**3. BEF experiments on benthic communities**

Benthos includes all organisms that live in or on water-covered substratum. These can be coastal and intertidal areas as well as limnic systems, including different types of substrates such as seagrass meadows, rocks, sand or mud. Sandy beaches, mudflats and other ecosystems found in intertidal zones act as buffer zones between the terrestrial and marine biomes, hosting a high diversity of species, which play a signicant role in the global storage and cycling of nutrients (Covich et al., 2004; Snelgrove et al., 1997). In addition, some coastal ecosystems such as seagrass meadows provide a rich source of primary productivity, as well as habitats for numerous fishes and invertebrates (Hughes & Stachowicz, 2004), and are among the most 'valuable' ecosystems on earth (Constanza et al., 1997).

Most of the benthic fauna are small organisms indiscernible to the naked eye, but they play a fundamental ecological role. For example free-living marine nematodes decompose organic matter, and facilitate the recycling of nutrients (Heip et al., 1985). Benthic diatoms not only constitute an essential part of the fundament of the benthic food chain, but in intertidal sediments their excretions stabilize the substratum by effectively binding sediment particles together (Stal, 2010). The communities of these buffer zones are severely affected by anthropogenic impacts from both the terrestrial and the marine side. The combination of those impacts with increasing surface water temperature and long exposure to high air temperature during spring tides may exceed the tolerance of some intertidal organisms. Because of the serious consequences this may have for the biological and the physical integrity of the ecosystem, but also because of their accessibility, relatively easy handling, and because of the typically high local diversity, benthic communities have become important for BEF studies.

#### **3.1 Effects of genotypic diversity on benthic ecosystem functioning**

There are very few studies about the effects of genotypic diversity on benthic ecosystem functioning. The most common approach has concentrated on plants or algae that reproduce clonally. In these experiments, genetic diversity refers to genotypic richness and usually monocultures of single genotypes are compared to mixtures of different genotypes (Bell, 1991; Hughes & Stachowicz, 2004; Reusch et al., 2005). In this context, the cosmopolitan eelgrass *Zostera marina* has received special attention. In extensive field experiments, the different genotypes were planted singly and in combination in a random block design after removing pre-existing eelgrass (Hughes & Stachowicz, 2004; Reusch et al., 2005). Experiments conducted on this system revealed a positive relationship between genotypic diversity and ecosystem functioning. *Zostera marina* of higher genotypic diversity recovered more rapidly from natural disturbances such as grazing geese or algal blooms (Hughes & Stachowicz, 2004, 2011). Genotypic diversity was also crucial to recovery (resilience) from experimentally implemented disturbance (i.e., biomass removal; Hughes & Stachowicz, 2011). Also biomass production, plant density and faunal abundance were enhanced (Reusch et al., 2005), especially in winter, when eelgrass experienced stress from

with various proxies for ecosystem functioning across a range of deep-sea regions suggests that species diversity profoundly affects marine benthic ecosystem functioning (Danovaro et al., 2008). Therefore, mesocosm studies with benthic communities and systems provide

Benthos includes all organisms that live in or on water-covered substratum. These can be coastal and intertidal areas as well as limnic systems, including different types of substrates such as seagrass meadows, rocks, sand or mud. Sandy beaches, mudflats and other ecosystems found in intertidal zones act as buffer zones between the terrestrial and marine biomes, hosting a high diversity of species, which play a signicant role in the global storage and cycling of nutrients (Covich et al., 2004; Snelgrove et al., 1997). In addition, some coastal ecosystems such as seagrass meadows provide a rich source of primary productivity, as well as habitats for numerous fishes and invertebrates (Hughes & Stachowicz, 2004), and are

