times: interesting in live you May 2002. P. J. Collins, nat- of use the and ecology Evolutionary 1986. P. J. Collins, Princeton environment. and Adaptation 1990. N. R. Brandon, el .20.Nurlmcoclg.Science macroecology. Neutral 2001. G. Bell, noois .19.Twr omnt eeis Pages genetics. community Toward 1992. J. Antonovics, hw n .Wgnu.20.Cmuiygntc:ex- genetics: Community 2003. Wagenius. S. and Shaw, 230. 228, 225, 222, 220, 212±218, ino h niso h genus the of snails the of tion fEooyadSystematics and Ecology of Princeton, USA. Press, Jersey, University New Princeton biogeography. and Club Botanical Torrey 463±481. the of Bulletin association. plant Biology of History 303±312. the of Journal biology. evolutionary Science plants. British in time ¯owering ecology. urban new A 2000. Borer. Scientist T. American E. and Wu, J. Hope, S nertdRsac hlegsi niomna Bi- Environmental in Challenges Research Integrated NSF aionaPes ekly aiona USA. California, Berkeley, of Press, University California essays. Conservation 1: Volume amphibians. ilnr eerhcalnefrte2s etr.Pgs43± Pages interdis- century. 52 an 21st the declines: for challenge amphibian research Global ciplinary 2003. Storfer. A. BioScience sciences. life cope the to in change programs with interdisciplinary and multidisciplinary History the of Journal Biology theory. of ecological in selection ural USA. Jersey, New Princeton, Press, University 2418. n eeis nvriyo hcg rs,Ciao Illi- Chicago, Press, Chicago USA. of nois, University evolution, genetics. ecology and pathogens: and herbivores to resistance Sciences 426±429 in .J ano dtr ttsadcnevto fU.S. of conservation and Status editor. Lannoo, J. M. in 12 19 pr II):1±92. (part .S rt n .L im,eios Plant editors. Simms, L. E. and Fritz S. R. :257±288. A L 88 CKNOWLEDGMENTS ITERATURE :416±425. 11 :311±332. C Io. ITED clg,Vl 4 o 3 No. 84, Vol. Ecology, ainlAaeyof Academy National 296 :1689±1691. 52 npress. In 293 :75±83. :2413± 239 44 19 : : : March 2003 COMMUNITY GENETICS: COMMENTARY 577

panding the synthesis of ecology and genetics. Ecology 84: Whit®eld, J. 2002. Neutrality versus the niche. Nature 417: 545±558. 480±481. Odum, E. P. 1969. The strategy of ecosystem development. Whitham, T. G., W. P.Young, G. D. Martinsen, C. A. Gehring, Science 164:262±270. J. A. Schweitzer, S. M. Shuster, G. M. Wimp, D. G. Fischer, Palumbi, S. R. 2001. Humans as the world's greatest evo- J. K. Bailey, R. L. Lindroth, S. Woolbright, and C. R. lutionary force. Science 293:1786±1790. Kuske. 2003. Community genetics: a consequence of the Pimm, S. L. 1982. Food webs. Chapman and Hall, London, extended phenotype. Ecology 84:559±573. UK. Wilbur, H. M. 1972. Competition, predation, and the structure Roughgarden, J. 2002. Evolution reduced to ecology: history of the Ambystoma±Rana sylvatica community. Ecology 53: of con¯ict and cooperation between disciplines. Ecological 3±21. Society of America 2002 Annual Meeting Abstracts: 45. Williams, G. C. 1992. Natural selection: domains, levels, and Sober, E. 1984. The nature of selection. MIT Press, Cam- challenges. Oxford University Press, Oxford, UK. bridge, Massachusetts, USA. Wilson, D. S. 1980. The natural selection of populations and Wade, M. J. 1978. A critical review of the models of group communities. Benjamin/Cummings, Menlo Park, Califor- selection. Quarterly Review of Biology 53:101±104. nia, USA.

Ecology, 84(3), 2003, pp. 577±580 ᭧ 2003 by the Ecological Society of America

COMMUNITY ECOLOGY AND THE GENETICS OF INTERACTING SPECIES

PETER J. MORIN Feature Special Department of Ecology, Evolution, and Natural Resources, 14 College Farm Road, Rutgers University, New Brunswick, New Jersey, 08901 USA

INTRODUCTION al. 2003, Whitham et al. 2003). The key questions that Neuhauser et al. (2003) and Whitham et al. (2003) I want to address in this commentary are: (1) which importantly stress that the selective forces acting on species need to be included, and (2) when does the populations are complex, nonlinear, and the result of application of community genetics improve our un- multispecies interactions peculiar to the speci®c com- derstanding of community patterns and processes? It munities where populations occur. Obviously, all nat- is also important to keep in mind that although selection ural populations are embedded in multispecies com- occurs in a community context, communities are not munities of varying complexity. Population biologists likely to be units of selection, except under exceptional can create and study single-species populations in the- circumstances (Gilpin 1975). For that reason, some oretical or laboratory settings, often with fascinating closing caveats about terminology and concepts seem and illuminating results (e.g., Lenski et al. 1991, Buck- prudent. ling et al. 2000, Kassen et al. 2000). However, natural WHICH SPECIES TO INCLUDE? populations must evolve in response to a diverse array Both Neuhauser et al. (2003) and Whitham et al. of biotic and abiotic selective pressures in the context (2003) focus on strong interactions among a limited of complex communities. This crucial point is generally set of species embedded in a larger community. This not stressed in elementary treatments of theoretical approach is similar in spirit to the idea of community population genetics (e.g., Hartl 1980). Clearly, the tra- modules that Holt (Holt et al. 1994, Holt and Lawton ditional treatment of selection pressure in simple pop- 1994) has championed as a way to make the bewil- ulation genetic models as an invariant coef®cient called dering complexity of natural communities more ana- ``s'' is a pedagogically useful, but ecologically unre- lytically tractable. Indeed, the very few empirical stud- alistic, oversimpli®cation. ies of interaction strengths that we have for natural Understanding how evolution depends explicitly on communities (Paine 1992, Raffaelli and Hall 1996) the identities, densities, and genotypes of strongly in- suggest that most species interact strongly with few teracting species in moderately complex communities others, and that interactions with remaining species are is a major challenge (Antonovics 1992, Neuhauser et weak or nonexistent. If these studies are at all representative of the broad range of communities where the Manuscript received and accepted 15 July 2002; ®nal version received 15 August 2002. Corresponding Editor: A. A. Agrawal. distribution of interaction strengths remains unmea- For reprints of this Special Feature, see footnote 1, p. 543. sured, it may be reasonable to ignore the formidable 1 E-mail: [email protected] analytical problem of treating natural selection as a Special Feature rmsuiso h eetdcnegn vlto of evolution convergent of repeated sets similar the comes of agent studies selective strong from a inter- as for competition evidence speci®c Comparable con- 1999). against (Morin defenses sumers chemical, morphological of and examples the in behavioral, spectacular comes numerous selection of of agents form as importance enemies the natural for of but evidence indirect, compelling, More 2000). nonetheless Lenski al. potent and et Levin be Bohannan 1959, 1977, can (Ratcliffe selection enemies natural natural of agents that resistance changes host rapid of inter- in examples strong other extremely from know an We as action. well relevant of as lack variation a re¯ect genetic on may of amphibians pathogens the lack to of the part resistance the case, in this increase in observable Granted, any 2000). 1999, al. observable Cunningham et rapid, and Dazsak (Dazsak of change instead evolutionary interac- extinctions ecological cause strong illus- which tions pathogens in viral situation and one trate chytrid emerging response newly in to populations amphibian of ongoing extinctions The occur. local can change before genetic extinction meaningful to any rapidly driven are strong populations so be that are sometimes interactions may ecological Some selection misleading. of agents as interactions reasonable. eminently seems 2003) al. or et (Neuhauser 2003) modules community al. limited et within (Whitham species species keystone community of the inter- genetics on of emphasis sets an Consequently, limited actions. those evo- of the consequences on focus lutionary interacting can strongly genetics few community and a species, of modules many into broken entire be down an can communities in complex interactions Instead, web. indirect food and direct of product 578 icn eetv rsueo niiul nnatural in sig- individuals impose on to populations. pressure likely all selective are ni®cant mutualists and com- enemies, petitors, natural with Interactions 1998). al. et sos ml egahcdsacs ouain foefrog one of very populations over species, even distances, that geographic show to small ponds arti®cial in conducted experiments garden'' ``common used Carolina, Fauth Fauth North USA. in occur. living amphibians of example species populations empirical for intriguing one those described has which (1998) in different very communities the by imposed from result forces that selective species different differences same genetic the of and populations to ®rst among ecological is The considers research. section these for this areas of fertile of other goal some main their suggest and The plants enemies. between al. natural et interactions Whitham on and focus (2003) (2003) al. et Neuhauser by scribed W ti ot onigotta ou ntestrongest the on focus a that out pointing worth is It ayo h xmlso omnt eeisde- genetics community of examples the of Many E DOES HEN U DRTNIGOF NDERSTANDING uoamericanus Bufo C Anolis OMMUNITY AND cmrh niln ans(Lo- faunas island in ecomorphs P ROCESSES ifrdsrknl ncom- in strikingly differed , C OMMUNITY G ENETICS ? I P POEOUR MPROVE ATTERNS EE .MORIN J. PETER hte hyrglryitrce ihacompetitor, a with on interacted depended regularly apparently they that whether ways in ability, petitive lodsrbsastaini hc h rdtr sea predatory the which in (1980) star situation Paine a component. describes genetic clear also a has which of much interactions, plant±insect geographic coevolving extensive in Cun- variation described have and (2002) Thompson ningham interactions. community-level answered. be to in begs or fascinating that a ability question is 1983) competitive Morin (see in predators geo- to differ resistance the will of range. range parts their graphic different of from portions populations southern Whether number in that species times anuran 10 the and of in range, species their anuran of two parts or northern one perhaps with interact palustris Rana ftewdl itiue ml frog small populations distributed example, widely the For of 1991). (Currie spe- richness of gradients cies distributed latitudinal widely well-known less along of species interact numbers potentially different that very there am- with species that American widespread shows North many 1991) to are Collins guide and ®eld (Conant a phibians the in to maps Reference range structure. community intercommunity in with correlated of differences are are that sort there strength this interaction that in suspect of variation I examples not. other did that many one the than prey hwddfeecsi h mat ftosubspecies two newt, of predatory impacts the of the in differences showed httepeao useista eual occurred regularly that subspecies with predator the that uinr hne ilspotls iest hna than diversity less support evo- that will predicts changes competition simple lutionary a exploitative Although of 1972). de- Cox model cycles (Wil- and communities Ricklefs taxon island 1961, on the son birds and with ants for consistent scribed is in that turnover temporal species co- show The also strengths. communities interaction assembled evolving ®xed systems with than species species from fewer support onists col- coevolving with communities model 1983). Interestingly, Roughgarden the and during (Rummel coevolve assembly of that ge- process species (no from properties or change) ®xed netic essentially with assemble species they from whether addressed on differ- depending have fundamentally compositions ent have Models as- will species communities communities. whether which in into way the semble alter changes and lutionary Gaines (e.g., unresolved. rates remains settlement 1985) pro- Roughgarden by ecological purely driven or differ- topology, cesses populations, web food predator in the ences in com- the differences on genetic important it re¯ect impacts differences locations exceptional these other Whether no munity. in have whereas to range, appears its of parts some n fterptnilpe,tdoe ftewidespread the of tadpoles prey, frog potential their of one hr r te xmlso egahcvrainin variation geographic of examples other are There eodisecnen h xett hc coevo- which to extent the concerns issue second A iatrochraceous Pisaster uoamericanus Bufo Bufo a uhsrne e aiaipc on impact capita per stronger much a had iial,KraaadMrn(1994) Morin and Kurzava Similarly, . eeteitrsigptenwas pattern interesting the Here . oohhlu viridescens Notophthalmus csa esoepeao in predator keystone a as acts clg,Vl 4 o 3 No. 84, Vol. Ecology, suarscrucifer Pseudacris ,on March 2003 COMMUNITY GENETICS: COMMENTARY 579 community assembled from non-evolving species, even their component modules are the units of natural more complex evolutionary frameworks can lead to the selection. For that reason, it seems prudent to avoid promotion of extensive diversity through networks of terminology that even indirectly implies that natural intransitive competitive interactions. selection operates on entire communities. Consequent- One system in which community genetics may in- ly, I suggest avoiding the use of the terms ``extended teract with species composition to actually maintain phenotypes'' and ``community heritability.'' Both ideas high levels of diversity is the microbial communities can be readily expressed instead as consequences of of soils (CzaÂraÂn et al. 2002). Soil systems exhibit spec- natural selection acting on individuals. Unfortunately, tacular levels of microbial diversity that have been dif- these terms recall some of the discredited ideas of Fred- ®cult to explain via traditional approaches, such as dif- erick Clements (1916), who likened the development ferences in resource utilization (e.g., Tilman 1982). of natural communities to that of a superorganism. However, if soil bacteria interact via nontransitive, There are enough fascinating consequences of natural competitive networks of the sort envisioned by CzaÂraÂn selection operating on individuals in the larger context et al. (2002) and Kerr et al. (2002), then there may be of communities that community-level selection need a major role for community genetics in maintaining not be invoked as an explanation. diversity in natural communities. In these microbial ACKNOWLEDGMENTS systems, the evolutionary dynamics of genes coding for interspeci®c toxin production, resistance, and sus- I thank Anurag Agrawal for giving me the opportunity to comment on the stimulating papers by Neuhauser et al. and ceptibility drive the spatial distibution of diversity. In Whitham et al. My musings were supported, in part, by NSF turn, both diversity and the genetics of keystone species grant 9806427. can have important effects on ecosystem functioning, as pointed out by Whitham et al. (2003). LITERATURE CITED Antonovics, J. 1992. Toward community genetics. Pages SOME CAVEATS 426±449 in R. S. Fritz and E. L. Simms, editors. Plant

