Ecology, 89(7), 2008, pp. 1931–1940
? 2008 by the Ecological Society of America
COMMUNITY ASSEMBLY IN THE PRESENCE OF DISTURBANCE:
A MICROCOSM EXPERIMENT
LIN JIANG1AND SHIVANI N. PATEL
School of Biology, Georgia Institute of Technology, 310 Ferst Drive, Atlanta, Georgia 30332 USA
the likelihood of alternative community stable states and how the history of community
assembly affects the relationship between disturbance and species diversity. Using microbial
communities comprising bacterivorous ciliated protists assembled in laboratory microcosms,
we experimentally investigated these questions by independently manipulating the intensity of
disturbance (in the form of density-independent mortality) and community assembly history
(including a control treatment with simultaneous species introduction and five sequential
assembly treatments). Species diversity patterns consistent with the intermediate disturbance
hypothesis emerged in the controls, as several species showed responses indicative of a trade-
off between competitive ability and ability to recover from disturbance. Species diversity in
communities with sequential assembly, however, generally declined with disturbance, owing to
the increased extinction risk of later colonizers at the intermediate level of disturbance.
Similarities among communities subjected to different assembly histories increased with
disturbance, a result due possibly to increasing disturbance reducing the importance of
competition and hence priority effects. This finding is most consistent with the idea that
increasing disturbance tends to reduce the likelihood of alternative stable states. Collectively,
these results indicate the strong interactive effects of disturbance and assembly history on the
structure of ecological communities.
Ecologists know relatively little about the manner in which disturbance affects
hypothesis; protists; species diversity.
alternative stable states; community assembly; disturbance; intermediate disturbance
To what extent does the structure of ecological
communities depend upon their history? Despite the
early recognition of the potential importance of history
in shaping community structure (Gleason 1927, Egler
1954, Diamond 1975), this fundamental question
remains largely unresolved. On the one hand, ample
theoretical and experimental evidence suggests that
multiple community configurations that differ in species
composition (i.e., alternative stable states; Lewontin
1969) can arise from differences in the historical
sequence of species colonization, even under identical
environmental conditions and common species pools
(e.g., Drake 1991, Law and Morton 1993, Blaustein and
Margalit 1996, Price and Morin 2004, Fukami et al.
2007; reviewed by Samuels and Drake 1997, Chase
2003a). On the other hand, theory and experiments also
show that communities with different assembly histories
may nevertheless attain similar species composition and
abundance, suggesting little or no role of history (e.g.,
Sommer 1991, Law and Morton 1996, Chase and
Leibold 2003, Fukami 2004a). These conflicting patterns
have led ecologists to look into factors that may
influence the role of history in community assembly,
including the size of species pool (Law and Morton
1993, Chase 2003a, Fukami 2004b), productivity (Chase
2003a, b, Fukami and Morin 2003), predation (Morin
1984, Louette and De Meester 2007), ecosystem size
(Drake 1991, Fukami 2004a), dispersal rate (Robinson
and Edgemon 1988, Lockwood et al. 1997), and
disturbance (Inouye and Tilman 1995, Chase 2003a,
2007, Trexler et al. 2005). Here we examine the effect of
disturbance on the historical contingency of community
assembly, a topic that has received relatively little
theoretical or experimental treatment (but see Chase
Disturbance sets the stage for community assembly to
occur by killing or damaging resident species and
releasing resources for prospective colonizers. Further-
more, recurring disturbances that differ in intensity
and/or frequency may alter community membership and
species abundance, which may in turn affect the
likelihood of alternative stable states (Belyea and
Lancaster 1999, Chase 2003a, Didham et al. 2005,
Fukami and Lee 2006). Ecologists, however, disagree on
how disturbance affects the historical contingency of
community assembly. Some have proposed that differ-
ence in assembly history is more likely to lead to
alternative stable states in habitats experiencing low
rates of disturbance (Belyea and Lancaster 1999, Chase
2003a, Fukami and Lee 2006), whereas others have
Manuscript received 1 August 2007; revised 19 October 2007;
accepted 5 November 2007. Corresponding Editor: B. J. M.
proposed an increased likelihood of alternative stable
states in habitats with high rates of disturbance (Didham
et al. 2005, Didham and Norton 2006). Proponents of
both hypotheses have drawn observational evidence to
support their arguments (Chase 2003a, Didham et al.
