vol. 170, no. 1

the american naturalist



july 2007

Predation and Disturbance Interact to Shape Prey Species Diversity*

Romain Gallet,1,2,† Samuel Alizon,2,‡ Pierre-Arnaud Comte,1,§ Arnaud Gutierrez,1,k Frantz Depaulis,2,# Minus van Baalen,2,** Eric Michel,2,†† and Christine D. M. Mu¨ller-Graf1,2,‡‡

1. Laboratoire Parasitologie E´volutive, Centre National de la Recherche Scientifique Unite´ Mixte de Recherche (CNRS UMR) 7103, Universite´ Pierre et Marie Curie, Baˆtiment A, 7e`me e´tage, CC 237, 7 Quai Saint Bernard, F-75252 Paris Cedex 05, France; 2. Laboratoire Fonctionnement et E´volution des Syste`mes E´cologiques, CNRS UMR 7625, E´cole Normale Supe´rieure, 46 Rue d’Ulm, F-75230 Paris Cedex 05, France Submitted March 31, 2006; Accepted January 18, 2007; Electronically published May 11, 2007 Online enhancements: appendixes.

abstract: Though predation, productivity (nutrient richness), spatial heterogeneity, and disturbance regimes are known to influence species diversity, interactions between these factors remain largely unknown. Predation has been shown to interact with productivity and with spatial heterogeneity, but few experimental studies have focused on how predation and disturbance interact to influence prey diversity. We used theory and experiments to investigate how these factors influence diversification of Pseudomonas fluorescens by manipulating both predation (presence or absence of Bdellovibrio bacteriovorus) and disturbance (frequency and intensity of disturbance). Our results show that in a homogeneous environment, predation is essential to promote prey species diversity. However, in most but not all treatments, elevated diversity was transitory, implying that the effect of predation on diversity was strongly influenced by distur* Eric Michel and Christine D. M. Mu¨ller-Graf contributed equally to the manuscript. †

Corresponding author; e-mail: [email protected].

‡

E-mail: [email protected].

§

E-mail: [email protected].

k

E-mail: [email protected].

#

E-mail: [email protected].

** E-mail: [email protected]. ††

E-mail: [email protected].

‡‡

Present address: Bundesinstitut fu¨r Risikoforschung Alt-Marienfelde 17-21, D-12277 Berlin, Germany; e-mail: [email protected]. Am. Nat. 2007. Vol. 170, pp. 143–154. 䉷 2007 by The University of Chicago. 0003-0147/2007/17001-41730$15.00. All rights reserved.

bance. Both our experimental and theoretical results suggest that disturbance interacts with predation by modifying the interplay of resource and apparent competition among prey. Keywords: disturbance, predator-prey interactions, biodiversity dynamics, trade-off, Pseudomonas fluorescens, Bdellovibrio bacteriovorus.

Understanding the development and maintenance of biodiversity is a major ecological challenge. Factors such as predation (Menge and Sutherland 1976), productivity (Currie 1991; Abrams 1995; Waide et al. 1999), spatial heterogeneity (Chesson and Huntly 1997), and disturbance (Lenz et al. 2004) influence species diversity (Chesson 2000). However, interactions between these factors remain largely unexplored. Here, we show how predation and disturbance interact to influence the diversity of coexisting mutants of Pseudomonas fluorescens. Predation can favor coexistence of prey species that exploit the same resource via two different but nonexclusive mechanisms (Armstrong 1979; Paine 1980; Abrams 1993, 1999; Holt et al. 1994; Leibold 1996). The first mechanism operates if there is a trade-off between competitive abilities and resistance to predation (Bohannan and Lenski 2000). Then, the combination of direct and apparent competition (indirect competition between prey species through a shared predator [Armstrong 1979; Holt et al. 1994]) allows competitive but susceptible prey to coexist with less competitive but more resistant prey. The second mechanism operates via the reduction of prey population density due to predation, which decreases intra- or interspecific competition between prey and limits the risk of competitive exclusion (Paine 1966, 1969a, 1969b; Krivan 2003). However, predation does not always promote diversity. If predation is too intense, it may reduce prey species diversity (Sih et al. 1985; Cadotte and Fukami 2005). The effect of predation on prey species diversity can range from positive to negative, depending on various conditions (Chase et al. 2002). Most of the studies addressing the impact of predation on prey diversity assume constant conditions. However, it

144

The American Naturalist

is well known that disturbance can have various effects on diversity, and how the two factors interact is not at all clear. In the absence of predation, disturbance can also have positive or negative effects on diversity, depending on its frequency, intensity, and temporal pattern (Connell 1978; Pike et al. 2004; Massin and Gonzalez 2006). The intermediate disturbance hypothesis (IDH) captures some of this complexity by stating that diversity is maximized at intermediate frequencies or intensities of disturbance (Connell 1978; Floder and Sommer 1999; Morgan and Buckling 2004). However, most previous studies of the IDH focus on a single trophic level (but see Sousa 1979), and testing the applicability of the IDH in communities with multiple trophic levels warrants further study. A recent study (Wootton 1998) suggested that the IDH might not always hold for communities consisting of multiple trophic levels. Disturbance might affect diversity by adding (Menge and Sutherland 1976, 1987) or removing trophic levels that are more sensitive than others to disturbance. Individuals of a disturbance-sensitive trophic level would experience more resource supply and consequently less competition, allowing coexistence as suggested by the IDH. Fluctuations might also reduce predation pressure and competition for resources for short periods of time, but overall competition would not be alleviated, precluding any beneficial effect of intermediate disturbance (Wootton 1998). However, the various effects of disturbance may propagate through the trophic network (Roughgarden 1979) and thus produce indirect effects that facilitate maintenance of diversity. In particular, depending on its frequency and intensity, disturbance can change mean fitness differences in competing prey by modulating the effects of direct and apparent competition (Kneitel and Chase 2004). Clarifying the effect of the interaction between predation and disturbance on prey diversity thus requires a better understanding of the trade-off between prey resistance and prey competitive abilities. Until now, the interaction between predation and disturbance has been examined only with relatively short-term studies (Lubchenco and Menge 1978; Kneitel and Chase 2004; Morgan and Buckling 2004), which limits conclusions about effects on long-term diversity maintenance. Moreover, these studies vary only disturbance frequency, which leaves open the question of what effect disturbance intensity may have. To better understand the impact of disturbance on prey diversity in a multiple trophic-level community, we carried out an integrated theoretical and experimental analysis. We modeled a prey population facing a constant predation pressure under various disturbance intensities and frequencies. This analysis showed that different competitive ability-susceptibility trade-offs (affecting the cost of resis-