resistance to herbivores and pathogens. University of Chi- Feature Special Some of the examples given by Neuhauser et al. cago Press, Chicago, Illinois, USA. (2003) and Whitham et al. (2003) focus on relatively Bohannan, B. J. M., and R. E. Lenski. 2000. Linking genetic low-diversity temperate systems in either natural or change to community evolution: insights from studies of agricultural settings. It is interesting to ask whether bacteria and bacteriophage. Ecology Letters 3:363±377. Buckling, A., R. Kassen, G. Bell, and P. B. Rainey. 2000. similar kinds of processes might operate in much more Disturbance and diversity in experimental microcosms. Na- diverse systems, especially if species in those systems ture 408:961±964. interact with a greater diversity of selective agents. Clements, F. E. 1916. Plant succession. Publication 242, Car- Novotny et al. (2002) suggest that the rarity and low negie Institution of Washington, Washington, D.C., USA. density of individual tree species in tropical forests Conant, R., and J. T. Collins. 1991. A ®eld guide to reptiles and amphibians: eastern/central North America. Houghton leads to the evolution of an insect fauna that is far more Mif¯in, Boston, Massachusetts, USA. generalized than the assemblage that one typically sees Currie, C. R., J. A. Scott, R. C. Summerbell, and D. Malloch. in temperate communities. If this is a general pattern, 1999. Fungus-growing ants use antibiotic-producing bac- the basic premise of community genetics described by teria to control garden parasites. Nature 398:701±704. Currie, D. J. 1991. Energy and large-scale patterns of animal- Neuhauser et al. (2003) and Whitham et al. (2003) may and plant-species richness. American Naturalist 137:27± not generalize well beyond low-diversity temperate 49. systems, where strong species-speci®c interactions pre- CzaÂraÂn, T. L., R. F. Hoekstra, and L. Pagie. 2002. Chemical vail. warfare between microbes promotes biodiversity. Proceed- Whitham et al. (2003) are correct in pointing out that ings of the National Academy of Sciences (USA) 99:786± 790. genetic variation in keystone species can have major Daszak, P., and A. A. Cunningham. 1999. Extinction by in- implications for community structure and ecosystem fection. Trends in Ecology and Evolution 14:279. functioning. It makes good sense to extend traditional Daszak, P., A. A. Cunningham, and A. D. Hyatt. 2000. population genetics to include the more complex in- Emerging infectious diseases of wildlifeÐthreats to biodiversity and human health. Science 287:443±449. teractions among species that doubtless occur in com- Davis, M. B. 1981. Quaternary history and the stability of munities. However, it is important not to con¯ate this forest communities. Pages 132±153 in D. C. West, H. H. useful framework with the far more controversial and Shugart, and D. B. Botkin, editors. Forest succession: con- problematic issue of selection acting on communities cepts and application. Springer-Verlag, New York, New or higher levels of ecological organization. It is worth York, USA. Fauth, J. D. 1998. Investigating geographic variation in in- pointing out that, with few known exceptions (e.g., teractions using common garden experiments. Pages 394± Currie et al. 1999), neither communities nor their dom- 415 in W. J. Resetarits, Jr. and J. Bernardo, editors. Ex- inant multispecies modules reproduce, disperse, or die perimental ecology: issues and perspectives. Oxford Uni- as units. Instead, communities seem to assemble ac- versity Press, Oxford, UK. Gaines, S., and J. Roughgarden. 1985. Larval settlement rate: cording to the individual properties of their component a leading determinant of structure in an ecological com- species (e.g., Davis 1981). This makes it dif®cult to munity of the marine intertidal zone. Proceedings of the imagine situations in which entire communities or their National Academy of Sciences (USA) 82:3707±3711. Special Feature ᭧ Ecology, 580 etr,sefont ,p 543. p. Special 1, this of footnote reprints see For Feature, Agrawal. A. A. Editor: Corresponding hta ta.20) em ob h rtclmissing critical the be 2003, to al. seems et 2003), (Neuhauser al. et feature the Whitham special in this on of elaborated papers and 1992), (Antonovics tonovics ehue,C,D .Adw .E eme,G a,R G. R. May, G. Heimpel, E. G. Andow, A. D. Science, C., Neuhauser, Blackwell ecology. Community 1999. J. P. Morin, compo- the and competition, Predation, 1983. J. P. Morin, and Queiroz, de K. Larson, A. Jackman, R. T. B., J. Losos, Resource- 1977. Chao. L. and Stewart, M. F. R., B. Levin, esi .E,M .Rs,S .Smsn n .C Tadler. C. S. and Simpson, C. S. Rose, R. M. E., R. Lenski, 03b h clgclSceyo America of Society Ecological the by 2003 hw n .Wgnu.20.Cmuiygntc:ex- genetics: Community 2003. Wagenius. S. and Shaw, USA. Massachusetts, Malden, Monographs Ecological guilds. 119±138. anuran larval of sition Sci- lizards. island of radiations ence adaptive replicated determin- in and ism Contingency 1998. Rodriguez-Schettino. bacteriophage. L. and bacteria and Naturalist with model American a studies predation: and experimental competition, growth, limited Naturalist American omnt eeis siiitdb oln n An- and Collins by initiated as genetics, Community 3 aucitrcie 1Jl 02 cetd1 uy2002; July 14 accepted 2002; July 11 received Manuscript 91 ogtr xeietleouinin evolution experimental Long-term 1991. uzv,L . n .J oi.19.Cneune of Consequences 1994. Morin. J. P. and M., L. Kurzava, Bohannan. M. J. B. and Feldman, W. M. Riley, 2000. A. M. Rainey. B., Kerr, B. P. and Bell, G. Buckling, con- A. ecological R., The Kassen, 1994. Lawton. H. J. and D., R. Holt, ot .D,J rvr n .Tla.19.Sml rules Simple 1994. Tilman. D. and Grover, J. D., R. Holt, Sinauer, genetics. population of Principles 1980. L. D. Hartl, com- predator±prey in selection Group 1975. E. M. Gilpin, coli -al [email protected] E-mail: egahcvraini h oysz fakytn pred- keystone a of size body ator, the in variation geographic real-life a Nature in rock±paper±scissors. biodiversity of promotes game dispersal Local 2002. laboratory Nature a in microcosm. productivity intermediate at peaks Diversity of Review Annual Systematics enemies. and natural Ecology shared of sequences 771. n paetcmeiin mrcnNaturalist exploitative American with competition. systems apparent in and dominance interspeci®c for USA. Massachusetts, Sunderland, Jer- New USA. Princeton, sey, Press, University Princeton munities. .Aatto n iegnedrn 00generations. 2000 during divergence and Adaptation I. . oohhlu viridescens Notophthalmus 279 43,20,p.580±582 pp. 2003, 84(3), :2115±2118. OMNT EEIS OADASYNTHESIS A TOWARD GENETICS: COMMUNITY eateto ilg,Wsigo nvriy an oi,Msor 33 USA 63130 Missouri Louis, Saint University, Washington Biology, of Department I 2 111 138 NTRODUCTION eateto olg,Uiest fFoia ansil,Foia361USA 32611 Florida Gainesville, Florida, of University Zoology, of Department 406 1 :3±24. :1315±1341. :508±512. 25 :495±520. OAHNM HS N IFN .KNIGHT M. TIFFANY AND CHASE M. JONATHAN J Oecologia . ONATHAN 418 :171±174. 99 .C M. :271±280. Escherichia HASE 144 53 :741± : 1,3 AND afel,D . n .J al 96 sesn h relative the Assessing 1996. Hall. J. S. and G., D. Raffaelli, oon,V,Y ast .E ilr .P ebe,B Bremer, B. Weiblen, P. G. Miller, E. S. Basset, Y. V., Novotny, an,R .18.Fo es ikg,itrcinstrength interaction linkage, webs: Food 1980. T. R. Paine, an,R .19.Fo e nlsstruh®l mea- ®eld through analysis web Food 1992. T. R. Paine, al ifrn rmtecretepai fmc of much of emphasis current the from fundamen- is different argument perspective tally compelling genetics They community a the species. provided that focal however, a not, of di- have selection genetic natural affect and species versity interacting how and prop- ecosystem erties), sometimes mem- (and the community other the in¯uence how of can bers showing species stories a of of diversity series genetic a et Neuhauser present and (2003) (2003) with al. al. biology et Whitham evolutionary Both and ecology. genetics linking piece motneo rpi ik nfo es ae 185±191 Pages webs. food in links trophic of importance 75. ioosiscsi rpclfrs.Nature her- forest. of tropical speci®city a cost Low in insects 2002. bivorous Drozd. P. and Cizke, L. Ecology genetics. and 545±558. ecology of synthesis the panding n omnt nrsrcue ora fAia Ecology Animal of 49 Journal infrastructure. community and ueeto e aiaitrcinsrnt.Nature strength. interaction capita per of surement acif,F .15.Terbi nAsrla ae 545±564 Pages Australia. in rabbit The 1959. N. F. Ratcliffe, isn .O 91 aueo h ao yl nteMela- the in cycle taxon the of Nature 1961. O. E. Wilson, Gehring, A. C. Martinsen, D. G. Young, P. W. G., community T. Whitham, and competition Resource 1982. D. Tilman, Geographic 2002. Cunningham. M. B. and L., J. Thompson, differences Some 1983. Roughgarden. J. and D., J. Rummel, ikes .E,adG .Cx 92 ao ylsi the in cycles Taxon 1972. Cox. W. G. and E., R. Ricklefs, in in nerto fpten n yais hpa n Hall, USA. and York, Chapman New York, dynamics. New and patterns of integration einatfua mrcnNaturalist American fauna. ant R. nesian C. the and of Woolbright, consequence Ecology S. a phenotype. genetics: Lindroth, extended Community L. 2003. R. Kuske. Fischer, Bailey, G. D. K. Wimp, M. G. J. Shuster, M. S. Schweitzer, A. J. Jer- New USA. Princeton, sey, Press, University Princeton structure. Nature selection. 417 coevolutionary of dynamics and structure anal- theoretical Oikos preliminary ysis. coevolution-structured a communities: and competitive invasion-structured between etIda vfua mrcnNaturalist American avifauna. Indian West Monographiae Netherlands. The Hague, Australia. The Junk, W. in Dr. VIII. ecology Biologicae and Biogeography :667±685. T IFFANY .A oi n .O ieilr dtr.Fo webs: Food editors. Winemiller, O. K. and Polis A. G. .Kat .L rce,adC .Crsin editors. Christian, S. C. and Crocker, L. R. Keast, A. :735±738. .K M. 41 :477±486. NIGHT 2 84 :559±573. clg,Vl 4 o 3 No. 84, Vol. Ecology, 95 :169±193. 416 106 :841±844. :195±219. 355 :73± 84 : March 2003 COMMUNITY GENETICS: COMMENTARY 581 evolutionary ecology, nor have they provided the nec- sects), will interact with completely different species essary framework for ecologists to use the community at different stages in their life cycle. genetics perspective within a synthetic approach to When a species has a large amount of genotypic questions involving many interacting species. variation in traits that play a strong role in interspeci®c In this response, we ®rst ask how community ge- interactions, then the community genetics approach, netics advances our understanding of fundamental eco- and the classi®cation of organisms by genotype rather logical questions, as well as more applied issues re- than by species, may be warranted. However, such garding the conservation of rare species, and responses guidance is not evident in the papers by Neuhaser et of species and communities to environmental change. al. (2003) and Whitham et al. (2003). For example, We then discuss reasons why empirical studies of se- Whitham et al. (2003) suggest that ecologists should lection in response to interspeci®c interactions often focus on measuring the genotypic variation in species do not connect with the theoretical studies on com- with disproportionate effects on the community/eco- munity genetics. Lastly, we suggest how empiricists system (i.e., keystone species). We would instead argue can better link their current research programs to the- that it is only necessary to measure genotypic variation oretical studies on community genetics. in keystone species when that variation directly affects its traits that are known, or suspected, to in¯uence the What does understanding community genetics do for community/ecosystem. That is, the trick is for the em- community ecology? piricist to identify those species within a community for which further classi®cation of organisms into ge- Over the past few decades, a majority of community notypes would provide a better understanding of the ecology studies have become highly reductionistic, and abundance and distribution of other species in the com- experiments focus primarily at the ®ne detail of species munity. interactions at local spatial scales. From this, many If the changes in the genetic structure of dominant community ecology studies have become mired in the or keystone species in the community have the potential complexity and intricacies of this detail, and have to affect the persistence of other interacting species (as Feature Special greatly lost the ability to provide any sort of general- suggested by Whitham et al. [2003]), then conservation ities (e.g., Lawton 1999, 2000). Community genetics efforts may need to be shifted. Speci®cally, conser- takes us one more step down the reductionistic ladder, vation genetics is almost exclusively studied at the pop- by adding genetic variation into the already complex ulation level, and focuses on the genetic variation of picture. When do we need to go down this extra step? rare species and questions involving inbreeding de- Empiricists interested in broader questions of species pression and loss of heterozygosity (Amos and Balm- diversity, distribution, and abundance will not be easily ford 2001). Such rare species are not likely to be key- convinced that studying the genetic variation within stone species within a community. Because species do species is important to their research program. At com- not occur in isolation, conservation of species may be munity and ecosystem levels of study, it is often dif- best addressed at the community level. When the con- ®cult enough to keep track of different species, much servation goals are at community and ecosystem levels, less different genotypes within species. Consider an instead of at the population level, perhaps conservation analogous type of reductionism: intraspeci®c stage (or geneticists should shift their focus to more dominant size, age) structure. It has been convincingly shown by species, as suggested by Whitham et al. (2003). many authors that intraspeci®c variation in the stage of an organism can have dramatic effects on the struc- The mismatch between theoretical and ture of a community (e.g., Werner and Gilliam 1984). empirical work For example, when prey species are vulnerable to pred- One of the best ways for community genetics to ators as juveniles, but invulnerable as adults, the nature achieve a synthetic framework is to develop a more of the entire food web can be very different than when intimate connection between theoretical and empirical prey are consistently vulnerable to predators (e.g., research. However, there is a current mismatch between Chase 1999). In these sorts of cases, then, considering the theoretical work on community genetics (e.g., the the complexity of stage structure can provide a much models described in Neuhaser et al. 2003), which ex- clearer understanding of the nature of interspeci®c in- plicitly considers the numerical responses of interact- teractions, as well as larger scale questions on the dis- ing species, and much of contemporary empirical work, tribution and abundance of organisms. However, this which often controls the density of one of the inter- does not mean that all species in a community should acting species as part of an experimental treatment. be classi®ed by stage or size, or that studies that ignore Experiments with biotic agents of selection are often stage structure are not adequate. The species within a conducted in a manner similar to those with an abiotic community that are best classi®ed by stage are obvious agent. However, mortality imposed by abiotic factors if one is looking for this. For example, species with represents a constant selective agent, whereas the mor- complex life cycles, such as those with aquatic juvenile tality imposed by biotic factors will be a function of and terrestrial adult stages (e.g., frogs and many in- the density of the interactor. In some circumstances, Special Feature et a evr ifrn hntepeitosof predictions the than experi- different the very of be conclusions may the ments Thus, to the responses treatments. numerical sometimes show the and and to allowed predators, cray®sh not the were and prey, cases, frogs, all zooplankton, In larval snails. and and midges larvae predator±prey phantom dragon¯y aquatic as of such studies systems, known better many include the the These web. of in food community responses a pheno- of numerical context prey predator on eliminate predators the types explore of that numeri- consequences studies respond empirical selective to of allowed majority than not A different cally. is very life- predator are optimal the prey the when the the and of prey, exerts phenotype its it history of that density pressure respond the selective to in allowed changes is to predator numerically a when that oretically, lifetime on herbivory dynamics. therefore of population or effects and ®tness the plants, about the little of ignore show also responses studies numerical these the that on Furthermore, selection species. of conclu- outcomes plant different predicted ultimate very the the about a is sion reach plant will herbi- the variable, of when response even ability respond, the to eliminate vores that Thus, 2000). studies Tif®n also empirical (see genotype tolerant effects less stronger a have on then can which herbivores, A of den- herbivore. sity the (2000) increases the actually genotype of al. plant response tolerant more was density et genotype the plant Chase by favored mediated by expected ex- the study For that showed constant. theoretical kept a is ample, damage herbivore those which from results in ex- different selection very cause in could respond periments Al- to possible. densities were herbivore herbivores re- lowing some numerical least which at of in rela- sponses conditions employed ®eld one natural highly tively only a whereas in manner, of simu- herbivory either controlled manipulated evolution used or the eight herbivory on 2000), lated al. review et recent (Stowe a tolerance in cited selection nine response the studies Of a ®tness. individual measure as and such of variable plants, levels clipping different by simulate herbivory a or in These herbivores experiment herbivory. of controlled density to the tolerance manipulate often plant studies on of studies evolution consider species, the interacting the of sponses role. selective inter- important the an play of will responses species density acting the eco- questions, genetic most for and logical However, year. density every reset population were crop structure sys- the the to because importance little a tem of affecting were pathogens species crop of target responses recognized numerical (1992) the be that Antonovics can example, agents For selective ignored. biotic of responses numerical 582 saohreape a ta.(02 hwd the- showed, (2002) al. et Day example, another As re- numerical of importance the of example an As OAHNM HS N IFN .KNIGHT M. TIFFANY AND CHASE M. JONATHAN rsue moe yitrpc® neatoscnbe can selective interactions gained. to interspeci®c evolutionarily by imposed responds pressures species a estimates how empirical of simu- realistic more explicit much and models, lation rates), interactions demographic natural (e.g., of the observations systems of response), aspects functional the key (e.g., on experiments term short- ex- step combining by hypothetico-deductive a example, For require traditional approach. perimental will the that from one but away way, microcosms. another in is times There generation rapid with experiments species small-scale soon (2) encom- only or which the generations, experiments many that pass long-term suggest re- (1) to numerical are: wish for lutions not allow do not we do sponses, are they ecology because evolutionary main limited the studies that empirical of argue jority we Although predictions? oretical nain pressures selective actual the ture. as well as theory, ms . n .Blfr.20.We osconservation does When 2001. Balmford. A. and W., Amos, noois .19.Twr omnt eeis Pages genetics. community Toward 1992. J. Antonovics, o a miia okb etrlne ihthe- with linked better be work empirical can How hta,T . .Yug .D atne,C .Gehring, A. C. Martinsen, D. G. Young, W. G., T. Whitham, niche ontogenetic The 1984. Gilliam. F. J. and E., E. Werner, evo- antibiosis and avoidance, tolerance, Are 2000. P. Tif®n, hs,J .19.Fo e fet fpe size-refugia: prey of effects web Food 1999. M. J. Chase, hs,J . .A ebl,adE .Sms 00 Plant 2000. Simms. L. E. and Leibold, A. M. M., J. Chase, atn .H 99 r hr eea asi clg?Oikos ecology? in laws general there Are of 1999. role H. J. The Lawton, 2002. Chase. M. J. and Abrams, A. P. T., Day, tw,K . .J aqi,C .Hcwne,adE L. E. and Hochwender, G. C. Marquis, J. R. A., K. Stowe, G. R. May, G. Heimpel, E. G. Andow, A. D. C., Neuhauser, world. changing a in ecology Community 2000. H. J. Lawton, eeismte?Heredity matter? genetics 426±449 .K aly .L idoh .Wobih,adC R. C. the and of Woolbright, consequence Ecology S. a phenotype. Lindroth, genetics: extended Community L. 2003. R. Kuske. Fischer, Bailey, G. D. Wimp, K. M. G. J. Shuster, M. S. Schweitzer, A. J. Systematics and Ecology of Review An- populations. nual size-structured in interactions species and Naturalist to plants American of herbivores? responses equivalent ecological and lutionarily aibeitrcin n lentv tbeequilibria. stable alternative Naturalist and American interactions variable Illi- Chicago, Press, Chicago USA. of nois, University genetics. evolution, ecology, and pathogens: and herbivores to resistance rdcin n vltoayipiain.Evolutionary community-level implications. webs: Ecology evolutionary food and in predictions resistance and tolerance 84 Evolution histories. diversi®cation life and prey evolution of the in predation size-speci®c im.20.Teeouinr clg ftlrneto tolerance System- and of atics Ecology of ecology Review Annual evolutionary damage. consumer The 2000. Simms. ex- Ecology genetics. genetics: and 545±558. ecology Community of synthesis 2003. the Wagenius. panding S. and Shaw, Institute, Ecology International Germany. Ecology, Oldendorf/Luhe, in Excellence :177±192. 31 :565±595. 14 in :289±314. .S rt n .L im,eios Plant editors. Simms, L. E. and Fritz S. R. L ITERATURE 154 :559±570. 87 :257±265. 84 C 56 :559±573. 155 ITED :877±887. clg,Vl 4 o 3 No. 84, Vol. Ecology, :128±138. 15 :393±425. 84 : March 2003 COMMUNITY GENETICS: COMMENTARY 583

