By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
1. Human activities are greatly changing the conditions of ecosystems. One of these major changes is that more and more ecosystems are facing increasing concentrations of highly toxic heavy metals like cadmium (Cd), lead (Pb) and mercury (Hg). Previous studies have shown this type of pollution may be associated with the loss of biodiversity, leaving some pollution-tolerant species to dominate the polluted ecosystems. To date, however, little is known about the extent to which the pollution caused by these toxins may affect the relationship between biodiversity and ecosystem function.
2. In this study, we aimed to assess the role of Cd pollution in determining biodiversity–productivity relationships in aquatic ecosystems, by using data from a laboratory microcosm experiment that tested 110 algal communities with three levels of Cd pollution.
3. Our results showed that the productivities of algal communities growing in non-Cd-polluted media did not significantly increase with increasing species richness, and this pattern did not change over time. In contrast, significantly positive biodiversity–productivity relationships were consistently found in algal communities growing in Cd-polluted media.
4. We revealed that the putative facilitation among algal species in response to Cd pollution was the main driver of the observed significantly positive biodiversity–productivity relationships. Most importantly, we found that not only Cd-tolerant algal species but also Cd-sensitive algal species were able to dominate Cd-polluted polycultures and contribute to the increased productivities of these polycultures.
5.Synthesis and applications. Collectively, our results provide the first explicit experimental evidence that Cd pollution can trigger a positive biodiversity–productivity relationship. We identify two profound implications: (i) conservation of biodiversity in all environments may reduce the future impacts of increasing environmental stresses (like Cd pollution) on ecosystem function (such as primary productivity); (ii) one possible approach to maintain or improve algal primary productivity in a polluted aquatic ecosystem may be to construct some suitable diverse algal assemblages that include not only pollution-tolerant species but also pollution-sensitive ones.
The unprecedented loss of biodiversity driven by rapid environmental changes has provoked a generation of ecologists to reconsider the relationship between biodiversity and ecosystem function (Grime 1997; Hooper et al. 2005). Over the past two decades, much empirical and theoretical work has assessed the effects of biodiversity on various aspects of ecosystem function (Loreau 2000; Hooper et al. 2005). Most of these studies have focused on understanding how plant biodiversity can affect ecosystem productivity (Yachi & Loreau 1999; Tilman et al. 2001; Bullock, Pywell & Walker 2007; Nyfeler et al. 2009). The predominant interest in plant species and productivity can be justified because plant species are primary producers in ecosystems and productivity is a fundamental ecosystem function closely related to a number of other potential services including carbon sequestration, nutrient cycling and as a resource for other trophic levels (Fridley 2001).
By synthesizing the results of previous studies on the relationship between biodiversity and productivity, recent meta-analyses have shown that a positive biodiversity effect was reported frequently (Cardinale et al. 2006, 2007). However, negative and weak effects of biodiversity on productivity were directly demonstrated in some empirical studies (Hooper & Dukes 2004; Bruno et al. 2005; Weis et al. 2007), especially those focusing on the linkage between biodiversity and productivity in aquatic ecosystems (for a recent review, see Stachowicz, Bruno & Duffy 2007). The contrasting evidence suggests that there is an urgent need to examine when and why a positive biodiversity–productivity relationship is not always found, as these patterns may have important implications not only for theory but also for management of ecosystems (Fridley 2001; Hooper et al. 2005; Duffy 2009).
A critical new direction in biodiversity–productivity research addresses the influence of environmental factors. These studies are of particular importance because the results have significant implications for the conservation of natural ecosystems (Stachowicz et al. 2008; Tylianakis et al. 2008; Wacker et al. 2009). Increased variability (e.g. drought perturbation or temperature anomaly) and severe pollution (SP) are the two most pronounced changes to the global environment (Kondratyev, Krapivin & Phillips 2002). However, there is a fundamental difference between the two components: environmental variability generates a dynamic stress on the ecosystem, whereas environmental pollution presents a constant stress to the ecosystem. A large number of studies have shown that environmental variability can strongly affect the relationship between plant diversity and productivity (Mulder, Uliassi & Doak 2001; Pfisterer & Schmid 2002; Allison 2004; Arnone et al. 2008; but see Wang, Yu & Wang 2007). In contrast, there have been relatively few studies exploring the role of environmental pollution in determining the biodiversity–productivity relationship, most of which have focused on nutrient enrichment (e.g. nitrogen deposition, Reich et al. 2001). Despite their wide distribution in the environment and threats to both ecosystem and human health (Kondratyev, Krapivin & Phillips 2002; McConnell & Edwards 2008), little is known about the effects of highly toxic heavy metals, such as cadmium (Cd), lead (Pb) and mercury (Hg), on the relationship between biodiversity and productivity in aquatic systems.