Most of the benthic fauna are small organisms indiscernible to the naked eye, but they play a fundamental ecological role. For example free-living marine nematodes decompose organic matter, and facilitate the recycling of nutrients (Heip et al., 1985). Benthic diatoms not only constitute an essential part of the fundament of the benthic food chain, but in intertidal sediments their excretions stabilize the substratum by effectively binding sediment particles together (Stal, 2010). The communities of these buffer zones are severely affected by anthropogenic impacts from both the terrestrial and the marine side. The combination of those impacts with increasing surface water temperature and long exposure to high air temperature during spring tides may exceed the tolerance of some intertidal organisms. Because of the serious consequences this may have for the biological and the physical integrity of the ecosystem, but also because of their accessibility, relatively easy handling, and because of the typically high local diversity, benthic communities have

There are very few studies about the effects of genotypic diversity on benthic ecosystem functioning. The most common approach has concentrated on plants or algae that reproduce clonally. In these experiments, genetic diversity refers to genotypic richness and usually monocultures of single genotypes are compared to mixtures of different genotypes (Bell, 1991; Hughes & Stachowicz, 2004; Reusch et al., 2005). In this context, the cosmopolitan eelgrass *Zostera marina* has received special attention. In extensive field experiments, the different genotypes were planted singly and in combination in a random block design after removing pre-existing eelgrass (Hughes & Stachowicz, 2004; Reusch et al., 2005). Experiments conducted on this system revealed a positive relationship between genotypic diversity and ecosystem functioning. *Zostera marina* of higher genotypic diversity recovered more rapidly from natural disturbances such as grazing geese or algal blooms (Hughes & Stachowicz, 2004, 2011). Genotypic diversity was also crucial to recovery (resilience) from experimentally implemented disturbance (i.e., biomass removal; Hughes & Stachowicz, 2011). Also biomass production, plant density and faunal abundance were enhanced (Reusch et al., 2005), especially in winter, when eelgrass experienced stress from

useful options to take BEF studies to the next level.

**3. BEF experiments on benthic communities** 

become important for BEF studies.

among the most 'valuable' ecosystems on earth (Constanza et al., 1997).

**3.1 Effects of genotypic diversity on benthic ecosystem functioning** 

abiotic and biotic factors (Hughes & Stachowicz, 2009). Similar results were obtained when genotypes of the green alga *Chlamydomonas reinhardtii* were set up in culture tubes in monocultures and all possible pair wise combinations: the mixtures were consistently more productive and less variable in their productivity under simulation of different environmental conditions (Bell, 1991).

To our knowledge, there are very few experimental studies on benthic animals investigating the role of genetic diversity on ecosystem function, two of which assessed the effects on intertidal invertebrates. Within-species diversity was manipulated in the barnacle *Balanus improvisus* by altering the number of parental broods (Gamfeldt et al., 2005). In a similar way, the number of maternal families was controlled in the amphipod *Gammarus duebeni* from intertidal rock pools (Gamfeldt & Kallstrom, 2007). Both experiments found positive effects of increased genetic diversity: larval settlement of *B. improvisus* increased and the more diverse amphipod assemblages exhibited more predictable (but not higher on average) survival when exposed to multiple stressors (Gamfeldt & Kallstrom, 2007). Apart from that, studies relating genetic diversity to some (though particular) measure of ecosystem functioning have mainly focused on tolerance to toxic substances, indicating that genetically diverse populations resist to or recover better from contamination (Pease et al., 2010; Phillips & Hickey, 2010).

The few studies relating genetic (genotypic) diversity to ecosystem functioning show that genetic diversity can be important for the stability and productivity of coastal ecosystems. Generally, increased genetic diversity enhances the continuity and reliability of ecosystems. The advantage of clonally reproducing organisms such as plants or algae is that they provide the relatively easy and thus straightforward possibility of setting up experimental groups of contrasting genotypic diversity. However, this method is not applicable to sexually reproducing organisms. For those, it is necessary to produce different levels of genetic diversity by controlled mating prior to the experiment. For example Gamfeldt & Kallstrom (2007) sampled natural populations of *G. duebeni* and then separated pairs of mating individuals to form different "families". With these families, they then formed inbred strains (lowest genetic diversity level) and strains of two to four different families (increasing diversity). This procedure provides reliable levels of contrasting genetic diversities, but it is time consuming and requires considerable effort and material. Nevertheless, it is probably the only reliable method of obtaining contrasting diversity levels for sexually reproducing organisms if natural inbred (representing low diversity) and outbred populations (high diversity) are unavailable. This is probably the main reason, why studies investigating the role of genetic diversity on the system functioning are still scarce.

