Research Article

Coping with Temperature at the Warm Edge – Patterns of Thermal Adaptation in the Microbial Eukaryote Paramecium caudatum

  • Sascha Krenek mail,

    Affiliations: Institute of Hydrobiology, Technische Universität Dresden, Dresden, Germany, Molecular Evolution and Animal Systematics, Institute of Biology, University of Leipzig, Leipzig, Germany

  • Thomas Petzoldt,

    Affiliation: Institute of Hydrobiology, Technische Universität Dresden, Dresden, Germany

  • Thomas U. Berendonk

    Affiliations: Institute of Hydrobiology, Technische Universität Dresden, Dresden, Germany, Molecular Evolution and Animal Systematics, Institute of Biology, University of Leipzig, Leipzig, Germany

  • Published: March 09, 2012
  • DOI: 10.1371/journal.pone.0030598



Ectothermic organisms are thought to be severely affected by global warming since their physiological performance is directly dependent on temperature. Latitudinal and temporal variations in mean temperatures force ectotherms to adapt to these complex environmental conditions. Studies investigating current patterns of thermal adaptation among populations of different latitudes allow a prediction of the potential impact of prospective increases in environmental temperatures on their fitness.

Methodology/Principal Findings

In this study, temperature reaction norms were ascertained among 18 genetically defined, natural clones of the microbial eukaryote Paramecium caudatum. These different clones have been isolated from 12 freshwater habitats along a latitudinal transect in Europe and from 3 tropical habitats (Indonesia). The sensitivity to increasing temperatures was estimated through the analysis of clone specific thermal tolerances and by relating those to current and predicted temperature data of their natural habitats.

All investigated European clones seem to be thermal generalists with a broad thermal tolerance and similar optimum temperatures. The weak or missing co-variation of thermal tolerance with latitude does not imply local adaptation to thermal gradients; it rather suggests adaptive phenotypic plasticity among the whole European subpopulation. The tested Indonesian clones appear to be locally adapted to the less variable, tropical temperature regime and show higher tolerance limits, but lower tolerance breadths.


Due to the lack of local temperature adaptation within the European subpopulation, P. caudatum genotypes at the most southern edge of their geographic range seem to suffer from the predicted increase in magnitude and frequency of summer heat waves caused by climate change.


Temperature is one of the most important environmental factors determining a variety of ecosystem elements, e.g. species ecophysiology, abundance and distribution, as well as species diversity and population dynamics [1][4]. Due to the current climate change, scientists started to re-evaluate the impact of elevated temperatures on the ecology of species. Here, ectothermic organisms are of special interest as their physiological performance is highly dependent on environmental temperature.

To make predictions of organisms' and population responses to global warming, studies on genetic and phenotypic diversity over a species' geographic range are important. Such investigations can unveil patterns of evolutionary temperature adaptation to the current thermal heterogeneity on Earth by determining which ectotherms have a high acclimatisation capacity and which only occur at specific temperatures. Adaptive phenotypic plasticity, for instance, may cause a higher tolerance to changing thermal conditions [5], [6], while local temperature adaptation might be detrimental.

Several studies could show a co-variation of latitude and thermal tolerance (e.g. [7], [8]) suggesting that organisms are adapted to the mean temperatures of their environment, but others failed (e.g. [9], [10]). Climate change is supposed to affect both climate averages and variability [11] and it has been shown that the thermal tolerance of many organisms is proportional to the magnitude of variation they are exposed to [12]. Organisms are also expected to be adapted to the thermal heterogeneity of their particular environment. This thermal heterogeneity increases with latitude. Therefore, organisms from variable climates, such as the temperate zone, should evolve a broad thermal tolerance resulting in thermal generalists. In contrast, tropical ectotherms, experiencing less variation in temperature, should be selected for narrow thermal niches resulting in thermal specialists [13], [14]. Consequently, analysing thermal niches of different populations along a latitudinal transect is necessary to understand the process of adaptation to novel thermal environments.

It has been shown that populations and individuals at the edge of the species range may suffer the most from increasing temperatures, because they often live close to the limit of their species' physiological thermal tolerance [15]. Therefore, it is not only important to investigate the intraspecific variation in species' thermal tolerance, but also to consider populations from the margins of their current distribution range. Especially if one would expect a thermal generalist pattern for ubiquitous species, populations at the ‘warm edge’ (such as the tropics or subtropics) might be most at risk due to global warming (cf. [16], [17]).

Because of the anticipated increasing risk of more intense, more frequent and longer-lasting heat waves during summer [18], species heat resistances are of particular significance [15], [19], [20], [21]. Here, thermal safety margins as well as the maximum warming tolerance are suitable characters to qualitatively elucidate the impact of climate change effects across latitude on different populations. These indicators are based on an organism's thermal tolerance and its relation to the local temperature regime [17]. Studies investigating species' current thermal adaptation patterns with respect to present-day and future environmental temperatures therefore allow predictions of species' and population responses to elevated temperatures.

Beside these patterns of evolutionary temperature adaptation obviously related to climate change, many other patterns are important in thermal adaptation with respect to species evolution and ecology. For example, the warmer is better hypothesis [22], [23], [24], which predicts a positive correlation between an organism's optimal temperature and its maximum performance; or the Jack-of-all-temperatures is a master of none hypothesis [25], which assumes an evolutionary trade-off between the performance breadth and the maximal performance of an organism, are controversially discussed. These patterns are relevant in a climate change context, too, but only few investigators have experimentally tested these basic ideas of evolutionary temperature adaptation 2629.

For the investigation of such elementary hypotheses, the determination of thermal performance curves (Figure 1) provides a suitable framework to evaluate an organism's thermal tolerance [30]. Thermal performance curves (TPCs) allow estimations on how basic physiological functions are influenced by environmental temperature [17]. Furthermore, TPCs permit the calculation of ecophysiological key characteristics like the lower and upper critical thermal limits (CTmin and CTmax) as well as the optimum temperature (Topt) and the maximum performance ([31]; cf. Figure 1]). Such key parameters are useful indicators for the thermal tolerance or thermal niche as well as for a potential environmental adaptation of different genotypes. As before mentioned, these ecophysiological characteristics can show a co-variation with latitude in metazoan species (e.g. [12], [32]), although other studies unveiled that the upper thermal limits of ectotherms vary little with latitude (e.g. [33], [34]). However, this has never been critically evaluated for microbial eukaryotes, which are not only important for aquatic ecosystems [35], [36], but also constitute well suited organisms for experimental evolution [37], [38].


Figure 1. General shape of a thermal performance curve.

Relationship between environmental temperature and a physiological rate of an ectotherm expressed as a thermal performance curve (grey line). The optimum temperature (Topt) specifies the temperature at maximum performance. The ecophysiological key characteristics critical thermal minimum (CTmin) and maximum (CTmax) delimit an organism's thermal tolerance.


While some recent studies have investigated the response of protozoan species to increasing temperatures (e.g. [39][42]), little is known about thermal adaptation patterns of globally distributed eukaryotic microbes and how temperature might affect the genetic diversity of natural populations. Furthermore, investigations on the intraspecific variation in species' thermal tolerance by considering populations from the margins of their current distribution range are rare as well.

In the present study, the microbial eukaryote Paramecium caudatum was used to investigate the intraspecific variation in temperature reaction norms of different genotypes. These were isolated from natural habitats along a latitudinal transect in Europe, while three genotypes from tropical habitats (Indonesia) served as a genetic and phenotypic outgroup. This globally distributed ciliate species inhabits the mud-water interface of littoral freshwater environments, which are considerably affected by atmospheric temperature changes [43], [44]. Consequently, P. caudatum has to cope with large temporal and spatial variations in temperature. It therefore constitutes a suitable ectotherm to test hypotheses in thermal adaptation as well as the consequences of climate change on such ubiquitous protists. Here, we performed temperature dependent growth experiments to (i) test for a hypothesized local temperature adaptation of different P. caudatum genotypes; (ii) investigate thermal constraints resulting from evolutionary temperature adaptation; and (iii) understand the sensitivity of P. caudatum to predicted future temperatures.

