Journal of Animal
Ecology 2007
76, 660–668
Temperature fluctuation facilitates coexistence of
competing species in experimental microbial communities
Blackwell Publishing Ltd
LIN JIANG*† and PETER J. MORIN*
*Department of Ecology, Evolution & Natural Resources, Cook College, Rutgers University, 14 College Farm Road,
New Brunswick, NJ 08901, USA; and †School of Biology, Georgia Institute of Technology, 310 Ferst Drive, Atlanta,
GA 30332 USA
Summary
1. Temperature fluctuation is a general phenomenon affecting many, if not all, species
in nature. While a few studies have shown that temperature fluctuation can promote
species coexistence, little is known about the effects of different regimes of temperature
fluctuation on coexistence.
2. We experimentally investigated how temperature fluctuation and different regimes
of temperature fluctuation (‘red’ environments in which temperature series exhibited
positive temporal autocorrelation vs. ‘white’ environments in which temperature series
showed little autocorrelation) affected the coexistence of two ciliated protists,
Colpidium striatum Stein and Paramecium tetraurelia Sonneborn, which competed for
bacterial resources.
3. We have previously shown that the two species differed in their growth responses
to changes in temperature and in their resource utilization patterns. The two species
were not always able to coexist at constant temperatures (22, 24, 26, 28 and 30 °C), with
Paramecium being competitively excluded at 26 and 28 °C. This indicated that resource
partitioning was insufficient to maintain coexistence at these temperatures.
4. Here we show that in both red and white environments in which temperature varied
between 22 and 32 °C, Paramecium coexisted with Colpidium. Consistent with the
differential effects of temperature on their intrinsic growth rates, Paramecium population
dynamics were largely unaffected by temperature regimes, and Colpidium showed more
variable population dynamics in the red environments.
5. Temperature-dependent competitive effects of Colpidium on Paramecium, together
with resource partitioning, appeared to be responsible for the coexistence in the white
environments; resource partitioning and the storage effect appeared to account for the
coexistence in the red environments.
6. These results suggest that temperature fluctuation may play important roles in
regulating species coexistence and diversity in ecological communities.
Key-words: competition, environmental fluctuation, environmental variability, red
noise, white noise.
Journal of Animal Ecology (2007) 76, 660–668
doi: 10.1111/j.1365-2656.2007.01252.x
Introduction
Simple equilibrium competition theory predicts that
the number of coexisting species cannot exceed the
number of limiting resources (MacArthur & Levins
1964; Levins 1968; Levin 1970; Tilman 1982; Grover
© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society
Correspondence: L. Jiang, School of Biology, Georgia
Institute of Technology, 310 Ferst Drive, Atlanta, Georgia,
GA 30332, USA. Tel.: 404 385 2514. Fax: 404 894 0519. Email: [email protected]
1997), and that species similar in their resource requirements will fail to coexist (Hardin 1960; MacArthur &
Levins 1967; May 1973; Tilman 1982; Grover 1997).
These predictions of equilibrium theory, supported by
many laboratory experiments (summarized in Grover
1997; Sommer & Worm 2002), suggest that species
diversity found in natural communities, where the
number of limiting resources is typically small, should
be generally low. In contrast, many natural communities
harbour a large number of species that appear to persist
without apparent competitive exclusion. Understanding
661
Temperature
fluctuation and
coexistence
© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society,
Journal of Animal
Ecology, 76,
660–668
coexistence-promoting mechanisms that could resolve
this discrepancy has been a subject of much interest in
community ecology.
The nonequilibrium view that temporal variation in
environmental factors may allow more species to coexist has enjoyed a long history in ecology (Hutchinson
1961; Connell 1978; Huston 1979; Chesson & Warner
1981), although the precise coexistence mechanisms
associated with variable environments were not well
understood until more recently (Chesson 1994, 2000).
Early conceptual work viewed fluctuation as a more or
less distinct mechanism contributing to coexistence.
For instance, Hutchinson (1961) hypothesized that
temporal fluctuation of appropriate frequencies may
maintain the diversity of phytoplankton assemblages
that would be lost at competitive equilibrium. Connell
(1978) proposed that ‘higher diversity of trees and
corals is maintained only in a nonequilibrium state’
where disturbance prevents competitive exclusion to
run its course.