#### **3.2 Effects of species diversity on benthic ecosystem functioning 3.2.1 Macrophyte vegetations**

Seagrass *Zostera marina*, pondweed or cordgrass *Spartina* ecosystems have been used widely to investigate the effect of species richness, species composition as well as food chain length and trophic cascades on ecosystem processes. For example all possible species combinations of four functionally and morphologically different submerged aquatic macrophytes (pondweed) were planted in artificial wading pools (Engelhardt & Ritchie, 2001). Although local wetlands are usually dominated by single vascular species, there was a clear effect of increased plant diversity leading to higher plant and algal biomass as well as lower loss of phosphorus (Engelhardt & Ritchie, 2001).

Integrating Different Organizational Levels in Benthic

(Emmerson & Raffaelli, 2000).

ecosystem processes in natural ecosystems (Stachowicz et al., 2008).

individual, different species at each site (sampling effect).

**3.2.2 Benthic invertebrate macrofauna from intertidal soft bottom areas** 

Biodiversity – Ecosystem Functioning (BEF) Studies 97

they are likely to be unable to capture seasonal environmental heterogeneity and population responses. Therefore, they are likely to underestimate the influence of diversity on

In a novel experimental approach designed to separate the effects of species richness and species identity, the dominant infaunal macrofaunal species of a mudflat community were collected in two distinct biogeographical areas: three species (one polychaete, an amphipod and a gastropod) in Scotland and five species (three gastropods, one ghost shrimp and one cockle) in South Australia. Null-models based on the response of monocultures (i.e., single species treatments) were constructed and then compared to the experimental results. It was found that individual species contributed idiosyncratically to the measured ecosystem process (ammonia release), and that the effects were biomass (density)-dependent

In a subsequent extended experiment, replicate species pools were sampled from three regions (Scotland, southwest Sweden and south Australia) and placed in mesocosms (Emmerson et al., 2001). The selected species (ranging from one to four species) were dominating the biomass at each of the three sites. Additionally, complete natural communities were sampled at three different sites in Scotland. These contained more species and included the small, delicate species that could not be manipulated directly. Thus two experiments were carried out: one with artificially assembled communities on one hand, and one with natural communities on the other. In the "artificial" assemblages, NH4-N production became more predictable (i.e., less variable) with increased diversity. In the experiment where natural communities were used, there was a significant positive effect of species richness on nutrient flux; however, the process was driven disproportionately by

The first field experiment on a benthic system was conducted on a mudflat in Scotland. Defaunated sediment was enclosed using cages with different mesh sizes. The cages were then allowed to be colonized by organisms from the surrounding sediments. Through this, groups of low, medium and high species richness as well as low, medium and high biomass of macrofaunal invertebrates were established (Bolam et al., 2002). In each experimental cage the following variables were measured as proxies of ecosystem functioning: phosphate and ammonium fluxes, community respiration, sediment shear strength, water content, water/silt content, organic content, redox potential, nitrate and nitrite. All but one variable remained unaffected by macrofaunal diversity or biomass; only oxygen consumption was

positively related to both, but in fact it was driven by the largest species in the study.

result in the additive effect one would have expected from single-species treatments.