Materials and Methods

Sampling sites and Organisms

Paramecium caudatum cells were isolated from freshwater samples of 12 different natural habitats along a north-south transect in Europe as well as from three tropical habitats in Indonesia, Sulawesi (see Figure 2 and Table 1 for specifications). No specific permits were required for the described field studies. In Europe and Indonesia, work with Paramecium does not require specific permission and samples were not taken from water bodies where private property was indicated or from nature reserves where sampling is prohibited. The field studies did not involve endangered or protected species.


Figure 2. Geographic origin of investigated Paramecium caudatum populations.

The small map shows the sampling points of all investigated P. caudatum clones within this study. The large map illustrates the sample sites within Europe in detail. Codes for clonal P. caudatum cultures refer to Table 1.


Table 1. Origin of Paramecium caudatum clones, genetic background and GenBank® accession numbers.


The food bacteria Enterobacter aerogenes were obtained from the American Type Culture Collection (ATCC 35028) and the kanamycin-resistant strain Pseudomonas fluorescens SBW25 EeZY-6KX [45] was acquired from the University of Oxford.

Paramecium Stock MAINTENANCE

The investigated P. caudatum stock cultures were maintained in a 0.25% Cerophyl infusion, prepared according to the methods of Sonneborn [46] with minor modifications [31]. Isolated cells were separated in 1 ml of filtrated habitat water in 24-well tissue culture plates (TPP® AG) to establish clonal cultures. Afterwards, cells were washed and maintained at 22°C in a Cerophyl infusion inoculated with Enterobacter aerogenes to establish mass cultures. Later, cultures were kept at 10°C, lowering the growth and ageing of P. caudatum. Previous to the start of the experiments, monoxenic P. caudatum cultures were established at 22°C in a Cerophyl infusion with Pseudomonas fluorescens serving as the only food bacteria (for details see [31]).

Temperature Dependent Growth Experiments

Cells from exponentially growing, monoxenic P. caudatum cultures were transferred to tissue culture flat tubes and acclimatised to experimental temperatures between 7°C and 35.5°C in steps of ±1.5 K d−1. Cultures were kept in exponential growth phase (500–1000 cells ml−1) during the acclimation period by doubling the culture volume with Pseudomonas fluorescens inoculated Cerophyl infusion (CMP; pH 7.0) as appropriate (1–5 ml per day). Due to the different acclimation phases from 22°C up to 35.5°C or down to 7°C, respectively, temperature-dependent experiments were conducted time-delayed. All experiments were performed in microprocessor-controlled, cooled incubators obtained from BINDER GmbH (Type KB 53).

Before the experimental start, acclimatised P. caudatum pre-cultures were adjusted to ~250 cells ml−1 with CMP. Two millilitres of these starting cultures were added to each microcosm containing 18 ml CMP and resulting in an initial abundance of ~25 cells ml−1. Growth experiments were run in triplicate in 60-ml tissue culture flasks with filter lids (TPP® AG) over two to eight days depending on the experimental growth temperature. The bacterial start density was regulated to a saturating prey level of about 2•108 cells ml−1. If the P. caudatum pre-culture densities were below 250 cells ml−1 because of growth-limiting temperatures (e.g. 7°C or ≥34°C), initial cell abundance was adjusted to the highest possible cell number (≥10 cells ml−1).

Paramecium cell abundance was estimated by sampling 1 ml every nine to 41 hours depending on the experimental growth temperature. This resulted in five to eight samples per replicate. For precise counting, cells were fixed by the addition of Bouin's solution [47] to a final concentration of 1%. Cell numbers were enumerated microscopically by threefold counting 100 µl to 300 µl subsamples using a dark field stereoscopic microscope (Olympus GmbH). The population growth rate (μ, d−1) for each replicate and at each experimental temperature was calculated over the period of exponential increase using the slope of the linear regression of loge-transformed cell densities versus time (t).

Intraspecific Differentiation

To identify and distinguish the individual, clonal cultures of natural P. caudatum genotypes from different geographic regions, the mitochondrial cytochrome c oxidase subunit I (COI) gene was sequenced following the protocol of Barth et al. [48]. Five cells from each stock culture were washed four times in sterile Eau de Volvic®and then incubated overnight with 100 µl of 10% Chelex®solution and 10 µl Proteinase K (10 mg ml−1) at 56°C. Afterwards, the mixture was boiled for 20 min and frozen at −20°C; the supernatant was used for subsequent PCR reactions. Each PCR reaction mix contained 10 µl of Chelex® extracted genomic DNA, 10 pmol of each primer, 1 U Taq-polymerase (SIGMA, Taufkirchen, Germany), 1×PCR buffer with 2 mM MgCl2 and 200 µM dNTPs in a total volume of 50 µl. PCR conditions were as follows: 5 min initial denaturation (95°C); 35 cycles of 1 min at 95°C, 1 min at 50°C and 45 s at 72°C; and a final extension step of 5 min (72°C). Using the primers CoxL11058 and CoxH10176 (see [48]), an 880-bp fragment of the mitochondrial COI gene was amplified. After purification with the Rapid PCR Purification System (Marligen Bioscience, Ijamsville, USA), PCR products were directly sequenced. Sequencing reactions were performed in both directions and analysed on an ABI 3100 Genetic Analyzer (Applied Biosystems).

Thermal Performance Curves

The calculation of thermal performance curves (TPCs) was performed to describe the temperature dependent growth rate data of the individual P. caudatum clones and to determine clone specific key ecophysiological characteristics. TPCs have a common general shape with a gradual increase from a lower critical temperature (CTmin) to a thermal optimum (Topt) where the investigated biological function reaches its maximum. With a further increase in temperature above Topt the TPCs show a rapid decline towards a critical temperature maximum (CTmax; Figure 1). It was shown that the nonlinear Lactin-2 optimum function [49] can adequately describe the temperature – growth rate relationship of Paramecium caudatum resulting in typically skewed TPCs with a right-shift towards warmer temperatures [31].

The TPC estimation was done by fitting nonlinear mixed-effects models [50] simultaneously to the whole data set. The mixed-effects models were compared with AIC based model selection at three hierarchical levels; the whole data set with common fixed effects for all 18 clones (null models nm0a and nm0b, cf. Table 2), with separate fixed effects for the two regions, Europe and Indonesia (model nm2, cf. Table 2) and with separate fixed effects for the four regions northern, central, southern Europe and Indonesia (model nm4, cf. Table 2). In all cases, all four original parameters of the Lactin-2 function (ρ, Tmax, Δ and λ) were used as fixed effects. In model nm0a, all four parameters were also used as random effects while for models nm0b, nm2, and nm4 only Tmax, Δ and λ were used because of the high correlation between the parameters ρ and λ resulting in a low model convergence. The decision which of the two parameters had to be omitted for nm2 and nm4 was made by comparing the respective AIC values (not shown). Model nm0b is shown for comparison only (cf. Table 2).


Table 2. Comparison of nonlinear mixed-effects models with different levels of spatial aggregation.