Recent theoretical work, however, suggests that
temporal fluctuation itself does not constitute a coexistence mechanism but creates conditions for coexistencepromoting mechanisms to operate (Chesson & Huntly
1997; Chesson 2000; Roxburgh, Shea & Wilson 2004;
Shea, Roxburgh & Rauschert 2004). Besides classic
mechanisms (i.e. niche differentiation) that can operate
regardless of the presence/absence of environmental
fluctuation (fluctuation-independent mechanisms),
two fluctuation-dependent mechanisms can contribute
to coexistence in variable environments (Chesson 1994,
2000). Relative nonlinearity operates when species
differ in the linearity of their responses to limiting
resources (Levins 1979; Armstrong & McGehee 1980;
Chesson 1994, 2000; Abrams & Holt 2002; Abrams
2004). For this mechanism to contribute to the coexistence of any two species, relative nonlinearity must be
sufficiently strong such that one species grows more
rapidly at some resource levels, and one grows more
rapidly at other resource levels. The storage effect, on
the other hand, operates in situations where species
differ in their responses to physical environmental
conditions and the strength of competition varies with
the fluctuating physical environment (Chesson 1994,
2000). Coexistence comes about when species not
favoured by the environment experience weak intraspecific competition due to their low population densities
and limited interspecific competition associated with
certain species life-history traits buffering populations
against extinction (buffered population growth;
Chesson 1994, 2000). In principle, these two fluctuationdependent mechanisms would work in concert with
fluctuation-independent mechanisms to regulate species
diversity in variable environments.
Most empirical studies of the environmental fluctuation effect on coexistence have focused on resource
variability (e.g. Sommer 1984, 1985; Grover 1988,
1989; Litchman 1998, 2003; Floder, Urabe & Kawabata
2002), reflecting the long-standing theoretical interests
in the role of resource fluctuations in species coexistence (Stewart & Levin 1973; Koch 1974; Levins 1979;
Armstrong & McGehee 1980; Hsu 1980; Grover 1990;
Chesson 1994, 2000; Abrams & Holt 2002; Abrams
2004); the effects of other environmental factors have
received relatively little attention. For instance, as one
of the most important environmental variables affecting
all levels of biological organization, temperature shows
strong diurnal and seasonal cycles as well as frequent
deviations from these trends. Given its constantly
varying nature, few studies have investigated the potentially important role of variation in temperature in
influencing species coexistence (but see Eddison &
Ollason 1978; Descamps-Julien & Gonzalez 2005).
Further, while almost all experimental studies of the
effects of environmental variability on coexistence have
considered periodic variations with fixed frequencies,
analyses of variation in many environmental variables
(including temperature) have revealed a mixture of
periodic components with various frequencies interposed upon each other. Some variables exhibit a ‘white’
colour with no dominant frequencies in the variance
spectrum (i.e. little autocorrelation in the time series),
whereas others demonstrate a ‘red’ colour where
the low frequency components dominate (i.e. positive
autocorrelation in the time series); the same variable
may also exhibit different colours in different habitats
(Steele 1985; Halley 1996; Cyr & Cyr 2003; Vasseur &
Yodzis 2004).
Our experiments investigated the effects of temperature fluctuation, including both white (temporally
uncorrelated) and reddened (temporally positively
autocorrelated) temperature regimes, on the coexistence and dynamics of two ciliated protists, Colpidium
striatum Stein and Paramecium tetraurelia Sonneborn,
which competed for multiple bacterial species. Previous work showed that coexistence between the
two species was not always possible at constant temperatures (Jiang & Morin 2004). Our experiments
addressed two sets of questions. First, can temperature
fluctuation lead to coexistence of the two competitors
that often failed to coexist at constant temperatures?
If coexistence does occur, do species coexistence
patterns differ in the two types of thermal environments? Second, which mechanisms, including both
fluctuation-dependent and -independent mechanisms,
could account for any observed coexistence? Do
coexistence mechanisms differ between the two
environments?
Materials and methods
Colpidium striatum and Paramecium tetraurelia are
free-swimming freshwater bacterivorous ciliates
belonging to the order Hymenostomatida. Colpidium
was obtained from Carolina Biological Supply Company
(Burlington, NC, USA), and Paramecium from the
American Type Culture Collection (Rockville, MD,
USA). Both had been maintained in laboratory stock
662
L. Jiang &
P. J. Morin
© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society,
Journal of Animal
Ecology, 76,
660–668
cultures separately on mixed bacteria (including
Serratia marcescens Bizio, Bacillus cereus Frankland &
Frankland, and Bacillus subtilis [Ehrenberg] Cohn) for
many years prior to the experiment.
Experimental microcosms consisted of 240 mL
capped Pyrex glass bottles, each filled with 100 mL of
aqueous growth medium, plus two wheat seeds as a
source of slow carbon release. All materials were
sterilized before use. The growth medium contained
0·55 g of protozoan pellet (Carolina Biological Supply
Company) per litre of well water. This medium, which
has been used in many of our previous studies (e.g.
Jiang & Morin 2004, 2005), provided an adequate
supply of energy and nutrition for bacterial growth.
The medium was first autoclave-sterilized, and then
inoculated with S. marcescens, B. cereus and B. subtilis.
