SEARCH

SEARCH BY CITATION

Keywords:

  • Biological pest control;
  • body size diversity;
  • Carabidae;
  • functional diversity;
  • niche complementarity;
  • redundancy;
  • sampling effect

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  1. Determining how multiple predators provide better prey suppression is a key step towards developing conservation biological control strategies. While numerous previous studies have demonstrated that diverse predator assemblages can be more effective in controlling pest populations, others have shown that it is the presence or absence of competitively superior species that is critical to pest biological control (i.e. selection effect).
  2. The present study investigated how increasing ground beetle body size diversity increases prey suppression. A mesocosm experiment was conducted to compare invertebrate prey suppression between nine created ground beetle assemblages. Size diversity of these assemblages was manipulated according to three diversity levels: low, medium, and high diversity.
  3. Partitioning of the diversity effects revealed that increasing the ground beetle size diversity had no effect on the strength of prey suppression. The absence of an effect of ground beetle size diversity may be because of the absence of resource partitioning among different-sized ground beetles. The amount and range of prey consumed increased with increasing ground beetle body size. Thus, prey suppression was strongly strengthened by the presence of large ground beetles in the assemblages.
  4. The present results suggest that for biological pest control, Agri-managers should emphasise practices that promote the presence of large carabids. This is not only because promoting the presence of large carabids could be at least as effective as conserving a diverse ground beetle community, but also because large ground beetles are more vulnerable to environmental disturbances and to predation than ground beetles of the other size classes.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

From a conservation biological control perspective, the attention of agroecologists has increasingly focused on the relationship between arthropod predator diversity and pest suppression. Numerous studies have provided consistent evidence that arthropod predators can effectively suppress populations of crop pests (Symondson et al., 2002; Snyder et al., 2005). However, the understanding of how and whether pest regulation is enhanced by changes in predator diversity remains limited (Straub et al., 2008).

Theory suggests that there are a variety of mechanisms by which changes in predator diversity could enhance prey suppression (Sih et al., 1998; Ives et al., 2005) and strengthen the top-down biological pest control. The sampling (Huston, 1997), or positive selection (Loreau, 2000) effect refers to the probability that a diverse predator community will include key species with unusually high consumption rates. Therefore, the performance of the predator community with regard to prey suppression may be driven primarily by whether these best-performing species are present. Changes in predator diversity can enhance prey suppression also through complementary resource use (Tilman et al., 1997; Loreau, 2000). Different predator species often consume different prey species (Duffy, 2002) or different life-history stages of a single prey species (Wilby et al., 2005). Such complementarity in resource use suggests that the effect of predators is additive if the prey mortality that results from the combined action of different species is equal to the summed mortality caused by each predator species on its own (Snyder et al., 2005).

In addition to its positive effect on prey suppression enhancement, changes in predator diversity sometimes leads to increasing intra-guild predation among predator species, which reduces their collective effect on prey suppression (Straub et al., 2008; Letourneau et al., 2009). The effects of changes in predator diversity on prey suppression could also be neutral. This occurs when multiple predators have ‘redundant’ or ‘compensatory’ effects, i.e. their combined effects are the average of the corresponding single-species effect (Sokol-Hessner & Schmitz, 2002; Schmitz, 2007). Thus, the removal of one species is compensated by increases in the prey consumption by another species (Navarrete & Menge, 1996; Otto et al., 2008). This effect is encompassed in the insurance hypothesis (Tilman, 1996), which states that redundancy stabilises the aggregate predation rates in the face of environmental changes (Wellnitz & Poff, 2001).