Ecology, 84(3), 2003, pp. 583±585 ᭧ 2003 by the Ecological Society of America

COMMUNITY GENETICS AND SPECIES INTERACTIONS

MICHAEL J. WADE1 Indiana University, Department of Biology, Jordan Hall 142, 1001 E. Third Street, Bloomington, Indiana 47405-3700 USA

INTRODUCTION expression determines genotype ®tness (cf., Schlicht- Whitham et al. (2003) and Neuhauser et al. (2003) ing and Pigliucci 1998). Clearly, context extends be- advocate the ``marriage of ecology and genetics'' into yond the individual to include conspeci®cs, e.g., in kin a new ®eld of community genetics, but do so in dif- selection (Wade 1980a), and the surrounding ecological ferent ways. Whitham et al. (2003) emphasize the com- community (cf. examples in Whitham et al. 2003). With munity-shaping effect of genetic variation in keystone G ϫ E in metapopulations, different demes can ex- species, connected ecologically to other community perience different contexts, environmental and/or ge- members, whereas Neuhauser et al. (2003) emphasize netic, so that evolution can occur at different rates or strong selection in nonequilibrium, genetically subdi- different directions in each local deme (Goodnight vided communities. Both papers present compelling 2000, Wade 2001, 2002). As a result, G ϫ E and epis- evidence from different systems to illustrate that ge- tasis are fundamental to speciation and the origins of netic variation has detectable effects on species inter- biodiversity (Wade 2002). Whenever the environment actions and the composition of ecological communities. itself contains genes, as in ecological communities, The genetically variable keystone species range from context itself can evolve (Wolf et al. 2003). The stan- aspens to microbial pathogens and the community con- dard conceptual framework, which assumes not only Feature Special sequences can occur at trophic levels other than that weak selection (as per Neuhauser et al. 2003), but also of the focal species. With ``community epistasis'' contextual variation independent of genetic change in (Whitham et al. 2003), a QTL (Quantitative Trait Lo- an evolving species, must be altered. This is the foun- cus) of a keystone species may affect the phenotypes dation of Thompson's (1994) geographic mosaic hy- of other species in the community with which the key- pothesis, in which ecological communities are inte- stone interacts. Indeed, these kinds of community-level grated by the reciprocal coevolution of their member effects, if as common as Whitham et al. (2003) argue, species. The evolution of an allele depends not only will require the study of QTLs in a much broader nat- on the context that it experiences, but also on the evo- ural context than is typically considered in molecular lutionary trajectory of that context, i.e., the ecological evolutionary genetic studies, whose ``gene for'' results community. are often viewed as independent of context. Whitham et al. (2003) extend the minimum viable COEVOLUTION IN SINGLE COMMUNITIES population size (MVP) in conservation genetics to the community level as ``the minimum viable interacting If the two species mix and interact randomly with population'' (MVIP). This requires preserving key- one another, the strength and direction of selection on stone genetic diversity (even speci®c genotypes). They one species is dependent upon the mean value of the also advocate determining whether global ecological context provided by the other species. Keister et al. changes might be ampli®ed by genetic interactions be- (1984) modeled this kind of within-community recip- tween species. Like Neuhauser et al. (2003), they are rocal coevolution and noted that: (1) coevolution takes concerned with genetic subdivision and apply concepts place between traits in two species and not, strictly from multilevel selection theory like ``community her- speaking, between species; and (2) the random diver- itability'' and ``community epistasis.'' Do the examples si®cation of coevolving characters depends on the presented constitute the foundation of a new ®eld of smaller of the two effective populations sizes. The ``community genetics,'' or do they emphasize the need MVP for a particular species may not be its own size to reintroduce genetics into community ecology? but rather the smaller effective size(s) of its ecological partners. Differently put, if a keystone species is large RECIPROCAL GENETIC EFFECTS WITHIN but numerically rare, then its effective size is critical EVOLVING COMMUNITIES not only for maintaining the keystone itself, but also With gene interaction (epistasis) and genotype-by- for maintaining coevolving traits in the myriad of other environment interaction (G ϫ E), the context of gene species with which it interacts. This is a more speci®c, theoretical rationale for the MVIP proposed by Whi- Manuscript received 26 June 2002; accepted 14 July 2002; Corresponding Editor: A. A. Agrawal. For reprints of this Special tham et al. (2003). Feature, see footnote 1, p. 543. Neuhauser et al. (2003) call into question the ``time 1 E-mail: [email protected] scale argument'' that has served as a barrier between Special Feature rblu castaneum Tribolium versa. vice and demo- equilibrium genetic without 1994), achieved be cannot (Charlesworth single-spe- equilibrium In graphic theory 1970s. the life-history concep- in was cies as it becomes as permeable barrier tually disciplinary dy- the nonequilibrium and namics, selection strong With genetic in models. ecology of absence evolutionary the justi®ed of has processes gradualism Darwinian re- treating the constants; justi®ed genetic ciprocally, equivalent, has as processes members ecological species' of pace fast- relatively er The decades. for Evolution and Ecology 584 rmteGongtadCag(96 td fcom- of beetles ¯our study the of (1996) meta-communities in Craig ability comes petitive and heritability Goodnight the community existing from only of The estimate 1996). coevolving empirical Craig affecting and variance (Goodnight genetic traits among-community the the of as fraction de®ned been ``meta'' has her- or ``community itability'' meta-community, subdivided a of In communities. genetics how- the them, before consider (1994) Neuhauser ever, Thompson (2003), and ge- (2003), al. of al. et et mark Whitham the fre- coevolution. is netic species, endosymbionts, and interacting hosts of correspondence quently phylogenies which the in et 1994), between Clark Thompson (e.g., 1992, interactions al. species of taxonomy ative omnt eeto ol ev oitgaecom- integrate to among- serve function. munity where could selection 1991), community (Goodnight the at level heritable becomes community deme as a within experienced environmental a variation Across however, variation. environmental meta-community, effect as competitor average com- its the of Within experiences species interaction. each two-species munities, not average and the communities among in variance ev- the in was only genetics'' ident arising ``community outcome the competitive association; mean from in change any ®nd community. wid- the er affects diversity genetic noting species' from com- keystone that concept ecological different entire qualitatively to a munities, extended within be selection should multilevel of that species claim (2003) the al. supports extinction et study to This Whitham time species. for losing and the species) of winning the of (iden- tity outcome competitive for found heritability and together, community coexisting species both i.e., vision, types omnt eeister hudb odetermine community to the to be extends principle should this whether theory genetics community isaoe(.. ae1980 spe- Wade each (e.g., for alone cies subdivision population contrasted cally h araeo clg n vlto ilb endur- be will Evolution and ing. Ecology of marriage the ovlto oa eispiaiyo h compar- the on primarily relies today Coevolution o neatn species interacting For not did (1996) Craig and Goodnight Interestingly, Z X and C EOUININ OEVOLUTION Z Y epciey eitn h ecological the mediating respectively, , and M X .confusum T. b n ftemjrgasof goals major the of One ETACOMMUNITIES and ihcmuiysubdi- community with ) Y ihma pheno- mean with , hyspeci®- They . IHE .WADE J. MICHAEL fso, If . ie rmrl yitrcinwt h te species. of other deter- individual the an is with Let interaction ®tness by individual primarily mined that imagine interaction, context, G where a sfrspecies for es species in only selection that (species herbivores G selection. some after is, negative That or positive be may it selection, if evolve may tv ausof pos- values reciprocal, itive mean may coevolved'' ``Tightly nent. ih ecrlalnt n ogelnt nteco- example. on for the differential bee, selection pollinating in a The length and plant tongue a and of evolution length corolla be might oritrcinwt species with interaction poor imi vltoaygntcter,slcincreates selection theory, `` genetic evolutionary in rium netepeec fspecies of presence the ence anan t iia oteaogdm opnn of component among-deme the to similar dispersal it, nonrandom maintains whereas disequilibrium, munity essential. estimation empirical and spe- much prediction complicated third making a more 2003), of al. et addition Whitham the (cf. with cies sign a change interac- can across species tions diversity pairwise However, more species be meta-community. maintaining may [2002]) to al. et important Geffeney races; arms prey 03;ad()( (2) and 2003); compounds, seietin evident is tesehnigitrcin o xml,anonzero a example, favorable, For a have interaction. will ®tness-enhancing others whereas ®tness, low have tas fet eeto.Cery h tesfunction ®tness the Clearly, species selection. for affects also it variance, on oladBy 19]frba eghadhs plant host ®tness mean and species because length of Overall, bugs.) beak soapberry for in type [1992] Boyd and roll 1 qa ozr oepaieitrcin n h coef- the and interaction) emphasize ®cient, to zero to equal n osbyopsn,wy.Tecovariance, The ways. opposing, possibly and etaino eodr compounds, secondary of centration omnt disequilibrium community ®tness, ie yspecies by vided hs in those acstebtengnrto rnmsino pos- en- a association, that of itive ecology transmission community between-generation the the hances of feature any vary, zZ z stecvrac of covariance the is XY admdseslo ihrseisdmnse com- diminishes species either of dispersal Random X z s®ns needn fspecies of independent ®tness is , X ol cu rmannadmdsrbto of distribution nonrandom a from occur could eed pnteaeaelclcontext, local average the upon depends V S w ␣ X ( Z Y z stevrac of variance the is , XY ( G Y XXX hc a naogcmuiiscompo- among-communities an has which , z oaiswt trait, with covaries , atrsteitrcinefc.Tetraits The effect. interaction the captures , ) X X zZ Y XY ,eult ( to equal ), S z ϭ ϭ ih eedupon depend might nrae when increases X ( X ihngtv vltoayconsequenc- evolutionary negative with , z mn otpat (species plants host among ) Y X ( cov( X ␣ G niiul ileprec relatively a experience will individuals ␣ yaaoywt ikg disequilib- linkage with analogy By . :()( (1) ): XY niiul ayi o hyexperi- they how in vary individuals u eaievle lk predator± (like values negative but , YYXX Z z X XY X Y XY G Y )( z )( zZ Z XY X Kitre l 94 ofe al. et Wolf 1984, al. et (Keister Z X , X z ␣ )( sfvrdb eeto.Note selection. by favored is , X w ihpeoyi value, phenotypic with )( XY a ihtlrnefrsecondary for tolerance with and G [ V 1 z )( zZ z XY X Z ]) ) 'btengnsin genes between '' Z ϩ␣ z Y hw ht fma local mean if that, shows ) ϩ Y z )( Y X Y z otx-pc® ®tness Context-speci®c . X X X mn niiul,and individuals, among .If V z n,cneunl,will consequently, and, ( in X and XY ␣ X coscommunities, across , Z ih euti co- a in result might clg,Vl 4 o 3 No. 84, Vol. Ecology, hw htselection that shows ) z X X G X and )( z z zZ Z Y is: XY Z Y .Te®s term, ®rst The ). Y .(e loCar- also (See ). Y adaptively )( G szr before zero is G wihIset I (which ndifferent, in Y XY ihcon- with z (1) ) Z X X Y have , pro- , G with zZ co- XY , March 2003 COMMUNITY GENETICS: COMMENTARY 585 linkage disequilibrium (Wade and Goodnight 1991, Goodnight, C. J. 1991. Intermixing ability in two-species 1998). The comparative taxonomic approach misses communities of ¯our beetles. American Naturalist 138: 342±354. this important aspect of the evolutionary dynamic, Goodnight, C. J. 2000. Quantitative trait loci and gene in- whereas the geographic mosaic hypothesis (Thompson teraction: the quantitative genetics of metapopulations. He- 1994) is founded on it. By focusing at the species level, redity 84:589±600. the comparative studies account for adaptation between Goodnight, C. J., and D. M. Craig. 1996. The effect of co- species, but not the underlying coevolutionary dynamic existence on competitive outcome in Tribolium castaneum and T. confusum. Evolution 50:1241±1250. that causes it. The origin and maintenance of heritable Keister, A. R., R. Lande, and D. W. Schemske. 1984. Models covariation between two or more ecologically inter- of coevolution and speciation in plants and their pollinators. acting species, i.e., community disequilibria, in re- American Naturalist 124:220±243. sponse to community subdivision and within- and Neuhauser, C., D. A. Andow, G. E. Heimpel, G. May, R. G. among-community selection is a critical theoretical Shaw, and S. Wagenius. 2003. Community genetics: expanding the synthesis of ecology and genetics. Ecology 84: and empirical task for community genetics. 545±558. As per Whitham et al. (2003) and Neuhauser et al. Schlichting, C. D., and M. Pigliucci. 1998. Phenotypic evo- (2003), a large number of ecological processes, espe- lution: a reaction norm perspective. Sinauer Associates, cially those involving keystone species, affect the with- Sunderland, Massachusetts, USA. in-community mean ®tness of many species. Thus, if Thompson, J. N. 1994. The coevolutionary process. Univer- sity of Chicago Press, Chicago, Illinois, USA. the genotype of a keystone species varies among local Wade, M. J. 1980a. An experimental study of kin selection. communities, it would result in locally variable evo- Evolution 34:844±855. lution across the meta-community and, consequently, Wade, M. J. 1980b. Group selection, population growth rate, in a geographic mosaic of character states in many other and competitive ability in the ¯our beetle, Tribolium spp. species. Thus, signi®cant subdivision of one species Ecology 61:1056±1064. Wade, M. J. 2001. Epistasis, complex traits, and rates of may create the necessary genetic covariance across spe- evolution. Genetica 112:59±69. cies that makes ``community genetics'' a novel and Wade, M. J. 2002. A gene's eye view of epistasis, selection, important area of study. Some of the empirical methods and speciation. Journal of Evolutionary Biology 15:337± Feature Special for estimating community heritability and community 346. disequilibrium can be found in the multilevel selection Wade, M. J., and C. J. Goodnight. 1991. Wright's shifting balance theory: an experimental study. Science 253:1015± studies of metapopulations (e.g., Wade 1980a, Wade 1018. and McCauley 1980, Goodnight 1991, Goodnight and Wade, M. J., and C. J. Goodnight. 1998. Genetics and ad- Craig 1996, Wade and Griesemer 1998). aptation in metapopulations: when nature does many small experiments. Evolution 52:1537±1553. LITERATURE CITED Wade, M. J., and J. R. Griesemer. 1998. Populational heri- Carroll, S. P., and C. Boyd. 1992. Host race radiation in the tability: empirical studies of evolution in metapopulations. soapberry bug: natural history with the history. Evolution American Naturalist 151:135±147. 46:1052±1069. Wade, M. J., and D. E. McCauley. 1980. Group selection: Charlesworth, B. 1994. Evolution in age-structured popula- the phenotypic and genotypic differentiation of small pop- tions. Cambridge University Press, Cambridge, UK. ulations. Evolution 34:799±812. Clark, M. A., L. Baumann, M. A. Munson, P. Baumann, B. Whitham, T. G., W. Young, G. D. Martinsen, C. A. Gehring, C. Campbell, J. E. Duffus, L. S. Osborne, and N. A. Moran. J. A. Schweitzer, S. M. Shuster, G. M. Wimp, D. G. Fischer, 1992. The eubacterial endosymbionts of white¯ies (Ho- J. K. Bailey, R. L. Lindroth, S. Woolbright, and C. R. moptera: Aleyrodoidea) constitute a lineages distinct from Kuske. 2003. Community genetics: a consequence of the the endosymbionts of aphids and mealybugs. Current Mi- extended phenotype. Ecology 84:559±573. crobiology 25:119±123. Wolf, J. B., E. D. Brodie III, and M. J. Wade. 2003. Geno- Geffeney, S., E. D. Brodie, Jr., P. C. Ruben, and E. D. Brodie type±environment interaction and evolution when the en- III. 2002. Mechanisms of adaptation in a predator±prey vironment contains genes. In T. DeWitt and S. Scheiner, arms race: TTX-resistant sodium channels. Science 297: editors. Phenotypic plasticity: functional and conceptual 1336±1339. approaches. Oxford University Press, Oxford, UK. In press. Special Feature ᭧ Ecology, 586 gaa.Frrpit fti pca etr,sefont ,p. 1, footnote see Feature, A. Special A. this Editor: 543. of Corresponding reprints 2002. For August Agrawal. 15 received version id,adtegnrlcnlso ht`saigu stud- up ``scaling seed-dispersing that conclusion of general the inclusion and birds, the junipers by and mutualistic mistletoe made the of between example relationship The 1996). parasitic Travis and Miller 1994, (Woot-ton been ecology already community in has important that as effects emphasized indirect of identical concept not the if similar, to, is phenotype'' ``extended an of apparent. becomes ecosys- ecology and tem ecology, community genetics, combine to ecological population need the that species, only within appreciate exists also exists we diversity diversity Once ecological species. that between homog- and are units species enous that assume frequently ecosystem equi- and ecologists rough community that a genetic is in message on time The but librium. over (2003), maintained is al. that et evolutionary diversity Neuhauser rapid in on as not change, ecosystem is im- emphasis and with The community processes. diverse, for genetically consequences are portant species single that from non-equilibrium. high- come ly be to examples tend best which all communities, in space. the human-in¯uenced evolution over of to outcomes some relevant of communities, are mosaic points a these that in Although complex result can suf®ciently they on are interactions species conjunction evolution of in intraspeci®c effects The studied other. each be with must the in ecology and species genetics community population other that by means in¯uenced which community, spe- highly single be a can in cies evolution non-equilibrium genetic in Moreover, especially on systems. place scales, takes that time process ecological rapid a is evolution genetic important genetics.'' ``community another of yet idea add the and to worth iden- themes meaning to but major attempt will the old, We twist. tify some new a new, with or some revisiting topics, a term cover of Collins' to diversity James 1992) use (Antonovics 2003) genetics'' ``community al. et Whitham 2003, 03b h clgclSceyo America of Society Ecological the by 2003 sivkdb hta ta.(03,teconcept the (2003), al. et Whitham by invoked As show to is (2003) al. et Whitham of point main The that show to is (2003) al. et Neuhauser of focus The al. et (Neuhauser discussion under papers two These aucitrcie uy20;acpe 4Jl 02 ®nal 2002; July 14 accepted 2002; July 1 received Manuscript 3 -al [email protected] E-mail: 43,20,p.586±588 pp. 2003, 84(3), 2 OMNT EEISADCMUIYSELECTION COMMUNITY AND GENETICS COMMUNITY eateto nhoooy igatnUiest,Bnhmo,NwYr 30-00USA 13902-6000 York New Binghamton, University, Binghamton Anthropology, of Department eateto ilg,Bnhmo nvriy igatn e ok19260 USA 13902-6000 York New Binghamton, University, Binghamton Biology, of Department 1 AI LA ISNADWLIMSWENSON WILLIAM AND WILSON SLOAN DAVID D AVID S LOAN W ILSON 1,2,3 AND tan Mi 1995). selected of (Muir eggs individually than strains more strain aggression lays less group-selected that exhibits developed and a been even and has Goodnight chicken 1997), by Stevens (reviewed experi- performed and group-selection been arti®cial taken have of has ments number evolution change genetic A a and place. by frequencies, caused gene is in presumably selection response group the Again, to the selection. group-level that to and heritable responds demonstrating partially is selection, size'' ``group trait of group-level direction the shifted phe- generation The in offspring groups. the of ``group distri- of generation distribution phenotypic trait new notypic the a the of create for to end bution d one selected 37 and after size,'' them bee- ¯our individuals. measured of than groups tles, created rather (1976) Wade groups example, are For selection of units place. taken frequencies, has gene in evolution change genetic a and by caused is response selection the to Presumably, heritable. and di- partially selection is the to trait response in the a is shifts there selection, offspring of the rection phe- of the If distribution generation. offspring notypic an create distribution to phenotypic the selected of is such end trait one a and size, for body measured as is individuals of population a ge- species. any within without changes place netic take can an enough, curiously genetics making community that, now of for concept we different complex, apologies very a more With identify even past. subject complex the already in been have they often as over- isolation, so in the pro- studied interspeci®c share be cannot and do cesses intraspeci®c they that theme articles, arching two these between processes. different ecosystem very and com- composition produce for munity consequences can important species with effects, same indirect individ- the different that of showing uals in but points, mak- these in not ing lies genetics'' ``community appropriate- term and the of novelty species ness The if even pro- uniform. right, genetically ecosystem own are their on in important the effects are Similarly, cesses indirect uniform. of genetically are consequences species state- the are even if conclusion'' applicable remain that basic effects indirect our about ments con- reverse environmental or may species more dition one include to ies o,cnie iia xeieti hc the which in experiment similar a consider Now, which in experiment selection arti®cial an Consider differences the distinguishing worth is it Although W ILLIAM S WENSON 1 clg,Vl 4 o 3 No. 84, Vol. Ecology, March 2003 COMMUNITY GENETICS: COMMENTARY 587

