Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Journal of Animal Ecology 2010, 79, 693–700 doi: 10.1111/j.1365-2656.2010.01662.x The interacting effects of temperature and food chain length on trophic abundance and ecosystem function Oliver S. Beveridge1*, Stuart Humphries2 and Owen L. Petchey1 1 Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield, S10 2TN, UK; and Department of Biological Sciences, University of Hull, Cottingham Road, Kingston-upon-Hull, HU6 7RX, UK 2 Summary 1. While much is known about the independent effects of trophic structure and temperature on density and ecosystem processes, less is known about the interaction(s) between the two. 2. We manipulated the temperature of laboratory-based bacteria-protist communities that contained communities with one, two, or three trophic levels, and recorded species’ densities and bacterial decomposition. 3. Temperature, food chain length and their interaction produced significant responses in microbial density and bacterial decomposition. Prey and resource density expressed different patterns of temperature dependency during different phases of population dynamics. The addition of a predator altered the temperature-density relationship of prey, from a unimodal trend to a negative one. Bacterial decomposition was greatest in the presence of consumers at higher temperatures. 4. These results are qualitatively consistent with a recent model of direct and indirect temperature effects on resource-consumer population dynamics. Results highlight and reinforce the importance of indirect effects of temperature mediated through trophic interactions. Understanding and predicting the consequences of environmental change will require that indirect effects, trophic structure, and individual species’ tolerances be incorporated into theory and models. Key-words: Colpidium striatum, decomposition, Didinium nasutum, ecosystem function, indirect and direct temperature effects, trophic interactions Introduction The temperature of aquatic systems is highly variable, both spatially (Botte & Kay 2000; Coats et al. 2006) and temporally (Moralesbaquero & Cruzpizarro 1995; Balistrieri et al. 2006). The direct effects of temperature on organisms are well documented, and include effects on: metabolic rate (Clarke 1991; Gillooly et al. 2001; Brown et al. 2004; Savage et al. 2004; Apple, Del Giorgi & Kemp 2006), nerve conduction (Clarke 1991; Moran & Melani 2001), digestion ⁄ gut passage time (Yee & Murray 2004) and cilial or flagellar activity (Sleigh 1956). Consequently, individual-level processes such as swimming (Winet 1976; Podolsky & Emlet 1993; Wilson, James & Johnston 2000; Wilson 2005), feeding (Bolton & Havenhand 1998; Yee & Murray 2004) and growth (Rose & Caron 2007) are directly affected by temperature. Temperature can affect populations indirectly via impacts on trophic interactions (Ives 1995). Populations can be con*Correspondence author. E-mail: [email protected] sidered to interact when actions of one population affect the characteristics of a second (Abrams 1987). Prey density may be indirectly affected by temperature via changes in predation rate, for example. In polychaetes a 10 C reduction in temperature is correlated with a 34–67% decrease in ingestion rate (Bolton & Havenhand 1998, 2005; Loiterton, Sundbom & Vrede 2004). Similarly bacterivore grazing rates decline with decreasing temperature (Delaney 2003). Indirect effects of temperature mediated by trophic interactions imply that the ecological effects of temperature will depend on trophic structure. Beisner, McCauley & Wrona (1996) investigated the impact of two vs. three trophic levels at 18 and 25 C on Daphnia pulex and phytoplankton population dynamics. At 18 C, the third trophic level (predator, Mesostoma ehrenbergii) had little impact on Daphnia density, but at 25 C the presence of the predator reduced the persistence of Daphnia and increased phytoplankton abundance. These observations suggest an interaction between trophic structure and temperature and thus the presence of indirect temperature effects. 2010 The Authors. Journal compilation 2010 British Ecological Society 694 O. S. Beveridge, S. Humphries & O. L. Petchey Predictions of mechanistic models of resource-consumer population dynamics support the intuitive importance of indirect effects (Vasseur & McCann 2005). A negative prey density-temperature trend is due to the positive effect of temperature on the consumer’s ingestion rate, which counteracts the direct positive influence of temperature on resource growth rate and density. However, there is little quantitative data to verify the presence or importance of indirect temperature effects via alterations in trophic structure. Temperature may have indirect effects by altering the strength of competitive interactions (Jiang & Morin 2004). Here, we focus on the temperature dependence of predator–prey