Waldbusser et al. (2004) examined the relationship between functional diversity and ecosystem processes by manipulating three functionally diverse polychaete worms from nearshore sediments. They established microcosms of one and three species and kept them for 4 months. They found that species as well as functional diversity had an effect on the measured proxies of ecosystem processes (i.e., oxygen and phosphate fluxes, profiles of oxygen and pH). Although it is difficult to separate species from functional diversity when each species represents a different function, it was clear that single species had disproportionate effects on selected variables and that the three-species assemblage did not

In several outdoor experiments, species diversity of different trophic levels was manipulated and the role of grazer diversity on plant biomass as well as grazer secondary production was investigated (Duffy et al., 2001). Grazers play an important role in the health of seagrass ecosystems because they feed preferentially on epiphytic algae and thus prevent eelgrass from becoming overgrown by epiphytes. In the experiment, the three grazer crustacean species (two isopods and one amphipod), which dominate the seagrass epifauna in the study area, were assembled in all possible species combinations (1 to 3 species and a grazer-free control), assessing the accumulation of epiphytic algae and eelgrass biomass as well as grazer secondary production. Results showed that all three grazer species differed substantially in their impact on ecosystem processes. For all processes measured, species composition was more important for eelgrass and grazer biomass accumulation than species richness. The study emphasized that species of one functional group need not necessarily be functionally redundant. In a follow-up experiment, six common grazer species (one gastropod, three amphipods and two isopods) were combined to further investigate the role of species diversity and species composition in relation to eelgrass ecosystem functioning (Duffy et al., 2003). Again, species were assembled in random combinations (one, three and six-species treatments) in outdoor mesocosms. The water supply was filtered, but allowed microscopic propagules of algae and sessile invertebrates to recruit. In addition to eelgrass, algal and grazer biomass, also measured was organic carbon in the sediment. Contrary to the previous study, increased biodiversity enhanced secondary production. Yet, increasing grazer diversity reduced total community diversity and allowed a sessile, grazing-resistant invertebrate to become dominant, presumably due to intraguild competition among grazer species.

In the previous experiment, interaction effects of species diversity and species composition were measured as deviations from expected additive effects based on single-species effects. In another attempt to separate the two effects, a mesocosm experiment was conducted where species richness and species composition were manipulated independently across multiple trophic levels including macrophytes, benthic grazers and carnivorous predators (Downing & Leibold, 2002). One, three or five species per functional group were assembled and within each level of richness, seven unique species compositions were nested and replicated. This being an extensive and statistically robust method, the results showed that species diversity as well as species composition had important effects on ecosystem functioning. Ecosystem productivity was highest at the highest diversity level, which implies that species within a functional group are not redundant, a result similar to that of Duffy et al. (2001). Not all measures of ecosystem functioning reacted similarly to species diversity or composition: phytoplankton biomass increased while zooplankton biomass decreased with total species richness. Moreover, there was a strong effect of species altering ecosystem processes indirectly by altering abundances of other species, implying that trophic interactions are highly important for the functioning of communities and ecosystems.

Experiments on macrophyte vegetations have successfully shown the effects of species diversity and identity on ecosystem processes. In general (but not always), a positive effect of biodiversity was found on ecosystem processes. However, these experiments were all conducted over a relatively short period of time. In contrast, the effects of seaweed species richness on biomass accumulation were stronger in long-term (3 years) than in short-term (2 months) field experiments (Stachovicz et al., 2008). Although short-term experiments are valuable in estimating the effects of species richness and identity on ecosystem processes,