In spite of the growing number of studies investigating how changes in predator diversity could enhance prey suppression, little has been done to examine the effects of variation in predator traits (functional diversity) that mediate such mechanisms. Previous research on the effects of predator diversity on prey suppression have focused almost exclusively on predator species richness, whereas biodiversity-ecosystem function studies suggest that functional diversity rather than richness per se drives ecological processes, such as prey suppression (Schmitz, 2007, 2009; Bruno & Cardianale, 2008). Predator body size is considered to be a key functional trait determining the strength and type of species interactions (Delclos & Rudolf, 2011; Rudolf, 2012) and, therefore, how predator diversity affects prey suppression (Brose, 2010; Rudolf, 2012). This attribute influences the habitat choice, prey size, range of prey, and consumption rate of the predator (Cohen et al., 1993). Predator size also determines the individual body mass and the biomass of a population as a whole (Woodward et al., 2010). Predator size variation can enhance prey suppression through an increase in the average body size that increases the per capita consumption rates of predators (Rudolf, 2012). Mechanisms such as complementary resource use could apply similarly to the ecological effects of body size variation among predator species (Woodward & Hildrew, 2002; Rudolf, 2012). Indeed, increasing the size range of predators has the potential to increase resource partitioning in terms of prey size (Woodward & Hildrew, 2002) and, therefore, enhance prey suppression through niche complementarity (Rudolf, 2012). The presence of different-sized species within a predator population could, however, lead to increasing the negative interferences between the predators (e.g. intra-guild predation). These antagonistic interactions should decrease the strength of top-down prey control (Delclos & Rudolf, 2011; Rudolf, 2012).

Given their cosmopolitan distribution, their polyphagy, and their taxonomic and functional diversity, ground beetles are especially suitable organisms for both ecological and entomological studies. Ground beetles are an important component of the ground-dwelling fauna of most of the world's terrestrial environments. They are predators of many invertebrate pests in agricultural ecosystems, such as aphids (Schmidt et al., 2004) and slugs (Mair & Port, 2001; Oberholzer & Frank, 2003). Thus, they may be potential biological control agents. With more than 40 000 described species, ground beetles are likely to have greater levels of diversity in traits associated with prey capture and consumption. Although superficially most ground beetle species seem to have a similar body shape, there are species-specific differences and the morphological particularities of each species reflect the special demands of its niche (Loreau, 1983; Bauer et al., 1998; Kromp, 1999). Ground beetles show considerable variation in their size. This trait diversity may lead to complementary resource use.

The objective of the present study was to determine how changes in the body size diversity of ground beetles influence prey suppression. To accomplish this objective and using a functional group approach, we have created ground beetle assemblages according to three levels of size diversity: low, medium, and high. Three species groups of a different size (large, medium, and small) were used as units of diversity in creating assemblages. Comparing prey suppression between the created assemblages enabled us to test whether increasing the ground beetles' body size diversity enhances prey suppression through a complementary resource use or through a sampling effect of the best performing size group of species. In addition, we used the variation in the species composition of the created assemblages to test if there is (or not) an additive effect among species of the best performing size group.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study system

This study was conducted at the INRA's Mirecourt Experimental Station in the northeastern France (48°29′48″N, 6°12′13″E). The choice of ground beetle species used in the study was based on a species list established during previous sampling of the site (Mignolet et al., unpublished). From this list, we classified the most common species into three size groups: small species (less than 9 mm), medium species (9.1–13 mm), and large species (over 13.1 mm). Nine species (three from each group) were chosen to minimise the possibility that species identity effects would be responsible for the differences between the functional diversity levels. Small species selected were: Anchomenus dorsalis Pontoppidan (6.8 mm), Agonum muelleri Herbst (8.1 mm), and Brachinus sclopeta Fabricius (6.3 mm); medium ones were Poecilus cupreus Linnaeus (12.1 mm), Nebria salina Fairmaire & Laboulbene (11.9 mm), and Limodromus assimilis Paykull (11.0 mm); and large species were Carabus auratus Linnaeus (23.5 mm), Abax parallelepipedus Piller & Mitterpacher (18.6 mm), and Pterostichus melanarius Illiger (15.7 mm). Individuals of these selected species were collected using dry pitfall traps. They were also hand caught and collected from underneath pieces of wood and cement that had been laid out in the fields to serve as hiding places for these animals during inactive periods. Ground beetle species caught were kept separate in plastic boxes (28 × 14 × 8 cm) filled with 2 cm of damp potting soil. They were fed apple slices. New individuals of all of the nine species were continuously trapped and added to the colonies to replace the ones that died and to ensure that healthy insects were available for the feeding trials. All of the experimental individuals of the nine species were assumed to be of a typical size, with the average body sizes cited by Hurka (1996). To prevent repeated measure data, each individual was used only once. The ground beetles were starved for 24 h prior to the experiment in boxes containing moistened cotton.