Then, the ecophysiological characteristics CTmin and CTmax were derived numerically as the intersection points of the resulting thermal performance curve with the temperature axis (μ = 0). The maximum growth rate (μmax, cal) was calculated analytically as the growth rate (μ) at Topt using the Lactin-2 function (Eq.1), while Topt was calculated using its first derivative (Eq.2) as follows:(1)
where, the parameter ρ is a constant influencing μmax and the slope of the low-temperature branch, Tmax is the maximum temperature, and Δ defines the temperature range of the thermal inhibition above Topt. Parameter λ is an intercept parameter that forces the curve to intersect the abscissa at low temperatures and allows the estimation of CTmin.

Finally, standard errors for both, the original Lactin-2 function parameters (see Table S1) and the derived ecophysiological key parameters (Table 3) were estimated by nonparametric residual bootstrapping [51] with 1000 bootstrap replicates. For all further analyses based on these key ecophysiological characteristics, estimated data derived from the nm0a mixed-effects model fitting and bootstrapping procedure were used if not otherwise stated.


Table 3. Ecophysiological characteristics of individual Paramecium caudatum clones and the two regions, Europe and Indonesia.


Climate Data

Investigating an organism's local temperature adaptation or its extinction risk due to climate change requires specific knowledge about the thermal conditions within its natural habitats. Here, we used site-specific temperature data to compare the clonal specific ecophysiological characteristics Topt and CTmax with the current climate conditions of the specific habitats. Climate data were obtained from nearby meteorological stations (Table S2) or derived from the WorldClim database [52] using the program DIVA-GIS. In case of the meteorological station data, daily mean and maximum air temperature data of the years 2000–2011 (if available) were used to calculate the mean surface air temperature (Thab, mean) as well as the mean maximal surface air temperature (Thab, max), both for the warmest three months of the specific habitat. The WorldClim database is a set of interpolated global climate layers considering monthly precipitation as well as mean, minimum and maximum temperatures of the years ~1950–2000. The database also provides 19 derived bioclimatic variables. The 2.5 arc-minutes resolution database was used to obtain the habitat mean temperature of the warmest quarter (bioclimatic variable 10 ≙ Thab, mean) and to calculate the habitat mean maximum temperature of the warmest three months (Thab, max). Furthermore, we used climate change data (2.5 arc-minutes) provided by DIVA-GIS to calculate future conditions for the respective habitats of the investigated P. caudatum clones. These data were derived from high-resolution simulations of global warming [53] using the CCM3 model and assuming a CO2 doubling until 2100 (see Table S2).

The use of such temperature data has proven controversial and it is well known that local and microhabitat temperature extremes and fluctuations can differ significantly from the regional average [54], [55]. However, it could be demonstrated that the summer lake surface water temperature of shallow lakes clearly correlates with the local air temperature [56]. Therefore, the use of local air temperature data seems to be a valid approach to estimate an aquatic ectotherm's performance temperature such as for P. caudatum that inhabits the littoral zone of freshwater environments.

Statistical Analyses

Correlation analyses between each key ecophysiological characteristic (CTmin, Topt, CTmax) and the latitude of the respective natural habitats were performed to compare these thermal adaptation indicators with the geographical origin of the different clonal P. caudatum cultures. We used a subset of data containing all investigated European clones and a second dataset including also the Indonesian paramecia. Additionally, the latitudes of the European habitats were corrected for the altitude by assuming that 100 m elevation translates into a ~100 km latitudinal increment within the temperate zone [57]. This correction was done to circumvent altitude effects on the thermal tolerance – latitude dependency. A correction for the Indonesian clones was disclaimed due to nonlinear and extraordinary steep elevational temperature gradients in the tropics [58]. The correlation analyses were performed with absolute values for the altitude-corrected latitude, where Spearman's correlation coefficient and the respective p-values were estimated by Pearson correlation on ranks.

Thermal safety margins (TSM = ToptThab, mean) as well as the warming tolerances of maximum temperatures (MWT = CTmaxThab, max) were calculated to correlate the clonal specific thermal adaptation indicators and the local climate conditions with the altitude-corrected latitudinal gradient (according to [17]).

A Spearman rank correlation analysis between all calculated maximum growth rates (μmax, cal) and Topt values as well as all μmax, cal data and thermal tolerance breadths (TTB = CTmaxCTmin) was used to test for the so-called warmer is better hypothesis [23] and the Jack-of-all-temperatures is a master of none hypothesis [25], respectively.

Additionally, the coefficients of variation (CV) of the mean growth rate of all investigated clonal P. caudatum cultures were analysed to estimate the intraspecific variation among all tested European clones as well as to assess intra-populational divergence. The calculation for each experimental temperature was as follows:(3)
where is the standard deviation of all growth rates (μi) at the investigated temperatures (Ti) and its arithmetic mean ().

To assess multivariate correlations between the genetic and the ecophysiological as well as geographic distances of the investigated P. caudatum clones, Mantel tests [59] were performed for the whole dataset as well as the European subset. Please note that this does not test for a causal link between the genetic variation of the COI gene and differing thermal tolerances as this gene was chosen to estimate the genetic differentiation within P. caudatum. It is a so called barcoding gene and facilitates comparisons of genetic variation within and among species [48], [60].

All statistical analyses were performed using the R system for statistical computing [61] with the add-on package nlme [62] for mixed-effects modelling and package vegan [63], [64] for Mantel tests.


Intraspecific Variation in Thermal Performance

Using the Lactin-2 model [49] to describe thermal performance curves (TPCs) of individual Paramecium caudatum clones resulted in typical left-skewed TPCs (Figure 3). These clonal specific TPCs allowed the calculation of ecophysiological key characteristics, which were qualitatively distinguishable between the different P. caudatum clones. For example, the Swedish clone SWV-2A possessed the lowest heat tolerance (CTmax = 32.13±0.05°C), while the clone from Greece (GRL-1) showed the highest (CTmax = 35.12±0.06°C) among all investigated European clones. Conversely, the Spanish clone (ESH-2) possessed a comparatively low heat tolerance (CTmax = 32.96±0.06°C) compared to another Swedish clone (CTmax, SWL-1A = 33.84±0.07°C). Not only the CTmax values showed high differences among the tested European clones, the calculated CTmin values were also considerably different (ΔCTmin, EU = 8.47±3.13°C, cf. Table 3). On the other hand, all European clones showed their highest growth rates (μmax) at the same experimental temperature of 28°C, while the calculated optimum temperatures (Topt) of the fitted TPCs ranged from 27.94±0.23°C (ESH-2) to 29.49±0.15°C (GRL-1).


Figure 3. Thermal performance curves.

Fitted thermal performance curves using the Lactin-2 model to describe the growth rates – temperature relationship of all investigated clonal P. caudatum cultures. Clones were arranged according to their geographic origin: A) Northern Europe, B) Central Europe, C) Southern Europe and D) Indonesia. Symbols represent the mean ± standard error of the mean (n = 3) of the determined growth rates at the respective temperatures. Lines define the fitted thermal performance curves. Clonal descriptions refer to Table 1.


The comparison between relative differences of intrinsic growth rate data of all European clones showed considerably larger differences with increasing distance from Topt. More precisely, large relative variations (CV) were obtained at low temperatures [CV(μ7°C) = 21.14%] and especially at temperatures above Topt [CV(μ32.5°C) = 91.41%], compared to the low relative variation at Topt [CV(μ28°C) = 7.50%]. While clones from the same habitat (GMA-1A, GMA-1B, GMA-2 and GMA-3) belonging to the same as well as to different COI haplotypes (see Table 1) showed seemingly similar reaction norms, we could detect some variation [CV(μ7°C) = 14.41%, CV(μ28°C) = 7.59%, CV(μ32.5°C) = 17.91%]. Performing one-way ANOVAs of clone specific growth rates at each experimental temperature suggests significant differences for the lowest and the highest temperatures tested (7°C, 32.5°C, 34°C; df = 3, p<0.001, with Bonferroni correction). This result indicates the existence of an intra-populational variation even though we found no significant differences for the intermediate temperatures.