2005). The presence of potential confounding factors in
observational studies, however, precludes an unambig-
uous evaluation of these contrasting ideas. The only
experimental work pertinent to this subject, conducted
in small freshwater ponds (Chase 2007), seems to
support the small likelihood of alternative stable states
in disturbed habitats. However, interpretation of this
experiment was clouded by the lack of control over
assembly history in disturbed ponds and over the degree
of environmental heterogeneity that may covary with
A well-known aspect of disturbance is its effect on
species diversity. Much research on this subject has
surrounded the intermediate disturbance hypothesis
(IDH), which predicts a maximum level of species
diversity at intermediate levels of disturbance (Connell
1978). The IDH, however, has received mixed empirical
support: besides the pattern of peaked diversity at
intermediate disturbance predicted by the IDH (Shea et
al. 2004), other disturbance–diversity relationships also
are common (Mackey and Currie 2001). Although
various hypotheses have been proposed to explain
deviations from the IDH (Huston 1994, Wootton
1998, Mackey and Currie 2000, Chase and Leibold
2003), none have taken community assembly into
account. The IDH often rests on the assumption that
species exhibit a trade-off between competitive ability
and ability to cope with disturbance (Connell 1978,
Petraitis et al. 1989, Chase and Leibold 2003; but see
Kadmon and Benjamini 2006). This trade-off, however,
could be weakened or disrupted in communities
demonstrating historical contingencies. For instance,
disturbance-tolerant inferior competitors may be unable
to establish in a locality where superior competitors, as
early colonizers, have already built large populations. In
this case, the lack of trade-off among coexisting species
may produce patterns inconsistent with the IDH.
Here we report the results of a laboratory microcosm
experiment that independently manipulated assembly
history and disturbance intensity to examine their effects
on community structure. We used fast-reproducing
ciliated protists as experimental organisms, whose short
generation times enabled us to collect multigenerational
data that permit more unambiguous identification of
alternative stable states (Connell and Sousa 1983). The
small scale of laboratory microcosms minimized spatial
heterogeneity that is often present in larger systems. The
experiment aimed to address two specific questions.
First, how does the rate of disturbance affect the
prevalence of alternative stable states associated with
the historical contingency of community assembly?
Second, how does the history of community assembly
affect the disturbance–diversity relationship?
MATERIALS AND METHODS
The experiment used a total of 10 ciliated protist
species, which were randomly assigned into five groups,
each containing two species (Table 1). Most of these
species were isolated from small freshwater ponds, and a
few were obtained from biological supply houses. All 10
species are bacterivores that can survive on bacteria
alone, although Blepharisma americanum and Tetrahy-
mena vorax can also form carnivorous individuals that
prey upon smaller conspecific and heterospecific indi-
viduals. Prior to the experiment, these species had been
separately cultured in the laboratory on three bacterial
species (Bacillus cereus, Bacillus subtilis, and Serratia
marcescens) for numerous generations.
Microcosms were 250-mL Pyrex glass bottles each
filled with 100 mL of medium. The medium was
prepared in large Erlenmeyer flasks using a formula of
0.55 g of protozoan pellet (Carolina Biological Supply,
Burlington, North Carolina, USA) per 1 L of deionized
water. This medium supported the growth of bacterial
food for the protists used in the experiment. Experi-
mental containers and medium were autoclaved, after
which the medium was inoculated with the bacterial
assemblage present in protist stock cultures. The
bacterial inoculum was prepared by mixing samples of
all protist stock cultures and running the mixed sample
through a 1.0-lm sterile filter to remove the protists. The
bacterized medium was distributed into individual
microcosms 24 h after bacterial inoculation. Each
microcosm also received two sterilized wheat seeds as
an additional source of carbon.
We used a two-way factorial design with disturbance
intensity and assembly sequence as the two main factors.
Disturbance, which took the form of nondiscriminative
density-independent mortality, had three levels: low
(10% weekly mortality), intermediate (50% weekly
mortality), and high (90% weekly mortality). Distur-
bance was imposed by sonicating different volumes of
medium (low, 10 mL; intermediate, 50 mL; and high, 90
mL) in experimental microcosms using a Sonic Dis-
membrator Model 100 (Fisher Scientific, Waltham,
Massachusetts, USA) at full power for 1 min. Micro-
scopic inspection indicated this procedure effectively
killed all living protists. The treated medium was
immediately returned to its original microcosm after
sonication. The assembly sequence factor had six
Species composition of each group used to assemble
Paramecium sp., Tetrahymena vorax
Paramecium aurelia, Blepharisma americanum
Paramecium bursaria, Colpidium kleini
Halteria sp., Spirostomum sp.
Glaucoma sp., Loxocephalus sp.
Note: Each of the five groups contained two species
randomly selected from the 10-species pool.