Anthropogenic inputs of heavy metals to the environment are presently 10- to 100-fold greater than those from natural sources (Nriagu & Pacyna 1988). In this context, Cd is considered as a priority pollutant (WHO 1992; Campbell 2006). In polluted aquatic systems, dissolved concentrations of Cd in water range from 0·005 to 1 mg L−1 (Szymanowska, Samecka-Cymerman & Kempers 1999; Smolders et al. 2003), but concentrations as high as 4–6·65 mg L−1 have also been reported (Luo et al. 2008; Rai 2009). Generally, Cd can disturb metabolism of algae, leading to significant growth inhibitions or even death. On the other hand, it has been also found that, at low levels (<0·5 mg L−1), Cd may increase the growth of some algal species (Payne & Price 1999). Moreover, some Cd-tolerant algal species that have evolved various mechanisms to deal with Cd can grow in culture solutions with Cd concentrations up to 50 mg L−1 (Torres et al. 1998; Brinza et al. 2009). Nonetheless, there is strong evidence that the presence of Cd pollution in aquatic ecosystems may lead to the selection of Cd-tolerant species that ultimately dominate these ecosystems at the expense of Cd-sensitive species (Monteiro, Oliveira & Vale 1995; Sabater 2000). In this study, we therefore aimed to explore whether and to what extent these widespread highly toxic heavy metals can alter biodiversity–productivity relationships in aquatic ecosystems, by using Cd as a representative. We focused on algae as they are primary producers and major suppliers of oxygen in aquatic ecosystems.
Given that the positive interactions (i.e. facilitations) among plant species may increase with increasing environmental stress (Callaway et al. 2002), we hypothesized that Cd pollution can strengthen or even trigger a positive biodiversity–productivity relationship. Thus, algal diversity may exert a positive influence on the productivity of Cd-polluted aquatic ecosystems that it would otherwise not exert under benign environmental conditions. To test our hypothesis, we constructed experimental algal communities with different levels of diversity by using 10 freshwater algal species and cultivated the communities in non-Cd-polluted or Cd-polluted media. We then detected possible changes in the biodiversity–productivity relationship and explored the underlying mechanisms.
Materials and methods
We selected 10 unicellular algal species [Ankistrodesmus falcatus (Af), Chlamydomonas eugametos (Ce), Chlamydomonas moewusii (Cm), Chlamydomonas reinhardtii (Cr), Chlorella pyrenoidosa (Cp), Scenedesmus dimorphus (Sd), Scenedesmus obliquus (So), Scenedesmus quadricauda (Sq), Selenastrum capricornutum (Sc) and Staurastrum polymorphum (Sp)] for the microcosm experiment. Three species (Cr, Cp and Sc) were obtained from Jinan University (Guangzhou, China) and the remaining species were provided by the Freshwater Algae Culture Collection of the Chinese Academy of Sciences (Wuhan, China). All study species are common in freshwater phytoplankton communities, grow well under laboratory conditions using common growth media, and are easily distinguished by their morphologies.
Assigning functional groups
We determined the Cd-tolerance levels of the 10 algal species according to the responses of their monocultures to Cd, one of the most ubiquitous and toxic pollutants affecting the environment (WHO 1992). Monocultures (five replicates for each species) of each focal species were incubated in 250-mL conical flasks containing 100 mL of autoclaved Bold’s Basal Medium (BBM) supplemented with 12 mg Cd L−1. The control experiment followed the same procedures using BBM without a Cd supplement. We removed a 10 mL subsample from each experimental unit weekly to measure the algal biomass (see PRODUCTIVITY MEASURE section below for details) and found that the biomass of most algal species reached a steady state within 7 weeks (growth rates of most focal species between the seventh and eighth week were close to zero). We calculated the Cd-tolerance index (TI) of each species using the algal biomass at the end of the seventh week of culture, according to the following formula (Baker 1987): TI = (biomass of alga growing in Cd-supplemented medium ÷ biomass of alga growing in the control medium) × 100%. We then assigned the 10 focal species to two functional groups: Cd-tolerant species with TI >100% [including Af (139%), Cm (371%), Cp (106%) and Cr (145%)] and Cd-sensitive species with TI <100% [including Ce (28·3%), Sc (2·56%), Sd (25·1%), So (20·1%), Sq (10·2%) and Sp (7·33%)].