Twenty-four hours after bacterial inoculation, we
introduced small Colpidium and Paramecium populations (0·5% of their respective carrying capacities at
the room temperature) from newly established stock
cultures into their designated microcosms. Prior to the
experiments, fresh Colpidium and Paramecium stock
cultures were established using a small number (< 5)
of individuals of each species after these individuals
were repeatedly washed in the bacterized medium.
The repeated-washing procedure effectively removed
many other bacterial species occurring in old stock
cultures and minimized differences in the initial
composition of bacterial communities associated
with the two protist species. All microcosms were
maintained in temperature-controlled incubators
without light.
Because competitive outcomes are best evaluated
using long-term data spanning multiple generations of
competing species, we ran the experiments long enough
(41 days compared with Colpidium and Paramecium
generation times of c. 6–12 h) to minimize transient
effects. The 41-day experiment used a set of temperature series in which temperature fluctuated between 22
and 32 °C. Because ambient temperature in nature
shows either ‘white’ (random) or ‘reddened’ (positively
autocorrelated) colour (Cyr & Cyr 2003; Vasseur &
Yodzis 2004), we considered both white and reddened
temperature series. A 3 × 4 factorial design was used:
three protist species treatments (Colpidium monocultures, Paramecium monocultures, Colpidium and
Paramecium polycultures), crossed with four variable
temperature treatments (white 1, white 2, red 1, red 2).
These temperature series were created with the spectral
mimicry algorithm (Cohen et al. 1998) by shuffling 82
temperatures uniformly distributed between 22 and
32 °C in different sequences (Fig. 1; see Petchey 2000
for more technical details). The realized temperature
series, controlled by four independent incubators, had
temperature changing continuously in a linear manner
during the 12 h interval between the current and next
temperature. Two different realizations for each
temperature ‘colour’ spectrum were used to separate
colour effects from potential idiosyncratic effects asso-
Fig. 1. Temperature series in (a) white (random) environments,
and (b) red (positively autocorrelated) environments. Two
independent series were used in each treatment.
ciated with any particular realization of a temperature
series. In the red 1 treatment, a programming error
resulted in the desired temperatures on days 15 and 16
being replaced by temperatures from the previous
2 days; this small error did not significantly affect the
autocorrelation structure of the temperature series, as
intended temperature values were fairly similar during
these days (Fig. 1b). We replicated the monocultures
and polycultures three and four times, respectively, for
a total of 40 microcosms.
During the experimental period, we replaced 10% of
the medium with sterile medium every week and sampled
protist densities every 2 days. The sampling procedure
involved first swirling the microcosm to mix the content
and removing c. 0·35 mL of the medium (exact volume
determined by weighing the sample on an analytic
balance) using a sterile Pasteur pipette; the number of
each protist species in the sample was then enumerated
with a Nikon SMZ-U stereomicroscope.
663
Temperature
fluctuation and
coexistence
Table 1. Intrinsic growth rates of Colpidium and Paramecium
at the six constant temperatures. Values are mean ± SE. Data
from 22 to 30 °C were previously reported in Jiang & Morin
(2004)
Temperature
(°C)
Colpidium
growth rate
(per hour)
Paramecium
growth rate
(per hour)
22
24
26
28
30
32
0·163 ± 0·004
0·151 ± 0·002
0·138 ± 0·003
0·133 ± 0·005
0·0948 ± 0·007
–0·018 ± 0·018
0·074 ± 0·006
0·056 ± 0·011
0·088 ± 0·008
0·100 ± 0·008
0·121 ± 0·013
0·100 ± 0·012
We used repeated-measures  (rm-) to
test for the effects of competition and temperature
regime on protist population densities over time.
Because competitive effects were most evident towards
the end of the experiment, we restricted rm-s to
the data collected between day 32 and day 40 (the last
sampling day). Prior to rm-s, all protist data
were log10 (abundance/mL + 1) transformed to improve
normality. We used  to test for the effects of
competition and temperature regime on protist
population variability, measured as standard deviation
of log10 [abundance/mL + 1] for the entire experimental
period. To quantify the effects of competition, we
calculated effect sizes (the degree to which competition
reduced population size) and per-capita impacts of
Colpidium competition on Paramecium. We focused on
effects of Colpidium on Paramecium because of the
asymmetric nature of their interaction in which Colpidium
did not appear adversely affected by Paramecium (see
Results). Effect sizes were calculated as average
Paramecium polyculture density from day 32 to 40
as a fraction of its corresponding monoculture density
in the same period on a logarithmic scale. The percapita impacts of Colpidium on Paramecium were
calculated as the competition effect size divided by
average Colpidium population size in polycultures
from day 32 to 40.