LIN JIANG AND SHIVANI N. PATEL1932 Ecology, Vol. 89, No. 7
treatments, including a control in which all protist
species were simultaneously introduced during week 1
and five sequential assembly treatments in which the five
groups were sequentially introduced in a five-week span
(Table 2). The inoculum of each protist species consisted
of ;100 individuals. Protist stock cultures were sampled
before species introduction to estimate population
densities and to determine volumes of the inocula
corresponding to 100 individuals. Because the age of
stock cultures may affect protist species traits (e.g., body
size, growth rate) and species introduction events were
distributed over a relatively long period of five weeks, it
was necessary to standardize the age of stock cultures
upon each species introduction. Fresh stock cultures
were thus established every week starting from two
weeks prior to the experiment until week 3, and two-
week-old stock cultures were always used to inoculate
Each treatment combination was replicated three
times, resulting in 54 microcosms (3 disturbance regimes
3 6 assembly sequences 3 3 replicates). During the
experiment, all microcosms were situated in a single
incubator at 228C with a 12/12 light/dark cycle. The
experiment lasted 11 weeks, including 46 d after the last
species introduction. Given the short generation times of
the protists (Blepharisma, 1–2 d; others, ,24 h), this
period encompassed dozens of complete turnovers of
even the slowest-growing community member, which
was sufficient for species interactions to shape the
structure of these fast-reproducing protist communities
Weekly samples were taken from each microcosm to
estimate the density of protist species. During sampling,
;0.4 mL of medium was withdrawn from each
microcosm after mixing and distributed as multiple
drops onto a preweighed petri dish; the exact mass
(’volume) of the sample was determined using an
analytic balance. The number of individuals of each
protist species in the sample was enumerated under a
stereoscopic microscope. To replenish nutrient and
remove excessive metabolic wastes, 10 mL medium in
each microcosm was replaced with fresh sterile medium
We used repeated-measures ANOVA (rm-ANOVA)
to test for the effect of disturbance intensity on
similarities between communities subjected to different
assembly histories, measured by the Bray-Curtis simi-
larity index (Bray and Curtis 1957). The Bray-Curtis
index ranges from 0 to 1, with increasing values
indicating increased similarities among communities.
Low index values, calculated for non-transient steady-
state communities, would point to the possible existence
of alternative stable states. In addition to similarity
indices, we also used MANOVAs to directly discern
whether assembly sequence and disturbance intensity
affected community structure. Dependent variables in
the MANOVAs were population densities of the 10
protist species at the last sampling date (week 11). We
also used rm-ANOVA to assess the effects of distur-
bance and assembly sequence on community species
richness. Within each assembly sequence, we conducted
Tukey’s hsd tests at each sampling date to uncover how
community similarity and species richness changed with
disturbance. Following Moran (2003), we did not adjust
probability values of these tests using the Bonferroni
procedure due to its overly conservative nature. We used
ANOVAs to assess the effects of assembly sequence and
disturbance on individual species densities, again using
only data from the final sampling date. Data on
population densities were log-transformed (log10[num-
ber of individiuals/mL þ 1]) before analyses.
Similarity among communities subjected to different
assembly histories, represented by the Bray-Curtis index,
differed significantly among disturbance levels (Fig. 1;
rm-ANOVA, F ¼ 13.94, df ¼ 2, 402, P , 0.0001). This
difference was mainly caused by significantly larger
similarity values at higher rates of disturbance during
early weeks (Fig. 1; Tukey’s hsd tests). Community
similarity showed different temporal patterns under
different disturbance intensities: it increased over time
under low disturbance, first increased then decreased
under intermediate disturbance, and remained about the
same level under high disturbance (Fig. 1), resulting in a
significant time effect (F¼20.67, df¼6, 397, P , 0.0001)
and a significant disturbance3time effect (F¼10.28, df
¼12, 794, P , 0.0001) in the rm-ANOVA. By the end of
the experiment, community similarities at the low and
high levels of disturbances were similar, whereas
similarity at the intermediate level of disturbance was
significantly lower (Fig. 1; Tukey’s hsd test). The decline
in similarity under intermediate disturbance was largely
due to the difference between the controls and sequential
assembly treatments that grew greater during the later
TABLE 2. Community assembly sequences used in the experiment.
Week ControlSequence 1Sequence 2 Sequence 3Sequence 4 Sequence 5
all groupsgroup 1
Note: See Table 1 for species composition of each group.
July 20081933COMMUNITY ASSEMBLY AND DISTURBANCE
stage of the experiment. When only the five sequential
assembly treatments were used for similarity index
calculations, no significant temporal decline in commu-
nity similarity was observed under intermediate distur-
bance and similarity for the three levels of disturbance
eventually converged (results not shown). Despite this
convergence in community similarity, MANOVA on
species abundances at the last sampling date revealed a
significant disturbance 3 sequence effect (F ¼ 1.49, df ¼
80, 192.15, P ¼ 0.0145), suggesting differential effects of
history under different levels of disturbance. MANOVA
conducted at each disturbance level revealed highly
significant differences under low disturbance (F ¼ 3.50,
df ¼ 35, 27.67, P ¼ 0.0006), significant differences under
high disturbance (F¼1.86, df¼30, 30, P¼0.0471), and
nonsignificant differences under intermediate distur-
bance (F ¼ 1.29, df ¼ 40, 20.589, P ¼ 0.2547). When
the controls were excluded from MANOVA, there were
highly significant differences under low disturbance (F¼
3.71, df¼28, 15.844, P ¼0.0042), significant differences
under intermediate disturbance (F ¼ 2.47, df ¼ 32,
12.659, P¼0.0445), and nonsignificant differences under
high disturbance (F¼1.73, df¼24, 18.653, P¼0.1159).
Thus both similarity indices and MANOVA indicated
that communities with different assembly histories
became more similar with increasing disturbance.