Our experimental design was a factorial manipulation of two independent factors: algal species richness × Cd pollution degree of culture medium. We established four different levels of species richness (1, 2, 4 and 8). Monocultures of all focal species were replicated five times. For each of the other species richness levels, there were 20 replicates that were different random draws from the 10 member species pool. The 110 randomly assembled algal communities were then crossed with three levels of Cd pollution: no pollution (NP: BBM without Cd supplement), moderate pollution (MP: BBM supplemented with 6 mg Cd L−1) and SP (BBM supplemented with 12 mg Cd L−1). This design yielded a total of 330 treatment replicates. We correspondingly constructed 330 experimental microcosms, which consisted of 250-mL conical flasks containing 100 mL of autoclaved BBM with various concentrations of Cd. We inoculated these BBM media with the randomly assembled algal communities and held the initial algal density constant at 104 cells per 100 mL media across all levels of species richness. We randomly placed the experimental microcosms on laboratory benches in a tissue culture room (20 °C) and repositioned them every 2 days. Cultures were illuminated with cool white light at a photon flux density of 30 μmol m−2 s−1, on a 14/10 h light/dark cycle. We terminated the experiment at the end of the seventh week because, as mentioned above, the biomass of most algal species reached a steady state within 7 weeks.
Theoretical (Kinzig, Pacala & Tilman 2002) and experimental evidence (Tilman et al. 2001; Hooper & Dukes 2004; Bell et al. 2005; Weis et al. 2007) suggest that the biodiversity–productivity relationship is always temporally dynamic. Therefore, besides assessing the role of Cd pollution in determining the biodiversity–productivity relationship, we also examined temporal dynamics. To detect possible changes in the biodiversity–productivity relationship from the beginning to the end of the 7-week microcosm experiment, we sampled all microcosms at the end of the first, fourth and seventh week, and determined the productivities of the randomly assembled algal communities individually. On each sampling date, we removed a 20 mL subsample from each microcosm and counted the cells of each algal species on a hemacytometer using a microscope at a magnification of 1000×. Before counting, the cells were preserved with 4% formaldehyde. The removed subsamples were compensated by the addition of 20 mL of their respective BBM. To calculate the productivity of each algal species in each microcosm, we multiplied the number of algal cells of the species by the average dry weight per cell of this species. To obtain the average dry weight per cell of each algal species, we constructed an additional 100 microcosms containing 100 mL of autoclaved BBM without Cd, where each monoculture of the focal species was replicated 10 times. At the end of the seventh week, a 20 mL subsample was removed from each microcosm for counting the algal cells, while the remaining culture medium in each microcosm was centrifuged at 104 rpm for 10 min at 4 °C. We collected the precipitates derived from each microcosm separately and allowed them to dry in an oven at 80 °C until we could measure a constant dry weight. For each algal species, we then estimated the average dry weight per cell by dividing the dry cell weight per microcosm by the cell number per microcosm.
We used a general linear model to examine the effects of algal species richness on productivities of algal communities under different Cd pollution levels. According to the additive partitioning method proposed by Loreau & Hector (2001), we calculated the net biodiversity effects (ΔY) in polycultures under different Cd pollution levels and then separated ΔY values into two components: complementarity effect and selection effect. All values from the additive partitioning method were tested against zero using a one-sample t-test. To gain additional insight into the mechanisms underlying the positive diversity–productivity relationships observed in this study, the expected yields of individual species in 8-species polycultures in the seventh week were calculated based on the method proposed by Loreau & Hector (2001) and were compared with their respective observed yields. If the observed yield of a species is higher than its expected yield, this species performs better in the polycultures than in a monoculture, and vice versa. A paired sample t-test was used to test for difference between observed yield and expected yield of each species.
Effects of pollution on the biodiversity–productivity relationship
In non-Cd-polluted microcosms, there were no significantly positive (P >0·05) relationships between algal species richness and community biomass (see NP in Fig. 1 and Table 1). In contrast, significantly positive (P <0·05) biodiversity–productivity relationships were consistently found in the microcosms that were moderately or severely Cd-polluted (see MP and SP in Fig. 1 and Table 1). In agreement with the above results, most net biodiversity effects recorded in non-Cd-polluted polycultures on different sampling dates were not significantly different from zero (P >0·05) and were seldom affected by changes in species richness (see NP in Fig. 2), while those of Cd-polluted polycultures were significantly higher than zero (P <0·05) and tended to increase with species richness (see MP and SP in Fig. 2).