Table 2. Persistence times of Colpidium and Paramecium in the competition-free
controls and competition treatments at the six constant temperatures. Values are
mean ± SE. Values of 44 ± 0 indicate that the species persisted throughout the 44-day
experiment in all replicates
Temperature
(°C)
Colpidium
persistence time (day)
Paramecium
persistence time (day)
Control
Competition
Control
Competition
44 ± 0
44 ± 0
44 ± 0
44 ± 0
44 ± 0
13·50 ± 0·50
44 ± 0
44 ± 0
44 ± 0
44 ± 0
44 ± 0
44 ± 0
44 ± 0
43·50 ± 0·50
35·50 ± 0·96
22·50 ± 2·63
44 ± 0
44 ± 0
© 2007 The Authors.
22
Journal
compilation44 ± 0
24
44 ± 0
© 2007 British
26
44 ± 0
Ecological Society,
28
44 ± 0
Journal of Animal
30
44 ± 0
Ecology, 76,
32
4·67 ± 0·67
660–668
Results
     
   
For comparison, we briefly summarize the results of
competition experiments conducted at six constant
temperatures (22, 24, 26, 28, 30, 32 °C); main results
from 22 to 30 °C were previously reported in Jiang &
Morin (2004). Changes in temperature had opposite
effects on the intrinsic growth rates of Colpidium and
Paramecium: Colpidium growth rate declined (32 °C was
lethal for Colpidium) and Paramecium growth rate
increased with temperature (Table 1). Differences in the
bacterial community structure in Colpidium and Paramecium cultures indicated that Colpidium and Paramecium
used bacterial resources differently, independent of
temperature (Jiang & Morin 2004). Competition between
Colpidium and Paramecium was asymmetric: while
Colpidium reached either equal or higher densities in the
presence of Paramecium relative to when it was alone,
Paramecium suffered significantly in the presence of
Colpidium with substantially reduced densities at 22, 24 and
30 °C and observed extinctions at 26 and 28 °C (Table 2).
The effect size of Colpidium competition on Paramecium differed significantly among temperatures
(Fig. 2a), showing its largest values at 26 and 28 °C
(where Paramecium extinction occurred), and smallest
values at 22 and 30 °C (where Paramecium coexisted
with Colpidium). The per-capita impact of Colpidium
on Paramecium increased with temperature (Fig. 2b;
: F4,15 = 4·26, P = 0·0168), suggesting that each
Colpidium individual had a significantly larger impact
on Paramecium at higher temperatures. This increase in
the per-capita effect presumably reflected the increase
in the resource uptake of an average Colpidium individual at higher temperatures.
    - 
 
In the two white temperature treatments, Colpidium
population sizes did not track rapid temperature
fluctuation and were relatively constant over time
(Fig. 3a,b), a pattern consistent with the theoretical
prediction that populations are able to average across
rapid environmental fluctuation (May 1976). In the
two red temperature treatments, however, Colpidium
showed large declines in abundance that coincided with
periods of high temperatures (compare Fig. 3c,d with
Fig. 1b), resulting in significantly higher variability
in population densities in the reddened than white
environments (two-way , temperature regime:
F3,19 = 35·16, P < 0·0001). Colpidium also showed greater
variability in population dynamics in polycultures than
monocultures (two-way , competition: F1,19 =
11·09, P = 0·0035), largely due to the difference in the
reddened but not white environments (competition ×
temperature regime: F3,19 = 11·22, P = 0·0002).
664
L. Jiang &
P. J. Morin
Fig. 2. (a) Effect size of Colpidium competition on Paramecium.
Competition effect size was calculated as Paramecium
polyculture density average over days 32 – 40 as a proportion
of its corresponding monoculture density in the same period
on a logarithm scale. (b) Per-capita Colpidium impact on
Paramecium. Per capita interaction strength was calculated as
the competition effect size divided by average Colpidium
nontransformed population size from day 32 to 40.
Calculations were not done for 32 °C, where no Colpidium
individuals remained during days 32– 40. The black bar
represents the expected (e) per capita Colpidium impact
averaged over the entire temperature gradient (with impact
= 0 at 32 °C). Error bars represent ± 1 SE. Different letters
indicated significant differences at P = 0·05 level in a Tukey’s
Honest Significant Difference test.
© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society,
Journal of Animal
Ecology, 76,
660–668
Focusing on the last five sampling dates (days 32 – 40),
there was a significant interactive effect of competition
and temperature on Colpidium density (rm-:
F3,19 = 5·56, P = 0·0061), corresponding to the opposite
effect of Paramecium on Colpidium in the two types
of thermal environments (larger Colpidium densities
in polycultures relative to monocultures in the
white environments vs. smaller Colpidium densities
in polycultures relative to monocultures in the red
environments). Colpidium extinction occurred in one
polyculture replicate of the red 1 treatment.
