community succession and assembly - lab home...

Community Successionand AssemblyComparing, Contrasting andCombining Paradigms in the Contextof EcoloNcal Restorationby Truman P. Young, Jonathan M. Chaseand Russell T. Huddleston

Three ecologists

exptore two ways of

thinking about the

devetopment of eco-

logical communities

and the retationships

of these approaches

to each other and

to restoration.

{f~’~/V/hen we look at the plants andVV bushes clothing an entangled

bank, we are tempted to attribute their pro-portional number and kinds to what we callchance. But how false a view this is! Everyone has heard that when an American for-est is cut down a very different vegetationsprings up; but it has been observed thatancient Indian ruins in the SouthernUnited States, which must formerly havebeen cleared of forests, now display thesame beautiful diversity and proportion ofkinds as in the surrounding forests."

--Charles Darwin,1872 (1958, pg. 84)

Restoration ecology is rooted in commu-nity ecology (Palmer and others, 1997;Young, 2000). One cannot restore individ-ual bits of biodiversity unless appropriatecommunities and soil exist. There are ofcourse population-restoration projects, orreintroductions, some of which overlayfunctional ecosystems. Nonetheless, it is inthe restoration of complex communitiesthat restoration ecology finds its greatestchallenges and opportunities, and also itsfullest expression (Young, 2000).

Therefore restoration ecology, in itssearch for conceptual bases, looks to theo-ries of community structure (Luken, 1990;Packard, 1994; Lockwood, 1997; Pritchett,1997). In particular, two conceptual mod-els in community ecology have relevanceto ecological restoration: 1) communitysuccession, which dates back more than a

century (Cowles, 1899) and 2) the morerecently developed ideas of communityassembly and priority effects (Palmer andothers, 1997; Lockwood, 1997). Briefly,succession refers to an orderly, more or lesspredictable turnover of species composi-tion at a site that has been cleared ofspecies or otherwise disturbed, often backtoward a predisturbance state. Communityassembly is similarly intended to explainhow communities form after a site iscleared of species. In assembly, however,community development is determined byrandom variation in species’ colonizationrates and the subsequent likelihood of theirestablishment and persistence in the com-munity. Although, as we will discuss, theseviews are not mutually exclusive, there area number of underlying differencesbetween succession and assembly as theyhave developed historically (Table 1).

Both succession and assembly haveenjoyed considerable theoretical andempirical attention from communityecologists, and have obvious connectionsto each other, tn addition, both of theseconceptual frameworks can be useful inproviding a scientific foundation for theemerging field of restoration ecology.However, despite the inherent similaritybetween community succession andassembly, there has been little cross-fertil-ization between these concepts (but seeLawton, 1987; Drake, 1990; Walker,1997; Mclntosh, 1999). In this paper weexamine both models, comparing and

ECOLOGICAL RESTORATION 19:1 ¯ 2001 5

Table 1. A summary of differences between succession and assembly as conceptual models in community ecology.

CharacteristicCommunities studiedDominant taxaEcosystems

Types of modelsEnd point(s)

Reasons foralternative states

Process vs. product

Life history differences

Processes

Dispersal limitationEstablishment and growth

Species’ residence times

Nature of facilitation

Number of trophic levels

SuccessionNatural, semi-natural

Vascular plantsMainly terrestrial (mainland)

Conceptual (verbal)A single "climax," or a few alternativestable statesArrested (truncated) succession, disclimax,cyclic succession, exogenous disturbancePrimary interest in the mechanisms of change,only secondary interest in how this results ina "final" state.Central (colonization/competition tradeoffs)

More deterministic than random

Sometimes considered, species-specificIndividual phenomena

Not much more than individual life span,often barely exceeding age of first reproduction

Within trophic levels, causes temporal changeOne (except as modifiers)

AssemblyArtificial (controlled), virtual(Invertebrate) animals and protistsMainly aquatic and islandSimulation (mathematical)Theoretically many alternatlve stable states.In empirical experiments, only a few.Priority effects and niche preemption, sometimeswith a cascading effectPrimary interest in explaining the final state(s),with the process seen mostly as a means to getto this state (but otherwise ignored).Invoked only after species arrival, and only insome modelsMore random than deterministicCentral, but stochasticPopulation phenomenaUsually far longer than individual life spans

Across trophic levels, causes alternative statesUp to several

contrasting them from the perspective ofrestoration. At the heart of this discussionis a question that is at the very root ofrestoration ecology (Palmer and others,1997): Do altered communities have aninherent ability to repair themselves andreturn to a structure and composition sim-ilar to the original, or can historicalevents and contingencies allow for morethan one, and perhaps an indefinite num-ber of (stable) community outcomes?

But this review is not simply about sin-gle as opposed to multiple stable states incommunities. Succession theory has beendealing with multiple stable states nearlysince its inception (see below). Assemblytheory represents a particular frameworkfor multiple stable states, but does so froman ecological perspective that differs inmany ways from most succession theory. Itis these differences that we will explore.

Historical contextCommunity SuccessionInterestingly, these two views of commu-nity dynamics recall a classic dichotomyestablished among early plant communityecologists. Darwin wrote about succes-

sional phenomena (see opening quote),and indeed, some of the first research inmodern ecology was based on the idea ofsuccession (Cowles, 1899). Henry Cowles’sclassic research on the dynamics of vege-tation on dunes on the shores of LakeMichigan helped to establish the idea thatthe development of a plant communitycould best be described as a sequence--orsuccession--of species and associations ofspecies replacing one another throughtime in a more or less orderly and pre-dictable way. Building on these and otherobservations, Frederick Clements (forexample, 1916, 1936) championed theview that communities represent analliance of species that reaches maturity ina stable "climax" community. Althoughthere were several underlying hypothesesin Clements’s work, the predominant par-adigm to emerge was that communities"succeed" along a fixed trajectory toward asingle, more or less well-defined end state.

Two kinds of observations of disturbedplant communities have inspired much ofthe research on succession. The first is thatthey progress through a set of different com-munity states, with certain common char-acteristics ("early and late successional") in

a wide variety of ecosystems. The second isthat they tend to develop back to a statesimilar to the original one, even from verydifferent starting points. Despite the cur-rent aversion to the classic notion of com-munity "climax" (see below), the factremains that many terrestrial communitiesdo tend to return to predisturbance states ina more or less predictable way (for example,van Breeman, 1995; Sheil, 1999; Buddleand others, 2000; Leps and others, 2000).Conceptual models of succession date fromthe early days of plant ecology (Clements,1916), and continue today (Pacala andRees, 1998). These models focus on mech-anisms of change within successionalcommunities, and have been reviewedelsewhere (Pickett and others, 1987;Luken, 1990; Van Andel and others, 1993;McCook, 1994).

In the context of successional theory,ecological restoration can be seen to rep-resent an attempt to accelerate or jump-start the successional sequence, bypassingearlier successional stages (Palmer andothers, 1997). Understanding the processesthat drive succession, we can suggest prac-tical ways to speed the rates at which thedesired states are achieved (see below).

6 ECOLOGICAL RESTORATION 19:1 ¯ 2001

While Clements emphasized thecoherence and unity of the community,his primary antagonist, Henry Gleason(for example, 1927), was impressed by theindividuality of species within communi-ties, and placed a greater emphasis on therole of species-specific processes and his-tory in determining eventual communitycomposition. Gleason was the first in along line of ecologists to propose and doc-ument a more pluralistic view of plant suc-cession. Interestingly, Gleason presented asurprisingly modem view of the role ofcommunity assembly history on the devel-opment of community composition in thispassage describing how emergent wetlandplants colonize and come to dominatesmall ponds in a region: "Only the chancesof seed dispersal have determined the allo-cation of species to different pools, but in

the course of three or four years, each poolhas a different appearance, although theenvironment, aside from the reaction ofthe various species, is precisely the samefor each." (Gleason, 1927). Tansley (1935)later coined the term "polyclimax" todescribe such alternative stable states.