In several outdoor experiments, species diversity of different trophic levels was manipulated and the role of grazer diversity on plant biomass as well as grazer secondary production was investigated (Duffy et al., 2001). Grazers play an important role in the health of seagrass ecosystems because they feed preferentially on epiphytic algae and thus prevent eelgrass from becoming overgrown by epiphytes. In the experiment, the three grazer crustacean species (two isopods and one amphipod), which dominate the seagrass epifauna in the study area, were assembled in all possible species combinations (1 to 3 species and a grazer-free control), assessing the accumulation of epiphytic algae and eelgrass biomass as well as grazer secondary production. Results showed that all three grazer species differed substantially in their impact on ecosystem processes. For all processes measured, species composition was more important for eelgrass and grazer biomass accumulation than species richness. The study emphasized that species of one functional group need not necessarily be functionally redundant. In a follow-up experiment, six common grazer species (one gastropod, three amphipods and two isopods) were combined to further investigate the role of species diversity and species composition in relation to eelgrass ecosystem functioning (Duffy et al., 2003). Again, species were assembled in random combinations (one, three and six-species treatments) in outdoor mesocosms. The water supply was filtered, but allowed microscopic propagules of algae and sessile invertebrates to recruit. In addition to eelgrass, algal and grazer biomass, also measured was organic carbon in the sediment. Contrary to the previous study, increased biodiversity enhanced secondary production. Yet, increasing grazer diversity reduced total community diversity and allowed a sessile, grazing-resistant invertebrate to become dominant, presumably due to intraguild

In the previous experiment, interaction effects of species diversity and species composition were measured as deviations from expected additive effects based on single-species effects. In another attempt to separate the two effects, a mesocosm experiment was conducted where species richness and species composition were manipulated independently across multiple trophic levels including macrophytes, benthic grazers and carnivorous predators (Downing & Leibold, 2002). One, three or five species per functional group were assembled and within each level of richness, seven unique species compositions were nested and replicated. This being an extensive and statistically robust method, the results showed that species diversity as well as species composition had important effects on ecosystem functioning. Ecosystem productivity was highest at the highest diversity level, which implies that species within a functional group are not redundant, a result similar to that of Duffy et al. (2001). Not all measures of ecosystem functioning reacted similarly to species diversity or composition: phytoplankton biomass increased while zooplankton biomass decreased with total species richness. Moreover, there was a strong effect of species altering ecosystem processes indirectly by altering abundances of other species, implying that trophic interactions are highly important for the functioning of

Experiments on macrophyte vegetations have successfully shown the effects of species diversity and identity on ecosystem processes. In general (but not always), a positive effect of biodiversity was found on ecosystem processes. However, these experiments were all conducted over a relatively short period of time. In contrast, the effects of seaweed species richness on biomass accumulation were stronger in long-term (3 years) than in short-term (2 months) field experiments (Stachovicz et al., 2008). Although short-term experiments are valuable in estimating the effects of species richness and identity on ecosystem processes,

competition among grazer species.

communities and ecosystems.

they are likely to be unable to capture seasonal environmental heterogeneity and population responses. Therefore, they are likely to underestimate the influence of diversity on ecosystem processes in natural ecosystems (Stachowicz et al., 2008).

#### **3.2.2 Benthic invertebrate macrofauna from intertidal soft bottom areas**

In a novel experimental approach designed to separate the effects of species richness and species identity, the dominant infaunal macrofaunal species of a mudflat community were collected in two distinct biogeographical areas: three species (one polychaete, an amphipod and a gastropod) in Scotland and five species (three gastropods, one ghost shrimp and one cockle) in South Australia. Null-models based on the response of monocultures (i.e., single species treatments) were constructed and then compared to the experimental results. It was found that individual species contributed idiosyncratically to the measured ecosystem process (ammonia release), and that the effects were biomass (density)-dependent (Emmerson & Raffaelli, 2000).

In a subsequent extended experiment, replicate species pools were sampled from three regions (Scotland, southwest Sweden and south Australia) and placed in mesocosms (Emmerson et al., 2001). The selected species (ranging from one to four species) were dominating the biomass at each of the three sites. Additionally, complete natural communities were sampled at three different sites in Scotland. These contained more species and included the small, delicate species that could not be manipulated directly. Thus two experiments were carried out: one with artificially assembled communities on one hand, and one with natural communities on the other. In the "artificial" assemblages, NH4-N production became more predictable (i.e., less variable) with increased diversity. In the experiment where natural communities were used, there was a significant positive effect of species richness on nutrient flux; however, the process was driven disproportionately by individual, different species at each site (sampling effect).