tance) may result in different effects of disturbance on prey diversity. We then investigated experimentally how predation (presence or absence of Bdellovibrio bacteriovorus) and six combinations of disturbance frequency and intensity influence the diversity of coexisting mutants of P. fluorescens. This bacterium is known to rapidly diversify into multiple mutants when ecological opportunities are available (Rainey and Travisano 1998). These mutants are functionally analogous to species, and invasion of adaptative mutations can be followed by tracking specific colony morphologies (see “Material and Methods”). This makes P. fluorescens a very good model system (Rainey and Travisano 1998; Buckling et al. 2000, 2003). Bdellovibrio bacteriovorus was chosen as the predator in our experiment first of all because the impact of this predator on bacterial diversity has received little study. Previous studies used a bacteriophage to examine the impact of predation on P. fluorescens diversity (Brockhurst et al. 2004; Morgan and Buckling 2004). Another reason is that predators with different behaviors can have different impacts on prey species diversity (Kurzava and Morin 1998; Sommer 1999). Bdellovibrio bacteriovorus is more like a conventional predator or parasitoid than the bacteriophages in that a successfully attacked prey produces only a modest number of predators (Stolp and Starr 1963). Disturbance consisted of periodic dilution events (varying in frequency and intensity) that affected the two trophic levels in our microcosms. Our experiment ran for 40–80 days (corresponding to approximately 200–266 prey generations), which was sufficient to observe long-term dynamics of diversity. Our results suggest that prey diversity was maintained under our low frequency and mild intensity disturbance regime, but increased diversity was transient under other disturbance treatments. This pattern concurred with the pattern predicted by a model assuming a trade-off between prey resistance and prey resource utilization.

Theory Theoretical approaches can be useful in studying the interaction between a predator and communities of prey species (Vance 1978; Leibold 1996). Some authors have focused on the influence of disturbance on prey diversity (Wootton 1998; Jansen and Mulder 1999). However, few models are tailored to Pseudomonas fluorescens (Buckling et al. 2000; Morgan et al. 2005), despite the growing number of experimental studies with this bacterium. We constructed a model that took particularities of the P. fluorescens– Bdellovibrio bacteriovorus interaction into account.

Predation and Disturbance Affect Prey Diversity The Model Our model is based on a standard Lotka-Volterra competition model for two prey species with a constant predation term added. We assumed that predator density is effectively constant not only because it simplified the model analysis but also because available data suggested that B. bacteriovorus grows much faster than P. fluorescens (Seidler and Starr 1969) and thus will tend to be at equilibrium. With these simplifications but taking mutations into account, dynamics of the two strains (denoted S and R) are given by the following differential equations: dX S p [(1 ⫺ m)bS ⫺ d ⫺ PSY ]X S ⫹ mb R X R, dt dX R p [(1 ⫺ m)b R ⫺ d ⫺ PRY ]X R ⫹ mbS X S , dt where XS and XR represent the densities of the two prey and Y the density of the predator. The mutation rate between the two strains was m, bS and bR were their growth rates, d was the background death rate, and PS and PR were the predation rates of the common predator on the two strains. Prey rates of reproduction were assumed to be limited by the total prey density. Thus, bS p C S[1 ⫺ K S (X S ⫹ X R)], b R p C R[1 ⫺ K R (X S ⫹ X R)], where CS and CR are the intrinsic growth rates of the two strains and KS and KR are constants that measure sensitivity to density dependence (see app. A in the online edition of the American Naturalist for further details). In the following, we assume that strain S was sensitive to predation and that strain R was resistant to predation (and thus bore the cost of resistance). Comparing Trade-Offs Experimental results indicated that among P. fluorescens mutants, there may be 10-fold differences in resistance to predation by B. bacteriovorus (fuzzy spreader [FS] mutants are more resistant than smooth morph [SM] or wrinkly spreader [WS] mutants). Our key assumption in formulating the model was that such resistance to predation is costly. It is important to realize that such costs may manifest themselves in more than one way. Here, we consider two ways of being competitive that lead to two possible trade-offs. To be competitive, a bacterium could either grow rapidly or better exploit the resources in the environment and thus achieve higher equilibrium densities. A similar trade-off between resource utilization and stress