Community AssemblyThe development of assembly theory didnot occur until almost half a century later.It was based on Jared Diamond’s (1975)studies of bird communities on Pacificislands and was developed by several zool-ogists, who did not recognize that Gleason,and later Frank Egler (1954), had antici-pated many of the main points of theassembly model. This conceptual frame-work attempts to explain the existence in

sites with similar environmental condi-tions of communities composed of differ-ent individual species (or relativeabundances of those species), eventhough a much broade~ pool of species hasaccess to the community. As we define ithere, community assembly refers to theprocess by which species colonize, interactwith other species, and establish a com-munity in such a way that multiple stablecommunity states may result. We are notreferring here to the related theory of"assembly roles," which seeks to explainnon-random similarities in guild structureacross communities differing in their com-ponent species (Weiher and Keddy, 1999).

When assembly biologists use theterm "multiple stable states" (Lewontin,1969; May, 1977; Drake, 1991; Knowlton,1992; Law and Morton, 1993; Drake and

Are aquatic ecosystems, such as this vernal pool, more likely to develop alternative stable states than are the more terrestrial systems in the

surrounding uplands of Oregon’s Agate Desert? Such differences between systems may account for different ideas of system dynamics ecolo-

gists have developed over the years. Photo by R.T HtJcJdleston

ECOtOOICAt RESTORATION 19:1 ¯ 2001 7

others, 1996; Samuels and Drake, 1997;Chase, 1998, 1999a) they restrict itsmeaning to a process in which the finaloutcome depends heavily on historicalcontingency--that is, all else being equal,more than one final (stable) compositioncan result, depending only on the detailsof the community’s history.

In assembly theory, "multiple stablestates" refers only to situations where allspecies have equal access to communities,and where differences in eventual commu-nity composition are due to variations in

Do altered communitieshave an inherent abilityto repair themselves andreturn to a structure andcomposition similar tothe original, or canhistorical events andcontingencies allow for.more than one, andperhaps an indefinitenumber of (stable)community outcomes?

the timing of colonization. Thus, it is notcorrect to invoke assembly theory to saythat a community with, for example, somelarge herbivore species exists in an "alter-native stable state" to one that does not(Strohmayer and Warren, 1997; Rietkerkand van de Koppel, 1997; Kim, 1997).Rather, in assembly theory alternative eco-logical states driven by herbivores existonly when the herbivore has equal accessto both communities, but plays a function-ally different role depending on the timingof its arrival in the community, which isconsidered to be random (see for example,Dublin and others, 1990). Succession bio!-

ogists recognize a wider range of causes formultiple stable states (see below).

In contrast to restoration efforts basedon the assumption that degraded commu-nities tend to proceed through successionback toward the predisturbance state,restoration based on the idea that a com-munity that can exist in multiple stablestates will be less straightforward, and canrepresent a severe limit to our ability toappropriately restore ecological communi-ties. Nevertheless, if we can understandexplicitly the nature and structure of thesemultiple stable states, restoration recom-mendations will be much better informed(see also "Implication for ecologicalrestoration" below).

Evidence for Stability inSuccession and AssemblyBoth succession and asserably havereceived considerable theoretical atten-tion from community ecologists in recentyears. However, the current dearth ofempirical, and particularly experimental,evidence that alternative stable commu-nity configurations are common makes itdifficult to evaluate the relative utility ofthese two conceptual frameworks, partic-ularly as a foundation for restoration ecol-ogy. Connell and Soma (1983) pointedout a number of stringent requirementsthat are needed to show stability in acommunity, and especially to determinewhether a community can exist in multi-ple stable states. They also suggested that,as of 1983, few experimental studies metthose requirements. This is part of abroader debate in community ecologyconcerning the importance of inherentvariability in community structure thatdates back to Watt (1947).

In order to demonstrate the existenceof the kinds of multiple stable states pro-posed by assembly theory, it is necessary toshow that different communities candevelop under conditions that are essen-tially identical except for variations in thetiming of species arrival. Since such con-trol is difficult to achieve in the field, themost successful early experimental studiesexploring the existence of one or morestable states have involved microcosms(see for example Robinson and Dickerson,

1987; Robinson and Edgemon, 1988, 1989;Drake, 1991; Drake and others, 1993;Weatherby and others, 1998). Neverthe-less, some experimental studies in naturalcommunities have also provided evidencefor multiple stable states (Sutherland,1974; Paine and others, 1985; Bazely andJefferies, 1986; Sutherland, 1990; Chase,1998, 1999b and unpublished manuscript;Petraitis and Latham, 1999). Both of thesekinds of studies (virtually all of whichwere carried out in aquatic systems) havedemonstrated that differences in the rela-tive arrival times of species can producealternative communities that are rela-tively stable. In addition, a variety ofmodels (both analytical and computersimulation) have been developed to sim-ulate the assumptions and outcomes ofassembly theory (Drake, 1990; Luh andPimm, 1993; Law and Morton, 1993,1996). It remains to be seen, however,how often and under what conditionscommunities converge toward a singlestable state (whether or not we call thisthe climax), and when priority and his-torical legacy will lead to multiple stablestates under identical environmental con-ditions (Ludwig and others, 1997).

Comparison and ContrastOne of the earliest successional models,the Initial Floristics Model of Frank Egler(1954), anticipated many of the basicconcepts underlying assembly theory.Egler proposed a model in which variationin initial composition, unrelated to abi-otic site differences, resulted in the estab-lishment of different stable communitytypes. Among late-successional species,those already present in the seed bank orarriving shortly after a disturbance eventwere able to establish in sufficient num-bers that later arrivals were not able tochange the course of community develop-ment (essentially a priority effect). Someof the earliest and best work on alterna-tive ecological states within the succes-sional context (see below) was done byEgler (1949, 1975) and his colleagues,most notably Bill Niering (Niering andEgler, 1955; Niering and Goodwin, 1974;Niering and others, 1986; Niering, 1987;Fike and Niering, 1999).


The paradigms of succession andassembly have much in common:1.Both seek to explain community com-

position.2.Both suggest that there is a historical

explanation to this composition.3. Both acknowledge that communities

develop through time toward relativelystable states.

4. Both assume the importance of bioticinteractions, especially competition.

Despite these similarities, these twoapproaches also differ in many ways. Table1 summarizes some of the contrasting fea-tures of these two paradigms as they areconceived today. These are not intended asstrict dichotomies, but rather as strong ten-dencies that may at least partly reflect thedifferent characters of the ecological sys-tems in which they are studied. It is strikingthat ecologists who work in terrestrial plantcommunities have generally thought inten~s of succession, while most modemresearch on assembly theory has been car-ried out in aquatic (invertebrate) systems,and in virtual (computer) "communities."The reasons for this are not clear, althoughwe suggest one possibility below (under"Mechanisms of establishment"). Eco-logical restoration has a primarily botanicalorientation (Young, 2000), and so perhapsit is natural that succession theory has beena dominant paradigm in restoration (forexample, see Packard, 1994; Anderson andothers, 2000; Kettle and others, 2000).

Another key difference is that assem-bly theory is far more amenable to mathe-matical modeling than is succession. Thisapparent rigor makes assembly theoryattractive to modem ecologists, but alsodepends on the simplicity of its underlyingassumptions, such as the ecological nicheequivalence of species and the absence ofcolonization-competition trade-offs. Indeed,if these computer models were to add theelement that the better competitors in acommunity on average arrive later, the vir-tual communities these models producemight more closely resemble convergentsuccessional comnctunities.

Process versus ProductTheories of succession are designed toexplain the changes in species composition

that occur through time following a distur-bance. They seek to explain the mecha-nisms that limit the establishment andpersistence of species, and thus determinethe relative timing of species’ appearancein, and disappearance from, the commu-nity. They also seek to identify the traits ofspecies that place them at different stagesof a successional sequence, and the pat-terns or trends that characterize such asequence. While it is true that successiontheory often treats this as a process con-verging toward a predisturbance state, itdoes not usually attempt to explain thisstate except as the product of the sequence.

Assembly theory, in contrast, is pri-marily concerned with the explanation ofstable community composition, and inparticular, the similarities and differencesamong communities that explicitly arenot due to transient dynamics. Simulationmodels do implicitly incorporate process(they "assemble" communities), but theirexplicit goal is to explain differences ineventual community composition. Cer-tainly, priority effects are inherently tem-poral, but there is little else in assemblytheory that explicitly addresses theprocess or mechanisms of communitychange or development through time.

Alternative StableStates or Convergence?