interactions in non-omnivorous food chains. Decomposition is a particularly important function in allochthonous systems where organisms are dependent on energy from the breakdown of material entering the system (Ribblett, Palmer & Coats 2005). An increase in temperature results in an increase in bacterial decomposition because of a net increase in bacterial metabolism (Abrams & Myron 1980). Variation in food chain length might also indirectly affect decomposition rates. For example, the addition of bacterivorous grazers can increase the rate of bacterial decomposition (Ribblett et al. 2005; Jiang & Krumins 2006a; Krumins et al. 2006; Jiang 2007). Less well documented are possible interactive effects of temperature and trophic structure on bacterial decomposition. Petchey et al. (1999) investigated the response of both high and low complexity microbial communities to warming environments (+0Æ1–0Æ2 C per generation). Petchey et al. (1999) found a significant and positive effect of warming, which interacted with community complexity, on decomposition rate. Microbial activity (ammonium production) has also been observed to be dependent on the interaction between warming and species richness (Newsham & Garstecki 2007). Here we investigate indirect temperature effects via a factorial manipulation of environmental temperature and food chain length. We measured the impact of temperature on density within each trophic level of communities containing different number of trophic levels. Analyses were repeated for two phases of population dynamics, as microbial community density can be variable and cyclic (Holyoak & Lawler 1996). Consequences of community response were evaluated by estimating their ability to decompose wheat grains. Microbial microcosms form an ideal system to test the interactive impacts of temperature and food chain length. The systems are readily replicable and straight forward to manipulate. Microcosms are closed systems with most environmental factors easily controlled, thus allowing responses to be confidently attributed to experimentally manipulated variables. Bacteria and protists have short generation times allowing several generations to be investigated within a matter of weeks. Effectively long term effects of temperature and community structure can be realized in a relatively short experimental duration. Decomposition within protists microcosms can be readily assayed (Petchey et al. 1999; Jiang & Krumins 2006a), allowing treatment effects on ecosystem function to be realized. Materials and methods STUDY ORGANISMS AND CULTURING Experiments were carried out using microbial microcosms consisting of a 200-mL sterile jar, 100-mL media and foil lid allowing exchange of gases but reducing the risk of contamination. Microcosm media was Chalkley’s solution (Tompkins et al. 1995) with 0Æ55-g L)1 protist pellet (Carolina T.M. Protozoan pellets, Burlington, NC, USA). To allow the assessment of decomposition provide a long term energy source, each microcosm also contained two wheat grains. All work was conducted using standard aseptic techniques. The total duration of experimentation was 6 weeks. Communities consisted of a resource bacterial fauna (Bacillus cereus Frankland and Frankland, Bacillus subtilis Ehrenberg and Serratia marcescens Bizio), consumer bacterivorous protist (Colpidium striatum Stokes), and a predatory protist (Didinium nasutum Muller), that readily consumes Colpidium. All organisms were obtained from stock cultures stored at experimental temperatures to eliminate shock effects. Bacterial inoculation took place on day 0, Colpidium inoculation on day 1 and Didinium inoculation on day 14. This inoculation protocol ensured that consumers and predators had sufficient resources available for successful establishment. There were two experimental treatments: temperature and food chain length. Food chain length had four levels: zero (sterile to assay abiotic mass loss of wheat grains); one, bacteria alone; two, bacteria + Colpidium; and three, bacteria + Colpidium + Didinium. Temperature had six levels: 5, 9, 13, 17, 21 and 25 C (temperatures were chosen, so as to encompass a natural range, without risking cell freezing). Microcosm temperature was controlled via immersion in a water bath. The experiment was factorial, with each treatment combination replicated five times. SAMPLING Population density of each species was estimated every 7 days, via the removal of aliquots (microcosms were not removed from water baths). Microcosms were agitated prior to sampling to homogenize population density and reduce sampling error. Decomposition was assayed via dry mass loss of wheat grains during the experiment, converted to proportion of initial mass lost. Protist densities were estimated via direct counting under a dissection microscope (Nikon SMZ1000). Only active motile cells were counted. Bacteria were assayed via serial dilution (·50, ·2500, ·125 000) and agar plating (Mikrobiolgie Nutrient Agar). Plates were incubated at room temperature for 48 h, and total bacterial density estimated via colony counts. Bacterial species identification was not possible. DATA AND ANALYSES Mean population densities were calculated for two periods for each population in each microcosm. The first period was over weeks 3 and 4, during which population densities were relatively constant through time. The second period was over weeks 5 and 6, during which population densities were beginning to decline. Data before day 14 were excluded as protist populations were in their growth phase or effectively absent. Species density was modelled (via a linear model) as a function of temperature, food chain length, temperature2, and all two-way interactions (following Crawley 2007). A separate analysis 2010 The Authors. Journal compilation 2010 British Ecological Society, Journal of Animal Ecology, 79, 693–700 Temperature and food chains 695 was conducted for each species in each of the two time periods. Temperature was as a continuous explanatory variable, and food chain length a categorical variable. The squared term was required to model the observed non-linear (unimodal) effects of temperature (Crawley 2007). For Didinium the food chain length parameter was omitted from the model as Didinium was only present in the three trophic level system. A similar model was fitted to proportional wheat seed mass loss. All analyses were conducted using the statistical package R (R Development Core Team 2007). The data were normally distributed and conformed to assumptions of parametric analysis. Results POPULATION DENSITY All populations increased in density from inoculation densities. In the final sampling event (day 42), several microcosms had zero counts of Didinium and Colpidium (despite previously having positive counts). Further inspection of these microcosms indicated that active Colpidium and Didinium were present, but at densities too low to be detected by the assay method. Aliquots of sufficient size to sample protists at very low densities would have been large enough to constitute a disturbance. Although reliable bacterial species identification was not possible via plating, both pink (Serratia) and white (Baccilus) colonies were observed throughout the experiment. No extinctions of any species were observed, despite some data containing estimates of zero density. A summary of population density statistics is given in Table 1. Bacterial densities showed a non-linear response to temperature that depended on the presence or absence of higher trophic levels (Fig. 1a,d and Table. 1). In the absence of other species bacterial densities were highest at intermediate temperatures, for weeks 3–4 (Fig. 1a). In weeks 5–6, this trend was a simple negative function of temperature (Fig. 1d). The presence of Colpidium resulted in an overall lower bacterial density in all temperatures (except 5 C) in data from weeks 3–4. As with bacteria alone, this trend was reversed in data from weeks 5–6, where the addition of Colpidium resulted in a positive relationship between bacterial density and temperature (Fig. 1d). In the three trophic level system, in both periods of data, bacterial density exhibited a negative unimodal trend, with lowest density at intermediate temperatures (Fig. 1a,d). The overall density of bacteria in cold treatments was higher in the three trophic level system, than the one trophic level system. Colpidium density also exhibited a non-linear response to temperature that depended on the presence or absence of Didinium (Fig. 1b,e and Table 1). In the absence of Didinium, Colpidium density was a positive and saturating function of temperature (Fig. 1b,e). In the three trophic level system, Colpidium density exhibited a unimodal trend to increasing temperature. As with bacterial density, Colpidium densities were increased by the presence of Didinium in cold environments, but only during weeks 3–4. The opposite was observed at the highest temperature (25 C), where the addition of Didinium reduced the density of Colpidium. This reduction in density was greater during weeks 5–6 (Fig. 1e) than during weeks 3–4 (Fig. 1b). Didinium population density was only significantly affected by temperature, in weeks 5–6 (Table 1 and Fig. 1c,f). In weeks 5–6, Didinium density exhibited a unimodal trend to increasing temperature, with maximum density occurring around 15 C. DECOMPOSITION Decomposition rate had a non-linear response to temperature that depended on food chain length (Fig. 2 and Table 2). Abiotic mass loss was small ( 7%) with little temperature response. The addition of bacteria resulted in a small increase in mass loss compared with sterile microcosms, indicating that biotic decomposition was occurring. The addition of a second trophic level caused a significant increase in decomposition