While we could obviously detect only slight differences among the performance curves of different European P. caudatum clones (cf. Figure 3A–C), the Indonesian clones showed considerably different reaction norms compared to the European clones. For these paramecia we could identify significantly higher CTmin, Topt, CTmax as well as μmax values compared to the key characteristics of all European P. caudatum clones (Mann-Whitney U-test; U = 0, p = 0.002, n1 = 15, n2 = 3). In addition, an AIC based comparison of nonlinear mixed-effects models with different levels of spatial aggregation (Table 2) identified the model with separate fixed effects for the two regions, Europe and Indonesia (model nm2) as the second best model. While the null model nm0a with common fixed effects for all 18 individual clones was the most parsimonious and significantly superior model, the model with two regions was significantly better than the four regions model (nm4, cf. Table 2). This indicates a clear separation between the European and Indonesian clones, but not within the European paramecia.

Correlation between Ecophysiology and Latitude

As illustrated in Figure 4, we obtained significant negative correlations for CTmin (rs = −0.795, p<0.001) and CTmax (rs = −0.596, p<0.01) with latitude using the complete dataset. When the correlation analyses were restricted to the European subset, we could only detect a significant correlation for CTmin (rs = −0.647; p<0.01; see Figure 4). These results were also supported by Mantel tests for the correlation between geographic and ecophysiological (CTmin, Topt, CTmax) distances. Here, analyses on the complete dataset revealed highly significant correlations for all tests (distance matrices of CTmin, Topt and CTmax vs. geographic distance matrix), while Mantel tests on the European subset showed non-significant relationships at all (cf. Table S3).


Figure 4. Latitude-dependent ecophysiology.

Dependency between latitude and ecophysiological key characteristics: A) critical thermal minimum (CTmin), B) thermal optimum (Topt) and C) critical thermal maximum (CTmax). Symbols represent the mean ± standard error of the mean derived from the nonlinear mixed-effects model nm0a with residual bootstrapping. Latitudes of the European P. caudatum clones were corrected for altitude assuming that 100 m elevation translates into a 100 km latitudinal increment within the temperate zone. Spearman's rank correlation coefficients and the respective p-values are as follows for the whole dataset (n = 18): CTmin (rs = −0.795, p<0.001), Topt (rs = −0.409, p = 0.092), CTmax (rs = −0.596, p<0.01); and for the European subset (n = 15): CTmin (rs = −0.647, p<0.01), Topt (rs = 0.027, p = 0.924), CTmax (rs = −0.299, p = 0.279).


Our results further showed that thermal safety margins increased with latitude as well as altitude (Figure 5A–C). All low-latitudinal European paramecia from low altitude possessed considerably smaller thermal safety margins than all other clones tested. Interestingly, thermal safety margins derived from nearby weather stations were on average 1.36±0.65 K lower than data derived from the WorldClim database (cf. Figure 5A,B). This is potentially due to differently estimated time scales (years 2000–2011 versus ~1950–2000) indicated by the fact that the first decade of the 21st century apparently was the warmest since climate records began and by holding at least two summers most likely been the warmest in Europe since year 1500 [65], [66]. Taking this into account, the lowest observed thermal safety margin was 2.79±0.45 K for clone POE-1 from Portugal and the highest for clone NOE-1 from Norway (16.51±0.16 K).


Figure 5. Latitudinal trends in thermal safety margin and maximum warming tolerance.

Thermal safety margins (A–C) and maximum warming tolerances (D–F) of the investigated Paramecium caudatum clones were calculated using habitat temperatures from global climate layers (A+D), near-by meteorological station data (B+E) and from climate change projections (C+F). Symbols represent the mean ± standard error of the mean. Latitudes of the European clones were corrected for altitude assuming a 100 km increase in latitude for a 100 m increase in altitude. Arrows indicate P. caudatum clones from high altitude (cf. Table 1).


As shown in Figure 4C, we could not detect a significant decrease in CTmax with increasing latitude by analysing the European dataset. Therefore, the decrease in maximum air temperature with increasing latitude was considerably higher than the decrease in CTmax of the respective P. caudatum clones. This resulted in a steep increase in maximum warming tolerance with increasing latitude for the European P. caudatum clones (Figure 5D–F). Here, the low-latitude clones from Spain and Portugal showed the lowest tolerance window for extreme temperatures (Figure 5D). Analysing maximum warming tolerances derived from modelled climate change scenarios further revealed that these two low-latitudinal European clones (ESH-2, POE-1) would show negative maximum warming tolerances (Figure 5F). The three investigated Indonesian clones, which served as a tropical outgroup, showed higher critical maximum temperatures than all tested European clones (Figure 4C). Hence, these clones showed positive maximum warming tolerances even for the predicted future maximum temperatures in their current habitats (Figure 5D–F).

Thermal Constraints

Analyses of all optimum temperatures (Topt) and calculated maximum growth rates (μmax, cal) revealed a significant positive correlation of Topt and μmax, cal (n = 18, rs = 0.756, p<0.001; Figure 6A). This pattern was not significant among the European subset.


Figure 6. Thermal constraints.

A) Relationship between calculated optimum temperature (Topt) and maximum growth rate (μmax, cal) for all investigated P. caudatum clones supporting warmer is better. Symbols represent the mean ± standard error of the mean derived from the nonlinear mixed-effects model nm0a with residual bootstrapping. Significance was tested with Spearman's rank correlation (n = 18, rs = 0.775, p<0.001). B) Trade-off between calculated maximum growth rate (μmax, cal) and thermal tolerance breadth (TTB = CTmaxCTmin). Symbols represent the mean ± standard error of the mean derived from the nonlinear mixed-effects model nm0a with residual bootstrapping. Significance was tested by using Spearman's rank correlation (n = 18, rs = −0,554, p<0.05).


A negative correlation between the maximum performance (μmax, cal) and thermal tolerance breadth was detected using the complete dataset and performing Spearman's rank correlation (n = 18, rs = −0,546, p<0.05; Figure 6B). This indicates that higher maximum growth rates at higher optimum temperatures resulted in a narrower thermal tolerance (Indonesian clones), while a broader thermal tolerance was connected with lower maximum performance (European clones).

Genetic and Ecophysiological Distances

Performing Mantel tests [59] on the complete dataset revealed highly significant correlations between ecophysiological (CTmin, Topt, CTmax) and genetic distances. Excluding the Indonesian clones from the analyses resulted in non-significant relationships (see Table 4). This indicates a missing correlation of the ecophysiology and the genetic distance within the investigated European P. caudatum clones using the mitochondrial COI gene as a phylogenetic marker and temperature dependent population growth rates as a fitness component. However, please note that the Indonesian clones possess both high geographical distances (10,715 km–13,261 km) as well as high genetic distances (0.073–0.087 substitutions per site) compared to the European P. caudatum clones, which exhibit comparatively low genetic distances among each other (0–0.02 substitutions per site). Performing Mantel tests for the genetic and geographic distances resulted in significant correlations for both datasets (Table S4).


Table 4. Mantel test for the correlation between genetic (x matrix) and ecophysiological distances (y matrix).