145

resistance has been shown experimentally in Escherichia coli (King et al. 2004). We varied the intensity of the tradeoff by increasing or decreasing the growth rate (Ci) or the equilibrium density (Xˆ i, included in the Ki term) of the resistant species (see app. A for further details). These equations describe the dynamics in between dilution events. We analyzed the consequences of these two different trade-offs by following the population densities of the sensitive and the resistant strains in time. We assumed that initial densities of both strains were equal, because it allowed us to easily compare the fates of sensitive and resistant strains. The impact of the type of trade-off on the final diversity was very clear when we compared density dynamics between two dilution events. When resistance to predation was traded off against the growth rate and when the cost of resistance was intermediate, the sensitive strain dominated until t p 48 h, while the resistant strain dominated after t p 96 h (fig. 1B). When the cost of resistance was high (i.e., when the growth rate of strain R was too low compared with that of strain S), then the sensitive strain dominated from the beginning (fig. 1C), whereas when the cost of resistance was low, the resistant strain dominated from the beginning (fig. 1A). This result is important since a strain that dominates when the dilution events occur will have an advantage in the next round. Thus, our model suggests that, after several dilution events, several outcomes might be observed, depending on the frequency of dilution (i.e., of disturbance) and on the intensity of the trade-off. When the trade-off was between resistance to predation and resource utilization, there were also three possible outcomes. For intermediate costs of resistance, the resistant strain dominated at t p 48 h, while the sensitive strain came to dominate at t p 96 h (fig. 1E). As in the previous case, when the cost of resistance was high, the sensitive strain dominated all over (fig. 1F), whereas when the cost of resistance was low, the resistant strain dominated from the beginning (fig. 1D). The two trade-offs thus predict opposite patterns for intermediate costs of resistance. When resistance decreases the prey growth rate, our model predicts a possible change in dominance from sensitive to resistant strains when decreasing disturbance frequency, whereas if resistance decreases resource utilization, the model predicts the reverse. Influence of Disturbance on Prey Diversity We used the same model to study the impact of disturbance intensity and frequency on long-term prey diversity. We did this by continuing the numerical solution of the system of equations but periodically reducing densities to

146

The American Naturalist

Figure 1: Influence of the trade-off function on the relative densities of the sensitive strain, XS (dotted line), and of the resistant strain, XR (dashed

line), along time. Assuming a weak trade-off between resistance and growth rate (A; CR p 0.25 , Xˆ R p 3 # 108 ), an intermediate trade-off between resistance and growth rate (B; CR p 0.2, Xˆ R p 3 # 108 ), a strong trade-off between resistance and growth rate (C; CR p 0.15, Xˆ R p 3 # 108), a weak trade-off between resistance and resource utilization (D; CR p 1.0 , Xˆ R p 1.6 # 108 ), an intermediate trade-off between resistance and resource utilization (E; CR p 1 , Xˆ R p 1.5 # 108 ), and a strong trade-off between resistance and resource utilization (F; CR p 1 , Xˆ R p 1.4 # 108 ). In all these plots, at t p 0, XS p XR p 106, m p 10⫺5, d p 0.01, PS p 0.5, PR p 0.053, CS p 1, and Xˆ S p 3 # 108 bacteria. See appendix A in the online edition of the American Naturalist for further details.

mimic disturbance events of varying frequency and intensity. Results obtained when assuming an intermediate trade-off between resistance and growth rate or between resistance and resource utilization are shown in figure 2. They indicate that in the presence of predators, longterm maintenance of diversity may depend on the disturbance regime. Moreover, disturbance intensity and disturbance frequency might influence diversity in different ways. For instance, in figure 2B, decreasing disturbance frequency first leads to higher species diversity, but after a given threshold, diversity decreases. Comparison of figure 2A and 2B confirmed that the pattern of diversity that is expected under a particular disturbance regime may strongly depend on the underlying trade-off between resistance and competitive ability. Observing a system response to changes in disturbance regime may thus give insight into the underlying trade-offs. Experimental Approach Material and Methods Bacterial Strains. Pseudomonas fluorescens SBW25 was the prey in our experiment. In static liquid medium, this

Gram-negative bacterium diversifies into three morphs adapted to the different niches of the environment (Rainey and Travisano 1998; Rainey 2005): (i) the SM (the ancestral morph) mutant is adapted to the liquid medium; (ii) the WS mutant is adapted to the air-liquid interface, where it builds a biofilm; (iii) the FS mutant is adapted to conditions prevailing at the bottom of the culture. In unstructured environments (shaken liquid medium), diversification does not occur, and only the ancestral morph (SM) is detected. Bdellovibrio bacteriovorus 109J (American Culture Type Collection) was the predator. This very small (0.2–0.5 mm wide and 0.5–2.5 mm long) Gram-negative bacterium is a ubiquitous obligate predator hunting a wide range of Gram-negative bacteria (Stolp and Starr 1963; Rendulic et al. 2004). The life cycle of this predatory bacterium is composed of (i) an attack phase (seeking after a prey) and (ii) a growth phase in the periplasm of the prey. During this latter phase, the predator kills the prey by ingesting its cytoplasm, becomes filamentous, and gives rise to four to five daughter cells, which emerge from the prey cell, ready to find new prey. Experimental Design. Thirty-six 60-mL polycarbonate bot-

Predation and Disturbance Affect Prey Diversity

147

Disturbance was operationally defined as a dilution event during which only a small part of the population survives. We created six different disturbance regimes by modulating two components of this factor: (i) disturbance intensity with two different levels, mild (100-fold dilution p 1% of survival) and strong (1,000-fold dilution p 0.1% of survival); and (ii) disturbance frequency with three different levels, high (dilution every 2 days), intermediate (every 3 days), and low (every 4 days). All cultures were propagated (i.e., diluted into fresh medium) 20 times, that is, for 40, 60, and 80 days (depending on the transfer treatment). Thus, we followed the diversity dynamics of 36 populations (18 with or 18 without predators), which were cultivated under six different disturbance regimes, using three microcosms per treatment. After an episode of predation, prey cells become temporarily (plastically) resistant (until the next exponential phase after which they become sensitive again) to B. bacteriovorus (Shemesh and Jurkevitch 2004) and multiply again by 10-fold. There was only one episode of predation per transfer (occurring between day 1 and day 2), which means that all populations underwent the same number of episodes of predation, whatever the disturbance regime.