The first field experiment on a benthic system was conducted on a mudflat in Scotland. Defaunated sediment was enclosed using cages with different mesh sizes. The cages were then allowed to be colonized by organisms from the surrounding sediments. Through this, groups of low, medium and high species richness as well as low, medium and high biomass of macrofaunal invertebrates were established (Bolam et al., 2002). In each experimental cage the following variables were measured as proxies of ecosystem functioning: phosphate and ammonium fluxes, community respiration, sediment shear strength, water content, water/silt content, organic content, redox potential, nitrate and nitrite. All but one variable remained unaffected by macrofaunal diversity or biomass; only oxygen consumption was positively related to both, but in fact it was driven by the largest species in the study.

Waldbusser et al. (2004) examined the relationship between functional diversity and ecosystem processes by manipulating three functionally diverse polychaete worms from nearshore sediments. They established microcosms of one and three species and kept them for 4 months. They found that species as well as functional diversity had an effect on the measured proxies of ecosystem processes (i.e., oxygen and phosphate fluxes, profiles of oxygen and pH). Although it is difficult to separate species from functional diversity when each species represents a different function, it was clear that single species had disproportionate effects on selected variables and that the three-species assemblage did not result in the additive effect one would have expected from single-species treatments.

Integrating Different Organizational Levels in Benthic

competitive behaviour among species occur.

reflect what is actually happening in nature.

**4.1 Trophic cascades** 

**4. Two "special topics" in benthic BEF studies** 

Malmqvist, 2000).

Biodiversity – Ecosystem Functioning (BEF) Studies 99

Other stream invertebrate larvae feed on leaf litter as "shredders" (Jonsson & Malmqvist, 2000). In a laboratory experiment, three species of stoneflies (*Plecoptera*), belonging to this leaf-eating feeding guild, were placed in one-, two- and three-species treatments to investigate the effect of species richness on leaf mass loss. Ecosystem process rates (leaf mass loss) increased significantly with species richness, but were not dependent on species identity, suggesting that the different species were not functionally redundant (Jonsson &

An extension of the previous experiment investigated the effect of intraguild species diversity for other feeding guilds (Jonsson & Malmqvist, 2003). In several controlled laboratory experiments, multiple species of the following functional groups were collected: filter feeders (six blackfly larvae *[Diptera]*), grazers (two species of mayfly larvae *[Ephemeroptera]* and one snail), and predators (two stonefly species [*Plecoptera*] and one caddisfly species [*Trichoptera*]). Each functional group was set up in single-, two- and three species combinations and provided with their respective food source: dry yeast for the filter feeders, algae for the grazers and blackfly larvae for the predators. There was a strong effect of species identity, and species combinations, whereas species richness *per se* did not affect

These four examples show that stream invertebrate larvae form a complex system consisting of different feeding guilds exhibiting many different functions. Apparently the extent to which species richness affects ecosystem functioning depends greatly on the identity and the function of the species present. In general, however, like marine soft-bottom invertebrates, lotic larval communities seem to contain species exhibiting disproportionate effects on the system. The effects of species loss may be predictable when their functional role is accurately ascertained, but can become idiosyncratic when interactions such as

As mentioned in the former sections, most BEF studies assess ecosystem functioning in terms of biomass production, effects on community composition or element cycling. However, there are other possible effects with important implications for the whole ecosystem. We choose two examples to illustrate this: trophic cascades and invasion success. We chose these two special topics because the former shows that slightly different levels of intraguild diversity can lead to completely opposed effects, whereas the latter illustrates that although experimental approaches meet hypotheses based on theory, they sometimes do not

Trophic species-level cascades occur when a change in predator abundance induces changes in the biomass of primary producers, due to a control of the abundance of grazers, thereby releasing lower trophic levels from grazing pressure (Polis et al., 2000). Predator diversity may reduce trophic cascades and may therefore be important for the population development of lower trophic levels and primary production. The importance of such intraguild diversity was studied by manipulating predator species richness (1, 2 or 3 species). The predator species pool consisted of several invertebrates such as hunting and web-building spiders preying on an arthropod (planthopper) assemblage inhabiting *Spartina* cordgrass (Finke & Denno, 2005). The occurrence of trophic cascades was

process rates in any of the three functional groups (Jonsson & Malmqvist, 2003).