What a struggle must have gone onduring the long centuries betweenthe several kinds of trees each annu-ally scattering its seeds by the thou-sand; what war between insect andinsect--between insects, snails, andother animals with birds and beastsof prey--all striving to increase, allfeeding on each other, or on thetrees, their seeds and seedlings, or onthe other plants which first clothedthe ground and thus checked thegrowth of the trees!

--Charles Darwin,1872 (1958, pg. 84)

The most fundamental difference betweensuccession and assembly as conceptualbases for community and restoration ecol-ogy is how they treat alternative stablestates. The stereotype of community suc-

cession theory is that it predicts a singlestable state--the climax. The stereotypeof assembly theory is the existence ofmany stable states. Superficially, theserepresent highly divergent views, withprofoundly different implications forrestoration. However, empirical evidencefrom actual studies of successional com-munities and assembly experiments sug-gest that this difference is less dramatic.

In reality, the difference is not, assome have suggested, that assembly the-ory recognizes or predicts many alterna-tive stable states, while successionrecognizes only one. Successional theoryhas long recognized alternative stablestates, and assumes that communities notreturning to the predisturbance state havefailed to reach it for reasons that can beexplained (for example, see Reynolds andPacala, 1993; see also below). Assemblymodels may predict highly divergent com-munity compositions, but both in the fieldand in experimental ecosystems the num-ber of alternative states produced is usu-ally very limited. In practice, bothsuccession and assembly theories are asso-ciated with a small number (one to a few)of stable ecological states.

Succession theory arose in part toexplain a perceived convergence of dis-turbed sites to a single community type.When ecologists began to recognize alter-native "stable" states, they responded notby discarding the idea of succession, butby elaborating and refining it. Alternativestates were considered merely variationson the theme of succession, not failures ofsuccession theory. These states have beencalled arrested succession, truncated suc-cession, polyclimax, and disclimax.Modern succession theory offers severalclasses of explanations for alternative sta-ble states:1. A disturbance generated by commu-

nity traits keeps the community in anarrested successional state. This forcecould be fairly constant (herbivory) orsporadic, but predictable (fire).

2. An exogenous random disturbance(droughts, storms, wave action, land-slides) can occur often enough that thecommunities of a natural landscape dis-play a variety of successional states, butrarely reach local stability.


Planting needlegrass (Nasella pulchra) at a grassland restoration site near Davis, California. Does planting of particular species establish their corn-petitive priority in ways that affect long-term community dynamics, thus running the risk of short-circuiting succession? Photo by Truman Young

3.Cyclic succession occurs when all suc-cessional states are intrinsically unsta-ble, with each yielding to anotherwithin the cycle (Watt, 1947; Olff andothers, 1999). Some examples involvesingle-aged stands of woody plants thatare not self-replacing (Yeaton, 1978;Agnew, 1984; Young and Lindsay,1988), but which can establish amongother species in successional communi-ties (Young and Peacock, 1992).

4.Succession theory recognizes that thereare some early-successional (or exoticinvasive) species that are highly sup-pressive of later-successional species,and that when these get established insufficient densities early in succession,succession can be arrested for long peri-ods of time, or even halted in an alter-native state. Natural examples includeshrublands, balds and fernlands inforests (Tappeiner and others, 1991;

Maxwell and others, 1993; Mallik,1995; Young, 1996; Holl, 1999; Fikeand Niering, 1999; Chapman andChapman, 1999), fertility mosaics ingrasslands and savannas (McNaughton,1983; Belsky, 1986; Young and others,1995), forest mosaics (Davis and oth-ers, 1998) and marine mussel and algalbeds (Petraitis and Latham, 1999).Egler (1949, 1975) provided manage-ment examples in rights-of-waythrough forests.

5. Egler’s Initial Floristics successionmodel explicitly considered the possi-bility that stochastic differences in thearrival and establishraent of differentlate successional species could lead toalternative ecological states (see alsoGleason, 1927; Tansley, 1935).

It is these last two processes thatmatch most closely the multiple stablestates of assembly theory, although with a

very limited number of alternative stablestates. Modem succession theory is lessfixated on the concept of a single "climax"but nonetheless still considers both pro-gression and convergence to be commonin regenerating communities in nature.

Assembly theory, on the other hand,arose out of a set of observations opposite tothose that inspired succession theory. Here,it was the multiple stable states themselvesthat drove theory (Diamond, 1975). In the1970s, plant ecologists were increasinglyrecognizing multiple stable states in somesystems (suppressive shrubs in forests beinga classic example) and were elaboratingsuccessional theory to account for them. Atthe same time animal ecologists, in theirdevelopment of assembly theory, were pay-ing little attention to succession theory intheir attempts to account for variations incommunity dynamics. Assembly theoryassumes that there is considerable niche


similarity among species, and that randomdifferences in colonization and establish-ment at a given site can result in alternativemixes of species that are resistant to inva-sion by others.

State-transition ModelsFaced with the realities of disclimax,arrested succession, and habitat degrada-tion, some ecosystem managers respondedwith a new kind of community model~the state-transition model (Westoby andothers, 1989; Lockwood and Lockwood,1993; Allen-Diaz and Bartolome, 1998).State-transition models were initiallydeveloped by rangeland ecologists, andstill are used mainly in that context. Likeassembly models, they are based on theidea of alternative ecological states.However, these states are not merely theresult of priority effects, but are often theresult of different forcing factors in theenvironment: grazing, drought, fire, or soilmodification, for exaraple. It is the pur-pose of these state-transition models notonly to identify the factors causing alter-native states, but also to identify remedialactions to bring about transitions fromless desirable states to more desirablestates. Because of their emphasis onexogenous processes, state-transitionmodels do not relate directly to our com-parison of succession and assembly, buttheir emphasis on management alterna-tives does make them valuable models forrestoration ecologists.

Priority Effects andNiche SimilarityCentral to assembly theory is the idea ofpriority effects (Belyea and Lancaster1999). Priority is established by the speciesthat are the first to arrive and can thereforebecome dominant in the community.These species then inhibit invasion byother species, particularly those with simi-lar niches. The strength of the prioritydepends on 1) arrival time, and 2) theattributes of species after arrival (for exam-ple, competitive ability, fecundity, or pop-ulation growth rate). Temporal prioritymay confer a competitive advantage on aspecies that is normally an inferior com-

petitor (Hodge and others, 1996). Multiplestable states are possible because differentspecies may by chance arrive earlier at dif-ferent sites, establish themselves, andexclude later arrivals.

Because established plants tend tohave local priority in any community,succession occurs when this priority (inhi-bition, or resistance to invasion) is over-

The stereotype ofcommunity successiontheory is that it predictsa singte stabte state--the climax. The stereo-type of assemb[y theoryis the existence ofmany stable states.

come either by more competitive species,or by changes in the environment broughtabout by the plants themselves (facilita-tion). But succession theory has alsoincorporated the idea of priority effectsmore directly. Working in an easterndeciduous forest where he saw the effec-tiveness of clonal shrubs in suppressingforest succession, Egler (1954) was one ofthe first to suggest that the initial speciescomposition on a site would determine theeventual outcome of community develop-ment (but see also Gleason, 1927). Twodecades later, Connell and Slatyer (1977)elaborated this idea in their inhibitionmodel, which presents priority as a mecha-nism driving successional processes. Initialfloristic composition and inhibition byearly-arriving species are important pro-cesses in succession, but these representonly a few of the many mechanisms thatinteract during succession. Competitiveinhibition may delay succession and createtemporary alternative states, but the gen-eral assumption is that given enough time,the community will progress toward thepredisturbance state.

Both succession and assembly theoriesrecognize that some species will always bebetter competitors regardless of the orderin which they arrive, and others will alwaysbe poor competitors, even if they have hada long time to establish in a community. Inassembly, priority effects can explain whyin some areas species coexist and in othersthere is competitive exclusion. In areaswhere the poor competitors arrive first,become established and are subsequentlyinvaded by a better competitor, their pop-ulations are significantly reduced, but thespecies is able to persist in the community.However, if the better competitor arrivesfirst, the less competitive species are unableto invade the community and are thereforeexcluded (Drake, 1991).