Latitude-dependent Ecophysiology

Generally, high-latitude Paramecium caudatum populations encounter lower temperatures than low-latitude populations, which should select for higher growth rates at low or high temperatures, respectively. Northern European P. caudatum clones should therefore exhibit lower critical minimum temperatures (CTmin) to which the southern European clones were barely exposed to, while southern European clones should possess higher critical maximum temperatures (CTmax). Our data partly support this general assumption of a co-variation of the critical thermal limits (CTmin, CTmax) and the optimum temperature (Topt) with the latitude when analysing the European dataset. Performing correlation analyses on this dataset revealed a significant dependence of CTmin with altitude-corrected latitudes, although we partly obtained high standard errors based on the used fitting procedure (cf. Figure 4A, Table 3). In addition, we could not detect such a correlation for Topt and CTmax using the European dataset (Figure 4B,C). Our data therefore indicate a potential thermal adaptation of the northern European P. caudatum clones to lower winter temperatures only. These results are in accordance with a number of investigations on terrestrial ectotherms, which revealed that the lower critical temperatures significantly decline with increasing latitude while the upper thermal limits do not or are less variable [12], [33], [34], [67], [68]. This might be due to different costs for cold and warm tolerance, but up to now there is no clear evidence whether cold or warm adaptation came at a higher cost or if high heat tolerance is the ancestral state with low retaining costs (cf. [33], [69], [70], [71]).

When using the complete dataset that includes experimental data of all tested European as well as the three Indonesian clones, significant co-variations of CTmin as well as CTmax with latitude were obtained (Figure 4A,C). Additionally, the results of the model selection approach (Table 2) identified the model nm2 with two regions (Europe and Indonesia) as significantly superior compared to the four regions model nm4 (with regions for northern, central, southern Europe and Indonesia). These results imply a stronger impact of latitude on the thermal performance of the eukaryotic microbe P. caudatum when comparing data on a large inter-continental geographic scale, in contrast to the analyses on the intra-continental European scale. Further, significant positive correlations of all ecophysiological distances with genetic distances could be identified only when using the complete dataset, but not among the European subset (Table 4). Hence, the detected high genetic distances and significantly different ecophysiological characters of the Indonesian compared to the European clones as well as the results of the model selection approach suggest a large-scale biogeographic diversification within Paramecium caudatum on the phenotypic as well as the genetic level.

Phenotypic Plasticity AND Thermal Adaptation

In general, the thermal performance of all investigated European Paramecium caudatum clones is indicative for a high phenotypic plasticity of this freshwater ciliate. Exemplified by the northern European clones, they showed a higher physiological optimum (around 29°C) compared to the temperatures they experienced in their natural habitats (Figure 5A). The optimum temperatures and the shapes of the thermal performance curves (TPCs) were reasonably similar for all genetically distinct clones from Europe (cf. Figure 3, 4B). Further, all European clones showed a general broad thermal tolerance and non-covarying CTmax as well as Topt values with latitude (Figure 4B,C). These results disapprove the hypothesised latitudinal clines for thermal adaptation indicators such as heat tolerance and optimum temperature in European P. caudatum genotypes, but are in agreement with the climatic variability or the seasonal variability hypotheses [13], [72], [73]. These hypotheses claim that greater environmental variability at higher latitudes, for example due to seasonal changes, select for a more ‘generalist’ climatic tolerance and favours phenotypic plasticity within populations. Such a scenario would be also supported by the low genetic differentiation among the investigated European P. caudatum clones. This fact could be interpreted as an increased gene flow facilitated by the high phenotypic plasticity and the resulting low dispersal costs [13]. In that case, the high gene flow and high dispersal rates would also limit a potential local adaptation to specific habitat temperatures.

As our data revealed a comparatively broad thermal tolerance for all European clones with exceptional high Topt and CTmax values, we would argue for an adaptive phenotypic plasticity among the investigated European populations resulting in thermal generalists. On the other hand, arguments for a potential thermal adaptation of the European P. caudatum to local microhabitat conditions are obvious due to the co-varying CTmin values with latitude (Figure 4A), but also because of large intraspecific growth rate variations at temperatures above Topt (Figure 3). These differences at low and high temperatures indicate the existence of various ecotypes and a potential microadaptation to the local microclimate within the European P. caudatum clones. Temperature adaptation to microhabitat conditions has also been shown by several studies for ectothermic metazoans [74][77].

Along the three tested Indonesian clones, it was remarkable that in comparison to the European clones the averaged lower critical temperature of these tropical paramecia was shifted by more than +6 K, whereas Topt and CTmax were only shifted by approx. +1.5 or +2.1 K, respectively. This means a reduction of their thermal tolerance breadth compared to the European P. caudatum clones. Given that in general the Indonesian clones were hardly ever stressed by temperatures below 10°C in their natural habitats, these tropical populations have not needed to adapt to lower temperatures. They have either lost the low-temperature tolerance or have never had this ability, depending on whether the tropical or the European clones represent the most ancestral phenotype. On the other hand, the Indonesian populations from low altitudes experience higher frequencies of hotter daily maximum temperatures. For example, the maximum air temperature for the natural habitat of clone INP-3 could reach temperatures of up to 43°C. However, the mean maximum air temperature, which corresponds arguably better to the maximum water temperature because of the buffering capacity of water [78], is around 33°C. Here, we could show that the CTmax values of all investigated Indonesian P. caudatum clones (cf. Table 3) were somewhat beyond this temperature regardless of the elevation of their natural habitats. This result is indicative for a common thermal adaptation of the investigated Indonesian P. caudatum clones to the tropical temperature regime.

These findings support the general assumption that acclimatization and a high phenotypic plasticity is more likely in temperate species or populations at higher latitudes because the overall temperature variation increases with latitude. Thus, the evolution of broad thermal tolerances is needed that temperate organisms can cope with a large seasonal variation while organisms from less variable tropical environments should have evolved narrow thermal tolerances and reduced acclimation responses ([13], [32], [33]; but see [79]), as shown by our study. To generalise this pattern for P. caudatum, more clones from other tropical habitats need to be investigated.

Thermal Constraints

In this study, a significant correlation between optimum temperature and maximum population growth rates could be shown (Figure 6A), which supports the so-called warmer is better hypothesis [22], [23], [24]. This thermodynamic-constraint hypothesis argues for a dependence of maximum performance on optimum temperature because of the thermodynamic properties of biochemical and physiological systems [24], [80]. Our results document that P. caudatum clones with higher optimum temperatures (Topt) have generally higher maximum growth rates (μmax) giving evidence for warmer is better concerning the growth performance as an important component of overall fitness. Nevertheless, we could not observe a significant correlation of Topt and μmax for the European subset, which indicates that the scale at which investigations are performed is of importance. Further studies on geographically well described organisms are necessary to understand if this finding is of general relevance for a variety of ectotherm species, for microbial eukaryotes only, or just for Paramecium caudatum.

The present study also provides an indication of an evolutionary trade-off between the performance breadth and maximum performance, which is known as the Jack-of-all-temperatures is a master of none hypothesis [25]. This result indicates that the selection for a broad thermal tolerance could result in a lower peak performance [81] while the selection for greater performance at a higher temperature would cause a correlated decrease in performance at lower temperatures [20]. Several studies provided mixed or no support of such specialist-generalist tradeoffs that constrain TPCs (e.g. [82][85]). Here, we could demonstrate such a negative correlation between maximum growth rates (μmax) and thermal tolerance breadths (Figure 6B). Especially the three tested Indonesian paramecia possessed higher μmax values but narrower thermal tolerances than the European clones, mainly caused by a considerably larger shift of CTmin to higher temperatures compared to CTmax. This result supports the above formulated suggestion that the European P. caudatum clones could be indicated as thermal generalists with a maximised performance breadth while the Indonesian clones seem to be thermal specialists with a maximised peak performance at higher temperatures.