Figure 2: Influence of both dilution frequency and intensity on prey diversity. Diversity (in the form of the Simpson index) is predicted by the mathematical model for 20 dilutions as a function of dilution intensity and frequency. High-diversity areas are in white, and low-diversity areas are in black (the dominant strain is indicated in those areas). The white and black circles indicate the six combinations of values that were tested experimentally (see “Experimental Approach”). Resistance to predation is traded off against the growth rate (A; CR p 0.2, Xˆ R p 3 # 108) or resource utilization (B; CR p 1 , Xˆ R p 1.5 # 108 ). In both these plots, parameter values are CS p 1, Xˆ S p 3 # 108, m p 10⫺5, d p 0.01, PS p 0.5, and PR p 0.053.

tles (Nalgene) containing 6 mL of diluted nutrient broth (DNB) were inoculated with 107 P. fluorescens isolate SBW25 prey bacteria. Eighteen of these populations received 105 B. bacteriovorus isolate 109J predatory bacteria. Populations were propagated at 28⬚C under constant agitation (200 revolutions/min) to ensure homogeneity of the environment and thus in conditions under which P. fluorescens diversification is not expected.

Competition Assay. To test for a trade-off between resistance and competitive abilities in P. fluorescens, we performed competition experiments following the protocol used by Brockhurst et al. (2004) to evaluate the invasive abilities of FS mutants when rare (1 : 100). Competitors (SM and FS ancestral mutants) were grown separately for 24 h at 28⬚C and shaken at 200 revolutions/ min in King’s broth (KB) medium. After centrifugation (5 min at 4,000 revolutions/min), bacteria were resuspended in fresh DNB medium. Ten 60-mL polycarbonate bottles containing 6 mL of DNB were each inoculated with 107 SM and 105 FS mutants. Bdellovibrio bacteriovorus were added to five of the 10 replicate cultures. The experiment was run for 96 h (4 days), the duration of the longest transfer in our experiment. In order to survey population dynamics of both competitors, 100 mL were removed every day from the cultures, diluted, and plated on four KB agar plates. Bacterial Counts and Culture Media. Bacteria were grown in DNB liquid medium: 0.8 g of nutrient broth, 1,000 mL H2O, MgCl2 3 mM, CaCl2 2 mM, pH 7.2 (Jurkevitch et al. 2000). Predators were counted as plaque-forming units (PFU) developing on a lawn of ancestral prey cells on DNB plates (0.8 g of nutrient broth, 15 g agar, 1,000 mL H2O, MgCl2 3 mM, CaCl2 2 mM, pH 7.2). Prey mutant concentrations were estimated by counting a minimum of 100 colonies (colony-forming units) on four replicate KB agar plates (20 g proteose peptone 3, 15 g agar [for plating

148

The American Naturalist

medium], 10 mL ethylene glycol, 1,000 mL H2O quantity sufficient for 8.6 mL K2HPO4 [1M], 6.1 mL MgSO4 [1M] added after autoclaving) 2 days after plating. Measurement of Diversity. The different mutant (SM, WS, and FS) morphologies allowed us to determine each mutant’s frequency in a population (the 36 replicate bottles were considered as 36 independent populations) by counting more than 100 colonies on four replicate KB agar plates. Our diversity measures were based on the probability that two randomly chosen colonies were morphologically different. Sympatric diversity (within population), ds, was calculated as the unbiased estimate of the complement of Simpson’s index,

(

1⫺lp 1⫺

冘 i

pi2

) N N⫺ 1 ,

where pi is the morph proportion and N the sample size (Lande 1996; Brockhurst et al. 2004). Allopatric diversity d (i.e., diversity between replicate populations from the same disturbance regime; population divergence) was then given by d2 p

冘冘 i

(pij ⫺ p¯ i)2,

j

where p¯ i is the mean proportion of the ith morph across all the j populations (Lande 1996). Statistical Analyses Diversity Dynamics. In order to understand whether the frequency or intensity of disturbance favored the appearance of diversity, we performed a two-way ANOVA on the number of transfers necessary to detect diversification beyond the initial ancestral morph (analogous to the date of appearance). The number of transfers during which mutants were detected was used to examine whether frequency or intensity of disturbance favored their maintenance (JMP 5.1 software). We chose transfer number as a variable instead of time because the number of prey generations and the number of predation episodes were related to transfer events. Sympatric Diversity. We performed a two-way ANOVA with repeated measures on our sympatric diversity measures (Super ANOVA software, Abacus Concepts). The model took into account the effects of disturbance intensity, disturbance frequency, transfer number, and all interactions between these parameters. To meet the assumption of the homoscedasticity, we performed the