Successional models generally explainthe coexistence of early- and late-succes-sional species at the landscape scale as a

"result of disturbance. Stochastic distur-bance events in the community removesome of the more competitive species andallow the less competitive species to existlocally and temporarily. Disturbance mayoccur on a large scale, returning large areasto an earlier state of succession, or on asmall scale, creating a mosaic of differentsuccessional stages within the larger com-munity. In addition, successional modelsexplicitly incorporate the idea of a colo-nization-competition trade-off. The speciesthat are most likely to arrive early are alsocompetitively subordinate, and eventuallyare replaced by late-arriving species that arecompetitively dominant. This coloniza-tion/competition trade-off is a centralprocess in successional change. Althoughcurrently absent from assembly models, itcould easily be incorporated, and wouldprobably result in more robust models thatmore closely resemble succession.

Recruitment LimitationRecruitment can be limited at either of twostages: dispersal or establishment. Recruit-ment limitation is central to restorationecology (Bakker and others, 1996; Strykstraand others, 1998; Bakker and Berendse,1999). In many communities, the availabil-ity of propagules is a limiting factor insuccessful restoration (Robinson andHandel 1993; Gorchov and others, 1993;


Guariguata and others, 1995; Keenan andothers, 1997; Parrotta and others, 1997b;Lamb and others, 1997; Stampfli andZeiter, 1999). One of the primary activitiesin ecological restoration is planting andassisting in the establishment of nativespecies. This work is a specific attempt toovercome the factors that limit recruit-ment. Conversely, the control of invasiveexotic species often occurs at the stages ofdispersal and establishment. We will com-pare how succession and assembly theoriestreat both stages of recruitment limitation.

Dispersal LimitationDispersal limitation was recognized as cen-tral to community development in theearly theories proposed to explain succes-sional processes. Clements (1916) andEgler (1954) were among the first to rec-ognize the influence of dispersal limitationon plant communities by suggesting thatthe failure of potentially dominant speciesto arrive at a given site could dramaticallyalter the trajectory of community develop-ment. Later models such as Connell andSlatyer’s (1977) tolerance, inhibition andfacilitation model make only passing refer-ences to dispersal limitation, and focusinstead on the processes occurring afterpropagules arrive at a given site. Recentlythere has been a renewed interest in therole of dispersal limitation in communitydynamics. The proximity of seed sources,differential timing of maturity and age ofreproduction, annual seed production, andtiming of disturbance have all been foundto influence the spatial and temporalaspects of community development (DelMoral, 1998; Hughes and Fahey, 1988;Fastie, 1995; Holl, 1999).

The life history traits of individualspecies can also play a central role in suc-cessional processes. Ridley (1930) sug-gested that dispersal strategy involves acompromise in energy expendituresbetween growth and reproduction. Specieswith short generation times, prolific repro-duction and efficient dispersal mecha-nisms tend to be favored in openenvironments where they can grow andreproduce before slower-growing but morecompetitive species become dominant.Some species are adapted to shift resources

from reproduction into vegetative growthas resources became limiting (Houssardand Escarre, 1995). The consequence ofsuch tradeoffs is increased variability inthe seed rain and seed bank, which can

Assembly theoryemphasizes the randomtiming and order ofarrival of species, likenumbers coming up ina game of roulette.Succession, in contrast,may be likened.to stowtydeveloping a winninghand in gin rummy.

ultimately influence successional dynam-ics (Clark and Yuan, 1995).

Assembly theory is typically based onthe assumption that all species have anequal probability of colonization, and thatthe long-term absence of a species frorn agiven site is not due to lack of access.These models typically ignore species-spe-cific differences in colonization ability, asGleason (1927) noted in an early discus-sion of how community composition canbe determined by the stochastic processesof species colonization history, and as isimplicit in some recent models of theassembly process (for example, see Drake,1990; Law and Morton, 1993, 1996;Morton and others, 1996). The emphasis ison the random timing and order of arrivalof species, like numbers coming up in agame of roulette. Succession, in contrast,may be likened to slowly developing a win-ning hand in gin rummy.

Ecologists have examined assemblytheory using numerous models and experi-ments designed to reveal the effects ofarrival order, invasion intensity (the

amount of time between successive inva-sion events) and species attributes (Drake,1991; Lockwood and others, 1997; Weiherand others, 1998). However, to our knowl-edge the effects of species’ dispersal abilitiesand/or colonization competition trade-offshave not been examined in the context ofassembly theory. It is also interesting tonote that one of the studies often cited insupport of the multiple stable states pre-dicted by assembly theory is a study byMcCune and Alien (1985) of the develop-ment of different coniferous forest commu-nities on similar sites in Montana. Theycite several historical factors that mayaccount for the different forest communi-ties, including variation in seed produc-tion, timing of seed crops relative todisturbance, differential tolerances of thespecies to climate conditions, herbivory,and local availability of seed source. Giventhese variables it seems unlikely that allspecies had an equal opportunity to colo-nize all of the study sites.

We suggest that the assumption ofequal probability of colonization is unreal-istic in many natural situations, and thatfuture models of community assemblyshould allow for the different dispersal abil-ities of species. In addition, if such modelswere also constructed to include species’life history attributes and differential colo-nization and competitive abilities, the dif-ferences between succession and assemblymay not be as great as they now appear.

Mechanisms of EstablishmentJust as they treat dispersal differently, theo-ries of succession and assembly deal in verydifferent ways with the process of establish-ment and growth following arrival. Inassembly theory, growth and establishmentare regarded as population-level phenom-ena. The basic model is one of rare propag-ules arriving, and either reproducing toestablish a growing population or failing todo so. Differences in population growthrates are allowed, and priority effects areachieved when the differences in arrivaltimes are sufficient to overcome differencesin subsequent population growth.

In succession theory growth and estab-lishment are usually treated as properties ofindividuals. Multiple propagules arrive, and


The persistence of shrub cover in this right-of-way suggests that restorationists may sometimes have opportunities to select from alternativestable states in defining management objectives. In the case of this site in western Kentucky, the herb and grass cover under the power linesmay be as stable as the surrounding forest. Photo by J.O. Luken

either succeed in genrtinating and growingpast vulnerable stages, or they do not. Formost successional species, populations havealready begun to decline or are at least sta-ble before individuals reach reproductivematurity. One of the definitions of climax isthat it is a community in which populationsreplace themselves locally. This is in con-trast to early successional stages, whichrarely last much longer than the generationtimes of their component species.

Variations in the post-dispersalgrowth rates of individual plants are cor-related with both dispersability and lifespan along a gradient from early- to late-successional species. Temporary priorityeffects are achieved when the differencesin arrival times are sufficient to overcomedifferences in the growth rates of compet-ing individuals. Indeed, many alternativeecological states in terrestrial plantecosystems are relatively stable becausethe individuals that make up the alterna-tive community are themselves clonal and

long-lived--good examples would beferns or clonal shrubs in forest ecosystems(Tappeiner and others, 1991; Maxwelland others, 1993; Mallik, 1995; Young,1996; Holl, 1999; Fike and Niering, 1999;Chapman and Chapman, 1999).

This difference in mechanisms ofestablishment may be partly due to inher-ent differences between terrestrial plants(succession) and aquatic invertebrates,microbes, and algae (assembly). The latterare characterized by determinate growth,high intrinsic rates of increase, and gener-ation times usually much less than thespecies’ residence time in the community.In contrast, vascular plants in terrestrialsuccession have indeterminate growth,relatively low population growth rates,and species’ residence times in the com-munity that are often not much longerthan individual life spans.

However, when dispersal is morelimiting and dispersal events much rarer,species establishment may be driven by

local population growth through repro-duction of earlier colonists, in a modemore like that assumed in assembly mod-els. This can occur in primary succession(Fastie, 1995; Del Morel, 1998), under-story herbs (Matlack, 1994), and perhapssome suppressive shrubs (Tappeiner andothers, 1991; Maxwell and others, 1993).

Invasive Exotics(All species) are not as perfect asthey might have been in relation totheir conditions; and this is shownto be the case by so many nativeforms in many quarters of the worldhaving yielded their places tointruding foreigners.

~Charles Darwin,1872 (1958, pg. 196)

The success of invasive exotic species isfundamentally a result of a breakdownin geographical isolation--the (usually

ECOtOOICAt RESTORATION 19:1 ¯ 2001 13

human-caused) elimination of geographiclimitations to dispersal. In addition, distur-bances of the native community, such asthose caused by fire, cultivation, or grazing,can provide more opportunities forinvaders to take advantage of the removalor breakdown of barriers to dispersal.