with a strong temperature interaction effect. In the presence of Colpidium, at 25 C, the rate of decomposition was c. 3Æ5 times higher than in the bacteria alone treatment. The further addition of the third trophic level had little impact on decomposition. Discussion Here we demonstrate how the density of organisms and the functioning of an ecosystem depends on both biotic (community structure) and abiotic (temperature) components and their interaction. The strong interaction of temperature and food chain length on decomposition highlights the importance of indirect effects on ecosystem functioning. Thus, in order to understand or predict the response of an organism to a change in temperature, we must consider both the organism’s position within a community, and the possible effects of temperature on species interactions. The temperature dependency of microbial metabolic rate is well documented (Fenchel & Finlay 1983). Energy or resources entering this microbial system do so via the activity of bacteria, and their decomposition of wheat seeds. The positive trend between bacterial decomposition and temperature make our experiments similar to studies of the direct and indirect effects of resource availability. Kaunzinger & Morin (1998) manipulated microbial food chain length across a nutrient gradient. Their results indicate that an organism’s response to a change in nutrient concentration depends on the number of trophic levels present in the community. Similarly, our results suggest that the response of an organism to temperature depends on the number of trophic levels present in an ecosystem. Our results for bacterial (Fig. 1a) and Colpidium density (Fig. 1b;), from the two trophic level system, resemble current model predictions (Vasseur & McCann 2005, their fig. 4). Vasseur & McCann (2005) modelled the equilibrium densities of an invertebrate consumer and a unicell resource along a temperature gradient (0–40 C) in a two tropic level system. Temperature dependence was model via impacts on 2010 The Authors. Journal compilation 2010 British Ecological Society, Journal of Animal Ecology, 79, 693–700 696 O. S. Beveridge, S. Humphries & O. L. Petchey Table 1. Summary of statistical models analysing species’ population densities (log10 cell per mL). Model structure and interpretation are detailed in the text Effect d.f. F-value Temperature Chain length Temperature2 Temperature : chain length Temperature2 : chain length Error Model coefficients Chain length = 1 Chain length = 2 Chain length = 3 1 2 1 2 2 81 Intercept 5Æ2 6Æ7 6Æ7 16 6Æ7 0Æ19 4 18 0Æ023 <0Æ001 Temperature 0Æ2 )0Æ068 )0Æ16 Temperature2 )0Æ0074 0Æ0011 0Æ0056 Effect d.f. F-value P-value Temperature Chain length Temperature2 Temperature : chain length Temperature2 : chain length Error Model coefficients Chain length = 1 Chain length = 2 Chain length = 3 1 2 1 2 2 81 Intercept 6Æ5 5 6Æ5 0Æ45 7Æ43 5Æ1 20 3Æ7 <0Æ001 0Æ03 Temperature )0Æ046 0Æ043 )0Æ15 Temperature2 0Æ00034 )0Æ00022 0Æ00039 Effect d.f. F-value P-value Temperature Chain length Temperature2 Temperature : chain length Temperature 2 : chain length Error Model coefficients Chain length = 2 Chain length = 3 1 1 1 1 1 54 Intercept 0Æ63 0Æ32 2Æ2 7Æ8 9Æ4 75 12 <0Æ001 <0Æ001 Temperature 0Æ26 0Æ25 Temperature2 )0Æ0067 0Æ00095 Effect d.f. F-value P-value Temperature Chain length Temperature2 Temperature : chain length Temperature2 : chain length Error Model coefficients Chain length = 2 Chain length = 3 1 1 1 1 1 52 Intercept 2 1Æ7 10 67 14 71 7Æ2 <0Æ001 0Æ0096 Temperature 0Æ11 0Æ13 Temperature2 )0Æ0022 )0Æ011 Effect d.f. F-value P-value Temperature Temperature2 Error Model coefficients Chain length = 3 1 1 21 Intercept 0Æ045 1Æ3 0Æ52 0Æ28 0Æ48 Temperature )0Æ042 Temperature2 0Æ002 Effect d.f. F-value P-value 1 2Æ2 0Æ15 Temperature P-value Data set Species Weeks 3–4 Weeks 3–4 Weeks 3–4 Weeks 3–4 Weeks 3–4 Weeks 3–4 Bacteria Bacteria Bacteria Bacteria Bacteria Bacteria Weeks 3–4 Weeks 3–4 Weeks 3–4 Bacteria Bacteria Bacteria Weeks 5–6 Weeks 5–6 Weeks 5–6 Weeks 5–6 Weeks 5–6 Weeks 5–6 Bacteria Bacteria Bacteria Bacteria Bacteria Bacteria Weeks 5–6 Weeks 5–6 Weeks 5–6 Bacteria Bacteria Bacteria Weeks 3–4 Weeks 3–4 Weeks 3–4 Weeks 3–4 Weeks 3–4 Weeks 3–4 Colpidium Colpidium Colpidium Colpidium Colpidium Colpidium Weeks 3–4 Weeks 3–4 Colpidium Colpidium Weeks 5–6 Weeks 5–6 Weeks 5–6 Weeks 5–6 Weeks 5–6 Weeks 5–6 Colpidium Colpidium Colpidium Colpidium Colpidium Colpidium Weeks 5–6 Weeks 5–6 Colpidium Colpidium Weeks 3–4 Weeks 3–4 Weeks 3–4 Didinium Didinium Didinium Didinium Weeks 5–6 Didinium 2010 The Authors. Journal compilation 2010 British Ecological Society, Journal of Animal Ecology, 79, 693–700 Temperature and food chains 697 Table 1. (Continued) Effect d.f. F-value P-value Data set Species Temperature2 Error Model coefficients Chain length = 3 1 21 Intercept )1Æ3 18 <0Æ001 Weeks 5–6 Weeks 5–6 Didinium Didinium Temperature 