ANOVA on the logarithm of sympatric diversity, log (d s ⫹ 1). Results In the treatments without predators, the ancestral morph occurred in 17 out of the 18 cultures, and low diversity was observed in the remaining one (data not shown). In sharp contrast, increased prey diversity was observed in all transfer treatments (at least temporarily) containing predators (fig. 3). Increased diversity was due to the emergence of two Pseudomonas fluorescens mutants called WS and FS (Rainey and Travisano 1998) in populations initially composed of the ancestral mutant called SM. Bdellovibrio bacteriovorus was never washed out, and its population densities remained roughly constant (ranging from 107 to 108 PFU) in all transfer treatments (data not shown), which led us to consider predation pressure as a constant parameter in the interpretation of our results and their modeling. Population Dynamics Figure 3 shows prey dynamics through time in homogeneous environments with predators. Wrinkly spreader and FS mutants appeared during the adaptive radiation episode at the beginning of the experiment, but WS mutants were not detected at the end of the experiment in any of the 18 populations. Fuzzy spreader mutants were present at the end of the experiment in 15 out of 18 populations. More precisely, FS were found in all high disturbance frequency treatments, in all mild disturbance treatments, and in three out of the six cultures in strong intensity treatments at intermediate and low frequencies of disturbance. Smooth morph mutants were present at the end of the experiment in six out of 18 populations. They went to fixation in three out of six populations in strong intensity treatments at intermediate and low frequencies of disturbance and coexisted with FS in the three other populations in low frequency disturbance treatments. Diversity Dynamics Diversity appeared faster under low disturbance frequency (disturbance frequency effect: F p 9.0, df p 2, P p .004), whatever the disturbance intensity (disturbance intensity effect: F p 3.0, df p 1, P p .11; interaction: F p 0.0, df p 2, P p 1; see figs. 3, 4). Among the 20 transfers of the experiment, increased sympatric (within-population) diversity was maintained only when the disturbance regime was of mild intensity with a low frequency (fig. 4E, solid lines). Under the other disturbance regimes, we observed only a transient increase

Predation and Disturbance Affect Prey Diversity

149

Figure 3: Population dynamics in our 36 populations. Each graph shows the prey mutant concentrations (in colony-forming units [CFU]) through time (blue, smooth morph; red, fuzzy spreader; green, wrinkly spreader). Populations were transferred following six different disturbance regimes: D2d100 p dilution by a 100-fold every 2 days, D2d1000 p dilution by a 1,000-fold every 2 days, D3d100 p dilution by a 100-fold every 3 days, D3d1000 p dilution by a 1,000-fold every 3 days, D4d100 p dilution by a 100-fold every 4 days, D4d1000 p dilution by a 1,000-fold every 4 days.

in sympatric diversity. We performed a two-way ANOVA with repeated measures to understand how parameters influence the maintenance of sympatric diversity. The results show significant effects of (i) interaction between frequency and intensity of disturbance (F p 6.689, df p 2, P p .011), (ii) disturbance frequency (F p

4.484, df p 2, P p .035), (iii) disturbance intensity (F p 8.551, df p 1, P p .013), and (iv) transfer (our measure of time; F p 6.406, df p 20, P p .0001). The interaction between frequency and intensity of disturbance was explained by the fact that disturbance frequency had an effect on sympatric diversity (the less frequent the dis-

150

The American Naturalist

Figure 4: Sympatric and allopatric diversity dynamics along the 20 transfers of the experiment. Diversity as a function of time under different disturbance intensities (mild, 100-fold dilution [black] and strong, 1,000-fold dilution [white]) and different disturbance frequencies (A, B: high frequency [every 2 days]; C, D: intermediate frequency [every 3 days]; E, F: low frequency [every 4 days]). In A, C, and E, each symbol (circle, triangle, square) corresponds to a different replicate.

turbances, the higher the diversity) only when the intensity of disturbance was mild. When the intensity of disturbance was strong, sympatric diversity was low whatever the disturbance frequency. On the contrary, allopatric diversity (diversity between replicate populations from the same disturbance regime) was maintained in three disturbance treatments: when cultures were disturbed at a low frequency under both mild and strong disturbance intensity (fig. 4F) and when cultures were diluted at intermediate frequency under strong disturbance intensity (figs. 3, 4D). This maintenance of allopatric diversity was due to the fixation of different mutants in various replicate populations. Disturbance intensity had antagonistic effects on sympatric and allopatric diversity: the more intense the disturbance, the lower the

sympatric diversity, but the higher the allopatric diversity (fig. 4). Experimental Demonstration of a Trade-Off in Pseudomonas fluorescens Mutants We tested the hypothesis of a trade-off between resistance to predation and competitive abilities in P. fluorescens to explain the maintenance of diversity through frequencydependent selection. The results showed that FS mutants were able to invade SM populations when rare (ratio ≈ 1/100) in the presence of predators but not in the absence of predators (data not shown). Measurement of resistance of SM and FS mutants to predation showed that FS are 10 times more resistant than SM (data not shown), in-

Predation and Disturbance Affect Prey Diversity dicating that FS mutants were able to invade SM populations because of resistance and not because their low density protected them from predation. This result shows that a single mutant cannot dominate in both the presence and the absence of predators, which supports the idea of a trade-off between resistance to predation and competitive abilities. Discussion In the absence of predators, competition was the only selective pressure in our experiment, and no diversity appeared because populations were dominated by the SM mutant. However, when both competition and predation affected prey populations, diversity was detected in all cases. This positive impact of predation on diversity may be due to a trade-off in prey between resource utilization and resistance to predation, as was consistent with the predictions of the model. Disturbance frequency and intensity modulated the impact of predation by favoring different mutants in different disturbance regimes. Competition and Predation Rainey and Travisano (1998) demonstrated that competition drives the adaptive radiation of Pseudomonas fluorescens only when ecological opportunities such as environmental heterogeneity are available (i.e., in unshaken microcosms). In our experiment, we observed a rapid diversification of P. fluorescens despite prevention of environmental heterogeneity by constant shaking but only when predators were present. We explain this diversification as a consequence of the selection of costly mutations conferring resistance to predation. We observed three mutants that used three different resistance strategies. The strategy of the SM mutant appeared to be to divide as rapidly as possible and to pay the price of predation. The strategy of the FS mutant was to divide more slowly but to be more resistant to predation. Finally, the WS mutant’s strategy was to exploit a physical refuge (by growing on the walls of the bottles). We never observed fixation or long-term maintenance of WS mutants, probably because this “refuge” strategy was less efficient because of (i) a very low transmission rate due to our transfer protocol, (ii) a trade-off between reproduction and polymer production necessary to form a biofilm (Spiers et al. 2003), and (iii) a limited surface of the bottle walls. Thus, in our experiment, it was not competition but predation that drove the emergence of increased diversity. However, the impact of predators on diversity may vary between systems (Kurzava and Morin 1998; Sommer 1999). In a similar experiment by Brockhurst et al. (2004),