Asserably theory suggests that prior-ity effects have a strong influence on thesuccess of late arrivals and on eventualcommunity composition. From this per-spective, invasive exotics are an excep-tion-species that aggressively displacespecies that established earlier. This looksmore like succession than assembly and,in fact, there is a growing consensus thatthe relative success of invading exoticspecies depends more on their life historytraits and edaphic limits than on theniche saturation of the invaded site(Rejmanek, 1996; Rejmanek andRichardson, 1996; Moyle and Light,1997).

Nonetheless, disturbed systems aremore susceptible than undisturbed ones toinvasion by exotics (Groves and Burdon,1986; Mclntyre and Lavorel, 1994; Pyle,1995; Kotanen, 1997), and invasion suc-cess seems to be negatively related to arti-ficial variation (Tilman 1997, Symstad2000), but not to natural variation(Stohlgren and others, 1999) in thespecies diversity of a community. Theseresults are consistent with assembly the-ory. The newly invaded community isoften relatively stable, but this is an alter-native stable state only in a trivial sense.

In summary, succession and assemblyoffer only limited insight into the problemof invasive exotics that is of so much con-cern to restorationists. To quote JamesDrake and his colleagues (1996, pg. 673):"Non-native species are typically notbound by the rules of assembly that oper-ate in the system they invade."

FacilitationSome successional models explicitlydescribe facilitative relationships betweenplant species (Connell and Slatyer, 1977).Positive interactions between plants areusually mediated by the edaphic environ-ment (Bertness and Callaway, !994;Callaway and Walker, 1997). In succes-

sional models, early-successional plantsmay modify the microclimate or soil in away that allows later successional plantsto enter the community. Assembly modelsusually allow only for negative interac-tions within guilds or trophic levels, andso do not incorporate facilitation in thesuccessional sense. Assembly models dosometimes include positive trophic inter-actions, both direct (plants provide foodfor herbivores) and indirect (predatorsallow certain prey species to co-exist bysuppressing competitive dominants).Both of these mechanisms have long beena part of community ecology theory.

Interspecific facilitation can be animportant force in ecological restora-tion, as when one suite of plants, includ-ing exotic species, is used to facilitatethe regeneration of natural forest inboth tropical (Geldenhuys, 1997;Oberhauser, 1997; see reviews in Lamb,1998 and Parrotta and others, 1997a)and temperate forests (Choi and others,

Both succession andassembly theories havedeveloped with LittLecross-fertiLization. Wesuggest that the timeis ripe for a synthesis.

1988; see also Rousset and Lepart, 1999;Fike and Niering, 1999). These facilita-tions can be in the form of amelioratedmicroclimate (Scowcroft and Jeffrey,1999), soil changes (Whisenant andothers, 1995; Choi and others, 1998;Rhoades and others, 1998), or assistedseed dispersal (Robinson and Handel,1993; da Silva and others, 1996; Parrottaand others, 1997a).

Prospects for a SynthesisBoth succession and assembly theoriesprovide conceptual frameworks for under-standing and studying how communities

are put together. Furthermore, theseframeworks are useful to those involved inthe restoration and management of eco-logical communities. However, these twotheories have developed with little cross-fertilization. We suggest that the time isripe for a synthesis, which we briefly out-line here.

Succession and assembly theoriesshare many underlying patterns and puta-tive mechanisms. Indeed, the basicassumptions and conclusions of assemblytheory were anticipated by plant ecolo-gists, such as Gleason, Tansley and Egler,who were interested in modifying the suc-cessional paradigm. In recent years, how-ever, succession ecologists have tended toexplain multiple stable states by mecha-nisms that only occasionally refer theseearly explorations of random colonizationand priority effects.

Succession theory focuses on theprocesses leading back toward a predistur-bance community composition, and mayassume a trajectory that will go towardsthat state, or examine reasons why thetrajectory does not (resulting in multiplestable states). Assembly theory, on theother hand, not only explicitly allows forthe possibility that multiple stable statescan exist but typically downplays the roleof transitional processes. Sometimes com-munities proceed toward a predisturbancestate regardless of historical conditions(that is, community convergence), where-as at other times historical legacy and pri-ority effects occur (i.e., communitydivergence) such that the eventual com-munity configuration results from a com-plex combination of the species’ biologyand historical contingency (Belyea andLancaster, 1999).

A more realistic view integratesrevised versions of both the succession andassembly processes to allow for the possi-bility of multiple stable states along a tra-jectory of community change. This viewalso allows for the possibility that one sta-ble state may be "stronger" or more resis-tant to perturbation than the others,rather like the climax of early successiontheory. However, along the successionaltrajectory, as a result of historical contin-gency, several other states can be achievedand can remain stable indefinitely, or until


some large perturbation allows continua-tion along the successional trajectory.

Implications for RestorationRestoration attempts based only on thesimplistic versions of ecological succes-sion may fail or be less likely to succeed,if, in fact, multiple stable states occur insome systems, regardless of whether wethink of these multiple states in the con-text of succession or assembly. Whenalternative states are rare, as they are inmany terrestrial plant communities, thenrestoration may appropriately be viewedas the task of simply assisting succession.On the other hand, in systems for whichmultiple ecological states are more likely,our search to determine the "reference"state to which to restore a communitymay be a difficult exercise involving sev-eral choices (Luh and Pimm, 1993). Inaddition, if multiple stable states exist astroughs along a successional trajectorytoward the desired community composi-tion, we cannot simply plant an assort-ment of species thought to be appropriateat a site and trust the process of succes-sion to push the community toward a pre-disturbance composition (Holl, 1999).The desired state may be only one ofmany possible outcomes, and we mustmore fully understand the role of bothasselnbly and succession if we are todevise successful restoration strategies.The existence of multiple stable statesand priority effects should not be viewedas a barrier to restoration, but they do callfor understanding of a more complex setof phenomena than is traditionally asso-ciated with simplistic (and outdated)forms of succession theory.

Once understood, the existence ofalternative ecological states can evensuggest management alternatives. Whereecological restoration must be sensitiveto particular land uses, the practitionermay be able to choose the most appropri-ate from among several alternative stablecommunities, all of which may be consid-ered ecologically and historically appro-priate. For example, either shrub- orgrass-dominated communities may appro-priate in forested ecosystems where thereis a need for visibility along transporta-

tion rights of way, such as or for fire con-trol and access along power line corridors(Niering and Goodwin, 1974; Egler,1975; Luken, 1990; Brown, 1995; Youngand Chan, 1998).

Restoration: ExperimentalTool for EcologyEcological restoration projects are anexperimental ecologist’s dream. Theyprovide opportunities for large-scalemanipulation of the very characteristicsthat we are beginning to suspect are fun-damental to the ecology of populations,communities and ecosystems (Jordan andothers, 1987), including those addressedin this review. For each of the followingresearch questions, restoration projectsoffer ideal contexts for rigorous experi-mental tests, and the chance to integrateassembly and succession theory into anintegrated concept of community struc-ture and development. In addition, theconcepts and theories involved are cen-tral to the development of restoration asan effective conservation and manage-ment tool.1.How important are priority effects in

different systems, and are there pre-dictable patterns in the role they play indetermining community structure?Most experimental research on thisquestion has been carried out’in micro-cosms and mesocosms (Weat~erby andothers, 1998), and most field work hasbeen done in aquatic systems (Suther-land, 1974, 1990; Paine and others,1985; Bazely and Jefferies, 1986;Petraitis and Latham, 1999; Chase,1998 and unpublished manuscript).The latter represents a rare attempt toanswer this important ecological prob-lem in a rigorous way. Unfortunately forecological restoration, none of theseexperiments has been done on terres-trial plant communities.

2.How important are the relative effectsof individual growth and populationgrowth in determining the species com-position of regenerating communities?

3. How great do differences in arrivaltimes and establishment rates need tobe to overcome differences in competi-tive ability and establish priority effects ?

Does the incorporation of competition-colonization trade-offs into assemblymodels produce transient dynamicsmore similar to those described by suc-cession models?How important is facilitation in thedevelopment of community structureand the maintenance of biodiversity?How important is dispersal limitation indetermining the rate and trajectory ofcommunity development? This longneglected determinant of communitystructure is currently being rediscov-ered, often in the context of restorationresearch (Robinson and Handel, 1993;Gorchov and others, 1993; Guariguataand others, 1995; Bakker and others,1996; Keenan and others, 1997;Parrotta and others, 1997; Clark andothers, 1998; Stampfli and Zeiter 1999).