Table 1. Statistical results from linear regression analysis of the diversity–productivity relationship
No Cd pollution (NP)
Moderate Cd pollution (MP)
Severe Cd pollution (SP)
Roles of complementarity and selection effects in contributing to biodiversity effects
Partitioning of the net biodiversity effect, we found that the complementarity effects observed in Cd-polluted polycultures were always significantly positive (P <0·05) and showed a slight tendency to increase with time and Cd pollution level (see MP and SP in Fig. 3), while no clear trend was found for the complementarity effects of non-Cd-polluted polycultures (see NP in Fig. 3). On the other hand, we revealed that most of the selection effects recorded in non-Cd-polluted and severely Cd-polluted polycultures were significantly negative (P <0·05) or close to zero (see NP and SP in Fig. 4), whereas those of moderately Cd-polluted polycultures were slightly or significantly >0 (P <0·05) (see MP in Fig. 4).
Performances of individual species in 8-species polycultures in the seventh week
In non-Cd-polluted polycultures, Af, Cp, Sq and Sc significantly (P <0·05) underyielded, whereas Sd and So greatly (P <0·05) overyielded; Sd, Ce and So were the most dominant species (Fig. 5a). In moderately Cd-polluted polycultures, the performances of individual species were similar to those of non-Cd-polluted polycultures (Fig. 5b). Remarkably, the dominant species Cm and Sd produced greater biomasses in these polycultures than in non-Cd-polluted ones (Fig. 5a vs. 5b). In severely Cd-polluted polycultures, Cm (Cd-tolerant), and Ce (Cd-sensitive) were the most dominant species, and significantly (P <0·05) overyielded (Fig. 5c). The biomasses of Cm, Cr and Ce were higher in severely Cd-polycultures than in non-Cd-polluted and moderately Cd-polluted ones (Fig. 5). In addition, the results from the other sampling dates (first or fourth weeks) also showed that some Cd-sensitive algal species significantly (P <0·05) overyielded and were able to become dominant species in Cd-polluted polycultures (see Figs S1 and S2, Supporting information).
Earlier studies have shown that biodiversity effects on productivity can strengthen with time (Tilman et al. 2001; van Ruijven & Berendse 2005; Cardinale et al. 2007; Stachowicz et al. 2008), although various biodiversity–productivity relationships were found (Cardinale et al. 2006, 2007). In this study, the population doubling times for monocultures of focal species were <1·5 days (10–32 h, detailed data not shown), which suggested that the 7-week time span of our experiment was long enough for most focal species to experience more than 30 generations. We therefore had expected a greater probability that significantly positive biodiversity–productivity relationships would be observed in this study, as compared to previous experiments that lasted for less than three generations of the study organisms (Cardinale et al. 2007). Surprisingly, in non-Cd-polluted microcosms, no significantly positive (P >0·05) relationships between algal species richness and productivity were found (see NP in Fig. 1 and Table 1). Furthermore, this pattern did not change throughout the duration of our experiment (first vs. fourth vs. seventh week). There were at least three possible explanations for these results. One was that our species pool (10 members) was not large enough. However, the available evidence strongly argued against such an explanation. For example, the results of previous studies focusing on algal diversity have shown that a species pool containing no more than six members was sufficient to attain positive diversity–productivity relationships in non-polluted ecosystems (Bruno et al. 2005; Zhang & Zhang 2006). The second explanation was that algal species in this system all had highly overlapping niches. On such a neutral view (Hubbell 2001), species performances in the polycultures would be similar to that expected from monocultures. However, our results argued against neutrality (Fig. 5a). The third explanation lay in a neutralization of positive and negative interactions among species, i.e. the contrasting mechanistic effects of biodiversity neutralized each other (Hooper & Dukes 2004; Bruno et al. 2005), here resulting in weak effects of biodiversity on productivity. This suspicion was confirmed by comparisons of the complementarity effects and the selection effects of biodiversity in non-Cd-polluted polycultures. For example, although a significantly positive (P <0·05) complementarity effect did occur in the non-Cd-polluted 8-species polycultures in the seventh week (Fig. 3c), it was largely offset by a significantly negative (P <0·05) selection effect (Fig. 4c) that was probably caused by the relatively low yields (low dominances) of the highly productive algal species (such as Af, Cp and Sq) in the polycultures (Fig. 5a; Loreau & Hector 2001), resulting in a weak net biodiversity effect (Fig. 2c).