151

diversity was detected in both the presence and absence of predators. Different predator-prey population dynamics can be induced when predators are either phages or bdellovibrios because (i) phages are generally known to attack dividing cells, whereas bdellovibrios attack bacteria in stationary phase; and (ii) the plastic resistance (Shemesh and Jurkevitch 2004) of residual prey after a bdellovibrio attack (possibly resistant because leaving stationary phase or because of some quorum sensing) tends to reduce the selective pressure induced by predators. A further factor may be productivity (higher in Brockhurst’s than in our experiment), which allows for faster appearance of resistant bacterial mutants (Bohannan and Lenski 1997). Trade-Off between Resistance and Competitive Abilities In our study, similar to the study by Brockhurst et al. (2004), the positive effect of predation on diversity was explained by a trade-off between resistance and competitive abilities in prey. Comparing our theoretical and experimental results helped us to understand the underlying mechanisms of this trade-off. One result was that treatments with low disturbance frequency sometimes led to the fixation of sensitive (i.e., SM) mutants, whereas treatments with high disturbance frequency always led to the fixation of resistant (FS) mutants. In our model, this pattern corresponds to a trade-off between resistance and resource utilization (fig. 1E). In the model, we never observed a change of dominance from resistant to sensitive mutant while decreasing disturbance frequency when the trade-off was only between the resistance to predation and the growth rate (see “The Model”). Simulations modeling transfer experiments assuming a trade-off between resistance and resource utilization (app. B in the online edition of the American Naturalist) also gave results similar to the patterns observed experimentally (fig. 3). To strengthen this theoretical result, we compared experimentally the equilibrium density of SM and FS mutants grown in isolation and without predation. We found that the SM mutant achieved significantly higher densities than the FS mutant (see app. A). Together, our theoretical and experimental results suggest that resource utilization could be a lifehistory trait traded off against resistance to predation. Such a trade-off allowed the coexistence of different prey mutants. However, the maintenance of this diversity was strongly dependent on disturbance regimes. Interaction between Predation and Disturbance Theory predicts that disturbance may have a different effect on diversity in multiple trophic-level communities, depending on how disturbance affects each trophic level (Wootton 1998). For instance, Menge and Sutherland

152

The American Naturalist

(1976, 1987) showed that disturbance can modulate the effects of predation if predators are more affected by disturbance than prey. However, because the predator population densities remained roughly constant during the course of our experiment in all transfer treatments (data not shown), we concluded that this was not the case in our predator-prey system. On the other hand, the hypothesis that disturbance can interact with predation was strongly supported by our finding of a trade-off between resistance and resource utilization in prey. Resistant mutants had an obvious selective advantage over sensitive mutants, but this was balanced under different disturbance regimes. Because prey grew again after an episode of predation, resistant mutants (which grow slower) were favored at high disturbance frequencies (i.e., when the growth period was reduced), while sensitive prey were favored at low disturbance frequencies. Sensitive prey were also favored under strong disturbance intensities because these treatments allow longer exponential phases. Disturbance reduces mean fitness differences in competing prey by tuning their competition and apparent competition interactions (Kneitel and Chase 2004). Intermediate Disturbance Hypothesis (IDH) Very different diversity dynamics were obtained under the six disturbance treatments (fig. 4). At low disturbance frequency, both sympatric and allopatric diversity were maximized. In contrast to disturbance frequency, disturbance intensity had different effects on sympatric and allopatric diversity in our experiment, as suggested by the model (fig. 2B). Mild disturbance intensity allowed the establishment of both sympatric and allopatric diversity. However, strong disturbance intensity tended to reduce sympatric diversity but to promote allopatric diversity. This may be explained by the fact that intense dilution (here a 1,000fold dilution) induced strong genetic drift, leading to a rapid fixation of a mutant in a population. Such fixation diminished sympatric diversity but could have promoted allopatric diversity if different mutants were fixed in different populations. Do these results support or give new insights into the IDH, which states that diversity is maximized at an intermediate level of disturbance? Our experimental range of disturbance frequency corresponded to a range of high to intermediate disturbance frequencies used in previous experiments specifically designed to test the IDH with P. fluorescens (Buckling et al. 2000; Morgan and Buckling 2004). Thus, our low disturbance frequency treatment actually corresponded to an intermediate frequency of disturbance in these previous studies. We found experimentally and theoretically that diversity