ConclusionAlthough theories of community succes-sion and assembly may at first seem tomake fundamentally different predictionswith profound implications for ecologicalrestoration, we propose that their real-world differences are not so large. It mighteven be argued that some lines of succes-sion theory not only anticipate assemblytheory, but fully incorporate it. We cer-tainly believe that differences in the his-torical development of these twoapproaches, and in the taxa and ecosys-tems in which they have been studied,have tended to obscure the similaritiesbetween them, but have also allowed eachto develop its own set of hypotheses. Foreach approach, comparison with theother often results in enlighteninginsights into particular restored commu-nities and ecological communities in gen-eral. These insights provide a road map forfuture studies of community ecology, espe-cially as they relate to restoration. Con-versely, restoration provides uniquelypowerful opportunities for experimentalapproaches to these same questions.

ACKNOWLEDGMENTSStudents in the Winter 1999 RestorationEcology Study Seminar at the University ofCalifornia at Davis helped define and refine theideas presented here. Additional insights came


frotn Marcel Rejmanek and Ted Foin. Threeanonymous reviewers greatly improved earlydrafts. J.M. Chase was supported by a Center forPopulation Biology Fellowship from UC Davisduring the early part of this collaboration.

REFERENCESAgnew, A.D.Q. 1984. Cyclic changes of vege-

tation in the plant communities of theAberdare Mountains, Kenya. Journal of theEast African Natural History Society75(183):1-12.

Allen-Diaz, B and J.W. Bartolome. 1998.Sagebrush-grass vegetation dynatnics: com-paring classical and state-transition models.8:795-804.

Anderson, R.C., J.E. Schwegman, and M.R.Anderson. 2000. Micro-scale restoration:A 25-year history of a southern Illinois bar-rens. Restoration Ecology 8:296-306.

Bakker, J.P. and E Berendse. 1999. Constraintsin the restoration of ecological diversity ingrassland and heathland communities.Trends in Ecology and Evolution 14:63-68

Bakker, J.P., P. Poschlod, R.J. Strykstra, R.M.Bekker, and K. Thompson. 1996. Seedbanks and seed dispersal: hnportant topicsin restoration ecology. Acta Botanica Neer.landica 45:461-490.

Bazely, D.R. and R.L. Jefferies. 1986. Changesin the composition and standing crop ofsalt marsh communities in response to theremoval of a grazer. Ecology 74:693-706.

Belsky, A.J. 1986. Population and communityprocesses in a mosaic grassland in theSerengeti, Tanzania. Journal of Ecology74:841-856.

Belyea, L.R. and J. Lancaster. 1999. Assemblyroles within a contingent ecology. Oikos86:402-416.

Bertness, M.D. and R. Callaway. 1994.Positive interactions in communities.Trends in Ecology and Evolution 9:191-193.

Brown, D. 1995. The impact of species intro-duced to control tree invasion on the veg-etation of an electrical utility right-of-way.Canadian Journal of Botany 73:1217-1228.

Buddle, C.M., J.R. Spence and D.W. Langor.2000. Succession of boreal forest spiderassemblages following wildfire and harvest-ing. Ecography 23:424-436.

Callaway, R.M. and L.R. Walker. 1997.Competition and t~acilitation: a syntheticapproach to interactions in plant commu-nities. Ecology 78:1958- 1965.

Chapman, C.A. and L.J. Chapman. 1999.Forest restoration in abandoned agricul-tural land: A case study from East Africa.Conservation Biology 13:1301-1311.

Chase, J.M. 1998. Size-structured interactionsand multiple stable states in pond foodwebs. Ph.D. dissertation, University ofChicago.

__. 1999a. Food web effects of prey size-refu-gia: Variable interactions and alternative

stable equilibria. American Naturalist 154:559-570.

__. 1999b. To grow or reproduce? The role oflife-history plasticity in food web dynamics.American Naturalist 154: 571-586.

Choi, K., J.-Y. Choi and B.M. Woo. 1988. Astudy on the plant succession and soil prop-erties of the erosion control area. ResearchReport of the Forestry Research Institute(Korea) 37:8-18.

Clark, J.S., E. Macklin and L. Wood. 1998.Stages and spatial scales of recruitment lim-itation in southern Appalachian forests.Ecological Monographs 68:213-235.

Clark, J.S. and J. Yuan. 1995. Fecundity anddispersal in plant populations: Implicationsfor structure and diversity. AmericanNaturalist 146:72-111.

Clements, EE. 1916. Plant Succession: AnAnalysis of the Development of Vegetation.Carnegie Institution of Washington.

¯ 1938. Nature and structure of climax.Journal of Ecology 24:252-282.

C0nnell J.A. and R.O. Slatyer. 1977. Mechan-isms of succession in natural communitiesand their role in community stability andorganization. American Naturalist 111:1119-1144.

Connell, J.H. and W.P. Sousa¯ 1983. On theevidence needed to judge ecological stabil-ity or persistence. American Naturalist121:789-824.

Cowles, H¯C. 1899. The ecological relations ofthe vegetation on the sand dunes of LakeMichigan. Botanical Gazette 27:95-117,167-202, 281-308, 361-391.

Darwin, C. 1872. The Origin of Species (6thEdition). 1958 Mentor Edition, PenguinBooks, New York.

Davis, M.B., R.R. Calcote, S. Sugita, and H.Takahara. 1999. Patchy invasion and theorigin of a hemlock-hardwoods forestmosaic. Ecology 79:2641- 2659.

Del Moral, R. 1998. Early succession on laharsspawned by Mount St. Helens. AmericanJournal of Botany 85: 820-828.

DaSilva, J.M.C., C. Uhl, and G. Murray, G.1996. Plant succession, landscape manage-ment, and the ecology of frugivorous birdsin abandoned Amazonian pastures. Conser.vation Biology 10:491-503.

Diamond, J.M. 1975. Assembly of species com-tnunities. Pp. 342-444 in: M.L. Cody andJ.M. Diamond (eds.) Ecology and evolutionof communities. Belknap Press, Cambridge,Mass.

Drake, J.A. 1991. Community-assembly mech-anics and the structure of an experimentalspecies ensemble. American Naturalist137:1-26.

__. 1990. The mechanics of communityassembly and succession. Journal ofTheoretical Biology 147:213-233.

Drake, J.A., T.E. Flum, G.J. Witteman, T.Voskuil, A.M. Hoylman, C. Creson, D.A.Kenny, G.R. Huxel, C.S. Larue and J.R.

Duncan.1993. The construction andassembly of an ecological community.Journal of Animal Ecology 62:117-130.

Drake, J. A., Huxel, G. R. and Hewitt, C. L.1996. Microcosms as models for generatingand testing community theory. Ecology77:670-677.

Dublin, H.T., A.R.E. Sinclair and J. McGlade.1990. Elephants and fire as causes of multi-ple stable states in the Serengeti-Marawoodlands. Journal of Animal Ecology59:1147-1164

Egler, EE. 1949. Right of way maintenance byplant community management. Acton Forest,Norfolk, Conn.

__. 1954. Vegetation science concepts. LInitial floristic composition~a factor inold-field vegetation development. Vegetatio4:412-417.

__. 1975. The plight of right of way domain.Futura Press, Mt. Kisco, NY.

Fastie, C.L. 1995. Causes and ecosystem con-sequences of multiple pathways of primarysuccession at Glacier Bay, Alaska. Ecology76:1899-1916.

Fike, J. and Niering, W.A. 1999. Four decadesof old field vegetation development andthe role of Celastrus orbiculatus in thenortheastern United States. Journal ofVegetation Science 10:483-492.

Geldenhuys, C.J. 1997. Native forest regener-ation in pine and eucalypt plantations inNorthern Province, South Africa. ForestEcology and Management 99:101-115.

Gleason, H.A. 1927. Further views on the suc-cession concept. Ecology 8: 299- 326.

Gorchov, D.L., E Comejo, C. Ascorra and M.Jaramillo. 1993. The role of seed dispersalin the natural regeneration of rain forestafter strip-cutting in the Peruvian Araazon.Vegetatio 108:339-349.