Sensitivity to Predicted Increasing Temperatures

The impact of the ongoing climate change on an ectotherm's fitness depends on numerous factors, including the community responses or the resource availability and the temperature-specific resource demand [86]. However, due to the expected temperature rise of up to 5.8°C by the year 2100 and the anticipated increase in diurnal variability of summer temperatures in the Northern Hemisphere [18], temperature is one of the most important factors which can drive shifts within the structure of natural populations [1], [2], [3]. Therefore, climate change effects are often predicted to give rise to species extinctions [82], [87]. The consequence of this temperature effect will considerably depend on the genotype specific Topt and CTmax values relative to the mean and extreme habitat temperatures [88]. Thermal safety margins as well as maximum warming tolerances are suitable characters to reveal the potential extinction risk of organisms due to increasing temperatures [17].

In the present study, positive thermal safety margins reflect the temperature range at which the respective P. caudatum clone may benefit from a future warming of the habitat due to increasing growth rates. Negative values are rather a measure of potential risk. Our analysis revealed considerably smaller thermal safety margins for the low-latitudinal and -altitudinal European and tropical paramecia compared to all other P. caudatum clones tested (Figure 5A). However, all investigated clones currently live in environments that are on average cooler than their physiological thermal optimum. All tested clones seem to potentially benefit from future increasing temperatures (Figure 5C) at least initially.

Nevertheless, a second important key characteristic especially in consideration of the expected increasing intensity, frequency and duration of summer heat waves is the maximum warming tolerance. This character illustrates the average increase in maximum temperatures which Paramecium can tolerate before harmful growth conditions will be reached. Unexpectedly, analyses of the maximum warming tolerance showed that the European paramecia from low latitude and altitude currently experience near- or even above-lethal temperatures during summer (Figure 5E). Furthermore, the predicted future maximum temperatures of the low-latitudinal habitats (Portugal and Spain) are higher than their critical maximum temperatures (Figure 5F). Climate change models further predict the highest warming rates for low-latitude European habitats based on our dataset. Consequently, if these genotypes cannot adapt to the expected higher temperatures in their current natural habitats, they will potentially suffer from global warming. This seems not to be the case for the Indonesian and all high-latitudinal and –altitudinal European clones, which possess higher maximum warming tolerances (Figure 5D–F).

Our results, therefore, only partly support the hypothesis that tropical ectotherms are most at risk due to novel as well as disappearing climates in the tropics and subtropics in consequence of climate change [17], [89], [90]. In terms of the microbial eukaryote Paramecium caudatum, only the most southern European populations seem to be adversely affected by global warming. While some of the high-latitude European populations may actually benefit from increased temperatures by an enhanced population growth, the investigated Indonesian clones seem also not to suffer from the expected temperature rise. This is due to their adaptation to higher temperatures as well as the fact that temperature increase in the tropics is expected to be less intensive compared to temperate habitats [18]. However, tropical P. caudatum populations may become also affected by global warming, because the predicted changing environmental conditions such as temperature and precipitation seem to be comparatively heterogeneous across latitude [90], [91] and we have only a small dataset of tropical paramecia. Further, how climate change affects species abundances, distribution or diversity depends on the multigenerational response of their survival and reproduction within ecosystems [92]. The magnitude of temperature effects on species hinge on several factors such as food-web interactions, altered competition as well as the species specific acclimation or adaptation capacity [88], [93], [94]. Hence, experimental selection and micro-evolutionary studies as well as competition experiments of artificial populations pose an interesting research outline for future studies.


All investigated European clones showed a broad thermal tolerance with high upper thermal limits as well as similar optimum temperatures, which is indicative for thermal generalists. These genetically closely related P. caudatum clones may have been selected to perform well in thermally variable environments utilising a high acclimation capacity. In contrast, the investigated Indonesian genotypes showed significantly higher optimum temperatures, maximum performances and critical thermal limits. This suggests local temperature adaptation of these tested clones to the less thermally variable tropical temperature regime. Further studies on additional ‘continental subpopulations’ of P. caudatum are needed to generalise the suggested thermal generalist and thermal specialist patterns in this ubiquitous microbial eukaryote.

Relating ecophysiological key characteristics such as optimum and maximum temperatures to current and predicted temperatures of all investigated P. caudatum clones suggest that only low-latitude European clones would be sensitive to global warming. During summer, they currently perform closer to their thermal limits than high-latitude European and tropical clones. Increasing temperature extremes, therefore, may severely affect their performance and fitness. Future investigations on experimental evolution would be useful to examine whether the different genotypes tested in this study can adapt to the predicted increase in temperatures and extreme events, or not.

Supporting Information

Table S1.

Parameter estimates of the Lactin-2 model.



Table S2.

Geographical and meteorological details for natural habitats of investigated Paramecium caudatum clones.



Table S3.

Mantel test for the correlation between geographic ( x matrix) and ecophysiological distances ( y matrix).



Table S4.

Mantel test for the correlation between genetic ( x matrix) and geographic distances ( y matrix).




First of all, we would like to thank D. Barth, M.U. Böhme, M.J. Caramujo, S.I. Fokin and Ch. Zschornack providing our study with freshwater samples from several habitats. This study would not have been possible without this support. SK especially would like to thank K. Stenchly for her help during his sampling trip in Sulawesi. We further thank Ch. Zschornack for her valuable laboratory and experimental assistance. All authors are also grateful to A. D. Sommerfeldt for finding and correcting faults in the manuscript and would like to thank anonymous reviewers for helpful comments on an earlier draft.

Author Contributions

Conceived and designed the experiments: SK TUB. Performed the experiments: SK. Analyzed the data: SK TP. Contributed reagents/materials/analysis tools: SK TP TUB. Wrote the paper: SK TP TUB.