In contrast, we found that the productivities of Cd-polluted algal communities significantly positively increased (P <0·05) with increasing species richness (Fig. 1 and Table 1). In theory, the observed positive effects of biodiversity on productivity could be driven by either the complementarity effect or the so-called ‘sampling effect’ (positive selection effect), solely or in combination (Loreau & Hector 2001). It was, however, necessary to test the relative importance of the two components of the biodiversity effect in promoting productivities (Cardinale et al. 2006, 2007), given that they have different implications for practice (Loreau & Hector 2001). Interestingly, the complementarity effects recorded in both moderately and severely Cd-polluted polycultures were always significantly positive (P <0·05) and tended to increase with time (see MP and SP in Fig. 3). These results supported the notion that the strength of complementarity effect can increase with time (Tilman et al. 2001; van Ruijven & Berendse 2005; Cardinale et al. 2007; Stachowicz et al. 2008). On the other hand, there were fundamental differences between moderately and severely Cd-polluted polycultures in the selection effect (Fig. 4). Except for those weak values, the selection effects found in moderately Cd-polluted polycultures were significantly positive (P <0·05). This result could be partly explained by the relatively high yields (high dominances) of highly productive species in these polycultures (Loreau & Hector 2001) and suggested that ‘sampling effects’ indeed contributed to some positive net biodiversity effects occurring in moderately Cd-polluted polycultures (Fig. 2). However, it is worth noting that the contributions of these ‘sampling effects’ were always exceeded by their corresponding complementarity effects (grey bars in Fig. 3 vs. those in Fig. 4). When the significantly negative (P <0·05) selection effects occurring in severely Cd-polluted polycultures were taken into account, they possibly resulted from the relative high yields of low-biomass species in the polycultures (Loreau & Hector 2001; Bruno et al. 2005). For example, the significantly negative selection effect observed in severely Cd-polluted 8-species polycultures in the seventh week (Fig. 4c) could be partly ascribed to the high dominance (high productivity) of Ce in these polycultures (Fig. 5c). Nonetheless, the negative selection effects of severely Cd-polluted polycultures (SP in Fig. 4) were often overcome by corresponding positive complementarity effects (SP in Fig. 3), leading to significantly positive (P <0·05) net biodiversity effects (SP in Fig. 2).
Despite its ability to partition complementarity and selection effects, the additive partitioning method is unable to distinguish niche complementarity from direct interspecific facilitation (Loreau & Hector 2001). This raised another important question: which plays a greater role in driving the complementarity effects in Cd-polluted polycultures, niche complementarity or facilitation? By comparing the productivities of Cd-polluted polycultures with those of their respective monocultures (Fig. 1), we found that some of the polycultures outperformed the most productive monocultures (i.e. transgressive overyielding; especially in the first week, Fig. 1a), which were strong signals of facilitation (Hooper & Dukes 2004). Furthermore, there were several lines of evidence indicating that facilitation may be important. Firstly, we found that some Cd-sensitive algal species in Cd-polluted polycultures not only performed better than expected from their respective monocultures but also attained equal (such as So in moderately Cd-polluted 8-species polycultures in the seventh week, Fig. 5b) or greater yields (such as Ce in severely Cd-polluted 8-species polycultures in the seventh week, Fig. 5c) as compared with those of their analogues in non-Cd-polluted polycultures. These results were largely inconsistent with those predicted from niche complementarity (Mulder, Uliassi & Doak 2001) but could well be explained by interspecific facilitation. It has been widely reported that some heavy-metal-tolerant algal species (such as Stigeoclonium tenue and Cyanidium caldarium) can bind more metals on cell walls or extracellular exudates and/or trap more metals in intracellular structures in response to increased concentrations of external heavy metals (Pawlik-Skowrońska 2003; Nagasaka et al. 2006), although little information on the mechanisms underlying the tolerances of Cd-tolerant species in this system is available. On this basis, the equal or greater yields of the Cd-sensitive algal species in Cd-polluted polycultures as compared with those in non-Cd-polluted ones could be assumed to be partly driven by the putative functions of Cd-tolerant algal species in sequestering Cd from the environment. Furthermore, a growing body of literature has shown that the growth of some algal species in the presence of heavy metals can be promoted