Unlike results observed in the constant temperature
experiment, Paramecium never went extinct under
temperature fluctuation (Fig. 3). Paramecium increased
more gradually in polycultures than in monocultures,
but eventually reached relatively steady population
sizes by the end of the experiment. Again focusing on
the last five sampling dates, there was a significant
competition effect (rm-: F1,19 = 16·91, P = 0·0006),
but no temperature regime effect (rm-: F3,19 =
0·17, P = 0·9151). The significant competition effect
arose from lower Paramecium densities in polycultures
relative to monocultures in the white environments,
but comparable densities in the red environments
(competition × temperature regime: F 3,19 = 5·52,
P = 0·0067).
Although the average temperature was 27 °C in both
white and red environments, the effect sizes of Colpidium competition on Paramecium in both environments
were considerably smaller than those under constant
environments at 26 and 28 °C (Fig. 2a). Competition
effect sizes did not differ significantly from zero in the
red environments (Fig. 2a), corresponding to reduced
Colpidium densities in response to periods of high
temperatures towards the end of the experiment
(Figs 1b and 3). Per-capita impacts of Colpidium on
Paramecium in the two white environments were less
than half of those estimated in thermally constant
environments at 26 and 28 °C, though these differences
were not statistically significant in a Tukey’s Honest
Significant Difference test (Fig. 2b). Substantial
changes in Colpidium densities between day 32 and 40
in the two red environments (Fig. 3c,d) prevented
accurate estimates of per-capita effects of Colpidium on
Paramecium in these treatments.
Discussion
Our experiment clearly showed that temperature
fluctuation, as a form of environmental variability, can
facilitate the persistence of competitively inferior
species. In the constant temperature environments,
Paramecium, often being the inferior competitor,
attained lower population densities in the presence
of Colpidium relative to when it was alone over much
of the temperature gradient, and was competitively
excluded at 26 and 28 °C (Jiang & Morin 2004). In the
variable temperature environments with an average
temperature of 27 °C, however, Paramecium always
persisted with Colpidium for the duration of the experiment, showing no trend of population decline that
preceded Paramecium extinction seen in some of the
constant temperature environments. Paramecium
population dynamics were largely unaffected by
temperature regimes, though it attained higher densities
in the red environments in which long-periods of high
temperatures reduced Colpidium densities. These findings support the idea that environmental variability
665
Temperature
fluctuation and
coexistence
Fig. 3. Population dynamics of Colpidium and Paramecium in the white 1, white 2, red 1, red 2 treatments. Error bars
represent ± 1 SE.
could permit more species to coexist and promote
greater diversity than stable environments will allow
(Hutchinson 1961; Connell 1978; Huston 1979; Chesson
1994, 2000). The question now is what mechanisms
accounted for the coexistence patterns in different environments? Below we evaluate relevant coexistence
mechanisms that may have been at work.
   
- 

© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society,
Journal of Animal
Ecology, 76,
660–668
Sufficient differences in resource use may allow species
to coexist in both stable and variable environments
(Chesson 2000). A certain degree of resource use
differentiation did exist between Colpidium and Paramecium, as indicated by their differential effects on bacterial communities (Jiang & Morin 2004). Jiang &
Morin (2004) have shown, however, that this difference
arose not because the two species fed on completely
different bacterial prey, but because Colpidium appeared
to specialize on Serratia with high affinity (lower R*,
sensu Tilman 1982), and Paramecium appeared to
prefer Serratia but was also able to use other bacteria
with less efficacy. As a result, the two species did not
always coexist at constant temperatures. By fitting
Lotka–Volterra competition models to long-term
population dynamics, Jiang & Morin (2004) also
found that the individual impact of Colpidium on
Paramecium (per capita interaction strength) increased
with temperature, presumably because each Colpidium
individual consumed more resources to balance its
elevated metabolic requirement at higher temperatures.
We came to the same conclusion here using simpler
calculations based on data near the end of the experiment (Fig. 2b). This increase in per capita interaction
strength, along with the decline in Colpidium density
with increasing temperature, accounted for the peculiar
pattern of Paramecium extinction along the temperature
gradient (coexistence at the 22 and 30 °C, and extinction
at the intermediate temperatures).