In all the above-mentioned studies, a disproportionate effect of single species could be detected. Unlike the community members of macrophyte systems, benthic invertebrates from soft bottom intertidal areas seem to contribute individually to ecosystem functioning with their impact strongly depending on their functional role. Therefore, it can be essential to include such critical species in experiments assessing the effect of species richness on the functioning of the system. Since many experimental approaches include only the dominant species, conclusions drawn may not necessarily hold for the natural system, especially since some key species are not necessarily very abundant. Generally, natural benthic invertebrate assemblages from marine soft bottoms can be sampled easily, and as such provide a good system for BEF studies. Moreover, they can be established in micro- or mesocosms in the laboratory or in the field, and respond rapidly to experimental treatments. However, these organisms can reach very high densities, and especially the members of the meiofaunal component are often difficult to count and identify. Therefore, the choice of working with "artificial" (i.e., fewer species) or "natural" assemblages needs careful planning, which often means facing a trade-off between "natural" and "doable" with the available resources. The decision is most dependent on the scientific question, but also on the possibilities in terms of expertise and time.

#### **3.2.3 Stream invertebrate larvae**

Many insects develop as larvae on the substrate of streams or ponds, where they exhibit different feeding strategies and play different functional roles in the benthic system. As such, they provide an interesting system to test hypotheses regarding intraguild diversity. Caddisfly larvae, for example, construct silk nets in the pore spaces of the streambed, and passively feed on suspended particulate matter. Their structures generate topographical features, which influence patterns of water flow and therefore food availability (Cardinale et al., 2002). Often, several species with anatomically different feeding structures co-occur. This has led to the hypothesis that increased diversity leads to facilitation of food uptake. The hypothesis has been tested in stream mesocosms, where caddisfly larvae assemblages were established, either with a single species (18 larvae) or with 3 species (6 larvae per species; Cardinale et al., 2002). Results showed that higher diversity led to facilitative interactions and an increase in the uptake of organic matter. However, the species building the largest tubes had the strongest physical impact on streambed water flow.

A follow-up experiment tested whether the diversity-functioning relationship would be different under conditions of regular disturbance (Cardinale & Palmer, 2002). The hypothesis was that disturbance would induce mechanisms that would interfere with ecological processes. Disturbance was simulated by mechanically removing a given number of larvae in randomly selected pore spaces. Three ecological processes were measured at the end of the experiment: respiration of the benthic biofilm, primary productivity of benthic algae and the flux of particulate organic matter (POM) from the water column to the streambed. The disturbance treatment resulted in the suppression of a dominant taxon, which had a particularly low rate of nutrient excretion. This led to a negative correlation between primary production and species richness in the undisturbed streams, as this taxon was included in multi-species treatments. Concordantly, POM flux increased with species richness under disturbance conditions. Disturbance thus favoured the co-existence of competitively superior and inferior species, enhancing ecosystem processes (Cardinale & Palmer, 2002).

In all the above-mentioned studies, a disproportionate effect of single species could be detected. Unlike the community members of macrophyte systems, benthic invertebrates from soft bottom intertidal areas seem to contribute individually to ecosystem functioning with their impact strongly depending on their functional role. Therefore, it can be essential to include such critical species in experiments assessing the effect of species richness on the functioning of the system. Since many experimental approaches include only the dominant species, conclusions drawn may not necessarily hold for the natural system, especially since some key species are not necessarily very abundant. Generally, natural benthic invertebrate assemblages from marine soft bottoms can be sampled easily, and as such provide a good system for BEF studies. Moreover, they can be established in micro- or mesocosms in the laboratory or in the field, and respond rapidly to experimental treatments. However, these organisms can reach very high densities, and especially the members of the meiofaunal component are often difficult to count and identify. Therefore, the choice of working with "artificial" (i.e., fewer species) or "natural" assemblages needs careful planning, which often means facing a trade-off between "natural" and "doable" with the available resources. The decision is most dependent on the scientific question, but also on the possibilities in terms of