was indeed maximized at an intermediate frequency of disturbance (our low frequency treatment under mild intensity). In a comparable configuration (two trophic levels, competitors at the bottom), Wootton (1998) found with a theoretical approach that intermediate levels of disturbance can potentially promote coexistence if the competitive dominant is more strongly affected by disturbance. Contrary to Wootton’s predictions, we observed that our results corroborate the IDH even though the competitive dominant mutant (FS resistant) is favored and not disfavored by disturbance. We explain this discrepancy by the fact that there are several mechanisms through which disturbance can promote diversity. Here, we show that the existence of a trade-off between resistance and resource utilization is a mechanism that had not been considered by Wootton. Experimentally, Morgan and Buckling (2004) showed in a heterogeneous environment that the IDH pattern was evident only in the absence of predators, whereas in our experiment, in a homogeneous environment the IDH was observed only in the presence of predators. These different outcomes require further research. Rigorously testing the IDH would require a treatment without disturbance. But the fact that disturbance is associated with resource supply (Shea et al. 2004) precluded the use of such a treatment because in a long-term experiment, the starvation effect would be confounded by the effect of the absence of disturbance. We chose to interpret diversity emergence in our system as a consequence of invariant ecological parameters alone. However, this may not be the only possible explanation of our results. We assumed implicitly in our interpretation that virulence and resistance stayed constant, but coevolution of prey and predators during the course of the experiment could lead to a similar outcome (Hochberg and van Baalen 1998). Also, frequency-dependent selection can explain the maintenance of diversity when assuming the specialization of different predator subpopulations on different mutants (Bohannan et al. 2002). Coevolution remains an additional factor, and this will be addressed in a subsequent study. Both our experimental and theoretical approaches led to the conclusion that disturbance interacted with predation by modifying the combination of resource and apparent competition between prey. A trade-off between resistance to predation and resource utilization could allow long-term coexistence of different mutants in a multiple trophic-level system. In a long-term experiment, predation had a positive effect on diversity, even if most of the time this effect was temporary. The effect of predation was strongly modulated by disturbance, and we observed longterm maintenance of diversity only when cultures underwent mild intensity and low frequency of disturbance,

Predation and Disturbance Affect Prey Diversity which seems to corroborate the IDH. These findings also suggest that (i) different parameters of disturbance should be tested instead of one and (ii) because predation and apparent competition are concomitant, future studies involving the impact of predation on diversity should always be conducted on a timescale long enough to test whether this observed diversity is transitory or whether it can be maintained. Acknowledgments We would like to thank A. Buckling and J. M. Ghigo for fruitful discussions; E. Jurkevitch for providing the Bdellovibrio bacteriovorus 109J strain; P. Rainey for letting us read his unpublished manuscript; and M. Danger, C. Fontaine, G. Lacroix, N. Massin, E. The´bault, T. Tully, and three anonymous reviewers for providing helpful comments that improved this manuscript. We also thank M. Huet for technical support and lab management. Literature Cited Abrams, P. A. 1993. Effect of increased productivity on the abundances of trophic levels. American Naturalist 141:351–371. ———. 1995. Monotonic or unimodal diversity productivity gradients: what does competition theory predict? Ecology 76:2019– 2027. ———. 1999. Is predator-mediated coexistence possible in unstable systems? Ecology 80:608–621. Armstrong, R. A. 1979. Prey species replacement along a gradient of nutrient enrichment: graphical approach. Ecology 60:76–84. Bohannan, B. J. M., and R. E. Lenski. 1997. Effect of resource enrichment on a chemostat community of bacteria and bacteriophage. Ecology 78:2303–2315. ———. 2000. The relative importance of competition and predation varies with productivity in a model community. American Naturalist 156:329–340. Bohannan, B. J. M., B. Kerr, C. M. Jessup, J. B. Hughes, and G. Sandvik. 2002. Trade-offs and coexistence in microbial microcosms. Antonie van Leeuwenhoek International Journal of General and Molecular Microbiology 81:107–115. Brockhurst, M. A., P. B. Rainey, and A. Buckling. 2004. The effect of spatial heterogeneity and parasites on the evolution of host diversity. Proceedings of the Royal Society B: Biological Sciences 271:107–111. Buckling, A., R. Kassen, G. Bell, and P. B. Rainey. 2000. Disturbance and diversity in experimental microcosms. Nature 408:961–964. Buckling, A., M. A. Wills, and N. Colegrave. 2003. Adaptation limits diversification of experimental bacterial populations. Science 302: 2107–2109. Cadotte, M. W., and T. Fukami. 2005. Dispersal, spatial scale, and species diversity in a hierarchically structured experimental landscape. Ecology Letters 8:548–557. Chase, J. M., P. A. Abrams, J. P. Grover, S. Diehl, P. Chesson, R. D. Holt, S. A. Richards, R. M. Nisbet, and T. J. Case. 2002. The interaction between predation and competition: a review and synthesis. Ecology Letters 5:302–315.