We believe that the low per capita impacts of Colpidium
on Paramecium (Fig. 2b), combined with resource use
differentiation between the two species, were responsible
for the coexistence in the temporally uncorrelated
‘white’ thermal environments. Two explanations may
account for the low per-capita interaction strength in
white environments. First, as Colpidium experienced
continuously changing temperatures but maintained a
relatively stable population size, the estimated per-capita
interaction strength would be a time-average over
all temperatures. Because the per-capita interaction
strength was much lower at the low (22 and 24 °C) and
high (presumably c. 0 at 32 °C, the lethal temperature
666
L. Jiang &
P. J. Morin
for Colpidium) ends of the temperature gradient,
when averaged over the entire temperature gradient
(22 –32 °C), the interaction strength in white environments would be much lower than those at intermediate
constant temperatures (Fig. 2b). Second, Colpidium
individuals may have been physiologically stressed by
rapid fluctuation in temperature, which could lead to
even smaller per capita impacts on Paramecium in white
environments. Comparison of per-capita interaction
strengths estimated for the white environments with
per-capita interaction strength averaged over the entire
temperature gradient indicates that this is indeed true
(Fig. 2b; t7 = 6·85, P = 0·0002). A recent study also
found that Tetrahymena thermophila, a ciliate belonging
to the same order as Colpidium, showed signs of physiological difficulty in white temperature environments
(Laakso, Loytynoja & Kaitala 2003). Physiological
stress experienced by Colpidium individuals reduced
their fitness and the fitness difference between Colpidium
and Paramecium individuals; this fitness equalizing
mechanism presumably allowed resource use differentiation (as a fitness stabilizing mechanism, which, if
strong enough, can alone lead to coexistence) to bring
coexistence in white environments (see Chesson 2000
for a discussion of fitness-equalizing and -stabilizing
mechanisms). Given the ubiquity of ambient temperature variability and the critical role of temperature in
regulating species physiology, physiological changes
associated with temperature fluctuation may be important in modulating species coexistence in many ecological
communities.
    

© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society,
Journal of Animal
Ecology, 76,
660–668
Two fluctuation-dependent coexistence mechanisms,
the relative nonlinearity and storage effect, may operate
in variable environments (Chesson 1994, 2000). While
models demonstrate the role of the relative nonlinearity
as a legitimate coexistence mechanism (Stewart &
Levin 1973; Koch 1974; Levins 1979; Armstrong &
McGehee 1980; Hsu 1980; Grover 1990; Chesson 1994,
2000; Abrams & Holt 2002; Abrams 2004), they also
show that, compared with the storage effect, the effect
of relative nonlinearity is generally weak (Chesson
1994, 2000; but see Huisman & Weissing 1999). Indeed,
the definitive role of relative nonlinearity in species
coexistence has yet to be empirically established. A
necessary condition for the relative nonlinearity to
operate is that oscillations in consumer (and resource)
population sizes must occur due either to exogenous
resource fluctuations (Grover 1990; Abrams 2004) or
endogenous resource fluctuations associated with
unstable consumer–resource interactions (Armstrong
& McGehee 1980; Abrams & Holt 2002). In the white
temperature treatments, Colpidium and Paramecium
showed relatively constant populations indicative of
stable consumer–resource interactions, suggesting a
lack of the relative nonlinearity effect. Colpidium
population sizes indeed fluctuated in the red temperature treatment, creating opportunities for the relative
nonlinearity to operate; our experimental data, however, did not allow quantitative assessments of the role
of the relative nonlinearity effect, which would require
further experimentation measuring Colpidium and
Paramecium growth rates along a bacterial resource
gradient at different temperatures.
The storage effect, perceived as a stronger coexistence
mechanism than the relative nonlinearity, has been
previously invoked to explain the coexistence of Daphnia
species (Caceres 1997), of forest trees (Kelly & Bowler
2002), and of prairies grasses (Adler et al. 2006). With
drastically different effects of temperature on intrinsic
growth rates of the two competitors, large-amplitude
fluctuations in Colpidium abundance (translating
into fluctuations in the intensity of competition on
Paramecium), and the relatively long periods of time
that it takes for competition to drive Paramecium to
extinction in environments that favours Colpidium
(buffered population growth; see Table 2), the storage
effect operated and was at least partly responsible for
the coexistence observed in the red temperature environments. In contrast, large fluctuations in Colpidium
and Paramecium abundances did not occur in the white
environments, implying that the storage effect probably
played a minimal role in this treatment.
   
Two other studies have investigated the effect of
temperature variation on species coexistence. Eddison
& Ollason (1978) subjected natural freshwater samples
to constant and periodically varying temperatures and
found that ciliate diversity tended to be higher in environments with temperature variation, but they did
not attempt to identify possible mechanisms leading to
higher diversity in the variable environment. Using
two freshwater diatoms exhibiting different responses to
changes in temperature, Descamps-Julien & Gonzalez
(2005) showed that sinusoidal temperature variation
promoted coexistence of the two diatoms, possibly
through the storage effect. Our study differs from them
in three important aspects. First, our experiment used
both white and reddened temperature series, which
are probably more realistic depictions of temperature
regimes in nature (Cyr & Cyr 2003; Vasseur & Yodzis
2004). Second, our work revealed the potential importance of species physiological changes associated with
rapid temperature fluctuation for regulating species
coexistence. Third, our work showed that the storage
effect may operate in the slow-changing reddened
environments but not in the rapid-changing white
environments.