Finally, consider an experiment in which the units species. When selection acts at the level of whole com- of selection are multispecies communities rather than munities, the community becomes analogous to an or- single-species groups. For example, Swenson et al. ganism and the constituents of the community become (2000b) created soil and aquatic microcosms inoculated analogous to genes within the organism. Populations with naturally occurring communities of microbes, of different species become roughly analogous to or- measured them after a period of time for plant biomass gans and chromosomes, interacting with each other to (in the soil microcosms) and pH (in the aquatic mi- produce the phenotypic properties that allow the whole crocosms), and selected from one end of the phenotypic community to survive the community-level selection distribution to inoculate a new set of microcosms. The process. These category shifts seem strange at ®rst, but phenotypic distribution of the offspring ``generations'' they follow directly from the concept of community- shifted in the direction of selection, demonstrating that level selection and are nicely illustrated by arti®cial ecosystem traits such as plant biomass production and selection experiments, which can be conducted with freshwater pH can respond to community-level selec- equal ease at the individual, group, and community/ tion. Just as group-level selection can be used for prac- ecosystem levels. The discussion by Whitham et al. tical purposes such as increasing egg productivity in (2003) of community-level selection, which they frame chickens, community-level selection can be used for in terms of the statistical method of contextual analysis, practical purposes such as developing microbial eco- makes the same points at a more abstract level. systems that degrade toxic compounds (Swenson et al. Even though community-level selection has been 2000a). demonstrated in the laboratory, it remains to show that Before addressing the question of whether commu- it operates in nature, requiring the expanded view of nity-level selection occurs in nature, let's examine the community genetics that we have outlined. We have response to selection in the laboratory experiments. In discussed this issue elsewhere (Wilson and Knollen- the case of individual-level and group-level selection, berg 1987, Wilson 1992, 1997, Swenson et al. 2000a, evolution at the phenotypic level is caused by a change b) and must be content to merely raise it here, alongside in gene frequencies. In the case of community-level the other meanings of the term ``community genetics'' Feature Special selection, evolution at the phenotypic level could be discussed in the target articles. caused by genetic changes in the component species, LITERATURE CITED changes in the species composition of the community, Antonovics, J. 1992. Toward community genetics. Pages or both. Goodnight (1990a, b) provides an example of 426±449 in R. S. Fritz and E. L. Simms, editors, Plant community-level selection resulting in genetic changes resistance to herbivores and pathogens: ecology, evolution, in the component species. He selected a two-species genetics. University of Chicago Press, Chicago, Illinois, ¯our beetle community for a number of traits, including USA. density of one of the species. There was a response to Goodnight, C. J. 1990a. Experimental studies of community evolution I: the response to selection at the community selection and the proximate mechanisms included level. Evolution 44:1614±1624. genes in both species that interacted with each other Goodnight, C. J. 1990b. Experimental studies of community to in¯uence the community-level phenotypic trait. In evolution II: the ecological basis of the response to com- our experiments, consider the hypothetical case in munity selection. Evolution 44:1625±1636. Goodnight, C. J., and L. Stevens. 1997. Experimental studies which the original microcosms start with a very large of group selection: what do they tell us about group se- pool of microbial species and the response to selection lection in nature? American Naturalist 150:S59±S79. is accomplished entirely by changing the frequencies Miller, T. E., and J. Travis. 1996. The evolutionary role of of the species without changing the frequencies of indirect effects in communities. Ecology 77:1329±1335. genes within species. Evolution has taken place, the Muir, W. M. 1995. Group selection for adaptation to multiple- hen cages: selection program and direct responses. Poultry communities have become ``designed'' by selection to Sciene 75:447±458. produce the selected phenotype, and the response to Neuhauser, C., D. A. Andow, G. E. Heimpel, G. May, R. G. selection has been caused by a change in the compo- Shaw, and S. Wagenius. 2003. Community genetics: ex- sition of the community. It seems like a trivial detail panding the synthesis of ecology and genetics. Ecology 84: 545±558. that the compositional change was in the proportions Swenson, W., J. Arendt, and D. S. Wilson. 2000a. Arti®cial of species rather than the proportions of genes within selection of microbial ecosystems for 3-chloroaniline bio- species. Note also that changes in species composition degradation. Environmental Microbiology 2:564±571. or population sizes within an ecosystem literally con- Swenson, W., D. S. Wilson, and R. Elias. 2000b. Arti®cial stitute changes in gene frequency at the community ecosystem selection. Proceedings of the National Academy of Sciences (USA) 97:9110±9114. level. Wade, M. J. 1976. Group selection among laboratory pop- This reasoning suggests that the concept of ``com- ulations of Tribolium. Proceedings of the National Acad- munity genetics'' (or ecosystem genetics, insofar as emy of Sciences (USA) 73:4604±4607. communities are selected on the basis of their ecosys- Whitham, T. G., W. Young, G. D. Martinsen, C. A. Gehring, J. A. Schwaitzer, S. M. Shuster, G. M. Wimp, D. G. Fischer, tem processes, as in our experiments) should be ex- J. K. Bailey, R. L. Lindroth, S. Woolbright, and C. Kuske. panded in certain contexts to include all changes in the 2003. Community genetics: a consequence of the extended composition of the community, between and within phenotype. Ecology 84:559±573. Special Feature ᭧ Ecology, 588 .Arwl o ernso hsSeilFaue e otoe1, footnote see Feature, Special this 543. of A. p. reprints Editor: For Corresponding 2002. Agrawal. August A. 15 received version ®nal opttradcnue ouain a decrease. in- population of alter can changes ®tness demographic populations resulting average The consumer the and interactions, re- competitor population biotic a to of makeup sponds genetic the When pop- selected ulation. the of ®tness average the increases se- ecosystem lection to response on evolutionary impacts de®nition, By major functioning. have can species ``key- of stone'' populations in genotypes alternative of lection and 1965). study Whittaker empirical 1926, of (Gleason by argument decades rejected after ago ecologists long most 1960), superorganism Dunbar as 1936, apparently community (Clements the the resurrects of it idea (2003), irrepressible al. espoused et is contemporary Whitham genetics community by to As an work ecology. of in relevance of issues the body diminish important to seems extensive, use genet- its ecological and and ics, population from genet- little community (2003), differs As al. ics 1980). et Janzen Neuhauser (e.g., by 429]'' applied [p. for discipline chose own would their ecol- coevolutionists reference, free that of to reciprocality, frame the needed restrictive overly was ``the [p. term from the ogists consequences that and felt evolutionary interactions He 448].'' and species of ecological genetics their of genetics study Anto- community ``the evolution. de®ned as and originally ecology or (1992) of is- discipline novics interface important new the identifying a at for make sues creating do not they for do as case (2003) subdiscipline al. strong et a Whitham such this and Neuhauser in by (2003) term papers al. the The et of me. usage trouble the Feature of Special aspects some us, with 03b h clgclSceyo America of Society Ecological the by 2003 ohesy usrb otento htsrn se- strong that notion the to subscribe essays Both stay probably will genetics'' ``community Although 1 aucitrcie 5Jn 02 cetd1 uy2002; July 14 accepted 2002; June 25 received Manuscript isn .S 97 ilgclcmuiisa functionally as communities Biological 1997. S. D. Wilson, on interactions complex of effect The 1992. S. D. Wilson, -al [email protected] E-mail: eateto ilg,Uiest fMsor,80 aua rdeRa,S.Lus isui61149 USA 63121-4499 Missouri Louis, St. Road, Bridge Natural 8001 Missouri, of University Biology, of Department raie nt.Ecology units. organized Ecol- selection. impli- of ogy with levels higher metacommunity, and biodiversity a for cations of units between variation 43,20,p.588±591 pp. 2003, 84(3), 73 EEIS VLTO,ADEOOIA COMMUNITIES ECOLOGICAL AND EVOLUTION, GENETICS, :1984±2000. 78 :2018±2024. OETE RICKLEFS E. ROBERT R OBERT .R E. fteeouino eitneo ouain opath- to populations of resistance of evolution examples the recent of several use evo- They slow responses. and lutionary selection with weak associate with they genetics evolution and population response, rapid community and selection associate strong Thus, change.'' authors as abundances scale numerical these time which same on the that on composition genetic genetics alters selection community when ``The understanding new that provides framework false. state notÐis al. does et old systems the Neuhauser nonequilibrium whereas with effects, strong deals and ®eld new the that geneticsÐ community and genetics ecological between sub- genetics. a ecological considered of be discipline suggested should genetics (1992) community Antonovics that Indeed, competition of predation. genetics the and pa- on classic (1968) Pimentel (1965) or Raven's a per), and nearly Ehrlich ``coevolution'' before word decade coevolutionary the on (coining (1958) on dynamics Mode (1960) predators, se- Sheppard by apostatic and lection on (1963) Clarke Owen of polymorphism, those mimicry as stud- through such ago ies decades tra- developed This richly was Ford (1951). dition B. Dobzhansky E. Theodosius of and genetics ecological (1971) the genetics by community nourished of are roots the that recognize (2003) desirable. and natural completely perspectives evolutionary ecological, and of genetic, integration continued the ecologists ®nd most also systems, would evo- natural and number in variation response growing genetic lutionary the assessing With for sound. techniques of be to community idea the of genetics foundation most the presented, trans- ®nd Thus would energy cyclers. ecologists as nutrient as structure, and trophic formers, commu- and of regulators diversity as asso- nity organisms functions living ecosystem with various ciated the and teractions ICKLEFS isn .S,adW .Kolneg 97 dpiein- Adaptive 1987. Knollenberg. G. W. and S., D. Wilson, otn .T 94 h aueadcneune findirect of consequences and nature The 1994. T. J. Wooton, h itnto htNuasre l 20)make (2003) al. et Neuhauser that distinction The al. et Neuhauser history. long a has integration This ietefcs h teso uyn ete ihadwith- Ecology and Evolutionary with mites. beetles phoretic their burying out of ®tness the effects: direct g n Systematics Ecol- and of Review ogy Annual communities. ecological in effects 1 25 :443±466. clg,Vl 4 o 3 No. 84, Vol. Ecology, 1 :139±159. March 2003 COMMUNITY GENETICS: COMMENTARY 589 ogens to support this dichotomy, but they might as well papers by Neuhauser et al. and Whitham et al. represent have turned to some of the earliest examples of pop- important areas of overlap between these disciplines, ulation and ecological genetics. Many of these involved involving genetic variation in consumer±resource re- strong selection and rapid evolutionary responses in lationships, especially defenses against herbivores and non-equilibrium systems, e.g., cyanide resistance in pathogens, that can in¯uence the composition of spe- scale insects (Quayle 1938, Dickson 1940) and indus- cies assemblages and ecosystem function. A number trial melanism (Kettlewell 1973), often in host±path- of related issues, which appear to me to be ripe for ogen systems, e.g., sickle-cell trait (Allison 1956), uni®cation of ecology and genetics at the community myxomatosis (Fenner and Ratcliffe 1965), and wheat level, exemplify the richness of this endeavor. In this rust (Williams 1975). Clearly, anthropogenic changes regard, the papers in this special feature make an im- in the environment can exert strong selection on pop- portant point. Although the mechanisms of evolution ulations and elicit rapid evolutionary responses that might be studied most ef®ciently by extracting evolv- might have important consequences for communities ing systems to the laboratory or to models, evolution and ecosystems (Palumbi 2001). Ecologists should pay takes place in natural systems and has consequence not close attention to these dynamics in the contexts of only for the gene pool and its phenotypic expression, such issues as emerging disease, changing trophic but also for the systems themselves. The ®eld of `èvo- structure of communities, and imbalances in the reg- lutionary ecology'' developed during the 1960s to pro- ulation of ecosystem function. This insight might have vide adaptive interpretations for patterns in nature, pri- been ignored by some ecologists, but it is not new. marily regarding life history and behavioral phenotypes Whitham et al. (2003) take the idea of community of organisms (Williams 1966, Roff 2002). No coherent, genetics a step further by arguing that the cascading parallel movement of `ècological evolution'' arose to effects of individual traits through the ecosystem (the provide a natural context for understanding the results `èxtended phenotype'') create heritable community of evolution. traits, which allow communities to respond to selection Four issues, beyond those raised by Whitham et al. as a unit. Few data support communities being inte- (2003) and Neuhauser et al. (2003), that interest me in Feature Special grated entities with discrete boundaries (i.e., units of particular are (1) the evolution of abundance and range selection). Even cases of close mutualism, such as size, (2) maintenance of genetic variation for traits that mimicry complexes and plant±pollinator relationships, have a strong in¯uence on population properties and break down as examples of tightly coupled coevolution community function, (3) formation of new species, and (Pellmyr and Thompson 1992, Thompson 1994, (4) evolutionary assembly of ecological communities. Thompson and Cunningham 2002), leaving a small Most of the variance in population density and range number of examples from obligate mutualisms and size resides at a low taxonomic level (Gaston 1998) host±parasite interactions (Hafner et al. 1994, Moran and would appear to re¯ect microevolutionary changes and Baumann 1994, Page and Hafner 1996; but see in population interactions, associated, for example, Ricklefs and Fallon 2002). Indeed, Antonovic's (1992) with genetic variation in pathogen virulence and host advocacy of ``community genetics'' appears to have resistance (Pimentel 1968, Ricklefs and Cox 1972, Van been partly a reaction against this type of community Valen 1973). Models of host±parasite interactions fea- thinking. ture the stable maintenance of variation in resistance Evolution follows upon the existence of heritable and virulence alleles with limit-cycle like dynamics in variation in ®tness. Even if communities did exist as both population size and allele frequency with time discrete units, evolution of populations within com- constants on the order of tens of generations (May and munities would weaken the heritability of community Anderson 1979, Antonovics 1992: Fig. 18.6). The lon- traits (Lewontin 1970, Wilson 1976; but see Gould ger time scales of the dynamics of range expansion and 1999, Johnson and Boerlijst 2002). Although strong contraction, on the order of 105 generations in Lesser interdependencies occur and undoubtedly guide evo- Antillean birds (Ricklefs and Bermingham 1999, lution, and although genetically determined qualities 2001), imply a potential role played by novel genetic in¯uence the array of species with which an individual variation through mutation. Understanding the dynam- interacts, these ``community'' qualities can be under- ics of this process will require detailed genetic and stood and communicated by the conventional vocab- ecological comparisons of closely related species with ulary of ecology, population genetics, and evolution. contrasting range sizes. Terms such as `èxtended phenotype'' and ``community Both Neuhauser et al. (2003) and Whitham et al. genetics'' evoke a structure that scarcely exists in na- (2003) emphasize the importance of genetic variation ture. within populations, yet the maintenance of such vari- Although I have complained (perhaps even whined ation, especially for traits potentially under strong se- a bit) about ``community genetics'' and its associated lection, has been a long-standing problem (Lewontin terms as unnecessary and potentially misleading, I also 1974). Population geneticists believe that most varia- share the belief of most ecologists in the integration of tion is maintained by spatial variation in the environ- ecology and evolution. The studies described in the ment and by frequency-dependent selection mediated Special Feature edaseiltr o hssnhss e' utgton get just Let's synthesis. it! this genetics with for injecting term special don't of we a but favor need versa, produce vice in and that ecology, much into evolution processes very and of the patterns am interpret I of to them. terms able in be will biodiversity we is it that coexistence, unlikely secondary of establishment habitat synthesize and formation, we species shift, of Until evolution 2002). and ecology Grant the and Grant et 1999, (Barraclough al. divergence ecological of involves evolution also the locally, increased be by 1993). speciation, may by diversity Latham which formed taxa and sister Ricklefs of coexistence 1987, The Saenger (Hutch- plants and of ings lineages terrestrial mangrove of by invasion the environments of sometimes case the environment, in as the dramatically, in conditions physical resources, or evolutionary food pathogens, with involve relationships species in might change of assemblages) withdrawal local and from contraction exten- also of (and process This sion assemblages. established local also of are ad- invaders members species or the localities as where other habitats up jacent from built distributions are their assemblages extend local reality, 1996). al. et In species (Morton pools ``non-evolving'' species external by from drawn invasion communities the of discrete models on of assembly largely communities, built been of has concept theory aban- unitary ago the long assemblages, doned species of structure open for- 1998). species Walker incipient and of (Avise patterns mation study to popu- efforts ecologists, continue team biologists as evolutionary I and come geneticists, 2000). will lation al. progress et Moritz that 1999, suspect May and Magurran Berlocher 1998, and Hauser (Howard understood and not is factors (Heard 1995), lineage-speci®c interac- intraspeci®c including and tions), mechanisms (ge- (genetic in- external and ternal interactions) of interspeci®c and importance com- habitat, relative ography, geographic the or but ecological ponents, include models speciation Most for- 2001). of (Hubbell the speciation diversity with the drives theories, begins process some In 2000), species. new Schluter of 1998, mation al. Lo- et 1997, (Givnish sos radiation adaptive i.e., semblages, distri- 1994). back- (Thompson the ecological the ground to against variation popu- attention genetic of a close bution within requires variation evolves, genetic lation of the how pattern and environmental resulting variation, genetic the maintains that across template How distributed 1997). are Clark interac- and individuals social (Hartl populations or within pathogens, tions predators, by primarily 590 aenmru epu ugsin ntemanuscript. the on suggestions and helpful discussion numerous insightful made provided Renner Susanne and quis, oahnCae yvaFlo,Jnta oo,BbMar- Bob Losos, Jonathan Fallon, Sylvia Chase, Jonathan ial,atog edeooit,rcgiigthe recognizing ecologists, ®eld although Finally, as- species within roles ecological of Diversi®cation A CKNOWLEDGMENTS OETE RICKLEFS E. ROBERT noois .19.Twr omnt eeis Pages genetics. community Toward Scienti®c 1992. evolution. J. and cells Antonovics, Sickle 1956. C. A. Allison, lre .A,adP .Sepr.16.Teeouinof evolution The 1960. Sheppard. M. P. and A., C. Clarke, rn,P . n .R rn.20.Aatv aito of radiation Adaptive 2002. Grant. R. B. and R., P. Grant, anr .S,P .Sda,F .Vlalna .A Spra- A. T. Villablanca, X. F. Sudman, D. P. S., M. Hafner, er,S . n .L asr 95 e vltoayin- evolutionary Key 1995. Hauser. L. D. and B., S. Heard, population of Principles 1997. Clark. G. A. and L., D. Hartl, ubl,S .20.Teuie eta hoyo biodiversity of theory neutral uni®ed The 2001. P. S. Endless Hubbell, 1998. editors. Berlocher, H. S. and J., D. Howard, arcog,T . .P olr n .H avy 1999. Harvey. H. P. and Vogler, P. A. G., T. Barraclough, phylogeo- Pleistocene 1998. Walker. D. and C., J. Avise, atn .J 98 pce-ag iedsrbtos products distributions: size Species-range 1998. Chap- J. edition. K. Gaston, Third genetics. Ecological Cam- 1971. B. Myxomatosis. E. 1965. Ford, Ratcliffe. N. F. and plants: F., and Fenner, Butter¯ies 1965. Raven. H. P. marine and in R., P. stability Ehrlich, of evolution The 1960. J. M. species. Dunbar, of origin the and Genetics 1951. T. Dobzhansky, hydrocyanic climax. to resistance of the Inheritance of 1940. C. structure R. Dickson, and Nature 1936. E. F. Clements, ins,T .19.Aatv aito n oeua sys- molecular and radiation Adaptive 1997. J. T. Givnish, ol,S .19.Glie' ute rvl:tenecessity the travels: further Gulliver's 1999. J. S. Gould, plant the of concept individualistic The 1926. A. H. Gleason, 429±449 American iir ntebutter¯y the in mimicry Oxford, Press, University Oxford UK. diversity. biological of awns®ce.Aeia Scientist American ®nches. Darwin's Ox- UK. Press, University ford, Oxford diversity. biological of lution ae fmlclreouini opcaighssadpar- Disparate and hosts 1994. Science cospeciating Nadler. in asites. evolution A. molecular S. of and rates Demastes, W. J. dling, oain n hi clgclmcaim.Hsoia Bi- Historical mechanisms. ology ecological their and novations Sunderland, Associates, Sinauer USA. Massachusetts, edition. Third genetics. n igorpy rneo nvriyPes Princeton, USA. Press, Jersey, University New Princeton biogeography. and Press, University USA. York, Oxford New York, speciation. New and Species forms. clg,eouin n genetics. and evolution, Ecology, eeln h atr htpooeseito.Pgs202± Pages speciation. promote 219 that factors the Revealing Series London, of Sciences Society Biological Royal B: the of pro- Proceedings speciation the cess. and populations avian on effects graphic Il- Chicago, Press, USA. Chicago linois, of University editors. Simms, rnatoso h oa oit fLno B London of Society Royal the of Philosophical Transactions transformation. and extinction speciation, of UK. London, Hall, and man UK. London, Press, University bridge Evolution coevolution. in study a ecosys- the of Naturalist level American the tem. at selection Natural environments. New York, USA. New York, Press, University Columbia edition. Third Hilgardia scale. red California 515±522. the in fumigation acid Ecology of Journal 163±173. eais susadapoce.Pgs1±54 Pages approaches. and issues tematics: 230. n i®ut faheacia hoyo eeto.Pages selection. of theory hierarchical 220±235 a of dif®culty and Club Botanical Torrey the 26. of Bulletin association. UK. Cambridge, Press, University adaptive and Cambridge evolution radiation. Molecular editors. Sytsma, J. K. and in 10 .E aurnadR .My dtr.Evolution editors. May, M. R. and Magurran E. A. :151±173. in in 195 .E aurnadR .My dtr.Evo- editors. May, M. R. and Magurran E. A. ln eitnet ebvrsadpathogens. and herbivores to resistance Plant :87±94. 265 L :1087±1090. TRTR CITED ITERATURE 24 :252±284. 265 aii dardanus Papilio 94 :457±463. :129±136. In 18 clg,Vl 4 o 3 No. 84, Vol. Ecology, 90 :586±608. .S rt n .L. E. and Fritz S. R. :130±139. Heredity . in .J Givnish J. T. 353 :219± 53 13 14 :7± : : March 2003 COMMUNITY GENETICS: COMMENTARY 591