0Æ22 Temperature2 )0Æ0068 Weeks 5–6 Didinium metabolic rates and consequently growth and ingestion rates. One scenario modelled resulted in a linear negative resource density–temperature relationship, and an increasing saturating relationship between consumer density and temperature. Our findings of bacterial density-temperature trend in the one and two food chain systems, lends empirical support to the prediction of the Vasseur & McCann model. Our experiment allowed us to extend the temperature dependence of resource–community dynamics from a two to a three trophic level system. This enabled us to test for the presence of trophic cascades caused by the addition of a top predator, and whether they were temperature dependent. Theory on trophic level cascades would predict an increase in the density of resource species in the presence of a predator which reduced the density of consumers (Carpenter, Kitchell & Hodgson 1985). Our data suggests the presence of a classical trophic cascade only at temperatures above 17 C. Surprisingly, the density of Colpidium was greater in the three trophic level system, compared with the two trophic level system, below 9 C in weeks 3–4. Additionally the density of bacteria was greatest in the three trophic level system (despite the presence of Colpidium) compared with the two or one trophic level system, below 9 C (in both periods of data). These observations suggest that in cold environments the overall effect of predation is to increase the density of prey and resources. This seems counter intuitive as the effect of predation on prey is to increase mortality. Abrams (2009) discusses three scenarios where an increase in mortality can lead to greater density (the ‘Hydra effect’): (1) A population with fluctuating dynamics may experience a greater average density because predation reduces the amplitude in population cycles. (2) Complex organisms with multiple life stages may experience an increase in average density with an increase in mortality at one life stage resulting in greater success at another life stage. (3) An increase in consumer mortality in a resource-limited population can lead to more prudent resource exploitation (i.e. reduced attack rate of consumers upon resources), which in turn reduces competition for resources and thus increases population density. Our experimental protocol was designed to highlight the effects rather Fig. 1. Mean density of (a,d) bacteria, (b,e) Colpidium and (c,f) Didinium across temperature treatments. Left hand column (a–c) are data collected from weeks 3 to 4. Right hand column (d–e) data collected from weeks 5 to 6. Light grey points and lines indicate data collected from bacteria alone microcosms; dark grey, bacteria–Colpidium microcosms; and black, bacteria–Colpidium–Didinium microcosms. Lines are the output of the statistically significant model predictions discussed in the text. An artificial horizontal offset has been applied to data points to aide interpretation. 2010 The Authors. Journal compilation 2010 British Ecological Society, Journal of Animal Ecology, 79, 693–700 698 O. S. Beveridge, S. Humphries & O. L. Petchey Fig. 2. Decomposition response to temperature and food chain length, measured as proportion dry mass loss (both wheat grains within microcosm). Light grey points and lines indicate data collected from bacteria alone microcosms; dark grey, bacteria–Colpidium microcosms; black, bacteria–Colpidium–Didinium microcosms; white circles with dashed black line, sterile microcosms (abiotic mass loss). Lines are the output of the statistically significant model predictions discussed in the text. An artificial horizontal offset has been applied to data points to aide interpretation. than the mechanisms of responses, thus the exact mechanism underlying the hydra effect cannot be tested. However, our microbial organisms are simple, with short generation times, thus differing mortality at various life stages will not be a mechanism for the hydra effect in our communities. Future studies may wish to employ the following modifications to our methods: (1) greater temporal resolution in population density, and increased experimental duration. This will allow observations of population density dynamics and the impact of predation on population cycling and crashing. (2) Decomposition rate should be assayed through time, in conjunction with bacterial density. Therefore, an accurate measurement of per capita bacterial decomposition rate is enabled. This will highlight whether or not consumer predation does cause a more prudent wheat seed exploitation by bacteria. Decomposition was significantly lower in sterile microcosms than those containing bacteria. This indicates that bio- logical decomposition took place in all microcosms containing bacteria. As expected, decomposition was positively related to temperature (Abrams & Myron 1980). The presence of bacterivores