Conclusions
Our study provides direct experimental evidence that
temporal variation in temperature could promote the
667
Temperature
fluctuation and
coexistence
coexistence of competing species (also see Eddison &
Ollason 1978; Descamps-Julien & Gonzalez 2005).
The competitively inferior species, which was driven to
extinction by its competitor at several constant temperatures, always managed to coexist with its competitor
in variable temperature environments. The relative
nonlinearity and storage effect contributed little to
coexistence in the white temperature environments.
Rather, differential resource use and temperaturedependent interactions appeared to account for
coexistence in white environments. Differential resource
use and the storage effect appeared to account for
coexistence in the red temperature environments.
These findings, if general, would suggest that temperature fluctuation may play an important role in regulating
species coexistence and diversity in many ecological
communities. In particular, that different coexistence
mechanisms may function in white and red environments points to the need for a better understanding of
species coexistence mechanisms and diversity patterns
in systems exhibiting various autocorrelation structures. The need for more work on this topic is further
underscored by the fact that human activities are altering
autocorrelation structures of important climate variables
(Wigley, Smith & Santer 1998).
Acknowledgements
The authors thank Aabir Banerji, Shibi Chandy, Peter
Chesson, Sebastian Diehl, Jim Drake, Valerie Fournier,
Andy Gonzalez, Jennifer Krumins, Carolyn Norin
and Owen Petchey for comments. This project was
supported by US NSF grant DEB-9806427.
References
© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society,
Journal of Animal
Ecology, 76,
660–668
Abrams, P.A. (2004) When does periodic variation in resource
growth allow robust coexistence of competing consumer
species? Ecology, 85, 372 – 382.
Abrams, P.A. & Holt, R.D. (2002) The impact of consumerresource cycles on the coexistence of competing consumers.
Theoretical Population Biology, 62, 281 – 295.
Adler, P.B., HilleRisLambers, J., Kyriakidis, P.C., Guan, Q.F.
& Levine, J.M. (2006) Climate variability has a stabilizing
effect on the coexistence of prairie grasses. Proceedings of
the National Academy of Sciences USA, 103, 12793 – 12798.
Armstrong, R.A. & McGehee, R. (1980) Competitive
exclusion. American Naturalist, 115, 151 –170.
Caceres, C.E. (1997) Temporal variation, dormancy, and
coexistence: a field test of the storage effect. Proceedings of
the National Academy USA, 94, 9171 – 9175.
Chesson, P. (1994) Multispecies competition in variable
environments. Theoretical Population Biology, 45, 227 –
276.
Chesson, P. (2000) Mechanisms of maintenance of species
diversity. Annual Review of Ecology and Systematics, 31,
343 – 366.
Chesson, P. & Huntly, N. (1997) The roles of harsh and fluctuating conditions in the dynamics of ecological communities.
American Naturalist, 150, 519 – 553.
Chesson, P.L. & Warner, R.R. (1981) Environmental variability promotes coexistence in lottery competitive systems.
American Naturalist, 117, 923 – 943.
Cohen, A.E., Gonzalez, A., Lawton, J.H., Petchey, O.L.,
Wildman, D. & Cohen, J.E. (1998) A novel experimental
apparatus to study the impact of white noise and 1/f noise
on animal populations. Proceedings of the Royal Society of
London Series B, 265, 11– 15.
Connell, J.H. (1978) Diversity in tropical rain forests and
coral reefs. Science, 4335, 1302 –1310.
Cyr, H. & Cyr, I. (2003) Temporal scaling of temperature
variability from land to oceans. Evolutionary Ecology
Research, 5, 1183 – 1197.
Descamps-Julien, B. & Gonzalez, A. (2005) Stable coexistence
in a fluctuating environment: An experimental demonstration. Ecology, 86, 2815 – 2824.
Eddison, J.C. & Ollason, J.G. (1978) Diversity in constant and
fluctuating environments. Nature, 275, 309 – 310.
Floder, S., Urabe, J. & Kawabata, Z. (2002) The influence
of fluctuating light intensities on species composition and
diversity of natural phytoplankton communities. Oecologia,
133, 395 – 401.
Grover, J.P. (1988) Dynamics of competition in a variable
environment: experiments with two diatom species. Ecology,
69, 408 – 417.
Grover, J.P. (1989) Effects of Si : P supply ration, supply
variability, and selective grazing on the plankton: an experiment with a natural algal and protistan assemblage.
Limnology and Oceanography, 34, 349 – 367.
Grover, J.P. (1990) Resource competition in a variable environment: Phytoplankton growing according to Monod’s
model. American Naturalist, 136, 771– 789.
Grover, J.P. (1997) Resource Competition. Chapman & Hall,
London.
Halley, J.M. (1996) Ecology, evolution and 1/f-noise. Trends
in Ecology and Evolution, 11, 33 – 37.