Groves, R.H. and J.J. Burdon, eds. 1986.Ecology of biological invasions: An Australianperspective. Australian Academy ofScience, Canberra.

Guariguata, M.R., R. Rheingans and EMontagnini. 1995. Early woody invasionunder tree plantations in Costa Rica:Implications for forest regeneration.Restoration Ecology 3:252-260.

Hodge S., W. Arthur , P. Mitchell . 1996.Effects of temporal priority on interspecificinteractions and community development.Oikos 76:350-358.

Holl, K.D. 1999. Tropical forest recovery andrestoration. Trends in Ecology and Evolution14:378-379.

Houssard, C. and J. Escarre. 1995. Variationand covariation among life-history traits inRumex acetosella from a successional old-field gradient. Oecologia 102:70-80.

Hughes J.W. and T.J. Fahey. 1988. Seed dis-persal and colonization in a disturbednorthern hardwood forest. Bulletin of theTorrey Botanical Club 115:89-99.

Jordan, W.R., M.E. Gilpin and J.D. Aber. 1987.


Restoration ecology: Ecological restorationas a technique for basic research. Pp. 3-22 inW.R. Jordan, M.E. Gilpin and J.D. Aber(eds.) Restoration ecology: A syntheticapproach to ecological research. CambridgeUniversity Press, Cambridge.

Keenan, R. D. Lamb, O. Woldring, T. Irvineand R. Jensen. 1997. Restoration of plantbiodiversity beneath tropical tree planta-tions in northern Australia. Forest Ecologyand Management 99:117~ 131.

Kettle, W.D., P.M. Rich, K. Kindscher, G.L.Pittman and P. Fu. 2000. Land-use historyin ecosystem restoration: A 40-year studyin the prairie-forest ecotone. RestorationEcology 8:307-317.

Kim, J.H. 1997. The role of herbivory, anddirect and indirect interactions, in algalsuccession. Journal of Experimental MarineBiology and Ecology 217:119-135.

Knowlton, N. 1992. Thresholds and multiplestable states in coral reef dynamics.American Zoologist 32:674-682.

Kotanen, P.M. 1997. Effects of experimentalsoil disturbance on revegetation by nativesand exotics in coastal California meadows.Journal of Applied Ecology 34:631-644.

Lamb, D., Parrotta, J., Keenan, R., Tucker, N.,1997. Rejoining habitat fragments:Restoring degraded rainforest lands. In:Laurance, W.E, Bierregaard, R.O. (Eds.),Tropical Forest Remnants: Ecology, manage-ment, and conservation of fragmented com-munities, University of Chicago Press,Chicago, pp. 366-385.

Lamb, D. 1998. Large-scale ecological restora-tion of degraded tropical forest lands: Thepotential role of timber plantations.Restoration Ecology 6:271- 279.

Law, R. and R.D. Morton. 1993. Alternativepermanent states of ecological communi-ties. Ecology 74:1347-1361.

__. 1996. Permanence and assembly of eco-logical communities. Ecology. 77:762-775.

Lawton, J.H. 1987. Are there assembly rulesfor successional communities? Pp. 225-244in: A.J. Gray, M.J. Crawley and P.J.Edwards (eds.) Colonization, succession andstability. Blackwell Scientific, Oxford.

Leps, J., J. Michalek, O. Rauch and P. Uhlik.2000. Early succession on plots with theupper soil horizon removed. Journal ofVegetation Science 11: 259- 264.

Lewontin, R.C. 1969. The meaning of stabil-ity. Pp. 13-24 in Diversity and stability in eco-logical systems. Brookhaven Symposia inBiology Number 22, BrookJaaven, NY.

Lockwood, J.A and D.R. Lockwood. 1993.Catastrophe theory--a unified paradigmfor rangeland ecosystem dynamics. JournalOf Range Management 46:282-288.

Lockwood, J.L. 1997. An alternative to suc-cession: Assembly rules offer guide torestoration efforts. Restoration and Manage-ment Notes !5:45-50.

Lockwood, J.L., R.D. Powel!, M.P. Nott, S.L.

Pimm. 1997. Assembling ecological coramu-nities in time and space. Oikos 80:549-553.

Luken, L.O. 1990, Directing ecological succes.sion. Chapman and Hall, NY.

Ludwig, D., B. Walker, and C. S. Holling.1997. Sustainability, stability, andresilience. Conservation Ecology [online]1 (1): 7. Available from the Internet. URL:http://www.consecol.org/voll/issl/art7

Luh, H.K. and S.I. Pimm. 1993. The assemblyof ecological communities--a minimalistapproach. Journal of Animal Ecology 62:749-765.

Mallik, A.U. 1995. Conversion of temperateforests into heaths - role of ecosystem dis-turbance and ericaceous plants. Environ.mental Management 19: (5) 675-684.

Matlack, G.R. 1994. Plant species migration ina mixed-history forest landscape in easternNorth America. Ecology 75:1491-1502.

Maxwell, B.D., J.C. Zasada and S.R.Radosevich. 1993. Simulation of salmon-berry and thimbleberry population estab-lishment and growth. Canadian Journal ofForest Research. 23:2194-2203.

May, R. 1977, Thresholds and breakpoints inecosystems with a multiplicity of stablestates. Nature 269: 471-477.

McCook, L.J. 1994. Understanding ecologicalcommunity succession---causal models andtheories, a review. Vegetatio 110:115-147.

McCune, B. and T.EH. Allen. 1985. Will sim-ilar forests develop on similar sites?Canadian Journal of Botany 63:367-376.

Mclntosh, R.P. 1999. The succession of suc-cession: a lexical chronology. Bulletin ofthe Ecological Society of America 80:256-265

McIntyre, S. and S. Lavorel. 1994. Predictingrichness of native, rare, and exotic plants inresponse to habitat and disturbance vari-ables across a variegated landscape.Conservation Biology 8:521-531.

McNaughton, S.J. 1983 Serengeti grasslandecology: The role of composite environ-mental factors and contingency in commu-nity organization. Ecological Monographs53:291-320.

Morton, R.D., R. Law, S.L. Pimm and J.A.Drake. 1996. On models for assemblingecological communities. Oikos 75:493-499.

Moyle, P. B. and T. Light. 1996. Biologicalinvasion of fresh water: Empirical rules andassembly theory. Biological Conservation78:149-161.

Niering, W.A. 1987.Vegetation dynamics(succession and climax) in relation to plantcommunity management. ConservationBiology 1:287-295.

Niering, W.A., G.D. Dreyer, EE. Egler, and J.P.Anderson. 1986. Stability of a Viburnumlentago shrub community after 30 years.Bulletin of the Torrey Botanical Club 113:23-27.

Niering, W.A. and EE. Egler. 1955. A shrubcommunity of Viburnum lentago, stableafter 25 years. Ecology 36:356-60.

Niering, W.A. and R.H. Goodwin. 1974. Crea-tion of relatively stable shrublands withherbicides: Arresting succession on rights-of-way and pastureland. Ecology 55:784-795.

Oberhauser, U. 1997. Secondary forest regen-eration beneath pine (Pinus kesiya) planta-tions in the northem Thai highlands: Achronosequence study. Forest Ecology andManagement 99:171-183.

Olff, H., EW.N. Veto, J. Bokdam and others.1999. Shifting mosaics in grazed woodlandsdriven by the alternation of plant facilitationand competition. Plant Biology 1: 127-137.

Pacala, S.W. and M. Rees, 1998. Models sug-gesting field experiments to test twohypotheses explaining successional diver-sity. American Naturalist 152:729-737.

Packard, S. 1994. Successional restoration:Thinking like a prairie. Restoration andManagement Notes. 12(1):32-39.

Paine, R.T., J.C. Costillo and J. Cancino.1985. Perturbation and recovery patterns ofstarfish-dominated intertidal assemblagesin Chile, New Zealand, and WashingtonState. American Naturalist 125:679-691.

Palmer, M.A., R.E Ambrose and N.L. Poff.1997. Ecological theory and communityrestoration ecology. Restoration Ecology5:291-300.

Parrotta J.A., Turnbull J.W., Jones N. 1997a.Introduction--Catalyzing native forestregeneration on degraded tropical lands.Forest Ecology and Management 99:1-7.