by exudates (that are rich in thiolic compounds) from other algal species (see Vasconcelos & Leal 2008 and the references cited therein). We therefore suspected that such a type of allelopathic effect of exudates also occurred in Cd-polluted polycultures, which could explain why some Cd-sensitive algal species performed better in Cd-polluted polycultures than in non-Cd-polluted ones (Fig. 5). Likewise, the greater yields of some Cd-tolerant algal species in Cd-polluted polycultures than in non-Cd-polluted polycultures (for example, Cm in severely Cd-polluted polycultures in the seventh week, Fig. 5c) could to some extent also be explained by the tentative facilitation among algal species. Secondly, under niche complementarity, it was likely that the dominant species in Cd-polluted polycultures were Cd-tolerant species. However, we found that some Cd-sensitive species could dominate Cd-polluted polycultures (Figs 5; Figs S1 and S2, Supporting information). Finally, niche complementarity did not explain why the average yields of Cd-polluted 8-species polycultures were approximately twofold higher than those of non-Cd-polluted 8-species ones (Fig. 1), given that the presence of Cd was unlikely to increase the total amount of available nutrients in the microcosms. Instead, the greater yields could be derived from some tentative allelopathic effects of exudates as mentioned above. Collectively, these results were consistent with the idea that interspecific facilitation is important under stressful conditions (Mulder, Uliassi & Doak 2001; Callaway et al. 2002), although more studies are needed to clarify the physiological mechanisms behind the facilitative effects among algal species in Cd-polluted polycultures.
In conclusion, our results suggest that Cd pollution can trigger a positive biodiversity–productivity relationship. This study indicates that some species that seem to be redundant in non-Cd-polluted ecosystems can become complementary or facilitative of other species in Cd-polluted ones. On this basis, we therefore argue that conservation of biodiversity in all environments may help ecosystems to withstand increasing environmental stresses such as Cd pollution. On the other hand, at most polluted sites some pollution-sensitive species are likely to be replaced by pollution-tolerant ones, leading to decreases in biodiversity. In this study, it was also observed that some Cd-sensitive algal species tend to display reductions in their relative abundances (measured as yields) in Cd-polluted polycultures (Fig. 5b, c). More importantly, we found that some Cd-sensitive algal species were able to dominate Cd-polluted polycultures, an intriguing phenomenon that can largely be explained by direct interspecific facilitation but which has been poorly recognized in previous studies. This finding highlights that (i) some toxin-sensitive species defined by the performances of their monocultures in tolerating the toxin (a widely used method for assessing the toxin tolerances of species, see Baker 1987) do not necessarily show sensitivity to the toxin when growing in polycultures that contain toxin-tolerant species; and that (ii) the potential important roles of some toxin-sensitive species in maintaining and improving the functions (such as productivity) of toxin-polluted ecosystems may have been so far underestimated. We therefore propose that attempts to manage toxin-polluted ecosystems for conservation, restoration or production must consider not only the roles of toxin-tolerant species but also the contributions of toxin-sensitive species. This goes against conventional wisdom that only toxin-tolerant species can be used for the management of toxin-polluted ecosystems (Dobson, Bradshaw & Baker 1997; Mendez & Maier 2008). Under this notion, for example, one possible approach to maintain or improve the algal productivities in an aquatic ecosystem experiencing environmental stress is to construct suitable diverse algal assemblages that contain both stress-tolerant and stress-sensitive algal species. Nonetheless, as pointed out in recent literature (Weis et al. 2007; Weis, Madrigal & Cardinale 2008), the laboratory microcosm of algae used in this study is an oversimplification of the complexity of natural ecosystems, although it is useful for maximizing our understanding of the mechanisms underlying changes in the biodiversity–productivity relationship along a gradient of Cd pollution. Therefore, future studies should be expanded to include realistic ecosystems under different levels of Cd pollution or other abiotic stresses, to determine the extent to which the results of this study and their implications can be extrapolated to realistic ecosystems.
We thank Prof. Alan JM Baker, Prof. Yong-Guan Zhu, Dr. Yong-Fan Wang, the editor and two anonymous referees for their insightful and detailed suggestions that helped to improve the manuscript. Funding was provided by the National Natural Science Foundation of China (No. 30970548).