  1. 1. Cossins AR, Bowler K (1987) Temperature biology of animals. London: Chapman and Hall. 339 p.
  2. 2. Houghton JT, Jenkins GJ, Ephraums JJ (1990) Scientific assessment of climate change – Report of working group I. Cambridge: Cambridge University Press. 365 p.
  3. 3. Johnston IA, Bennett AF (1996) Animals and temperature: phenotypic and evolutionary adaptation. Cambridge: Cambridge University Press. 419 p.
  4. 4. Petchey OL, McPhearson PT, Casey TM, Morin PJ (1999) Environmental warming alters food-web structure and ecosystem function. Nature 402: 69–72.
  5. 5. Wang H, Lewis M (2011) “Invasive species show higher phenotypic plasticity than non-invasive species. Should this characteristic benefit invasive species…” Evaluation of: [Davidson AM et al. Do invasive species show higher phenotypic plasticity than native species and, if so, is it adaptive? A meta-analysis. Ecology Letters 14: 419–431]. F1000com/9268957.
  6. 6. Ghalambor CK, McKay JK, Carroll SP, Reznick DN (2007) Adaptive versus non-adaptive phenotypic plasticity and the potential for contemporary adaptation in new environments. Functional Ecology 21: 394–407.
  7. 7. Guerra D, Cavicchi S, Krebs RA, Loeschcke V (1997) Resistance to heat and cold stress in Drosophila melanogaster: intra and inter population variation in relation to climate. Genetics Selection Evolution 29: 497–510.
  8. 8. Sørensen JG, Dahlgaard J, Loeschcke V (2001) Genetic variation in thermal tolerance among natural populations of Drosophila buzzatii: down regulation of Hsp70 expression and variation in heat stress resistance traits. Functional Ecology 15: 289–296.
  9. 9. Elliott JM, Elliott JA, Allonby JD (1994) The critical thermal limits for the stone loach, Noemacheilus barbatulus, from three populations in north-west England. Freshwater Biology 32: 593–601.
  10. 10. Kimura MT, Ohtsu T, Yoshida T, Awasaki T, Lin FJ (1994) Climatic adaptations and distributions in the Drosophila takahashii species subgroup (Diptera: Drosophilidae). Journal of Natural History 28: 401–409.
  11. 11. Räisänen J (2002) CO2-induced changes in interannual temperature and precipitation variability in 19 CMIP2 experiments. Journal of Climate 15: 2395–2411.
  12. 12. Addo-Bediako A, Chown SL, Gaston KJ (2000) Thermal tolerance, climatic variability and latitude. Proceedings of the Royal Society of London B - Biological Sciences 267: 739–745.
  13. 13. Janzen DH (1967) Why mountain passes are higher in the tropics. American Naturalist 101: 233–249.
  14. 14. Dobzhansky T (1950) Evolution in the tropics. American Scientist 38: 209–221.
  15. 15. Parmesan C, Root TL, Willig MR (2000) Impacts of extreme weather and climate on terrestrial biota. Bulletin of the American Meteorological Society 81: 443–450.
  16. 16. Dillon ME, Wang G, Huey RB (2010) Global metabolic impacts of recent climate warming. Nature 467: 704–706.
  17. 17. Deutsch CA, Tewksbury JJ, Huey RB, Sheldon KS, Ghalambor CK, et al. (2008) Impacts of climate warming on terrestrial ectotherms across latitude. Proceedings of the National Academy of Sciences of the United States of America 105: 6668–6672.
  18. 18. IPCC (2007) Climate change 2007: the fourth assessment report. CambridgeNew York, NY, UKUSA: Cambridge University Press.
  19. 19. Schär C, Vidale PL, Lüthi D, Frei C, Häberli C, et al. (2004) The role of increasing temperature variability in European summer heatwaves. Nature 427: 332–336.
  20. 20. Angilletta MJ, Niewiarowski PH, Navas CA (2002) The evolution of thermal physiology in ectotherms. Journal of Thermal Biology 27: 249–268.
  21. 21. Buckley LB (2008) Linking traits to energetics and population dynamics to predict lizard ranges in changing environments. American Naturalist 171: E1–E19.
  22. 22. Bennett AF (1987) Evolution of the control of body temperature: is warmer better? In: Dejours P, Bolis L, Taylor C, Weibel E, editors. Comparative physiology: life in water and on land. Padova, Italy: Liviana Press, Springer Verlag. pp. 421–431.
  23. 23. Hamilton WJ (1973) Life's color code. New York: McGraw-Hill. 238 p.
  24. 24. Huey RB, Kingsolver JG (1989) Evolution of thermal sensitivity of ectotherm performance. Trends in Ecology & Evolution 4: 131–135.
  25. 25. Huey RB, Hertz PE (1984) Is a jack-of-all-temperatures a master of none? Evolution 38: 441–444.
  26. 26. Bennett AF, Lenski RE, Mittler JE (1992) Evolutionary adaptation to temperature. I. Fitness responses of Escherichia coli to changes in its thermal environment. Evolution 46: 16–30.
  27. 27. Frazier MR, Huey RB, Berrigan D (2006) Thermodynamics constrains the evolution of insect population growth rates: “warmer is better”. American Naturalist 168: 512–520.
  28. 28. Gilchrist GW (1996) A quantitative genetic analysis of thermal sensitivity in the locomotor performance curve of Aphidius ervi. Evolution 50: 1560–1572.
  29. 29. Knies JL, Kingsolver JG, Burch CL (2009) Hotter is better and broader: thermal sensitivity of fitness in a population of bacteriophages. American Naturalist 173: 419–430.
  30. 30. Cooper BS, Williams BH, Angilletta MJ (2008) Unifying indices of heat tolerance in ectotherms. Journal of Thermal Biology 33: 320–323.
  31. 31. Krenek S, Berendonk TU, Petzoldt T (2011) Thermal performance curves of Paramecium caudatum: a model selection approach. European Journal of Protistology 47: 124–137.
  32. 32. Ghalambor CK, Huey RB, Martin PR, Tewksbury JJ, Wang G (2006) Are mountain passes higher in the tropics? Janzen's hypothesis revisited. Integrative and Comparative Biology 46: 5–17.
  33. 33. Sunday JM, Bates AE, Dulvy NK (2011) Global analysis of thermal tolerance and latitude in ectotherms. Proceedings of the Royal Society of London B - Biological Sciences 278: 1823–1830.
  34. 34. Terblanche JS, Deere JA, Clusella-Trullas S, Janion C, Chown SL (2007) Critical thermal limits depend on methodological context. Proceedings of the Royal Society of London B - Biological Sciences 274: 2935–2942.
  35. 35. Foissner W (1999) Soil protozoa as bioindicators: pros and cons, methods, diversity, representative examples. Agriculture, Ecosystems & Environment 74: 95–112.
  36. 36. Sherr EB, Sherr BF (2002) Significance of predation by protists in aquatic microbial food webs. Antonie van Leeuwenhoek International Journal of General and Molecular Microbiology 81: 293–308.
  37. 37. Elena SF, Lenski RE (2003) Evolution experiments with microorganisms: the dynamics and genetic bases of adaptation. Nature Reviews Genetics 4: 457–469.
  38. 38. Jessup CM, Kassen R, Forde SE, Kerr B, Buckling A, et al. (2004) Big questions, small worlds: microbial model systems in ecology. Trends in Ecology & Evolution 19: 189–197.
  39. 39. Andrushchyshyn OP, Wilson KP, Williams DD (2009) Climate change-predicted shifts in the temperature regime of shallow groundwater produce rapid responses in ciliate communities. Global Change Biology 15: 2518–2538.
  40. 40. Fox JW, Morin PJ (2001) Effects of intra- and interspecific interactions on species responses to environmental change. Journal of Animal Ecology 70: 80–90.
  41. 41. Christoffersen K, Andersen N, Sondergaard M, Liboriussen L, Jeppesen E (2006) Implications of climate-enforced temperature increases on freshwater pico- and nanoplankton populations studied in artificial ponds during 16 months. Hydrobiologia 560: 259–266.
  42. 42. Norf H, Arndt H, Weitere M (2007) Impact of local temperature increase on the early development of biofilm-associated ciliate communities. Oecologia 151: 341–350.
  43. 43. Boucher P, Blinn DW, Johnson DB (1984) Phytoplankton ecology in an unusually stable environment (Montezuma Well, Arizona, USA). Hydrobiologia 119: 149–160.
  44. 44. Little C (1990) The terrestrial invasion: an ecophysiological approach to the origins of land animals. Cambridge: Cambridge University Press. 314 p.
  45. 45. Bailey MJ, Lilley AK, Thompson IP, Rainey PB, Ellis RJ (1995) Site directed chromosomal marking of a fluorescent pseudomonad isolated from the phytosphere of sugar beet; stability and potential for marker gene transfer. Molecular Ecology 4: 755–763.
  46. 46. Sonneborn TM (1970) Methods in Paramecium research. In: Prescott DM, editor. Methods in cell physiology. New York: Academic Press. pp. 241–339.