Species richness declined over time in all microcosms
(rm-ANOVA, time, F ¼ 48.12, df ¼ 6, 31, P , 0.0001;
Fig. 2), and the tempo of decline varied with assembly
sequence (sequence 3 time, F ¼ 2.15, df ¼ 30, 126, P ¼
0.0017) and disturbance intensity (disturbance 3 time,
F ¼ 5.65, df ¼ 12, 62, P , 0.0001). Both assembly
sequence (F ¼ 28.52, df ¼ 2, 36, P , 0.0001) and
disturbance (F ¼ 4.61, df ¼ 5, 36, P ¼ 0.0024) had
significant main effects on species richness, with no
significant interactions between the two factors (F ¼
1.30, df ¼ 10, 36, P ¼ 0.2661). While the effects of
assembly sequence appeared to be idiosyncratic, increas-
ing disturbance, in general, tended to reduce species
richness (Fig. 2). Despite this average negative effect of
disturbance and the nonsignificant sequence 3 distur-
bance term in the rm-ANOVA, richness patterns in the
controls differed markedly from the five sequential
assembly treatments. In the controls, richness under
intermediate disturbance was significantly higher than
those under low and high disturbances during weeks 8
and 9 and was significantly higher than those under high
but not low disturbances during the following weeks
(Fig. 2A; Tukey’s hsd tests). In the five sequential
assembly treatments, however, richness never peaked at
intermediate disturbance and generally declined with
disturbance (Fig. 2B–F).
Individual species patterns
Responses to experimental treatments differed con-
siderably among species. Tetrahymena vorax and Hal-
teria went extinct during early stages of the experiment
in all microcosms, regardless of the level of disturbance
or the history of community assembly (Appendix A:
Figs. A1–A6). By contrast, Colpidium persisted at
appreciable densities in all treatments (Fig. 3; Appendix
A: Figs. A1–A6), and its final density was significantly
affected by assembly sequence but not disturbance
(Appendix B: Table B1). Densities of the other species
were all affected by disturbance (all P’s , 0.05;
Appendix B: Table B1), though their responses to
disturbance differed. Many species declined in abun-
dance with increasing disturbance. For instance, Para-
mecium bursaria persisted in all microcosms under low
disturbance, persisted in some microcosms under
intermediate disturbance, but went extinct in all
microcosms under high disturbance (Fig. 3; Appendix
A: Figs. A1–A6). Blepharisma show extremely similar
patterns to P. bursaria, except it was present in one high-
disturbance microcosm at the end of the experiment. By
contrast, some species, including Loxocephalus and
Paramecium aurelia, appeared to have benefited from
disturbance. Loxocephalus was unable to persist in low-
disturbance environments; however, it was able to
sustain its populations in at least some replicates with
among communities subjected to different assembly regimes at
each of the three disturbance levels through time. Mean
similarity was obtained by averaging index values from all
possible pairwise comparisons of communities. Each commu-
nity consisted of 10 ciliated protist species, assigned to five
groups (each containing two species; see Tables 1 and 2) with
different assembly sequences of the groups. Communities
experienced three levels of disturbance (low, 10% weekly
mortality; intermediate, 50% weekly mortality; and high, 90%
weekly mortality). Different lowercase letters associated with
symbols represent significant differences at P ¼ 0.05 in a
Tukey’s hsd test conducted for each sampling date; insignificant
differences are not shown. Values are means 6 SE.
Similarity (measured by the Bray-Curtis index)
LIN JIANG AND SHIVANI N. PATEL1934 Ecology, Vol. 89, No. 7
increased rate of disturbances, especially in the high-
disturbance environments (Fig. 2; Appendix A: Figs.
A1–A6). Similar pattern held for P. aurelia, although it
was only completely eliminated under low disturbance in
the controls (Fig. 3; Appendix A: Figs. A1–A6).
There has been much recent debate regarding how
disturbance affects the likelihood of alternative stable
states in community structure (Chase 2003a, Didham et
al. 2005, Didham and Norton 2006, 2007, Fukami and
Lee 2006, Mason et al. 2007). This debate, however, has
been largely based on discussions of conceptual models
and observational evidence, and more objective evalu-
ations of these ideas using mathematical models or
manipulative experiments have been scarce (but see
Chase 2007). Here we provide a direct experimental test
of how disturbance affects the possibility of alternative
stable states associated with different assembly histories.
Similarity among communities subjected to different
assembly histories was higher in microcosms experienc-
ing higher rates of disturbance. This is most consistent
consisted of 10 ciliated protist species, assigned to five groups (each containing two species; see Tables 1 and 2) with different
assembly sequences of the groups. Communities experienced three levels of disturbance (low, 10% weekly mortality; intermediate,
50% weekly mortality; and high, 90% weekly mortality). Different lowercase letters represent significant differences at P¼0.05 in a
Tukey’s hsd test conducted for each sampling date; insignificant differences are not shown. Values are means 6 SE.