increased decomposition 3Æ5-fold with a strong, positive and non-linear response to temperature. A 1Æ5- to 3-fold increase in bacterial decomposition (wheat grain dry mass loss) has previously been observed with the addition of bacterivores to a bacterial fauna (Jiang & Krumins 2006a; Krumins et al. 2006; Jiang 2007). Similarly Ribblett et al. (2005) observed a 3Æ5- to 4-fold increase in the bacterial decomposition of leaf disks because of the addition of bacterivores. Our experimental design allowed us to uniquely test the impact on decomposition of food chain length, rather than total protist species diversity or absence and presence of other bacterivore species. Only the presence of a bacterial grazer, not total food chain length was important in determining decomposition rate in this system. Previous work has identified a positive correlation between Eukaryotic (including third trophic levels) species richness and decomposition (Krumins et al. 2006). Further work is required to investigate the importance of species richness at the second trophic level and independently at the third trophic level. It is important to note that these previous studies (Ribblett et al. 2005; Jiang & Krumins 2006a; Krumins et al. 2006; Jiang 2007) were conducted at a single environmental temperature. Evidence provided here suggests that the impact a bacterivore has on decomposition is greatest at higher temperatures. Various hypotheses have been proposed to explain how the presence of bacterivores facilitates bacterial decomposition. (1) Bacterivores may alter the composition of the bacterial fauna, and result in a fauna more favourable to decomposition (Krumins et al. 2006). (2) The addition of bacterivores may increase the recycling of minerals and nutrients, which would otherwise limit bacteria activity (Ratsak, Maarsen & Kooijman 1996). (3) The consumption of bacteria will likely maintain a bacterial population density below carrying capacity, within the more active exponential growth phase (Jurgens & Sala 2000). An alteration in the bacterial fauna is unlikely to explain the positive effects of bacterivores on decomposition in our experiments. Our bacterial fauna was simple from the outset and although reliable species identification was not possible, Table 2. Summary of the statistical model analysing wheat seed mass loss. Model structure and interpretation are detailed in the text Effect d.f. F-value P-value Temperature Chain length Temperature2 Temperature : chain length Temperature2 : chain length Error Model coefficients Sterile Chain length = 1 Chain length = 2 Chain length = 3 1 3 1 3 3 12 Intercept 0Æ074 0Æ11 0Æ20 0Æ33 170 40 55 )35 9Æ9 0Æ00040 0Æ0312 Temperature )0Æ0034 )0Æ012 )0Æ050 )0Æ13 Temperature2 0Æ00014 0Æ00057 0Æ0021 0Æ0023 2010 The Authors. Journal compilation 2010 British Ecological Society, Journal of Animal Ecology, 79, 693–700 Temperature and food chains 699 both white (Bacillus) and pink (Serratia) colonies were observed throughout the experiment. We were unable to ascertain the limiting factors for bacterial activity in this experiment, and so cannot draw conclusions on the impact of bacterivores on resources limiting bacterial activity. Our data best supports the hypothesis that bacterivores lowered the density of the bacterial fauna and maintained an active bacterial population in exponential growth phase. Furthermore, the impact of Colpidium on bacterial density increased with temperature (Fig. 1), and may explain the strong food chain length-temperature interaction observed for decomposition. A fourth mechanism may also exist for the positive impact of bacterivores on bacterial decomposition: Rozen et al. (2009) observed that bacterial consumption of lysed bacterial debris facilitates bacterial growth. As predation is unlikely to be 100% efficient, it seems plausible that Colpidium could increase the concentration of bacterial debris, thus increasing bacterial growth. However, we are unaware of any reported link between Colpidium grazing and bacterial debris concentration in the literature. The strength of predator ⁄ consumer impacts on prey ⁄ resources were positively temperature dependent. The addition of Colpidium altered the bacterial temperature–density relationship from one of a unimodal nature, to a negative one. This suggests that the direct positive influence of temperature on bacterial density is counteracted by the indirect negative impact of temperature on bacteria via its interaction with Colpidium. A similar relationship was observed with Colpidium density and Didinium. For Colpidium this overall increase in predator impact was most pronounced at later stages of population dynamics, suggesting that population dynamic phase may alter the importance of predator impact on prey numbers. A reduction in Colpidium density would be predicted to increase the density of bacteria because of the reduction in