Hutchings, P., and P. Saenger. 1987. Ecology of mangroves. Pellmyr, O., and J. N. Thompson. 1992. Multiple occurrences University of Queensland Press, St. Lucia, Australia. of mutualism in the yucca moth lineage. Proceedings of Janzen, D. H. 1980. When is it coevolution? Evolution 34: the National Academy of Sciences (USA) 89:2927±2929. 611±612. Pimentel, D. 1968. Population regulation and genetic feed- Johnson, C. R., and M. C. Boerlijst. 2002. Selection at the back. Science 159:1432±1437. level of the community: the importance of spatial structure. Quayle, H. J. 1938. The development of resistance in certain Trends in Ecology and Evolution 17:83±90. insects to hydrocyanic gas. Hilgardia 11:183±225. Kettlewell, H. B. D. 1973. The evolution of melanism. Ox- ford University Press, Oxford, UK. Ricklefs, R. E., and E. Bermingham. 1999. Taxon cycles in Lewontin, R. C. 1970. The units of selection. Annual Review the Lesser Antillean avifauna. Ostrich 70:49±59. of Ecology and Systematics 1:1±18. Ricklefs, R. E., and E. Bermingham. 2001. Nonequilibrium Lewontin, R. C. 1974. The genetic basis of evolutionary diversity dynamics of the Lesser Antillean avifauna. Sci- change. Columbia University Press, New York, New York, ence 294:1522±1524. USA. Ricklefs, R. E., and G. C. Cox. 1972. Taxon cycles in the Losos, J. B., T. R. Jackman, A. Larson, K. de Queiroz, and West Indian avifauna. American Naturalist 106:195±219. L. RodriguezSchettino. 1998. Contingency and determin- Ricklefs, R. E., and S. M. Fallon. 2002. Diversi®cation and ism in replicated adaptive radiations of island lizards. Sci- host switching in avian malaria parasites. Proceedings of ence 279:2115±2118. the Royal Society of London B 269:885±892. Magurran, A. E., and R. M. May, editors. 1999. Evolution Ricklefs, R. E., and R. E. Latham. 1993. Global patterns of of biological diversity. Oxford University Press, Oxford, UK. diversity in mangrove ¯oras. Pages 215±229 in R. E. Rick- May, R. M., and R. M. Anderson. 1979. Population biology lefs and D. Schluter, editors. Species diversity in ecological of infectious diseases: part II. Nature 280:455±461. communities. Historical and geographical perspectives. Mode, C. J. 1958. A mathematical model for the co-evolution University of Chicago Press, Chicago, Illinois, USA. of obligate parasites and their hosts. Evolution 12:158±165. Roff, D. A. 2002. Life history evolution. Sinauer Associates, Moran, N., and P. Baumann. 1994. Phylogenetics of cyto- Sunderland, Massachusetts, USA. plasmically inherited microorganisms of arthropods. Schluter, D. 2000. The ecology of adaptive radiation. Oxford Trends in Ecology and Evolution 9:15±20. University Press, Oxford, UK. Moritz, C., J. L. Patton, C. J. Schneider, and T. B. Smith. Thompson, J. N. 1994. The coevolutionary process. Univer- 2000. Diversi®cation of rainforest faunas: an integrated