153

Chesson, P. 2000. Mechanisms of maintenance of species diversity. Annual Review of Ecology and Systematics 31:343–366. Chesson, P., and N. Huntly. 1997. The roles of harsh and fluctuating conditions in the dynamics of ecological communities. American Naturalist 150:519–553. Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs: high diversity of trees and corals is maintained only in a nonequilibrium state. Science 199:1302–1310. Currie, D. J. 1991. Energy and large-scale patterns of animal species and plant species richness. American Naturalist 137:27–49. Floder, S., and U. Sommer. 1999. Diversity in planktonic communities: an experimental test of the intermediate disturbance hypothesis. Limnology and Oceanography 44:1114–1119. Hochberg, M. E., and M. van Baalen. 1998. Antagonistic coevolution over productivity gradients. American Naturalist 152:620–634. Holt, R. D., J. Grover, and D. Tilman. 1994. Simple rules for interspecific dominance in systems with exploitative and apparent competition. American Naturalist 144:741–771. Jansen, V. A. A., and G. Mulder. 1999. Evolving biodiversity. Ecology Letters 2:379–386. Jurkevitch, E., D. Minz, B. Ramati, and G. Barel. 2000. Prey range characterization, ribotyping, and diversity of soil and rhizosphere Bdellovibrio spp. isolated on phytopathogenic bacteria. Applied and Environmental Microbiology 66:2365–2371. King, T., A. Ishihama, A. Kori, and T. Ferenci. 2004. A regulatory trade-off as a source of strain variation in the species Escherichia coli. Journal of Bacteriology 186:5614–5620. Kneitel, J. M., and J. M. Chase. 2004. Disturbance, predator, and resource interactions alter container community composition. Ecology 85:2088–2093. Krivan, V. 2003. Competitive co-existence caused by adaptive predators. Evolutionary Ecology Research 5:1163–1182. Kurzava, L. M., and P. J. Morin. 1998. Tests of functional equivalence: complementary roles of salamanders and fish in community organization. Ecology 79:477–489. Lande, R. 1996. Statistics and partitioning of species diversity, and similarity among multiple communities. Oikos 76:5–13. Leibold, M. A. 1996. A graphical model of keystone predators in food webs: trophic regulation of abundance, incidence, and diversity patterns in communities. American Naturalist 147:784–812. Lenz, M., M. Molis, and M. Wahl. 2004. Testing the intermediate disturbance hypothesis: response of fouling communities to various levels of emersion intensity. Marine Ecology Progress Series 278:53–65. Lubchenco, J., and B. A. Menge. 1978. Community development and persistence in a low rocky intertidal zone. Ecological Monographs 48:67–94. Massin, N., and A. Gonzalez. 2006. Adaptive radiation in a fluctuating environment: disturbance affects the evolution of diversity in a bacterial microcosm. Evolutionary Ecology Research 8:1–11. Menge, B. A., and J. P. Sutherland. 1976. Species diversity gradients: synthesis of roles of predation, competition, and temporal heterogeneity. American Naturalist 110:351–369. ———. 1987. Community regulation: variation in disturbance, competition, and predation in relation to environmental stress and recruitment. American Naturalist 130:730–757. Morgan, A. D., and A. Buckling. 2004. Parasites mediate the relationship between host diversity and disturbance frequency. Ecology Letters 7:1029–1034. Morgan, A. D., S. Gandon, and A. Buckling. 2005. The effect of

154

The American Naturalist

migration on local adaptation in a coevolving host-parasite system. Nature 437:253–256. Paine, R. T. 1966. Food web complexity and species diversity. American Naturalist 100:65–75. ———. 1969a. A note on trophic complexity and community stability. American Naturalist 103:91–93. ———. 1969b. Pisaster-Tegula interaction: prey patches, predator food preference, and intertidal community structure. Ecology 50: 950–961. ———. 1980. Food webs: linkage, interaction strength and community infrastructure. Journal of Animal Ecology 49:667–685. Pike, N., T. Tully, P. Haccou, and R. Ferriere. 2004. The effect of autocorrelation in environmental variability on the persistence of populations: an experimental test. Proceedings of the Royal Society B: Biological Sciences 271:2143–2148. Rainey, P. B. 2005. Bacteria evolve and function within communities: observations from experimental Pseudomonas populations. Pages 83–102 in M. J. McFall-Ngai, B. Henderson, and N. Ruby, eds. The influence of cooperative bacteria on animal host biology: advances in molecular and cellular microbiology. Cambridge University Press, Cambridge. Rainey, P. B., and M. Travisano. 1998. Adaptive radiation in a heterogeneous environment. Nature 394:69–72. Rendulic, S., P. Jagtap, A. Rosinus, M. Eppinger, C. Baar, C. Lanz, H. Keller, et al. 2004. A predator unmasked: life cycle of Bdellovibrio bacteriovorus from a genomic perspective. Science 303:689–692. Roughgarden, J. 1979. Theory of population genetics and evolutionary ecology: an introduction. Macmillan, New York. Seidler, R. J., and M. P. Starr. 1969. Factors affecting the intracellular parasitic growth of Bdellovibrio bacteriovorus developing within Escherichia coli. Journal of Bacteriology 97:912–923. Shea, K., S. H. Roxburgh, and E. S. J. Rauschert. 2004. Moving from pattern to process: coexistence mechanisms under intermediate disturbance regimes. Ecology Letters 7:491–508.

Shemesh, Y., and E. Jurkevitch. 2004. Plastic phenotypic resistance to predation by Bdellovibrio and like organisms in bacterial prey. Environmental Microbiology 6:12–18. Sih, A., P. Crowley, M. McPeek, J. Petranka, and K. Strohmeier. 1985. Predation, competition, and prey communities: a review of field experiments. Annual Review of Ecology and Systematics 16:269– 311. Sommer, U. 1999. The impact of herbivore type and grazing pressure on benthic microalgal diversity. Ecology Letters 2:65–69. Sousa, W. P. 1979. Experimental investigations of disturbance and ecological succession in a rocky intertidal algal community. Ecological Monographs 49:227–254. Spiers, A. J., J. Bohannan, S. M. Gehrig, and P. B. Rainey. 2003. Biofilm formation at the air-liquid interface by the Pseudomonas fluorescens SBW25 wrinkly spreader requires an acetylated form of cellulose. Molecular Microbiology 50:15–27. Stolp, H., and M. P. Starr. 1963. Bdellovibrio bacteriovorus gen. et sp. n.: a predatory, ectoparasitic, and bacteriolytic microorganism. Antonie van Leeuwenhoek Journal of Microbiology and Serology 29: 217–248. Vance, R. R. 1978. Predation and resource partitioning in one predator–two prey model communities. American Naturalist 112:797– 813. Waide, R. B., M. R. Willig, C. F. Steiner, G. Mittelbach, L. Gough, S. I. Dodson, G. P. Juday, and R. Parmenter. 1999. The relationship between productivity and species richness. Annual Review of Ecology and Systematics 30:257–300. Wootton, J. T. 1998. Effects of disturbance on species diversity: a multitrophic perspective. American Naturalist 152:803–825.

Associate Editor: Peter J. Morin Editor: Donald L. DeAngelis