Species richness at each of three disturbance levels through time for the control and sequences 1–5. Each community
July 20081935 COMMUNITY ASSEMBLY AND DISTURBANCE
with the idea that alternative community stable states
are more likely to arise from different assembly histories
in environments characterized by lower rates of distur-
bance than by higher rates of disturbance (Belyea and
Lancaster 1999, Chase 2003a, Fukami and Lee 2006). In
an experiment conducted in small freshwater ponds,
Chase (2007) also showed that similarity was greater
among pond communities experiencing drought than
those lacking drought disturbance. Several observation-
al studies also reported findings consistent with this idea.
the control and sequences 1–5. Values are means þ SE with density data originally measured as no. individuals/mL and log10-
transformed prior to analyses.
Population density of each protist species at each of the three disturbance levels on the final sampling date (week 11) for
LIN JIANG AND SHIVANI N. PATEL 1936Ecology, Vol. 89, No. 7
For instance, Inouye and Tilman (1995) examined
similarity in plant community structure in several old
fields and in a savanna with no historical agricultural
practice. They found that communities within each of
the two recently disturbed old fields were more similar,
whereas communities that were in much older old fields
(i.e., longer time since disturbance) and native savanna
were less similar. Likewise, Chase (2003a) compared
animal species composition in freshwater ponds classi-
fied into different permanence categories (characterized
by different drying [disturbance] frequencies) and found
that similarity among pond communities increased as
ponds became less permanent (i.e., more frequent
drying). Trexler et al. (2005) obtained essentially the
same results when examining the effect of drying
frequency on small fish communities in the Everglades.
Given that communities at the end of our experiment
did not appear to have settled into distinct states
characterized by different species composition (Appen-
dix A: Figs. A1–A6), one could argue that increased
community similarity under a higher rate of disturbance
may not directly translate into the reduced likelihood of
alternative stable states under a higher rate of distur-
bance. Our experiment, however, may have yet reached
steady states. This is despite the fact that its duration
was sufficient for dozens of complete community
turnovers, longer than those of most existing experi-
ments on alternative stable states (reviewed by Schroder
et al. 2005). Nevertheless, the structure of communities
going through transient phases may simply differ as the
result of community divergence towards alternative
stable states, barring any stochastic events that can lead
to unexpected divergence in community structure
(random divergence, sensu Schroder et al. 2005). Thus
differences in transient community patterns could be
indicative of potential differences in alternative commu-
nity stable states (Didham et al. 2005). Indeed,
divergence in species composition occurred in the
absence of disturbance at the end of a 15-week assembly
experiment that used the same set of protist species,
albeit a different set of colonization sequences (L. Jiang
and Z. Pu, unpublished manuscript). Moreover, even if
communities may eventually converge to similar struc-
tures in the absence of alternative stable states, the
presence of possible long-term transients in many
systems (Hastings 2004) suggests the need to understand
the role of community assembly for long-term transient
dynamics (Fukami 2004a, Schroder et al. 2005). Fukami
(2004a) has experimentally demonstrated that assembly
history can affect species diversity for long periods of
time before its effect disappears. Our finding would be
qualitatively similar to that of Fukami (2004a) should
community convergence occur. Overall, our results
indicate that disturbance can interact with assembly
history to affect community structure for an ecologically
meaningful time span, regardless of whether observed
patterns represent transient or steady states.
Why were communities with different assembly
sequences more similar with increasing disturbance?
The answer seems to lie in the fact that competition
became less important in habitats characterized by high
disturbance rates, an idea first proposed by Grime
(1979). Recall that individual species responses to
disturbance were consistent with a trade-off between
species’ ability to compete and their ability to cope with
disturbance (Fig. 3). Increasing disturbance caused
density reduction and in many cases (especially under
high disturbance) extinction of disturbance-intolerant
species that also appeared competitively superior.
Consequently, communities under higher disturbance
were increasingly characterized by disturbance-tolerant
species that also appeared to be inferior competitors.
The reduced role of competition means that early
colonizing species had lesser effects on species that
arrive later and were less likely to preclude late-arriving
species, which should increase community similarity and
reduce the likelihood of alternative stable states (Chase
2003a, Fukami and Lee 2006). To directly examine the
manner in which competition varied with disturbance,
we conducted multiple regressions within each distur-
bance level to relate the density of each of the four
common species (Colpidium, Loxocephalus, P. aurelia,
and Glaucoma sp.) characteristic of high-disturbance
environments to densities of all its potential competitors.
Only data from the final sampling date were used in
multiple regressions, which employed the backward
elimination procedure and a significance level of 0.05.
Regressions revealed that Colpidium was not significant-
ly adversely affected by other species at all disturbance
levels; that Loxocephalus was not significantly adversely
affected by other species at intermediate and high levels
of disturbance (Loxocephalus was driven to extinction
everywhere in low-disturbance microcosms); that Glau-
coma was adversely affected by P. aurelia at the low level
of disturbance only; and that P. aurelia was adversely
affected by P. bursaria at low and intermediate levels of
disturbance only (Fig. 4). In addition, the per capita
(i.e., the slope of the regression) and overall (i.e., per
capita 3 P. bursaria density) negative effects of P.
bursaria on P. aurelia were greater at the low than at the
intermediate level of disturbance (Fig. 4). These results
thus provide suggestive, though not confirmatory,
evidence that the strength of competition may have
declined with increasing disturbance.