sity of Chicago Press, Chicago, Illinois, USA. Feature Special molecular approach. Annual Review of Ecology and Sys- Thompson, J. N., and B. M. Cunningham. 2002. Geographic tematics 31:533±563. Morton, R. D., R. Law, S. L. Pimm, and J. A. Drake. 1996. structure and dynamics of coevolutionary selection. Nature On models for assembling ecological communities. Oikos 417:735±738. 75:493±499. Van Valen, L. 1973. A new evolutionary law. Evolutionary Neuhauser, C., D. A. Andow, G. E. Heimpel, G. May, R. G. Theory 1:1±30. Shaw, and S. Wagenius. 2003. Community genetics: ex- Whitham, T. G., W. Young, G. D. Martinsen, C. A. Gehring, panding the synthesis of ecology and genetics. Ecology 84: J. A. Schweitzer, S. M. Shuster, G. M. Wimp, D. G. Fischer, 545±558. J. K. Bailey, R. L. Lindroth, S. Woolbright, and C. R. Owen, D. F. 1963. Polymorphism and population density in Kuske. 2003. Community genetics: a consequence of the the African snail Limicolaria martensiana. Science 140: extended phenotype. Ecology 84:559±573. 616±617. Whittaker, R. H. 1965. Dominance and diversity in land plant Page, R. D. M., and M. S. Hafner. 1996. Molecular phylog- communities. Science 147:250±260. enies and host±parasite cospeciation: gophers and lice as a model system. Pages 255±270 in P. H. Harvey, A. J. L. Williams, G. C. 1966. Adaptation and natural selection. Brown, J. Maynard Smith, and S. Nee, editors. New uses Princeton University Press, Princeton, New Jersey, USA. for new phylogenies. Oxford University Press, Oxford, UK. Williams, P. H. 1975. Genetics of resistance in plants. Ge- Palumbi, S. R. 2001. The evolution explosion: how humans netics 79:409±419. cause rapid evolutionary change. W. W. Norton, New York, Wilson, D. S. 1976. Evolution on the level of communities. New York, USA. Science 192:1358±1360. Special Feature ᭧ Ecology, 592 .Arwl o ernso hsSeilFaue e otoe1, footnote see Feature, Special this 543. of A. p. reprints Editor: For Corresponding 2002. Agrawal. August A. 15 received version ®nal vlto ftercnttetognss h genetic The organisms. al. constituent et continuing their and of Neuhauser history evolution evolutionary by the agri- re¯ect studied the (2003) Even communities 1996). evolutionary Miller cultural continuing and and (McPeek past processes also ecological present an- but only not one processes, of result with the interact are They potentially other. that species curring structure. per- in¯uence community that ecological processes the macroevolutionary synthetic both and of micro- a understanding an time, incorporates that ®rst spective the advances parallel for These allow, ecology. phylo- in community advances genetic allowed phylogenetic methods modern of comparative development and the has pos- so genetics change sible, community made evolutionary has populations examining within for ge- population techniques and quantitative netic of Just genetics. development community the of arisen as emergence has 2002) the al. to et parallel Webb and by McPeek 1991, reviewed 1996, 1996; McLennan Losos Miller 1993, and Schluter Brooks and Ricklefs (e.g., ecology munity com- integrating in ecology. of processes munity that macroevolutionary and view, micro- larger both even an for ge- and invites ecology variation netics of merging genetic This structure. intraspeci®c community far-reaching of the rapidly consequences demonstrate and in organisms processes evolving microevolutionary into provide they insights so, doing In communities. and within abundance popu- diversity species within of patterns variation in¯uence genetic can lations how and direction evolution, and of rates how in¯uence can show interactions approaches organism These ecology genetics. community population and combining an by in context process be evolutionary ecological can the that examining from insights gained powerful the demonstrate clearly 03b h clgclSceyo America of Society Ecological the by 2003 clgclcmuiisaeasmlgso co-oc- of assemblages are communities Ecological com- in analysis phylogenetic of incorporation The (2003) al. et Whitham and (2003) al. et Neuhauser 3 aucitrcie 9Jl 02 cetd2 uy2002; July 20 accepted 2002; July 19 received Manuscript -al [email protected] E-mail: 2 eateto raimcadEouinr ilg,HradUiest,Cmrde ascuet 23 USA 02138 Massachusetts Cambridge, University, Harvard Biology, Evolutionary and Organismic of Department NERTN IR-ADMCOVLTOAYPROCESSES MACROEVOLUTIONARY AND MICRO- INTEGRATING 43,20,p.592±597 pp. 2003, 84(3), mtsna niomna eerhCne,67Cne' hr od deae,Mrln 13 USA 21037 Maryland Edgewater, Road, Wharf Contee's 647 Center, Research Environmental Smithsonian 1 I NTRODUCTION J EANNINE NCMUIYECOLOGY COMMUNITY IN .CVNE-AE N .WILCZEK A. AND CAVENDER-BARES J. C AVENDER -B ARES iygntc oadessvrlknso questions. of kinds several commu- address with to genetics combined phy- nity be how (McPeek illustrate can present we information essay, not logenetic this are In 1996). species Miller and of and 2000), types (Schluter other possess why phe- they the have properties (Manos species notypic from these of why comes 2001), collection today Donoghue the and coexisting where see broader ask we This to species events. us historical allows of perspective about timing information provide relative also the can It prop- phenotypic species. and of genetic erties can the it understand and to or- us history, help evolutionary which shared to a extent have ganisms the reveals the information logenetic of Phy- dynamics. these investigation in¯uence that allows processes historical perspective ecology phylogenetic community into a community incorporating in¯uences dynamics, variation genetic how of (2003). present-day examination al. allows et genetics Neuhauser community and While detail (2003) rich al. in et illustrated Whitham as by species, of ecolog- interactions the pro- ical for evolutionary consequences far-reaching these have of cesses outcomes phenotypic and ovlto,o h otn fpedpe lineages? preadapted of evolution, sorting adaptive the of or result coevolution, com- the in assemblages characters munity ecological and Are [1996])? Jablonski Sepkoski sensu locking,'' (``ecological munities 1,3 ifrn eoye a liaeyrsl ntespe- the in result ultimately for may selection genotypes patchy different This members. community morpho- affect other that plastic 1989) of West-Eberhard cascade (sensu changes a logical to lead ge- can small changes apparently netic interactions, pinyon these in Through variation pine. genetic of maintenance re- the composition of inforce chemical genotypes contrasting with different pine in- and pinyon how moths example, for between 2001). describe, teractions (Wade (2003) species al. affected et even are Whitham second, species the mechanisms in affected isolating present the genetic of postmating speciation before to impact lead strong species' one may exert that other may end each on species one selection enough coexisting At another. spectrum, one the in¯uence of de- they the which in to vary gree communities in together living Species o ihl necnetdaeseiswti com- within species are interconnected tightly How AND D ISTINGUISHING A MITY W L ILCZEK INEAGE A DAPTIVE 2 S ORTING E clg,Vl 4 o 3 No. 84, Vol. Ecology, OUINFROM VOLUTION Ð March 2003 COMMUNITY GENETICS: COMMENTARY 593 ciation of pinyon pine (see Sultan 1995), depending on view of community evolution and to increase our abil- the spatial distribution and strength of different selec- ity to predict the future of communities. tion pressures (McPeek 1996). At the other end of the spectrum, species may simply INTRINSIC FEATURES OF LINEAGES be ``co-present'' (Bazzaz 1996), and while coexisting What role do intrinsic and idiosyncratic features of in a predictable fashion, they may not in¯uence one lineages play in in¯uencing diversity and other com- another in an evolutionary or selective sense. Coevo- munity features? Are some communities more diverse lution only occurs when species' interactions result in because they include lineages that are inherently more reciprocal genetic change. Species that do not currently likely to diversify or are less vulnerable to extinc- show a measurable in¯uence on one another may none- tion?ÐPotential to diversify and susceptibility to ex- theless have done so in the past. Such historical inter- tinction might be related to intrinsic features, such as actions may be elucidated by a broader view that in- population structure (Losos 1996), plasticity (e.g., Sul- corporates both an analysis of genetic variation within tan 1995, Schlichting, in press), or evolvability (Wag- and among populations and phylogenetic and ecolog- ner and Altenberg 1996), or alternatively, to differences ical information about related species in other com- in the strengths of selection pressures in different pop- munities (Losos et al. 1998). ulations resulting from differences in organism inter- Perhaps the most famous example of the importance actions in those populations (McPeek 1996). of considering a phylogenetic perspective when inter- Insights about such intrinsic features of lineages us- preting species interactions across communities is that ing phylogenetic approaches may inform studies of cur- of the Anolis lizards in the Lesser Antilles (summarized rent evolutionary processes, such as those examined in Losos 1996). Throughout the Lesser Antilles, wher- by Neuhauser et al. (2003). For example, in exploring ever two species of Anolis lizards coexist on an island, rates of evolution of resistance of the European corn there is one large and one small species; lizard species borer (Ostrinia nubilalis) to Bt corn, we might gain that live alone on islands are of intermediate size. Re- perspective by knowing something about rates of di- cent analyses have shown that this common pattern is versi®cation in the corn borer lineage in comparison Feature Special actually the result of two different processes. In the to rates in other butter¯y lineages and in other corn northern islands, the large and small species appear to pest lineages. Janz et al. (2001) observed that poly- have evolved in situ and are the result of character phagous butter¯y lineages are more speciose than those displacement resulting from sympatric evolution. How- that specialize for particular plant host lineages. This ever, in the southern islands the large and small species led them to postulate that expansion of insect ranges have not experienced predictable directional selection to other hosts, possibly through evolved resistance to following introduction. Instead, it appears that ecolog- new secondary compounds (e.g., Zangerl and Beren- ical sorting has occurred such that only species with baum 1993), may be linked to diversi®cation. Infor- signi®cantly divergent morphologies were able to co- mation about whether the European corn borer and as- colonize successfully. In the absence of phylogenetic sociated pests are found within polyphagous or spe- information, it would be impossible to distinguish be- cialized lineages may allow us to predict whether these tween the two different causes for the same pattern. organisms have the evolutionary potential to escalate Using a phylogenetic approach, Janz and Nylin resistance rapidly to a new toxin. (1998) reanalyzed Ehrlich and Raven's (1964) classic Similarly, phylogenetic information could reveal hypothesis of stepwise escape and radiation between whether corn smut (Ustilago maydis) shows potential butter¯ies and their host plants. By incorporating phy- for rapid evolution of increased virulence, based on logenetic and fossil evidence, they were able to show previous diversi®cation rates. How host speci®c is corn that butter¯y diversi®cation postdated the diversi®ca- smut, and did it arise within a diverse lineage? In other tion of their plant hosts, making the hypothesis of re- words, is the evolutionary ``cold spot'' that Neuhauser ciprocal diversi®cation unlikely. Their inclusion of the et al. (2003) hypothesize characteristic of the lineage, relative timing of diversi®cation in phylogenetic anal- or is this a unique pattern found only in relation to yses of these lineages enabled them to hypothesize that anthropogenic systems? butter¯y evolution is linked to colonization of new While community genetics emphasizes intraspeci®c plant lineages rather than to cospeciation. genetic variation of interacting organisms, the pheno- While community genetics approaches can reveal typic variation in traits of organisms in response to the possible mechanisms by which organism interactions environment (plasticity or polyphenism) is also likely might lead to speciation and how genetic variation to in¯uence ecological and microevolutionary pro- within species can in¯uence community composition, cesses (e.g., Sultan 1995). Whitham et al. (2003) point phylogenetic approaches have the potential to discern out the importance of genotype ϫ environment inter- the mechanisms by which past organism interactions actions in their examples of the polyphenism in pinyon or environmental changes have in¯uenced current di- pine that results from moth attack and the decreased versity or current community assemblages. Combined, resistance of willows to herbivores after fertilization. the two approaches are likely to offer a more synthetic There is a growing recognition that plasticity may be Special Feature Ws-brad18,Jn ta.20,Schlichting, 2001, al. et Janz lineages 1989, of (West-Eberhard diversity and speciation to linked positively 594 hs nKle n arl 98 n h historical the and as 1998) (such Farrell specialist and Kelley to in generalist those from and transition (Kelley both the consider age that analyses lineage Phylogenetic to- 1998). with either trend Farrell expansion a to to niche due or rate) ward taxa, extinction specialized (high factors highly intrinsic of ephem- the nature to attributable eral be may plasticity diversity lineage of and coincidence the more Alternately, broad their are distribution. to due plasticity events vicariance of experience levels to likely high exhibit if that means species ecological to through lead lineages may of Plasticity diversi®cation 2001). Agrawal (see unclear are press o o og(alnk n eksi19) hs id faaye hudgv sisgtit h evolutionary ecological the current into the insight on us and perspective past give offer the should and them. evo- the in analyses communities within over interacted of of occurring of together have members processes assembly kinds occurred species other microevolutionary the These have which and in of 1996). about species species dynamics involved traits evidence these Sepkoski newcomer mechanisms provide phenotypic whether and as and to about about well (Jablonski processes begin information and as long can scales pool, have how data are time species to for Fossil communities regional evolutionary scales. important Most over the time [2002]). is al. interacted in lutionary it et have present Webb cases, that not [2000], these lineage species lineage Schluter on the In of see of as members caveats, 1996). composite such extant (for a all systems, (Losos ``closed'' interpretation and be together safe In Parallel to evolved a pattern. have etc. likely is to a habitats resistance, likely this such contrasting are ®re present, generate for present and are species could specialization envi- desiccation all lineages, and of contrasting where in multiple differentiation islands, have maintenance convergence undisturbed across causes species the of displacement speciation related or level ultimately, character closely constraints high and, which why architectural the in explain or by radiation, not morphological suggested adaptive do as of panel). scenarios (right tolerances, these result values trait ronmental lineages, the similar within have be co-occur niches may that ecological species conservatism because ®ltering, trait environmental Although for The important repulsion. density-dependent levels. be phylogenetic decrease phylogenetic to of higher might appear pattern to species) a species related to could the (distantly contribute one of (Janzen- species thereby in¯uence then mortality level nonsusceptible and density-dependent here, also the and mortality increased presented beyond may susceptible to example 1971]) disease be leads of the Connell to species would interspersion in 1970, related resistance species) [Janzen as closely Similarly, hypothesis related conserved, of 1976). Connell highly (distantly co-occurrence Bazzaz were the depths that resistance and rooting related hypothesize disease (Parrish (closely contrasting If coexist depth with structure. rooting to similar community Species able with habitats. species and forcing different complementary by occupy repulsion to phylogenetic co-occurrence to species) hinder lead that and processes that structure with other those community correlation or and positive exclusion 2002), a 2002). competitive al. show al. for et that et important Webb Webb traits be attraction; repulsion; those may (phenotypic (phenotypic panel, correlation ®ltering positive right negative environmental a the a for with In show important those 1999). be panel, Ackerly may left e.g., for co-occurrence the (see, coef®cients In testing correlation traits. signi®cance negative the several co-occurrence, for a and shows for similarity with respectively, trait 1B those between panels, Fig. as (labile); right well 1998). as and Donoghue distance (Bo left phylogenetic conservatism and and trait in Ackerly similarity quantifying trait for also between between method correlation see relationship one panels). The the 1999; is species. right distance Oberrath these and phylogenetic among and and left factors) ing-Gaese difference) bottom habitat (or (and 1A, distances) similarity traits Fig. value these length phenotypic trait communities; branch of identify in conservatism (using can together or and pairs found Researchers current convergence species are either species. they between of related often distances result (how distantly a phylogenetic co-occurrence as phylogenetic contain of their possibly of correlations communities and case communities, examining individual the different by In that in patterns 2000). occur region so (Webb species geographic ®ltering competition, related given environmental past closely a for right), in important (top present traits repulsion savannas) share they swamps, because (forests, presumably types community major ( three of assembly munity ϳ ncnrs,tat uha r n eicto oeac hc r ovreti hseape(i.1,lf panel), left 1B, (Fig. example this in convergent are which tolerance desiccation and ®re as such traits contrast, In in¯uence may panel), left 1B, (Fig. conserved is example hypothetical this in which plants, of depth Rooting evolutionary the of examination an by part, in explained, be can repulsion and attraction of patterns contrasting These com- understanding for analysis phylogenetic of potential the illustrates 1) Fig. in (shown scenario hypothetical A ,atog h neligmcaim o this for mechanisms underlying the although ), 0 km 100 B OX 2 .I h aeo hlgntcatato Fg A o etdarm,coeyrltdseisocrtogether, occur species related closely diagram), left top 1A, (Fig. attraction phylogenetic of case the In ). .A lutaino o ri vlto a nunetepyoeei tutr fcommunities of structure phylogenetic the in¯uence can evolution trait how of illustration An 1. r au r osre.Dt r oprmti n ulmdl r eeal required generally are models null and nonparametric are Data conserved. are value .CVNE-AE N .WILCZEK A. AND CAVENDER-BARES J. in edo ohgmoprsadagoprs n both and angiosperms, and gymnosperms both that on members feed contain families Both the beetles (Chrysomelidae). world, leaf the and in (Curculionidae) weevils families of phytophagous speciose two most within currently diversi®cation the explored spectac- (1998) most kingdom. the Farrell animal the of in one documented ever in radiations role ular a play may extinction 2002). Spencer necessary and be Sultan not may 1995, habitats (Sultan different in formation ecotypes or of populations of differentiation types, habitat genetic of and range wide a colonize successfully individuals can plastic because speciation, provide to may alternative plasticity an hand, other the On between causes. distinguish these to help may species of distribution ohtetnec odvriyadssetblt to susceptibility and diversify to tendency the Both r au r convergent are value clg,Vl 4 o 3 No. 84, Vol. Ecology, Èhn- March 2003 COMMUNITY GENETICS: COMMENTARY 595 pca Feature Special