Colpidium impact on bacteria. However, the reduction in Colpidium density did not correspond with an increase in bacterial density (Fig 1a & d). This implies that grazing by Colpidium is not directly dependent on Colpidium population density. This suggests the presence of interference between individual Colpidium. Interference has been proposed for Colpidium and a competitor Paramecium caudatum (Jiang & Krumins 2006b). Further work is required to investigate whether there is intraspecies interference for Colpidium consumption of bacteria. Our results reveal an important role for temperature and community structure in controlling population size and decomposition rate. We also clearly show a strong interaction between temperature and community structure, affecting both populations and their ecosystem function. The importance of interactive effects between food chain length (or vertical diversity as discussed in Duffy et al. 2007) and temperature are made more compelling when one considers that in addition to climate change biodiversity is also rapidly declining (Ehrlich & Pringle 2008). Independently modelling either the direct impacts of temperature or extinctions on ecosystems may lead to incorrect predictions. Our work demon- strates that the biotic and abiotic components of an ecosystem interact strongly, with impacts observed at all levels. Acknowledgements OSB was funded by a NERC studentship, OLP is a Royal Society University Research Fellow. SH was supported by a NERC Advanced Fellowship (NE ⁄ B500690 ⁄ 3). We also thank two anonymous reviewers for their constructive and helpful comments. References Abrams, P.A. (1987) On classifying interactions between populations. Oecologia, 73, 272–281. Abrams, P.A. (2009) When does greater mortality increase population size? The long history and diverse mechanisms underlying the hydra effect. Ecology Letters, 12, 462–474. Abrams, B.I. & Myron, J.M. (1980) Role of nematode-bacterial interactions in heterotrophic systems with emphasis on sewage sludge decomposition. Oikos, 35, 404–410. Apple, J.K., Del Giorgi, P.A. & Kemp, W.M. (2006) Temperature regulation of bacterial production, respiration, and growth efficiency in a temperate salt-marsh estuary. Aquatic Microbial Ecology, 43, 243–254. Balistrieri, L.S., Tempel, R.N., Stillings, L.L. & Shevenell, L.A. (2006) Modeling spatial and temporal variations in temperature and salinity during stratification and overturn in Dexter Pit Lake, Tuscarora, Nevada, USA. Applied Geochemistry, 21, 1184–1203. Beisner, B.E., McCauley, E. & Wrona, F.J. (1996) Temperature-mediated dynamics of planktonic food chains: the effect of an invertebrate carnivore. Freshwater Biology, 35, 219–231. Bolton, T.F. & Havenhand, J.N. (1998) Physiological versus viscosity-induced effects of an acute reduction in water temperature on microsphere ingestion by trochophore larvae of the serpulid polychaete Galeolaria caespitose. Journal of Plankton Research, 20, 2153–2164. Bolton, T.F. & Havenhand, J.N. (2005) Physiological acclimation to decreased water temperature an the relative importance of water viscosity in determining the feeding performance of larvae of a Serpulid polychaete. Journal of Plankton Research, 27, 875–879. Botte, V. & Kay, A. (2000) A numerical study of plankton population dynamics in a deep lake during the passage of the spring thermal bar. Journal of Marine Systems, 26, 367–386. Brown, J.H., Gillooly, J.F., Allen, A.P., Savage, V.M. & West, G.B. (2004) Toward a metabolic theory of ecology. Ecology, 85, 1771–1789. Carpenter, S.R., Kitchell, J.F. & Hodgson, J.R. (1985) Cascading trophic interactions and lake Productivity. BioScience, 35, 634–639. Clarke, A. (1991) What is cold adaptation and how should we measure it? American Zoologist, 31, 81–92. Coats, R., Perez-Losada, J., Schladow, G., Richards, R. & Goldman, C. (2006) The warming of Lake Tahoe. Climatic Change, 76, 121–148. Crawley, M.J. (2007) The R Book. John Wiley & Sons Ltd, West Sussex. Delaney, M.P. (2003) Effects of temperature and turbulence on predator-prey interactions between a bacterotrophic flagellate and a marine bacterium. Microbial Ecology, 45, 218–225. Duffy, E.J., Cardinale, B.J., France, K.E., McIntyre, P.B., Thébault, E. & Loreau, M. (2007) The functional role of biodiversity in ecosystems: incorporating trophic complexity. Ecology Letters, 17, 522–538. Ehrlich, P.R. & Pringle, R.M. (2008) Where does biodiversity go from here? A grim business-as-usual forecast and a hopeful portfolio of partial solutions. Proceedings of the National Academy of Sciences of the United States of America, 105, 11579–11586. Fenchel, T. & Finlay, B.J. (1983) Respiration rates in heterotrophic, free-living protozoa. Microbial Ecology, 9, 99–122. Gillooly, J.F., Brown, J.H., West, G.B., Savage, V.M. & Charnov, E.L. (2001) Effects of size and temperature on metabolic rate. Science, 293, 2248–2251. Holyoak, M. & Lawler, S.P. (1996) Persistence of an Extinction-Prone PredatorPreyInteractionthroughMetapopulationDynamics.Ecology,77,1867–1879. Ives, A.R. (1995) Predicting the response of populations to environmentalchange. Ecology, 76, 926–941. Jiang, L. (2007) Negative selection effects suppress relationships between bacterial diversity and ecosystem functioning. Ecology, 88, 1075–1085. Jiang, L. & Krumins, J.A. (2006a) Consumer vs environmental productivity control of bacterial diversity and bacteria-mediated organic matter decomposition. Oikos, 114, 441–450. 