Our results did not lend much support to the view that
disturbance promotes alternative stable states (Didham
et al. 2005, Didham and Norton 2006). Didham and
colleagues (Didham et al. 2005, Didham and Norton
2006) argue that because species traits in habitats
characterized by high rates of disturbance are more
similar than those in habitats characterized by low rates
of disturbance, early colonizers are more likely to resist
invasion of later colonizers that share more similar
traits, resulting in higher probability of alternative stable
states with increasing disturbance. With the increasing
July 2008 1937 COMMUNITY ASSEMBLY AND DISTURBANCE
commonness of disturbance-tolerant species under
higher disturbance, species traits in habitats with higher
rates of disturbance did become more similar in our
experiment. However, the increased similarity in species
traits is not necessarily equivalent to increased compe-
tition among community members. In our experiment,
competition appeared to have weakened in habitats with
high disturbance despite the increase in species similarity
(see the previous paragraph). As suggested by Chase
(2003a) and Fukami and Lee (2006) and supported by
our experiment results, the more likely outcome of this
reduced role of competition in high disturbance habitats
is increased similarity among communities with different
assembly sequences and small likelihood of alternative
Species diversity pattern consistent with the IDH was
found only in the controls with simultaneous species
colonization, but not in the five sequential assembly
treatments comprising multiple species introduction
events. The deviation from the IDH in most assembly
treatments accords with the general finding that only a
small fraction of empirical studies testing the IDH
actually supported it (Mackey and Currie 2001). A
number of factors, such as environmental productivity
(Huston 1979, Kondoh 2001, Scholes et al. 2005),
predation (Kneitel and Chase 2004, Morgan and
Buckling 2004, Gallet et al. 2007), trophic positions of
study organisms (Wootton 1998), and the spatial scale at
which disturbance occurs (Chase 2003a, Chase and
Leibold 2003), are known to influence the relationship
between disturbance and diversity, and deviations from
the IDH could potentially arise from variations in any of
these factors. The different diversity responses to
disturbance in the controls and sequential assembly
treatments suggest that the history of community
assembly should be incorporated into the suite of factors
impacting the diversity–disturbance relationship.
The IDH often assumes a trade-off between compet-
itive ability and ability to cope with disturbance
(Connell 1978, Petraitis et al. 1989, Chase and Leibold
2003), and results in the controls were consistent with
the role of this trade-off in producing the unimodal
diversity–disturbance relationship. In the controls, Ble-
pharisma and P. bursaria, presumably strong competi-
tors, dominated under low disturbance but were driven
to extinction under high disturbance; Loxocephalus and
P. aurelia, presumably weak competitors, were driven to
extinction under low disturbance but were common
under high disturbance; only at the intermediate level of
disturbance did both types coexist (Fig. 3A; Appendix
A: Fig. A1). The lack of support for the IDH in the
sequential assembly treatments may be explained by the
elevated extinction rates of species at both ends of the
trade-off under intermediate disturbance. This resulted
because later colonizers, especially those at the ends of
the competition–tolerance axis, generally had difficulty
building viable populations under intermediate distur-
bance. Loxocephalus, a ‘‘tolerant’’ species, almost always
failed to establish under intermediate disturbance if it
was not the first to colonize habitats (Appendix A: Figs.
A1–A6); competition from early colonizers presumably
drove it to extinction. Competitive superiors, such as
Blepharisma and P. bursaria, also frequently experienced
extinction at the intermediate level of disturbance under
sequential assembly (Appendix A: Figs. A1–A6). These
disturbance-intolerant superior competitors were more
vulnerable to disturbance when introduced later, as
competition from early colonizers prevented them from
building sufficiently large populations to endure distur-
bance of the intermediate intensity. Note that this
absence of species at both ends of the trade-off did not
occur at low and high levels of disturbance, where
competitively superior (disturbance-intolerant) and in-
ferior (disturbance-tolerant) species dominated, respec-
tively. Using a similar experimental system, Cadotte
(2007) also showed that the lack of competition–
colonization trade-off among species, another trade-off
frequently invoked to explain the IDH, may cause
species diversity to decline with disturbance. The current
study and that of Cadotte (2007) thus support the
importance of ecological trade-offs for producing the
unimodal diversity–disturbance relationship (but see
Kadmon and Benjamini 2006). Our experiment further
suggests that the lack of competition–tolerance trade-off
associated with sequential community assembly may be
a likely explanation behind the low occurrence of IDH
patterns in empirical studies, since different species
bursaria density at each of the three disturbance levels on the
final sampling date (week 11). Symbols: black circles and solid
regression line, low disturbance; open circles and dashed
regression line, intermediate disturbance; gray circles, high
disturbance. Density data, originally measured as no. individ-
uals/mL, were log10-transformed prior to analyses.
Paramecium aurelia density as a function of P.
LIN JIANG AND SHIVANI N. PATEL 1938Ecology, Vol. 89, No. 7
generally colonize natural communities sequentially
rather than simultaneously.