FIG. 1. (A) Alternative scenarios for the phylogenetic structure of communities. (B) Metrics to examine convergence and conservatism in trait evolution (left panel) and to identify traits that may be important in the assembly of communities (right panel).

families were in existence before the putative appear- the current prevalence and distribution of phytopha- ance and rise of angiosperms. Every group in each of gous beetles in communities worldwide. these two families that switched from feeding on gym- Neuhauser et al. (2003) examine the effects of frag- nosperms to angiosperms underwent a pronounced ra- mentation in Midwestern prairies on persistence of pur- diation. Angiosperms tend to be heavily preyed upon ple cone ¯ower (Echinacea angustifolia) populations by herbivores, but they produce a great diversity of from two perspectives, which they distinguish as ge- defensive compounds that may allow them to escape netic (number of self-incompatibility alleles, rate of temporarily from their specialized herbivores. Given inbreeding) and ecological (dispersal of pollen and that beetles are markedly conservative in their asso- seeds, in¯uence of ®re). They suggest that cone¯ower ciations (Farrell 1998), those beetle lineages that had can serve as a model species for many prairie natives historically fed on angiosperms were most likely to because of shared life history characteristics. Although track their escaping hosts successfully and speciate in this may be true, distinct prairie lineages may have the process. Insect lineages show low extinction rates intrinsic properties with respect to inbreeding-related (Labandeira and Sepkoski 1993), and this may also characters. Different taxa show different rates of evo- contribute to the current extraordinary diversity of phy- lution and maintenance of self-incompatibility alleles tophagous beetles. Thus both ecological consequences (Lawrence 2000); these differences can be related to of conserved phenotypes (preference for angiosperm the type of incompatibility mechanism (e.g., sporo- hosts and host speci®city) and intrinsic properties of phytic vs. gametophytic), which is usually correlated beetle lineages (low extinction rates) have in¯uenced with taxonomic af®liation. An examination of these Special Feature xett hc ifrn pce a ecpbeof capable structure. be population may changing the to also species response in but different evolution model, which al. to et extent Neuhauser pa- the lineage-speci®c for only prairie rameters not other illuminate and Asteraceae may the lineages among traits of kinds 596 nesadn ri vlto n h eei under- expression? genetic trait the of from and pinnings arise evolution community trait a within understanding coexist to species of multiple allow history that mechanisms evolutionary into the insights What to species? linked are that munities lrseist cu nsmlrhbtt i environ- via habitats similar in occur to species of ilar sim- phenotypically conservation caused lineages the within that phenotypes oc- hypothesized related He closely (phylogenetic attraction). expected were than (2000) often that more Webb together species curred Borneo, tree in that communities found forest rain of interac- coexistence. biotic facilitate and that differentiation tions niche of current processes the into Meanwhile, using insight provide tested can approaches. genetics be community can modeling that and 1997) al. coexis- experimental et of Wills mechanisms (e.g., about tence used hypotheses be generate can patterns to such con- trait conservatism, of and analyses with vergence combined When 1). Fig. and Wb 00 rrplin(.Cvne-ae,D D. Bazzaz, D. A. F. Cavender-Bares, and Baum, (J. D. Ackerly, ap- repulsion or macroevolutionary 2000) attraction (Webb Such phylogenetic of patterns 1998). reveal may communities proaches al. of assembly et the (Weiher in are compet- interactions vs. ®ltering itive environmental important and how 1998), Donoghue (2) and evo- Ackerly (e.g., through how time are (1) lutionary phenotypes determine to conserved used or be convergent organisms can co-occurring relatives) evo- of their the (and traits and phenotypic 2000) of (Webb lution communities of structure context a processes. providing evolutionary in current critical for is multiple scales at phylogenetic processes indicate, evolutionary (2003) al. past et understanding Neuhauser and (2003) al. et Whi- tham as evolution, community community of understand to goals is central genetics the of one dynam- if community Moreover, for ics. important more across be variation may structure, species varia- community clearly phenotypic impact and can they tion genetic Although intraspeci®c species. that with show interact other species of one of individuals individuals which in the way in¯uence differences phenotypic dem- these and that species onstrate re- within phenotypes variation genetic in from variation sulting of em- (2003) importance al. the et phasize Whitham they environment. how their and to interact relate species how determine ganisms uscript r hr sebyrlsi h tutrn fcom- of structuring the in rules assembly there Are nhspoern td ntepyoeei structure phylogenetic the on study pioneering his In o xml,koldeaottephylogenetic the about knowledge example, For T RAIT mn ebr facmuiy(e o 1 Box (see community a of members among ) E S OUINAND VOLUTION TRUCTURING C A OMMUNITIES Tepeoye for- of phenotypes ÐThe SSEMBLY .CVNE-AE N .WILCZEK A. AND CAVENDER-BARES J. nulse man- unpublished R LSIN ULES vleadt lo st rdc hr hyaeheaded. are communities they where how predict to about us allow is insights to and ecology new evolve bring in to approaches likely phylogenetic and community genetics of interac- merging com- organism Such communities. and for within tions organisms consequences of have assembly munity that processes macroevolutionary for pro- explanations to microevolutionary begin the can vide on we traits approaches), of (phylogenetic evolution other long-term the and (community found hand genetics), be one the can on structure, links genetic over between If ¯exibility scales. their as time well behavior macroevolutionary as genetic traits, both of of properties study and al. the et Whitham allowing as out, an taxa, point Altenberg for and traits identi®ed of and been number have increasing (Wagner loci trait Important evolvability may 1996). traits certain greater of structure 2001). have genetic Shaw the and addition, Etterson In (e.g., more and labile constrained less evolutionarily linkage be to genetic likely by not are (either are pleiotropy) or traits and other loci to fewer linked by closely controlled are hypothesis that the traits including that possibilities, of There number evolution. a through are convergent or conserved are how communities. of within maintained understanding is our This diversity species. improve that of may assemblage traits information and and inferences sorting processes make the evolutionary in¯uence to past us the allow about ecology to- community approaches been phylogenetic to have 2002), al. they et long Webb (see how gether and lineages within species about information and suf®cient competition With past differentiation? of result the traits 1999). convergent traits (Ackerly Are ®ltering which environmental and for pattern, critical a are such a indeed for is explanation ®ltering likely evolution, environmental trait conver- whether correlated reveal of or can patterns as conservatism well their as of gence, and Examination than traits often attraction). phenotypic less niches (phylogenetic same the monocots expected occupy and to the found eudicots were At scale, repulsion scales. phylogenetic and phylogenetic (2001) attraction broadest different at al. both but et of emerged, patterns Silvertown com- that meadow Britain, on found Great study in related a munities In ®ltering. mental cel,D .19.Pyoeyadtecmaaiemethod comparative the and Phylogeny 1999. D. D. Ackerly, ie yCmWb n ihe Donoghue. Michael Synthesis) and orga- Webb Ecology and Cam Community by Analysis and nized Phylogenetics Ecological on for workshop NCEAS Center the in involved (National those Cam and Baum, Donoghue, David Michael Webb, Ackerly, David including colleagues with ial,w a r oeaiewypriua traits particular why examine to try can we Finally, cel,D . n .J oohe 98 efsz,sapling size, Leaf 1998. Donoghue. J. M. and D., D. Ackerly, npatfntoa clg.Pgs391±413 Pages ecology. functional plant in h huhsi hspprr¯c ueosinteractions numerous re¯ect paper this in thoughts The loer,adCre' ue:pyoeyadcorrelated and phylogeny rules: Corner's and allometry, UK. plant Oxford, Physiological editors. Scienti®c, Barker, Blackwell G. ecology. M. and Scholes, D. A L CKNOWLEDGMENTS ITERATURE C ITED clg,Vl 4 o 3 No. 84, Vol. Ecology, in .Pes J. Press, M. March 2003 COMMUNITY GENETICS: COMMENTARY 597

evolution in maples (Acer). American Naturalist 152:767± McPeek, M. A., and T. E. Miller. 1996. Evolutionary biology 791. and community ecology. Ecology 77:1319±1320. Agrawal, A. 2001. Phenotypic plasticity in the interactions Neuhauser, C., D. A. Andow, G. E. Heimpl, G. May, R. G. and evolution of species. Science 294:321±326. Shaw, and S. Wagenius. 2003. Community genetics: ex- Bazzaz, F. A. 1996. Plants in changing environments: linking panding the synthesis of ecology and genetics. Ecology 84: physiological, population, and community ecology. Cam- 545±558. bridge University Press, Cambridge, UK. Parrish, J. A. D., and F. A. Bazzaz. 1976. Niche separation BoÈhning-Gaese, K., and R. Oberrath. 1999. Phylogenetic ef- in roots of successional plants. Ecology 57:1281±1288. fects on morphological, life-history, behavioural and eco- Ricklefs, R. E., and D. Schluter. 1993. Species diversity in logical traits of birds. Evolutionary Ecology Research 1: ecological communities: historical and geographical per- 347±364. spectives. University of Chicago Press, Chicago, Illinois, Brooks, D. R., and D. A. McLennan. 1991. Phylogeny, ecol- USA. ogy and behavior. University of Chicago Press, Chicago, Schlichting, C. D. In press. The role of phenotypic plasticity Illinois, USA. in diversi®cation. In T. J. DeWitt and S. M. Scheiner, ed- Connell, J. H. 1971. On the role of natural enemies in pre- itors. Phenotypic plasticity: functional and conceptual ap- venting competitive exclusion in some marine animals and proaches. Oxford University Press, Oxford, UK. in rain forest trees. Pages 298±310 in P. J. den Boer and Schluter, D. 2000. Ecological character displacement in adap- G. R. Gradwell, editors. Dynamics of numbers in popu- tive radiation. American Naturalist 156:S4±S16. lations. Center for Agricultural Publishing and Documen- Silvertown, J., M. Dodd, and D. Gowing. 2001. Phylogeny tation, PUDOC, Wagenigen, The Netherlands. and the niche structure of meadow plant communities. Jour- Ehrlich, P. R., and P. H. Raven. 1964. Butter¯ies and plants: nal of Ecology 89:428±435. a study in coevolution. Evolution 18:586±608. Sultan, S. E. 1995. Phenotypic plasticity and plant adaptation. Etterson, J., and R. Shaw. 2001. Constraint to adaptive evo- Acta Botanica Neerlandica 44:363±383. lution in response to global warming. Science 294:151± Sultan, S. E., and H. G. Spencer. 2002. Metapopulation struc- 154. ture favors plasticity over local adaptation. American Nat- Farrell, B. D. 1998. ``Inordinate Fondness'' explained: why uralist 160:271±283. are there so many beetles? Science 281:555±559. Wade, M. J. 2001. Infectious speciation. Nature 409:675± Jablonski, D., and J. J. Sepkoski. 1996. Paleobiology, com- 677. munity ecology, and scales of ecological pattern. Ecology Wagner, G. P., and L. Altenberg. 1996. Complex adaptations 77:1367±1378. and the evolution of evolvability. Evolution 50:967±976. Feature Special Janz, N., K. Nyblom, and S. Nylin. 2001. Evolutionary dy- Webb, C. 2000. Exploring the phylogenetic structure of eco- namics of host-plant specialization: a case study of the tribe logical communities: an example for rain forest trees. Nymphalini. Evolution 55:783±796. American Naturalist 156:145±155. Janz, N., and S. Nylin. 1998. Butter¯ies and plants: a phy- Webb, C. O., D. D. Ackerly, M. A. McPeek, and M. J. Don- logenetic study. Evolution 52:486±502. Janzen, D. 1970. Herbivores and the numbers of tree species oghue. 2002. Phylogenies and community ecology. Annual in tropical forests. American Naturalist 104:501±528. Review of Ecology and Systematics 33:475±505. Kelley, S. T., and B. D. Farrell. 1998. Is specialization a dead Weiher, E., G. Clarke, and P. Keddy. 1998. Community as- end? The phylogeny of host use in Dendroctonus bard bee- sembly rules, morphological dispersion, and the coexis- tles (Scolytidae). Evolution 52:1731±1743. tence of plant species. Oikos 81:309±322. Labandeira, C. C., and J. J. Sepkoski. 1993. Insect diversity West-Eberhard, M. J. 1989. Phenotypic plasticity and the in the fossil record. Science 261:310±315. origins of diversity. Annual Review of Ecology and Sys- Lawrence, M. J. 2000. Population genetics of homomorphic tematics 20:249±278. self-incompatibility polymorphisms in ¯owering plants. Whitham, T. G., W. P.Young, G. D. Mortinsen, C. A. Gehring, Annals of Botany 85:221±226. J. A. Schweitzer, S. M. Shuster, G. M. Wimp, D. G. Fischer, Losos, J. B. 1996. Phylogenetic perspectives on community J. K. Bailey, R. L. Lindroth, S. Woolbright, and C. R. ecology. Ecology 77:1344±1354. Kuske. 2003. Community genetics: a consequence of the Losos, J., T. Jackman, A. Larson, K. DeQueiroz, and L. Ro- extended phenotype. Ecology 84:559±573. griguez-Schettino. 1998. Contingency and determinism in Wills, C., R. Condit, R. B. Foster, and S. P. Hubbell. 1997. replicated adaptive radiations of island lizards. Science Strong density- and diversity-related effects help to main- 279:2115±2118. tain tree species diversity in a neotropical forest. Proceed- Manos, P. S., and M. J. Donoghue. 2001. Progress in northern ings of the National Academy of Sciences (USA) 94:1252± hemisphere phytogeography: an introduction. International 1257. Journal of Plant Sciences 162:S1±S2. Zangerl, A. R., and M. R. Berenbaum. 1993. Plant chemistry, McPeek, M. A. 1996. Linking local species interactions to insect adaptations to plant chemistry, and host plant uti- rates of speciation in communities. Ecology 77:1355±1366. lization patterns. Ecology 74:47±54. Special Feature ᭧ Ecology, 598 ue e otoe1 .543. Fea- Special p. this 1, of footnote reprints see For Agrawal. ture, A. A. Editor: sponding etrdi hsiseof papers issue two the this on in the commentary featured decline a to provide excuse to an invitation ®nd couldn't enquiry, unfortunately scienti®c for I area rich and important an be may oosadcmiain nprdcnussand conquests inspired new combinations their symbols but were and sparse, those castle was In colors and content days. cross, syntactical olden of whose chevron, insignia the or lik- ¯ags days, be the can that often to phrases also ened and are words They new politics. by and instantiated hype, uncomfortable reality, an of by mix heralded always nearly are clear conceptual shifts Additionally, than discovery. rather new is known, for issue directions the the lead, of directions: to research reinterpretation likely obvious often are at harder ideas jump far to new is or these it where conceptual, pinpoint is to How- advance level. the ecosystem infer when to an ever, and at ratios processes isotope physiological stable measure spec- opened mass to of that application trometry the ecology is in directions new advance many up technical a ex- of Another eco- explicit. ample spatially previous in- made when huge are happens'' models a logical ``what fuelled seeing has in that terest computers access desktop ready the fast is to example algo- good almost a ecology, and In clear-cut rithmic. directions often the are DNA opportunities interference), (e.g., and RNA innovations technical PCR, of sequencing, result a and comes as excitement exciting about the in When invigorated ways. unpredictable and often science refreshed ad- of be feasibly still areas be can most can however, that Fortunately, the questions dressed. exceed the to of starts importance correctness technical im- of the portance and wilted, ini- have answered, dichotomies been contentious have tially pressing were once that Ques- tions disillusionment. and decline of periods including per- should issue. I latter useful! the is with really start it of haps in whether issue unifying and and the idea novel on anything this comment really to is there me part, whether asked his also on editor skepticism Perhaps healthy the theme! some unifying re¯ecting a part in as genetics community of 03b h clgclSceyo America of Society Ecological the by 2003 aigpstdteie htcmuiygenetics community that idea the posited Having 1 aucitrcie n cetd1 uut20.Corre- 2002. August 12 accepted and received Manuscript l cet® icpie aeterondynamics, own their have disciplines scienti®c All -al [email protected] E-mail: 43,20,p.598±601 pp. 2003, 84(3), ilg eatet nvriyo igna hrotsil,Vrii 20 USA 22904 Virginia Charlottesville, Virginia, of University Department, Biology Ecology OADCMUIYGENOMICS? COMMUNITY TOWARD htueteconcept the use that AI ANTONOVICS JANIS J ANIS A NTONOVICS ilg.Iwspriual tukb h on htge- that point the applied by for struck particularly relevance obser- was direct their I have biology. that out results point and papers vations featured that the relevant very of is both it context, this Sci- In National Foundation. the ence the and both Agriculature of from Department support U.S. received has weed and interactions, communities) crop±pathogen plant± interactions, (e.g., inter- levels insect many scientists at in pure interactions species founded and in ested applied Genetics integrated Community has 1994 for Min- the Center Minnesota, of nesota University the successes. at tangible example, very For some useful. been is have it the there whether whether (remember Already by by apposite more less or but biologists), dictated new, structural accurate, be is will discipline. label term our the the of of estab- context use become sociopolitical The is and the not, question in or correct accepted lished be The will question. or it it wrong whether need the we ``whether de- is think not'' job I that follow! hope will scriptions we Certainly genetics? community community and of genetics tenet addresses ecology. second directly the creed Indeed, the 2001). Lenski 1991, En- ecological 1976, dler Antonovics combining also (see of worldviews genetic challenges and the a of as provocative 1) statement hopefully (Table but Creed tongue-in-cheek, Geneticist's handed somewhat of have Ecological start I yr, the the 25 at past out the example, over For courses graduate ways. my many many in for and sporadically discussed years Whitham been 2003, have 2003) al. al. et et (Neuhauser these in papers raised are featured that two issues that The idea ¯ag. the own in its new deserves anything is there whether question usual! mea- large as in sure on carried and biologists'' ``molecular themselves as renamed simply sounding biochemists old-fashioned most discipline, whereas the their for to descriptor a resort as biology'' ``structural to been seems recourse have only to Their fashions. trumped semantic the solidly these at so by were action they and because folding level) biologists molecular protein molecular ``real'' study al- for actually have (who sorry I most biology. molecular felt of ways growth the in overtly sym- selfsame these bols. of incarnations previous trumped od ene e aeo e icpieof discipline new a or name new a need we do So certainly can we genetics, community to regard With most hype and reality of mixture this seen have We 1 clg,Vl 4 o 3 No. 84, Vol. Ecology, March 2003 COMMUNITY GENETICS: COMMENTARY 599