2010 The Authors. Journal compilation 2010 British Ecological Society, Journal of Animal Ecology, 79, 693–700 700 O. S. Beveridge, S. Humphries & O. L. Petchey Jiang, L. & Krumins, J.A. (2006b) Emergent multiple predator effects in an experimental microbial community. Ecological Research, 21, 723–731. 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. Jurgens, K. & Sala, M.M. (2000) Predation-mediated shifts in size distribution of microbial biomass and activity during detritus decomposition. Oikos, 91, 29–40. Kaunzinger, C.M. & Morin, P.J. (1998) Productivity controls food-chain properties in microbial communities. Nature, 395, 495–497. Krumins, J.A., Long, Z.T., Steiner, C.F. & Morin, P.J. (2006) Indirect effects of food web diversity and productivity on bacterial community function and composition. Functional Ecology, 20, 514–521. Loiterton, B., Sundbom, M. & Vrede, T. (2004) Separating physical and physiological effects of temperature on zooplankton feeding rate. Aquatic Science, 66, 13–129. Moralesbaquero, R. & Cruzpizarro, L. (1995) Effects of fluctuating temperatures on the population dynamics of Hexarthra bulgarica (Wiszniewski) from high mountain lakes in Sierra Nevada (Spain). Hydrobiologia, 313, 359–363. Moran, O. & Melani, R. (2001) Temperature-dependent conduction properties in Arctic fish peripheral nerves. Polar Biology, 24, 9–15. Newsham, K.K. & Garstecki, T. (2007) Interactive effects of warming and species loss on model Antarctic microbial food webs. Functional Ecology, 21, 577–584. Petchey, O.L., McPhearson, P.T., Casey, T.M. & Morin, P.J. (1999) Environmental warming alters food-web structure and ecosystem function. Nature, 402, 69–72. Podolsky, R.D. & Emlet, R.B. (1993) Separating the effects of temperature and viscosity on swimming and water movement by sand dollar larvae (Dendraster excentricus). The Journal of Experimental Biology, 176, 207–221. R Development Core Team (2007). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0. Available at http://www.R-project.org, accessed 12 October 2009. Ratsak, C.H., Maarsen, K.A. & Kooijman, S. (1996) Effects of protozoa on carbon mineralization in activated sludge. Water Research, 30, 1–12. Ribblett, S.G., Palmer, M.A. & Coats, D.W. (2005) The importance of bacterivorous protists in the decomposition of stream leaf litter. Freshwater Biology, 50, 516–526. Rose, J.M. & Caron, D.A. (2007) Does low temperature constrain the growth rates of heterotrophic protists? Evidence and implications for algal blooms in cold waters. Limnology and Oceanography, 52, 886–895. Rozen, D.E., Philippe, N., Arjan de Visser, J., Lenski, R.E. & Schneider, D. (2009) Death and cannibalism in a seasonal environment facilitate bacterial coexistence. Ecology Letters, 12, 34–44. Savage, V.M., Gilloolly, J.F., West, G.B. & Charnov, E.L. (2004) Effects of body size and temperature on population growth. The American Naturalist, 163, 439–441. Sleigh, M.A. (1956) Metachronism and frequency of beat in the peristomial cilia of Stentor. Journal of Experimental Biology, 33, 15–28. Tompkins, J., Deville, M.M., Day, J.G. & Turner, M.F. (1995) Culture Collection of Algae and Protozoa: Catalogue of Strains 1995. Titus Wilson & Son lmtd, Kendal. Vasseur, D.A. & McCann, K.S. (2005) A mechanistic approach for modeling temperature-dependent consumer-resource dynamics. The American Naturalist, 166, 184–198. Wilson, R.S. (2005) Temperature influences the coercive mating and swimming performance of male eastern mosquitofish. Animal Behaviour, 70, 1387– 1394. Wilson, R.S., James, R.S. & Johnston, I.A. (2000) Thermal acclimation of locomotor performance in tadpoles and adults of the aquatic frog Xenopus laevis. Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology, 170, 117–124. Winet, H. (1976) Ciliary propulsion of objects in tubes - wall drag on swimming Tetrahymena (Ciliata) in presence of mucin and other long-chain polymers. Journal of Experimental Biology, 64, 283–302. Yee, E.H. & Murray, S.N. (2004) Effects of temperature on activity, food consumption rates, and gut passage times of seaweed-eating Tegula species (Trochidae) from California. Marine Biology, 145, 895–903. Received 10 August 2009; accepted 15 December 2009 Handling Editor: Tom Webb 2010 The Authors. Journal compilation 2010 British Ecological Society, Journal of Animal Ecology, 79, 693–700