To allow for resource complementarity in terms of size, a mixture of four different-sized invertebrate prey species were used. The prey species included two insects: the cherry-oat aphid Rhopalosiphum padi Linnaeus (Homoptera: Aphididae) and larvae of the yellow mealworm beetle Tenebrio molitor Linnaeus (Coleoptera: Tenebrionidae), and two gastropods: adults and eggs of the grey field slug Deroceras reticulatum Müller (Pulmonata: Limacidae) and eggs of the land snail Helix aspersa Müller (Pulmonata: Helicidae). The aphids were collected from a colony maintained in culture on barley plants in a growth chamber (25 °C and LD 14:10 h photoperiod). The larvae of the yellow mealworm beetle and the adult slugs were provided respectively, by Animal Farming and Arbiotech; two societies specialise in the breeding and commercialisation of insects and other animals in France. The mealworm beetle larvae were kept in groups in plastic boxes and fed bran until they were required for the feeding trials. Before use in the experiments, the slugs were placed in sealed plastic boxes for egg production. Each box contained moist soil, and holes were made in the lid to allow air circulation. Slugs were fed leaves of two clover species (Taraxacum officinale L. and Trifolium pratense L.). The boxes were placed in a rearing chamber at 20 °C (LD 12:12 h photoperiod) and were examined twice per week. Laid eggs were transferred to Petri dishes filled with moist soil and were placed in a cold room at 4 °C to delay hatching. The eggs of H. aspersa were provided by a snail farmer and stored at 4 °C until required for the experiment.

Ground beetles assemblages

Ground beetle assemblages were created using a functional group approach, where the three size groups of ground beetles were treated as units of functional diversity. Thus, from the pool of the nine selected species, we have generated all possible combinations in order to create assemblages with the same species richness (three species in each combination) but different levels of functional diversity: low (including mono-group assemblages = all of the ground beetles belong to only one size group), medium (including bi-group assemblages = ground beetles were from two size groups), and high (including tri-group assemblages = ground beetles were added from three size groups). From the 84 possible combinations obtained, we chose to test three assemblages for each level of diversity (nine in total ‘A1–A9’; Table 1), so that each selected species was present at least once in each level of diversity (Table 1). All of the assemblages received a total of nine ground beetles (three individuals of each species present in the assemblage). Keeping the total abundance of individuals constant isolates the effect of increasing the diversity from that of increasing the total predator abundance (Straub & Snyder, 2006).

Table 1. Species composition, size diversity, and abundance of the nine ground beetle assemblages created to test for ground beetles size diversity on prey suppression
   Beetles number   
AssemblagesSize diversitySpecies compositionSmallMediumLargeTotal
  1. Each assemblage was tested five times, and five ground beetle-free treatments were used as controls (n = 50).

  2. A.D., Anchomenus dorsalis; B.S., Brachinus sclopeta; A.M., Agonum muelleri; P.C., Poecilus cupreus; N.S., Nebria salina; L.A., Limodromus assimilis; P.M., Pterostichus melanarius; A.P., Abax parallelepipedus; C.A., Carabus auratus.

A0 (control)0No beetle0000
A11A.M. + A.D. + B.S.9009
A21P.C. + N.S+L.A.0909
A31P.M.+C.A.+A.P.0099
A42N.S.+L.A. + A.P.0639
A52C.A. + P.M. + A.M.3069
A62B.S.+A.D. + P.C.6309
A73B.S. + P.C. + P.M.3339
A83A.M. + L.A. + A.P.3339
A93A.D. + N.S. + C.A.3339

This substitutive experimental design (which holds the total abundance of predators constant across the levels of diversity) enabled us to separate the effect of functional identity from that of functional diversity and detect both transgressive and non-transgressive overyielding (Loreau, 1998; Bruno et al., 2006). Non-transgressive overyielding occurs when the performance of the highly diverse assemblages exceeds that of the average low-functional-diversity (monogroup) assemblages. Transgressive overyielding occurs when the performance of the highly diverse assemblages exceeds that of the best-performing monogroup assemblage (see Fridley, 2001). All of the assemblages were replicated five times, and five ground beetle-free treatments were used as controls (n = 50).