Several caveats may potentially limit the generality of
our results. First, our experiment was conducted in
laboratory microcosms using species that may or may
not co-occur in nature. Based on this fact, one could
argue that our findings may not be directly generalized
to natural systems. The use of laboratory microcosms
and protists with simple life histories, however, allowed
us to examine relatively long-term community dynamics
that otherwise would not be possible with most natural
systems. The use of microcosms also minimized the
presence of environmental heterogeneity, which could
covary with disturbance in larger systems (e.g., Chase
2007). Second, in our experiment disturbance was
implemented using sonication, which mimicked nonse-
lective density-independent mortality that killed all
individuals affected by disturbance. While density-
independent mortality is not uncommon in nature
(Huston 1994), it remains to be seen whether our results
can apply to other types of mortality associated with
different disturbance regimes (e.g., density-, size-, or
species-dependent mortality, sensu Huston 1994). Third,
species may tolerate disturbance by being either resistant
during disturbance and/or being resilient after distur-
bance. The nondiscriminative nature of our disturbance
regimes, however, suggests that here species tolerance is
being measured by only resilience and that the
competition–tolerance trade-off is essentially a compe-
tition–resilience trade-off. Future experiments with
different disturbance regimes are necessary to test
whether competition–resistance trade-offs could pro-
duce similar results.
Despite these caveats, our results clearly demonstrate
that disturbance and community assembly interact to
affect community structure. Increasing disturbance led
to an increase in the similarity among communities
subjected to different assembly sequences, suggesting
that alternative stable states are less likely in habitats
characterized by high rates of disturbance. Community
assembly history had a major impact on the relationship
between disturbance and species diversity, with patterns
consistent with the IDH observed only for communities
assembled via a single colonization event, but not for
communities assembled via multiple sequential species
colonization events. Although much attention has been
given to the ecological consequences of disturbance and
community assembly, few studies have examined the
combined effects of both factors. The strong interactive
effects of disturbance and assembly history found in this
study suggest that further understanding of their
ecological roles may be attained by considering both
We thank Jon Chase, Marc Weissburg, and two anonymous
reviewers for comments that significantly improved the
manuscript. We also thank Peter Morin for discussions and
Jennifer Price for providing Blepharisma growth rate data. This
project was supported by Georgia Tech and NSF (DEB-
Belyea, L. R., and J. Lancaster. 1999. Assembly rules within a
contingent ecology. Oikos 86:402–416.
Blaustein, L., and J. Margalit. 1996. Priority effects in
temporary pools: nature and outcome of mosquito larva–
toad tadpole interactions depend on order of entrance.
Journal of Animal Ecology 65:77–84.
Bray, J. R., and C. T. Curtis. 1957. An ordination of the upland
forest communities of southern Wisconsin. Ecological
Cadotte, M. W. 2007. Competition–colonization trade-offs and
disturbance effects at multiple scales. Ecology 88:823–829.
Chase, J. M. 2003a. Community assembly: When should
history matter? Oecologia 136:489–498.
Chase, J. M. 2003b. Experimental evidence for alternative
stable equilibria in a benthic pond food web. Ecology Letters
Chase, J. M. 2007. Dought mediates the importance of
stochastic community assembly. Proceedings of the National
Academy of Sciences (USA) 104:17430–17434.
Chase, J. M., and M. A. Leibold. 2003. Ecological niches:
linking classical and contemporary approaches. University of
Chicago Press, Chicago, Illinois, USA.
Connell, J. H. 1978. Diversity in tropical rain forests and coral
reefs. Science 199:1302–1310.
Connell, J. H., and W. P. Sousa. 1983. On the evidence needed
to judge ecological stability or persistence. American
Diamond, J. M. 1975. Assembly of species communities. Pages
342–444 in M. L. Cody and J. M. Diamond, editors. Ecology
and evolution of communities. Harvard University Press,
Cambridge, Massachusetts, USA.
Didham, R. K., and D. A. Norton. 2006. When are alternative
stable states more likely to occur? A reply to T. Fukami and
W. G. Lee. Oikos 113:357–362.
Didham, R. K., and D. A. Norton. 2007. Alternative logical
states—A reply to N. W. H. Mason, J. B. Wilson and J. B.
Steel. Oikos 116:358–360.
Didham, R. K., C. H. Watts, and D. A. Norton. 2005. Are
systems with strong underlying abiotic regimes more likely to
exhibit alternative stable states? Oikos 110:409–416.
Drake, J. A. 1991. Community-assembly mechanics and the
structure of an experimental species ensemble. American
Egler, F. E. 1954. Vegetation science concepts I. Initial floristic
composition, a factor in old-field vegetation development.
Fukami, T. 2004a. Assembly history interacts with ecosystem
size to influence species diversity. Ecology 85:3234–3242.
Fukami, T. 2004b. Community assembly along a species pool
gradient: implications for multiple-scale patterns of species
diversity. Population Ecology 46:137–147.
Fukami, T., H. J. E. Beaumont, X. X. Zhang, and P. B. Rainey.
2007. Immigration history controls diversification in exper-
imental adaptive radiation. Nature 446:436–439.