  47. 47. Bouin PA (1897) Etudes sur l'évolution normale et l'involution du tube seminifère. Archives d'Anatomie Microscopique et de Morphologie Experimentale 1: 225–339.
  48. 48. Barth D, Krenek S, Fokin SI, Berendonk TU (2006) Intraspecific genetic variation in Paramecium revealed by mitochondrial cytochrome c oxidase I sequences. Journal of Eukaryotic Microbiology 53: 20–25.
  49. 49. Lactin DJ, Holliday NJ, Johnson DL, Craigen R (1995) Improved rate model of temperature-dependent development by arthropods. Environmental Entomology 24: 68–75.
  50. 50. Pinheiro J, Bates D (2000) Mixed effects models in S and S-Plus. New York: Springer. 530 p.
  51. 51. Efron B, Tibshirani RJ (1994) An introduction to the bootstrap. New York, London: Chapman and Hall/CRC. 456 p.
  52. 52. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25: 1965–1978.
  53. 53. Govindasamy B, Duffy PB, Coquard J (2003) High-resolution simulations of global climate, part 2: effects of increased greenhouse gases. Climate Dynamics 21: 391–404.
  54. 54. Bolstad PV, Bentz BJ, Logan JA (1997) Modelling micro-habitat temperature for Dendroctonus ponderosae (coleoptera: scolytidae). Ecological Modelling 94: 287–297.
  55. 55. Bryant SR, Shreeve TG (2002) The use of artificial neural networks in ecological analysis: estimating microhabitat temperature. Ecological Entomology 27: 424–432.
  56. 56. Livingstone DM, Lotter AF (1998) The relationship between air and water temperatures in lakes of the Swiss Plateau: a case study with palaeolimnological implications. Journal of Paleolimnology 19: 181–198.
  57. 57. Jump AS, Matyas C, Penuelas J (2009) The altitude-for-latitude disparity in the range retractions of woody species. Trends in Ecology & Evolution 24: 694–701.
  58. 58. Colwell RK, Brehm G, Cardelus CL, Gilman AC, Longino JT (2008) Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. Science 322: 258–261.
  59. 59. Mantel N (1967) Detection of disease clustering and a generalized regression approach. Cancer Research 27: 209–220.
  60. 60. Hebert PDN, Ratnasingham S, deWaard JR (2003) Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society of London B - Biological Sciences 270: S96–S99.
  61. 61. R Development Core Team (2008) R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.
  62. 62. Pinheiro J, Bates D, Debroy S, Sarkar D, the R Core Team (2008) nlme: linear and nonlinear mixed effects models. R package version 3.1–89.
  63. 63. Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, et al. (2011) vegan: community ecology package. R package version 2.0-1.
  64. 64. Legendre P, Legendre L (1998) Numerical ecology. Amsterdam: Elsevier. 853 p.
  65. 65. Cole S, McCarthy L (2010) NASA research finds last decade was warmest on record, 2009 one of warmest years. NASA press release: 10-017.
  66. 66. Barriopedro D, Fischer EM, Luterbacher J, Trigo R, Garcia-Herrera R (2011) The hot summer of 2010: Redrawing the temperature record map of Europe. Science 332: 220–224.
  67. 67. Gibert P, Huey RB (2001) Chill-coma temperature in Drosophila: effects of developmental temperature, latitude, and phylogeny. Physiological and Biochemical Zoology 74: 429–434.
  68. 68. Helmuth B, Kingsolver JG, Carrington E (2005) Biophysics, physiological ecology, and climate change: does mechanism matter? Annual Review of Physiology 67: 177–201.
  69. 69. Jackson MW, Stinchcombe JR, Korves TM, Schmitt J (2004) Costs and benefits of cold tolerance in transgenic Arabidopsis thaliana. Molecular Ecology 13: 3609–3615.
  70. 70. Kristensen TN, Hoffmann AA, Overgaard J, Sorensen JG, Hallas R, et al. (2008) Costs and benefits of cold acclimation in field-released Drosophila. Proceedings of the National Academy of Sciences of the United States of America 105: 216–221.
  71. 71. Hoffmann AA, Sorensen JG, Loeschcke V (2003) Adaptation of Drosophila to temperature extremes: bringing together quantitative and molecular approaches. Journal of Thermal Biology 28: 175–216.
  72. 72. Letcher AJ, Harvey PH (1994) Variation in geographical range size among mammals of the Palearctic. American Naturalist 144: 30–42.
  73. 73. Stevens GC (1996) Extending Rapoport's rule to Pacific marine fishes. Journal of Biogeography 23: 149–154.
  74. 74. Danks HV (2006) Insect adaptations to cold and changing environments. Canadian Entomologist 138: 1–23.
  75. 75. Edgerly JS, Tadimalla A, Dahlhoff EP (2005) Adaptation to thermal stress in lichen-eating webspinners (Embioptera): habitat choice, domicile construction and the potential role of heat shock proteins. Functional Ecology 19: 255–262.
  76. 76. Irwin JT, Lee RE (2003) Cold winter microenvironments conserve energy and improve overwintering survival and potential fecundity of the goldenrod gall fly, Eurosta solidaginis. Oikos 100: 71–78.
  77. 77. Sinclair BJ (2001) Field ecology of freeze tolerance: interannual variation in cooling rates, freeze-thaw and thermal stress in the microhabitat of the alpine cockroach Celatoblatta quinquemaculata. Oikos 93: 286–293.
  78. 78. Livingstone DM (1993) Temporal structure in the deep-water temperature of four Swiss lakes: a short-term climatic change indicator? Verhandlungen des Internationalen Verein Limnologie 25: 75–81.
  79. 79. Gaston KJ, Chown SL (1999) Why Rapoport's rule does not generalise. Oikos 84: 309–312.
  80. 80. Savage VM, Gillooly JF, Brown JH, West GB, Charnov EL (2004) Effects of body size and temperature on population growth. American Naturalist 163: 429–441.
  81. 81. Levins R (1968) Evolution in changing environments. Princeton, NJ: Princeton University Press. 132 p.
  82. 82. Anderson AR, Hoffmann AA, McKechnie SW (2005) Response to selection for rapid chill-coma recovery in Drosophila melanogaster: physiology and life-history traits. Genetical Research 85: 15–22.
  83. 83. Bennett AF, Lenski RE (1993) Evolutionary adaptation to temperature. II. Thermal niches of experimental lines of Escherichia coli. Evolution 47: 1–12.
  84. 84. Carriere Y, Boivin G (2001) Constraints on the evolution of thermal sensitivity of foraging in Trichogramma: genetic trade-offs and plasticity in maternal selection. American Naturalist 157: 570–581.
  85. 85. Leroi AM, Lenski RE, Bennett AF (1994) Evolutionary adaptation to temperature. III. Adaptation of Escherichia coli to a temporally varying environment. Evolution 48: 1222–1229.
  86. 86. Norf H, Weitere M (2010) Resource quantity and seasonal background alter warming effects on communities of biofilm ciliates. FEMS Microbiology Ecology 74: 361–370.
  87. 87. Pounds JA, Bustamante MR, Coloma LA, Consuegra JA, Fogden MPL, et al. (2006) Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439: 161–167.
  88. 88. Pörtner HO, Bennett AF, Bozinovic F, Clarke A, Lardies MA, et al. (2006) Trade-offs in thermal adaptation: the need for a molecular to ecological integration. Physiological and Biochemical Zoology 79: 295–313.
  89. 89. Kearney M, Shine R, Porter WP (2009) The potential for behavioral thermoregulation to buffer “cold-blooded” animals against climate warming. Proceedings of the National Academy of Sciences of the United States of America 106: 3835–3840.
  90. 90. Williams JW, Jackson ST, Kutzbacht JE (2007) Projected distributions of novel and disappearing climates by 2100 AD. Proceedings of the National Academy of Sciences of the United States of America 104: 5738–5742.
  91. 91. Bonebrake TC, Mastrandrea MD (2010) Tolerance adaptation and precipitation changes complicate latitudinal patterns of climate change impacts. Proceedings of the National Academy of Sciences of the United States of America 107: 12581–12586.
  92. 92. Harmon JP, Moran NA, Ives AR (2009) Species response to environmental change: impacts of food web interactions and evolution. Science 323: 1347–1350.
  93. 93. Tylianakis JM, Didham RK, Bascompte J, Wardle DA (2008) Global change and species interactions in terrestrial ecosystems. Ecology Letters 11: 1351–1363.
  94. 94. Frost TM, Carpenter SR, Ives AR, Kratz TK (1995) Species compensation and complementarity in ecosystem function. In: Jones CG, Lawton JH, editors. Linking species and ecosystems. New York: Chapman & Hall. pp. 224–239.