Hardin, G. (1960) The competitive exclusion principle.
Science, 131, 1292 – 1297.
Hsu, S.B. (1980) A competition model for a seasonally
fluctuating nutrient. Journal of Mathematical Biology, 9,
115 –132.
Huisman, J. & Weissing, F.J. (1999) Biodiversity of plankton
by species oscillations and chaos. Nature, 402, 407–
410.
Huston, M. (1979) A general hypothesis of species diversity.
American Naturalist, 113, 81–101.
Hutchinson, G.E. (1961) The paradox of the plankton. American Naturalist, 95, 137 –145.
Jiang, L. & Morin, P.J. (2004) Temperature-dependent interactions explain unexpected responses to environmental
warming in communities of competitors. Journal of Animal
Ecology, 73, 569 – 576.
Jiang, L. & Morin, P.J. (2005) Predator diet breadth influences
the relative importance of bottom-up and top-down control of prey biomass and diversity. American Naturalist,
165, 350 – 363.
Kelly, C.K. & Bowler, M.G. (2002) Coexistence and relative
abundance in forest trees. Nature, 417, 437 – 440.
Koch, A.L. (1974) Coexistence resulting from an alteration of
density dependent and density independent growth. Journal of Theoretical Biology, 44, 373 – 386.
Laakso, J., Loytynoja, K. & Kaitala, V. (2003) Environmental
noise and population dynamics of the ciliated protozoa
Tetrahymena thermophila in aquatic microcosms. Oikos,
102, 663 – 671.
Levin, S.A. (1970) Community equilibria and stability, and an
extension of the competitive exclusion principle. American
Naturalist, 104, 413 – 423.
Levins, R. (1968) Evolution in Changing Environments. Princeton University Press, Princeton, NJ.
Levins, R. (1979) Coexistence in a variable environment.
American Naturalist, 114, 765 – 783.
Litchman, E. (1998) Population and community responses of
phytoplankton to fluctuating light. Oecologia, 117, 247–257.
668
L. Jiang &
P. J. Morin
© 2007 The Authors.
Journal compilation
© 2007 British
Ecological Society,
Journal of Animal
Ecology, 76,
660–668
Litchman, E. (2003) Competition and coexistence of
phytoplankton under fluctuating light: experiments with two
cyanobacteria. Aquatic Microbial Ecology, 31, 241 – 248.
MacArthur, R.H. & Levins, R. (1964) Competition, habitat
selection, and character displacement in a patchy environment. Proceedings of the National Academy of Sciences
USA, 51, 1207 –1210.
MacArthur, R.H. & Levins, R. (1967) The limiting similarity,
convergence, and divergence of coexisting species. American
Naturalist, 101, 377 – 385.
May, R.M. (1973) Stability and Complexity in Model Ecosystems, 2nd edn. Princeton University Press, Princeton, NJ.
May, R.M. (1976) Theoretical Ecology: Principles and Applications. Blackwell Scientific, Oxford.
Petchey, O.L. (2000) Environmental colour affects aspects of
single-species population dynamics. Proceedings of the
Royal Society of London Series B-Biological Sciences, 267,
747 – 754.
Roxburgh, S.H., Shea, K. & Wilson, J.B. (2004) The intermediate disturbance hypothesis: Patch dynamics and
mechanisms of species coexistence. Ecology, 85, 359 – 371.
Shea, K., Roxburgh, S.H. & Rauschert, E.S.J. (2004) Moving
from pattern to process: coexistence mechanisms under intermediate disturbance regimes. Ecology Letters, 7, 491–508.
Sommer, U. (1984) The paradox of the plankton: fluctuations
of phosphorus availability maintain diversity of phytoplankton in flow-through cultures. Limnology and Oceanography, 29, 633 – 636.
Sommer, U. (1985) Comparison between steady state and
non-steady state competition: experiments with natural
phytoplankton. Limnology and Oceanography, 30, 335–346.
Sommer, U. & Worm, B., eds. (2002) Competition and Coexistence. Springer-Verlag, Berlin.
Steele, J.H. (1985) A comparison of terrestrial and marine
ecological systems. Nature, 313, 355 – 358.
Stewart, F.M. & Levin, B.R. (1973) Partitioning of resources
and the outcome of interspecific competition: a model
and some general considerations. American Naturalist, 107,
171 –198.
Tilman, D. (1982) Resource Competition and Community
Structure. Princeton University Press, Princeton, NJ.
Vasseur, D.A. & Yodzis, P. (2004) The color of environmental
noise. Ecology, 85, 1146 –1152.
Wigley, T.M.L., Smith, R.L. & Santer, B.D. (1998) Anthropogenic influence on the autocorrelation structure of
hemispheric-mean temperatures. Science, 282, 1676–1679.
Received 18 October 2006; accepted 27 March 2007