Feeding trials

Plastic mescosms (90 cm long, 40 cm wide, and 18 cm deep) that were filled with 3 cm of damp potting soil were used to compare prey suppression between the nine created assemblages. To prevent the slugs from climbing out, the inner rim of each mesocosm was painted with FluonTM, a slick material that prevents animals from climbing over barriers (Symondson, 1993). Mesocosms were placed in a controlled environment in a room with a LD 12:12 h light cycle and a temperature of 20 ± 2 °C. Two hours before starting the experiment, a mixed prey community including 5 adults and 20 eggs of D. reticulatum, 10 larvae of T. molitor, 10 eggs of H. aspersa, and 40 wingless R. padi was added to each mescosm. Aphids were transferred to the mesocosms onto four 10- to 15-cm tall barley plants (each plant had been previously infested with 10 mixed second and third instar aphids). Early instar aphids were used to prevent their multiplication during the experiment. Eggs of D. reticulatum and H. aspersa were placed separately in clusters and covered with vegetation to prevent their desiccation and to mimic as closely as possible the way that eggs are found in the field. Adults of D. reticulatum were released on fresh T. pratense leaves placed in the centre of each mesocosms as a food source for the slugs. Mesocosms' centres were watered to maintain adequate moisture conditions for slug activity. Ground beetles of each assemblage were released in the mesocosms at the start of the experiment and left undisturbed for 24 h. At the end of this period, the number of remaining prey was counted, and the mesocosms were hand-searched until all of the ground beetles or their remains were recovered. Prey suppression was calculated as the mean proportional mortality across the five prey species. Eggs presented with evidence of severe attacks (i.e. egg completely or partially emptied, and eggs with chorion broken open) but was not completely eaten, were considered as killed. Aphids were counted by carefully hand-searching the plants.

Statistical analysis

Because there was very little mortality in the ground beetle-free mesocosms (only three aphids were not found), the results of the control treatment were not included in any of the statistical analyses. All of the statistical analyses were conducted on proportional prey mortality (arcsine transformed).

We used a one-way analyses of variance anova including the nine assemblages (A1–A9) to test for differences in prey consumption among these assemblages as well as identify the best performing group among small (A1), medium (A2), and large (A3) ground beetles,

To distinguish the effects of ground beetles functional diversity versus ground beetles functional identity on total prey suppression, we performed three separate one-way (anova) followed by planned comparisons. Each anova included all of the three assemblages with low functional diversity (mono-group assemblages; A1 + A2 + A3) and one of the three assemblages with high functional diversity (tri-group assemblages; A7, A8, and A9). d.f. = 4 for each anova. After each anova, a planed comparison was done to compare between prey suppression in the highly diverse assemblages and prey suppression of the pooled assemblages with low functional diversity. This allowed us to test for functional diversity effect and for the non-transgressive overyielding (Fridley, 2001). The residual SS in each comparison is then attributable to the differences among mono-group assemblages, and constitutes a test of the functional identity effect (Bruno et al., 2005; Duffy et al., 2005). Transgressive overyielding, which identifies the effects of diversity that could not result from the dominance of any one of the size group, was also tested using planned comparisons between each one of the three assemblages with high functional diversity (A7, A8, and A9) and the best performing assemblage of low functional diversity.

Using the opportunity offered by our experimental design (i.e. the presence of three different species for each size group, and the variation in the species composition of the nine assemblages), we tested for species identity and additive effects among the three large ground beetle species. A test of these effects was done for species pairs (i.e. C. auratus vs. A. parallelepipedus, C. auratus vs. P. melanarius, and P. melanarius vs. A. parallelepipedus). For this, three separate one-way anovas were performed to compare prey suppression between the following treatments: (1) assemblages in which the two species were absent, (2) assemblages in which only species A was present, (3) assemblages in which only species B was present, and (4) assemblages in which both species A and B were present. Differences in prey suppression between assemblages in which the two species occurred separately were interpreted as indirect evidence for a species identity effect of either one or the other species. A significant increase of prey suppression in the assemblages in which both species were present compared with the assemblages in which the two species occurred separately was interpreted as a result of an additive effect among these two species. In contrast, a similarity in prey suppression between assemblages in which both species were present and assemblages in which the two species occurred separately (treatments two and three) was interpreted as an outcome of a redundant effect of the two species.