Many insects develop as larvae on the substrate of streams or ponds, where they exhibit different feeding strategies and play different functional roles in the benthic system. As such, they provide an interesting system to test hypotheses regarding intraguild diversity. Caddisfly larvae, for example, construct silk nets in the pore spaces of the streambed, and passively feed on suspended particulate matter. Their structures generate topographical features, which influence patterns of water flow and therefore food availability (Cardinale et al., 2002). Often, several species with anatomically different feeding structures co-occur. This has led to the hypothesis that increased diversity leads to facilitation of food uptake. The hypothesis has been tested in stream mesocosms, where caddisfly larvae assemblages were established, either with a single species (18 larvae) or with 3 species (6 larvae per species; Cardinale et al., 2002). Results showed that higher diversity led to facilitative interactions and an increase in the uptake of organic matter. However, the species building the largest

A follow-up experiment tested whether the diversity-functioning relationship would be different under conditions of regular disturbance (Cardinale & Palmer, 2002). The hypothesis was that disturbance would induce mechanisms that would interfere with ecological processes. Disturbance was simulated by mechanically removing a given number of larvae in randomly selected pore spaces. Three ecological processes were measured at the end of the experiment: respiration of the benthic biofilm, primary productivity of benthic algae and the flux of particulate organic matter (POM) from the water column to the streambed. The disturbance treatment resulted in the suppression of a dominant taxon, which had a particularly low rate of nutrient excretion. This led to a negative correlation between primary production and species richness in the undisturbed streams, as this taxon was included in multi-species treatments. Concordantly, POM flux increased with species richness under disturbance conditions. Disturbance thus favoured the co-existence of competitively superior and inferior species, enhancing ecosystem processes (Cardinale &

tubes had the strongest physical impact on streambed water flow.

expertise and time.

Palmer, 2002).

**3.2.3 Stream invertebrate larvae** 

Other stream invertebrate larvae feed on leaf litter as "shredders" (Jonsson & Malmqvist, 2000). In a laboratory experiment, three species of stoneflies (*Plecoptera*), belonging to this leaf-eating feeding guild, were placed in one-, two- and three-species treatments to investigate the effect of species richness on leaf mass loss. Ecosystem process rates (leaf mass loss) increased significantly with species richness, but were not dependent on species identity, suggesting that the different species were not functionally redundant (Jonsson & Malmqvist, 2000).

An extension of the previous experiment investigated the effect of intraguild species diversity for other feeding guilds (Jonsson & Malmqvist, 2003). In several controlled laboratory experiments, multiple species of the following functional groups were collected: filter feeders (six blackfly larvae *[Diptera]*), grazers (two species of mayfly larvae *[Ephemeroptera]* and one snail), and predators (two stonefly species [*Plecoptera*] and one caddisfly species [*Trichoptera*]). Each functional group was set up in single-, two- and three species combinations and provided with their respective food source: dry yeast for the filter feeders, algae for the grazers and blackfly larvae for the predators. There was a strong effect of species identity, and species combinations, whereas species richness *per se* did not affect process rates in any of the three functional groups (Jonsson & Malmqvist, 2003).

These four examples show that stream invertebrate larvae form a complex system consisting of different feeding guilds exhibiting many different functions. Apparently the extent to which species richness affects ecosystem functioning depends greatly on the identity and the function of the species present. In general, however, like marine soft-bottom invertebrates, lotic larval communities seem to contain species exhibiting disproportionate effects on the system. The effects of species loss may be predictable when their functional role is accurately ascertained, but can become idiosyncratic when interactions such as competitive behaviour among species occur.