TABLE 1. The ecological geneticist's creed.

Creed Explanation Explaining the abundance and distribution of organisms is The ecological amplitude of a species both within and a genetic problem. among communities has a genetic component. The forces maintaining species diversity and genetic diver- An understanding of community structure will come from sity are similar. considering how these kinds of diversity interact. Adaptation is a dynamic process, operationally de®nable, Fitness and the contribution of phenotypes to ®tness can and not just an emotional matching of the character to be measured in terms of the mortality and fecundity of the environment. individuals within populations. Environmental change will be accompanied by changes in Genetic response is likely to result in compensatory chang- both genetic composition and changes in numerical dy- es in ®tness and life-history components. namics. The distinction between `ècological time'' and `èvolu- Changes of both kinds may be on any time scale: in prin- tionary time'' is arti®cial and misleading. ciple, evolutionary and ecological changes are simulta- neous. The genetic quality of offspring is as important as the Sexual systems are concerned with regulating the genetic quantity. quality of offspring. The view that there is always an `èvolutionary play'' The `ècological play'' often happens in the `èvolutionary within an `ècological theater'' is arti®cial and mislead- theater.'' Selection at the genic or cellular levels may ing. have phenotypic effects with enormous ecological consequences. Genetic events may drive ecology, rather than vice versa. Speciation is an ongoing and commonplace process, occur- It is only deemed to be rare by taxonomists, and the use of ring constantly and persistently around us. Latin binomials by ecologists is at best a crude approximation. Environments are most appropriately de®ned by the ecolo- We can recognize three types of environments: external, gy and genetics of the organisms themselves, and only ecological, and selective. Their measurement and inter- indirectly by environmental measurements. pretation have important consequences for population and evolutionary dynamics.

A population to an ecologist is not the same as it is to a Understanding the contrasting way in which the term is Feature Special geneticist. used is essential for unifying ecology and genetics.

netic variation has impacts on communities that go well This interesting idea was presaged many years ago by beyond the species in which it is being measured. It is the work of Maddox and Root (1990), who showed that therefore likely that genetic variation is probably being clones of goldenrod plants could be characterized by quanti®ed (and certainly conceptualized) inappropri- their herbivores and by the genetic correlations among ately in conservation biology. It points out that con- the herbivore abundances. However, I was still left un- servation biologists must look beyond population ge- clear about whose phenotype was actually being ex- netics and perhaps more to community genetics in their tended. If genetic variation per se is the cause of new thinking about diversity. phenotypes at the community level, then is it the phe- From an academic perspective, the featured papers notype of the population that is being extended? How illustrate that the insignia of community genetics pro- the heritability of a population propertyÐas opposed claims that numerous questions remain unanswered to the heritability of, say, disease resistanceÐwould with regard to the role of genetic variation in the func- be estimated needs to be ¯eshed out. Although there tioning and composition of communities and ecosys- is no doubt that ®tness effects of genes can interact via tems. Both papers point out that we need new levels indirect community interactions, it may be premature of interpretation and new laws that scale to the level to transfer genetic terms to a community context with- of the community rather than to the level of the single- out the same rigor that has accompanied genetic think- species population. Neuhauser et al. (2003) contrast the ing on gene interactions, linkage, and their consequenc- classical ``evolutionary ecology'' approach of exam- es for genetic architecture. ining equilibrium/optimal situations with an approach It has obviously not been the intention of these pa- that focuses on genetic and ecological dynamics in non- pers to cover the ®eld of community genetics compre- equilibrium situations. I found their paper particularly hensively, and so it may be useful to point out some valuable in pointing out how explicitly manipulating other issues and approaches that may gain momentum the building blocks of community genetics can lead to in the future. Coming from population and ecological outcomes different from those in which we assume that genetics, two questions strike me as crucial. The ®rst evolutionary ecology is a long and tempered dance. is whether, and to what degree, genetic recombination Whitham et al. (2003) take a more holistic approach, (as actualized in outcrossing and sexual reproduction) and show that genetic variation within keystone or is responsible for maintaining population abundance. dominant species can have cascading effects on the Much of the focus on discussions of the evolution of associated community and the ecosystem. They posit sex has been on the adaptive signi®cance of sex, and these effects as representing an ``extended phenotype.'' on attempts to account for its maintenance, given its Special Feature atwudi e(n o ogwudi ess)i a in persist) it would long how (and community? be abun- it how would uniform, genetically dant species one make attention. no we almost If received has knowledge, interesting, my to equally but, is abundance species promotes sex how of question converse The disadvantage. ``twofold'' 600 susi h otx ftelre omnt patterns gradients). community latitudinal these larger (i.e., diversity the of species of of discussion context years hears the hardly in many issues one for but emphasized 2002), been (Lee also this biological in has exploring genetics invasions data of few importance The are and relationship. There diversity diversity. genetic between species relationship showing positive (1978) results Oka a and some Morishima presented of studies I the which from (An- on raised genetics and I community 1992) that of tonovics description question earlier a my is diversity in This species diversity. between genetic relationship and the crucial. is as me ®rst strike that The questions two posit again can I 1998). Hogeweg and (Savill coevolution- of processes simulations ary computer of context the in quasi- ``the quali®er the species any by use thoughts) they their that in binomial salutary least Latin (at be preface may to It ecologists variation. for genetic have all of and also amounts ecotypes, large ecologists or races uniform most of consist genetically to always a nearly familiar by more caused Species host, not entity. they a is within when infection infection HIV this that particular a say, acknowledging of, speak thereby particular 1993), (Ei- ``quasi-species'' a host, gen particular of a of genotypes within or face variable strain, the col- highly such in of term lections therapy They re- vaccine variation. strain interpret and within-pathogen to drug how to with sponses struggle as biologists signi®cance disease applied This tremendous gaining polymorphisms. is such species question involving than not invasibility) mutual interactions a and (vis stable coexistence in- more vis interactions are polymorphisms species genetic whether volving is com- question have the can level, reductionist more species a to Translated dominant consequences. munity a genetic that within studies, Lenski other's variation and ex- own and extensive their with from show, Bohannan amples (2003) al. also et Whitham (see 2000). cite phage/bac- interactions on they (2001) terial this, Lenski of of work support experimental the in and strong coexistence, with a associated species be make may polymorphism (2003) genetic al. that case et spe- Neuhauser maintaining diversity. for cies crucial is polymorphism genetic ahseis ol hscmuiyb ssal sone as quasi-species? stable of as consisting be community this would within species, sampled uniform each genetically randomly are that of individuals consisting asexual community a erated n o ih eicue hrceie n measure and characterize, include, we diversity, might community how measuring and for units right the cies oigmr rmacmuiyeooystandpoint, ecology community a from more Coming which to extent the addresses question second The ....''!Thetermis led ann acceptance gaining already fw gen- we If AI ANTONOVICS JANIS r spe- Are h us-pce opnn?Hwde iest at parameters? community diversity in¯uence does level How this component? quasi-species the iua oiin ihnfo esaemr ieyto likely more others. are par- than webs occupying evolve species food if within ask positions also ticular can (Caldarelli we evolution 1998), web al. food et in interest Given community? growing a se- the of particularly members dominant are the for that vere evolution on im- constraints community- interactions pose if multispecies ask through also feedbacks can by we level paper the (2003), of al. context et the Whitham In change 2002)? climatic of Shaw rate and the (Davis with changes pace evolutionary keep because to it unable oth- is are than Or why?)? less anal- so evolve the (if species ers some in because used it are Is yses)? evidence patterns other show with that species concordant those result only the (i.e., Their it sampling so. Is of do puzzle. a to remains pressure conservatism distri- under apparent new been the and have had and tolerances have butions, new species evolve Surely, to more. thou- opportunity or of years tens been over of change sands has climate record predicting in evolu- paleontological useful for the potential change, huge the tionary given why, been always eitne o l r hs ee?Wa fractions What genes? these pathogen in are involved old are How organisms resistance? host of in fractions genes What the ecology. brought and be- change changes feedback genomic the genomic tween and the interactions community about by ask about questions can many ``com- we indeed of that are discipline There an newer genomics.'' get even is munity perhaps the ``genetics'' should positing of I by that insignia edge and the fading, that rapidly notes itself the me course, Of in future. fasci- the cynic short- for other no directions many research and of are genetics age community there in that questions illustrate nating to brief- tried also have ly I control. disease and conservation, biology, insights for invasion and consequences These practical insights and ecology. real new have community to lead on in can perspectives contexts interact genetical species the which about broadly thinking how of ¯ag emphasize papers the featured under The brought genetics.'' ``community be can that questions lenging Presumably species Hawaiian relationships. determining the in species±area speciation process and that important idea diversity an the be of accepting may more are ogists xetmcaim fseito ttegntclevel what structure. genetic to community the into ask at back also speciation feed must of we mechanisms then extent likely, in- appears as deed patterns, macroecological spe- in¯uence If cases. does special ciation as dismissed previously were rica hne n fs,wa stercmuiycontext? global community their past is to what so, genetically if responded and change, not have and fseito edt oedvrecmuiisthan communities diverse more modes? to other lead speciation of ntrso lblcag,amjrpzl o ehas me for puzzle major a change, global of terms In ncnlso,teeaenmru xiigadchal- and exciting numerous are there conclusion, In ecol- (2001), Hubbell of work the through Largely Drosophila a eietf pce hthave that species identify we Can n h ihi se fAf- of ®shes cichlid the and clg,Vl 4 o 3 No. 84, Vol. Ecology, osm modes some Do March 2003 COMMUNITY GENETICS: COMMENTARY 601 in pathogen genes are involved in immune evasion? Endler, J. A. 1991. Genetic heterogeneity and ecology. Pages Are genes determining host±pathogen interactions 315±334 in R. J. Berry, T. J. Crawford, and G. M. Hewitt, editors. Genes in ecology. Blackwell, Oxford, UK. more duplicated and multiallelic than genes determin- Hubbell, S. P.2001. The uni®ed neutral theory of biodiversity ing predator±prey interactions? What is the role that and biogeography. Princeton University Press, Princeton, noncoding DNA plays in life history, phenology, and New Jersey, USA. community interactions (the community DNA para- Lee, C. E. 2002. Evolutionary genetics of invasive species. Trends in Ecology and Evolution 17:386±391. dox!)? And so on. . . but then maybe one commentary Lenski, R. E. 2001. Testing Antonovics' ®ve tenets of eco- is enough for now. logical genetics: experiments with bacteria and the interface of ecology and genetics. Pages 25±45 in M. C. Press, LITERATURE CITED N. Huntly, and S. Levin, editors. Ecology: achievement and challenge. Blackwell, Oxford, UK. Antonovics, J. 1976. The input from population genetics: Maddox, G. D., and R. B. Root. 1990. Structure of the en- ``The new ecological genetics.'' Systematic Botany 1:233± counter between goldenrod (Solidago altissima) and its di- 245. verse insect fauna. Ecology 71:2115±2124. Antonovics, J. 1992. Toward community genetics. Pages Morishima, H., and H. J. Oka. 1978. Genetic diversity in 426±449 in R. S. Fritz and E. L. Simms, editors. Plant rice populations of Nigeria: in¯uence of community struc- resistance to herbivores and pathogens: ecology, evolution, ture. Agro-ecosystems 5:263±269. genetics. University of Chicago Press, Chicago, Illinois, Neuhauser, C., D. A. Andow, G. Heimpel, G. May, R. Shaw, USA. and S. Wagenius. 2003. Community genetics: expanding Bohannan, B. J. M., and R. E. Lenski. 2000. Linking genetic the synthesis of ecology and genetics. Ecology 84:545± change to community evolution: insights from studies of 558. bacteria and bacteriophage. Ecology Letters 3(4):362±377. Savill, N. J., and P. Hogeweg. 1998. Spatially induced spe- Caldarelli, G., P. G. Higgs, and A. J. McKane. 1998. Mod- ciation prevents extinction: the evolution of dispersal dis- elling coevolution in multispecies communities. Journal of tance in oscillatory predator prey models. Proceedings of Theoretical Biology 193:345±358. the Royal Society of London Series B 265:25±32. Davis, M. B., and R. G. Shaw. 2002. Range shifts and adap- Whitham, T. G., W. Young, G. D. Martinsen, C. A. Gehring, tive responses to quaternary climate change. Science 292: J. A. Schweitzer, S. M. Shuster, G. M. Wimp, D. G. Fischer, 673±679. J. K. Bailey, R. L. Lindroth, S. Woolbright, and C. R. Eigen, M. 1993. Viral quasi-species. Scienti®c American Kuske. 2003. Community genetics: a consequence of the Feature Special 269(1):32±39. extended phenotype. Ecology 84:559±573.