Resource partitioning among (i) the three size groups of ground beetles (i.e. differences in the composition and relative abundances of prey types consumed between small (A1), medium (A2), and large (A3) ground beetles) and (ii) all nine ground beetle assemblages was tested using a permutational multivariate analyses of variance (PERMANOVA) (Anderson, 2001). For this, Bray–Curtis dissimilarity was used as a distance measure, and 4999 unrestricted permutations of the raw data were performed (Anderson, 2005).

Similarity percentage (SIMPER) analysis was used to determine which prey species contributed the most to the observed differences in total prey suppression between small (A1), medium (A2), and large (A3) ground beetles.

anovas and planned comparisons were performed using R software 2.15.2 (R Development Core Team, 2012). The permutational multivariate analyses were conducted using PERMANOVA 1.6 (Anderson, 2005). SIMPER was conducted using PAST (Hammer et al., 2001).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

At the end of the experiment, all of the prey from the ground beetle-free mesocosms were recovered (except three aphids), confirming that all of the prey losses were because of the ground beetles' consumption rather than other causes of mortality or escape. Additionally, no ground beetles died during the experiment. This indicates that there was no intra-guild predation.

Best performing ground beetles

Mean prey suppression significantly differed across the nine assemblages (F8.36 = 26.019, P < 0.001). The highest mortality of prey was always recorded in the assemblages with at least one large species (Fig. 1). The mean total prey suppression was also significantly different among the three monogroup assemblages (A1 to A3). Large ground beetles (A3) had greater total prey suppression than medium and small ones and, thereby, were the best performing group of species. However, the effect of medium (A2) and small (A1) species on prey suppression was not significantly different (Fig. 1).

image

Figure 1. Prey suppression by the ground beetles in the nine assemblages tested. The highest prey suppression was always recorded in the case of assemblages with at least one large species (A3, A4, A5, A7, A8, and A9). See Table 1 for assemblages' species composition and size diversity. Prey suppression was calculated as the mean proportional mortality (arc sin transformed) across the five prey species. Values are given as mean ± SE. Means with different letters are significantly different (P < 0.05).

Download figure to PowerPoint

Functional diversity versus functional identity effects

Although the ground beetles did exert significant prey mortality, increasing their functional diversity had no effect on the strength of prey suppression. Indeed, planned comparisons between the average prey suppression in the assemblages with low functional diversity (A1 + A2 + A3) and that in each one of the highly diverse assemblages (A7, A8, and A9) showed no significant effect of ground beetle functional diversity compared with their functional identity (Table 2). Planned comparisons testing for transgressive overyielding showed that prey suppression in the highly diverse assemblages (A7, A8, and A9) was never significantly greater than in the assemblage with only large species (A3) (Table 2). Thus, there was no evidence for transgressive overyielding. However, the effects of increasing ground beetle functional diversity can be explained by a sampling effect for larger species. Indeed, greatest suppression of prey was always recorded in the assemblages with at least one large species (Fig. 1).

Table 2. Results of statistical analysis separating the effects of ground beetle functional diversity and identity on prey suppression
 d.f.Sum sqMean sqF-valueP
  1. Data were analysed by one-way analyses of variance (anova) followed by orthogonal planned comparisons. The anovas were conducted on proportional prey mortality (arc sin transformed). Each anova and planed comparison compared between the pooled mean of prey suppression in all low functional diversity (monogroups) replicates (A1 + A2 + A3) and prey suppression in each one of the high-functional diversity assemblages (A7, A8, and A9).

  2. MS, mean square; SS, sum of squares.

monogroups vs.C7
Community190.91130.282556.684< 0.001***
Diversity (non-transgressive overyielding)10.02030.02030.41060.5297
Identity180.89100.049510.106< 0.01**
Transgressive overyielding10.15500.15503.690.0707
monogroups vs.C8
Community190.94730.287042.185< 0.001***
Diversity (non-transgressive overyielding)10.02840.02830.55540.4657
Identity180.91890.05107.6820<0.05*
Transgressive overyielding10.13890.13893.09210.0957
monogroups vs.C9
Community191.06150.335779.216<0.001***
Diversity (non-transgressive overyielding)10.18190.18193.72240.0696
Identity180.87960.048911,690<0.01**
Transgressive overyielding10.02620.02620.45590.5082