Fukami, T., and W. G. Lee. 2006. Alternative stable states, trait
dispersion and ecological restoration. Oikos 113:353–356.
Fukami, T., and P. J. Morin. 2003. Productivity–biodiversity
relationships depend on the history of community assembly.
Gallet, R., S. Alizon, P. A. Comte, A. Gutierrez, F. Depaulis,
M. van Baalen, M. Michel, and C. D. M. Mu ¨ ller-Graf. 2007.
Predation and disturbance interact to shape prey species
diversity. American Naturalist 170:143–154.
Gleason, H. A. 1927. Further views on the succession concept.
Grime, J. P. 1979. Plant strategies and vegetation processes.
John Wiley and Sons, Chichester, UK.
July 20081939 COMMUNITY ASSEMBLY AND DISTURBANCE
Hastings, A. 2004. Transients: The key to long-term ecological
understanding? Trends in Ecology and Evolution 19:39–45.
Huston, M. 1979. A general hypothesis of species diversity.
American Naturalist 113:81–101.
Huston, M. A. 1994. Biological diversity: the coexistence of
species on changing landscapes. Cambridge University Press,
Inouye, R. S., and D. Tilman. 1995. Convergence and
divergence of old-field vegetation after 11 yr of nitrogen
addition. Ecology 76:1872–1887.
Kadmon, R., and Y. Benjamini. 2006. Effects of productivity
and disturbance on species richness: a neutral model.
American Naturalist 167:939–946.
Kneitel, J. M., and J. M. Chase. 2004. Trade-offs in community
ecology: linking spatial scales and species coexistence.
Ecology Letters 7:69–80.
Kondoh, M. 2001. Unifying the relationships of species richness
to productivity and disturbance. Proceedings of the Royal
Society B 268:269–271.
Law, R., and R. D. Morton. 1993. Alternative permanent states
of ecological communities. Ecology 74:1347–1361.
Law, R., and R. D. Morton. 1996. Permanence and the
assembly of ecological communities. Ecology 77:762–775.
Lewontin, R. C. 1969. The meaning of stability. Pages 13–24 in
Diversity and stability in ecological systems. Brookhaven
Symposium in Biology 22. Brookhaven National Laboratory,
Upton, New York, USA.
Lockwood, J. L., R. D. Powell, M. P. Nott, and S. L. Pimm.
1997. Assembling ecological communities in time and space.
Louette, G., and L. De Meester. 2007. Predation and priority
effects in experimental zooplankton communities. Oikos 116:
Mackey, R. L., and D. J. Currie. 2000. A re-examination of the
expected effects of disturbance on diversity. Oikos 88:483–
Mackey, R. L., and D. J. Currie. 2001. The diversity–
disturbance relationship: Is it generally strong and peaked?
Mason, N. W. H., J. B. Wilson, and J. B. Steel. 2007. Are
alternative stable states more likely in high stress environ-
ments? Logic and available evidence do not support Didham
et al. 2005. Oikos 116:353–357.
Moran, M. D. 2003. Arguments for rejecting the sequential
Bonferroni in ecological studies. Oikos 100:403–405.
Morgan, A. D., and A. Buckling. 2004. Parasites mediate the
relationship between host diversity and disturbance frequen-
cy. Ecology Letters 7:1029–1034.
Morin, P. J. 1984. Odonate guild composition: experiments
with colonization history and fish predation. Ecology 65:
Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989.
The maintenance of species diversity by disturbance.
Quarterly Review of Biology 64:393–418.
Price, J. E., and P. J. Morin. 2004. Colonization history
determines alternate community states in a food web of
intraguild predators. Ecology 85:1017–1028.
Robinson, J. V., and M. A. Edgemon. 1988. An experimental
evaluation of the effect of invasion history on community
structure. Ecology 69:1410–1417.
Samuels, C. L., and J. A. Drake. 1997. Divergent perspectives
on community convergence. Trends in Ecology and Evolu-
Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The
combined effects of energy and disturbance on species
richness in protist microcosms. Ecology Letters 8:730–738.
Schroder, A., L. Persson, and A. M. De Roos. 2005. Direct
experimental evidence for alternative stable states: a review.
Shea, K., S. H. Roxburgh, and E. S. J. Rauschert. 2004.
Moving from pattern to process: coexistence mechanisms
under intermediate disturbance regimes. Ecology Letters 7:
Sommer, U. 1991. Convergent succession of phytoplankton in
microcosms with different inoculum species composition.
Trexler, J. C., W. F. Loftus, and S. Perry. 2005. Disturbance
frequency and community structure in a twenty-five year
intervention study. Oecologia 145:140–152.
Wootton, J. T. 1998. Effects of disturbance on species diversity:
a multitrophic perspective. American Naturalist 152:803–
Protist population dynamics in different treatments (Ecological Archives E089-112-A1).
Results of ANOVAs for the effects of disturbance and assembly sequence on protist abundances (Ecological Archives E089-112-
LIN JIANG AND SHIVANI N. PATEL1940 Ecology, Vol. 89, No. 7