Parrotta, J.A., Knowles, H.K., Wunderle, J.M.,1997b. Development of floristic diversity in10-year-old restoration forests on a bauxitemine in Amazonia. Forest Ecology andManagement 99: 21-42.

Petraitis, P.S. and R.E. Latham. 1999. Theimportance of scale in testing the origins ofalternative community states. Ecology80:429-442.

Pickett, S.T.A., S.L. Collins and J.J. Armesto.1987. Models, mechanisms and pathwaysof succession. Botanical Review 53:335-371.

Pimm, S.L. 1991. The balance of nature? Eco.logical issues in the conservation of species andcommunities. University of Chicago Press,Chicago.

Pritchett, D.A. 1997. Maybe, sometimes: Apractitioner replies. Restoration and Man-agement Notes 15 (1):51.

Pyle, L.L. 1995. Effects of disturbance on her-baceous exotic plant species on the flood-plain of the Potomac River. AmericanMidland Naturalist 13:244- 253.

Rejmanek, M. 1996. A theory of seed plantinvasiveness: The first sketch. BiologicalConservation 78:171-181.

Rejmanek, M. and D.M. Richardson. Whatattributes make some plant species moreinvasive? Ecology 77:1655-1661.

Reynolds, H. L. and S. W. Pacala. 1993. Ananalytical treatment of root-to-shoot ratioand plant competition for soil nutrient andlight. American Naturalist 141: 51-70.


Rhoades, C.C., G.E. Eckert and D.C. Cole-man. 1998. Effect of pasture trees on soilnitrogen and organic matter: Implicationsfor tropical montane forest restoration.Restoration Ecology 6: 262-270.

Ridley, H.N. 1930. The Dispersal of PlantsThroughout the World. L. Reeve, Ashford,England.

Rietkerk, M. and J. van de Koppel. 1997.Alternative stable states and thresholdeffects in semi-arid grazing systems. Oikos79:69-76.

Robinson, G.R. and S.N. Handel. 1993. Forestrestoration on a closed landfill: Rapid addi-tion of new species by bird dispersal.Conservation Biology 7:271-278.

Robinson, J.V. and J.E. Dickerson. 1987. Doesinvasion sequence affect community com-position? Ecology 68:587-595.

Robinson, J.V. and M.A. Edgemon. 1988. Anexperimental evaluation of the effects ofinvasion history on community structure.Ecology 69:1410-1417.

¯ 1989. The effect of predation on thestructure and invasibility of assembledcommunities. Oecologia 79:150- 157.

Rousset, O. and J. Lepart. 1999. Shrub facili-tation of Quercus humilis regeneration insuccession on calcareous grasslands. Journalof Vegetation Science 10:493-502.

Samuels, C.L. and J.A. Drake. 1997. Divergentperspectives on community divergence.Trends in Ecology and Evolution 12:427-432.

Scowcroft, P.G. and J. Jeffrey. 1999. Potentialsignificance of frost, topographic relief, andAcacia koa stands to restoration of mesicHawaiian forests on abandoned rangeland.Forest Ecology and Management 114:447-458.

Sheil, D. 1999. Developing tests of succes-sional hypotheses with size-structured pop-ulations, and an assessment usinglong-term data from a Ugandan rain forest.Plant Ecology 140:117-127.

Stampfli, A. and Zeiter, M. 1999. Plant speciesdecline due to abandonment of meadowscannot easily be reversed by mowing: Acase study from the southern Alps. Journalof Vegetation Science 10:151-164.

Sutherland, J.P. 1974. Multiple stable points inecological communities. AmericanNaturalist 108:859-873.

1990. Perturbations, resistance, andalternative views of the existence of multi-ple stable points in nature. AmericanNaturalist 136:270- 275.

Stohlgren, T.J., D. Binkley, G.W. Chong, andothers. 1999. Exotic plant species invadehot spots of native plant diversity. Ecologi-cal Monographs 69:25-46.

Strohmayer, K.A.K. and R.J. Warren. 1997.Are overabundant deer herds in the easternUnited States creating alternate stablestates in forest plant communities? WildlifeSociety Bulletin 25:227-234.

Stryksra, R.J., R.M. Bakker and J.P. Bakker.1998. Assessment of dispersule availability:Its practical use in restoration manage-ment. Acta Botanica Neerlandica 47:57-70.

Symstad, A.J. 2000. A test of the effects offunctional group richness and compositionon grassland invasibility. Ecology 81:99-109.

Tansley, A.G. 1935. The use and abuse ofvegetation concepts and terms. Ecology16:284-307.

Tappeiner, J., J. Zasada, P. Ryan and M. Newton.1991. Salmonberry clonal and populationstructure--the basis for a persistent cover.Ecology 72:609- 618.

Tilman, D. 1997. Community invasibility,recruitment limitation, and grassland bio-diversity. Ecology 78: 81-92.

Van Andel, J., J.P. Bakker and A.P. Grootjans.1993. Mechanisms of vegetation succession:A review of concepts and perspectives. ActaBotanica N eerlandica 42:413-433.

Van Breemen, N. 1995. How Sphagnum bogsdown other plants. Trends in Ecology andEvolution 10:270-275.

Walker, S. 1997. Models of vegetation dynam-ics in semi-arid vegetation: Application tolowland central Otago, New Zealand. NewZealand Journal of Ecology 21:129-140.

Watt, A.S. 1947. Pattern and process in theplant community. Journal of Ecology 35:1-22.

Weatherby, A.J., P.H. Warren. and R. Law.1998. Coexistence and collapse: An exper-imental investigation of the persistentcommunities of a protist species pool.Journal of Animal Ecology 67:554-566.

Weiher, E, G.D.P. Clarke, P.A. Keddy. 1998.Community assembly rules, morphologicaldispersion, and the coexistence of plantspecies. Oikos 81:309-322.

Weiher, E. and P.A. Keddy (eds.). 1999. Eco-logical assembly rules: Perspectives, advances,retreats. Cambridge University Press, Cam-bridge, New York.

Westoby, M., B.H. Walker and I Noy-Meir.1989. Opportunistic management forrangelands not at equilibrium. Journal of

Range Management 42:195- 205.Whisenant, S.G., T.L. Thurow and S.J.

Maranz. 1995. Initiating autogenic restora-tion on shallow semiarid sites. RestorationEcology 3:61-67.

Yeaton, R.I. 1978. A cyclic relationshipbetween Larrea tridentata and Opuntia lep.tocaulis in the northern ChihuahuanDesert. Journal of Ecology 66:651- 656.

Young, T.P. 1996. High mountain andafroalpine ecosystems. Pp. 401-424 in: T.R.McClanahan and T.P. Young (eds.)Ecosystems of East Africa and their Conser-vation, Oxford University Press, New York.

¯ 2000. Restoration ecology and conserva-tion biology. Biological Conservation 92:73-83.

Young, T.P. and E Chan. 1998. Establishmentof deergrass (Muhlenbergia rigens) at arestoration site, and its prospects for woodyplant suppression in transmission rights ofway. Grasslands 8(3):3-4,10.

Young, T.P. and W.K. Lindsay. 1988. Role ofeven-aged population structure in the dis-appearance of Acacia xanthophloea wood-lands. African Journal of Ecology 26:69-72.

Young, T.P., N. Patridge and A. Macrae. 1995.Long-term glades in acacia bushland andtheir edge effects in Laikipia, Kenya.Ecological Applications 5:97-108.

Young, T.P. and M.M. Peacock. 1992. Giantsenecios and the alpine vegetation of MountKenya. Journal of Ecology 80:141-148.

Truman P. Young is an associate professor in theDepartment of Environmental Horticulture at theUniversity of California at Davis, 95616, Tel.(530) 754-9925; Fax (530) 752-1819; e-mail [email protected].

Jonathan Chase is an assistant professor in theDepartment of Biological Science at the University ofPittsburgh, 4400 Fifth Avenue, Pittsburgh PA15260, Tel. (412) 624-4265; Fax (412) 624-4759; e-mail jchase [email protected] .

Russell T. Huddleston was a graduate student at UCDavis, and is now a biologist at CH2M Hill, 2485Natomas Park Drive, Suite 600, Sacramento, CA95833, Tel. (916) 920-0300; Fax (916) 920-8463; e-mail rhuddle [email protected].


community succession and assembly - lab home...

Documents