Species identity and additive effects among large ground beetles

Overall, the three anovas testing for the species identity and additive effects among large ground beetles were significant (F3,43 = 12.147 and P < 0.001 for C. auratus vs. the P. melanarius test, F3,43 = 30.701 and P < 0.001 for C. auratus vs. the A. parallelepipedus test, and F3,43 = 11.989 and P < 0.001for P. melanarius vs. the A. parallelepipedus test). Significance in all cases was mainly because of the low values of assemblages in which large species were absent (Fig. 2a–c). However, comparisons of prey suppression between assemblages in which species occurred separately (CA to PM in Fig. 2a, CA to AP in Fig. 2b, and PM to AP in Fig. 2c) revealed that there was no species identity effect among the large ground beetles. Prey suppression appeared to increase slightly in the assemblages that included C. auratus (CA in Fig. 2a,b), but this trend was not statistically significant. Also, there was no additive effect among the three large ground beetles, as prey suppression was the same even where assemblages included either one or two large species (CA & PM to both in Fig. 2a, CA to both in Fig. 2b, and PM & AP to both in Fig. 2c). One exception to this result concerns the comparison between C. auratus and A. parallelepipedus (Fig. 2b) where prey suppression was higher in the assemblages included the two species together (both), as opposed to those that included just A. parallelepipedus (AP).

image

Figure 2. Comparison of prey suppression among large ground beetle species. Comparisons were done for species pair [(a) C. auratus vs. P. melanarius, (b) C. auratus vs. A. parallelepipedus, and (c) P. melanarius vs. A. parallelepipedus], between assemblages in which the two species were absent (Absent), assemblages in which both species were present together (both), and assemblages in which the two species occurred separately (CA for Carabus auratus, AP for Abax parallelepipedus, and PM for Ptersostichus melanarius). Values are given as mean ± SE. Means with different letters are significantly different (P < 0.05).

Download figure to PowerPoint

Resource partitioning

Patterns in prey consumption were significantly different between small (A1), medium (A2), and large (A3) ground beetles (F2,12 = 6.7901, P < 0.001; from the PERMANOVA among the three size groups of ground beetles). Unlike medium and large ground beetles (A2, A3) that demonstrated an ability to consume all of the prey species, small ground beetles (A1) did not attack the adults of D. reticulatum and only weakly predated the mealworm beetle larvae (Fig. 3). Furthermore, prey suppression by the small and medium ground beetles was weak for all of the prey species attacked, whereas the large ground beetles showed maximum prey suppression and tended to prefer slugs (adults and eggs), as well as eggs of the land snail. Aphids were the less attacked by large ground beetles (Fig. 3). From the PERMANOVA among the nine ground beetle assemblages, clear differences in resource use patterns were evident between the assemblages with at least one large ground beetle species and those created with only small and/or medium ground beetles (F8,36 = 3.9212, P = 0.0006).

image

Figure 3. Consumption rates of the five prey species by ground beetles of the three size classes (small (A1), medium (A2), and large (A3) ground beetles). Ground beetles of the different size groups exhibited significantly different patterns of prey consumption. While adults of Deroceras reticulatum have never been attacked by small ground beetles, all of them were eaten by large ones. Values are given as mean ± SE. Prey species are labelled as follows D.r.: adults of Deroceras reticulatum; T.m.: Tenebrio molitor; R.p.: Rhopalosiphum padi; D.r. Eggs: eggs of Deroceras reticulatum; and H.a. Eggs: eggs of Helix aspersa.

Download figure to PowerPoint

SIMPER analysis revealed that the adults of D. reticulatum were the prey that contributed the most to dissimilarity between small and large ground beetles (A1 vs. A3; Fig. 4), as well as between medium and large ones (A2 vs. A3; Fig. 4). Tenebrio molitor contributed 32% (from the SIMPER analysis) to the observed difference in prey suppression between small and medium ground beetles (A1 vs. A2; Fig. 4).

image

Figure 4. Contribution (%) of the five prey species to the observed differences in total prey suppression between small (A1), medium (A2), and large (A3) ground beetles. See Results for SIMPER results. Prey species are labelled as follows D.r.: adults of Deroceras reticulatum; T.m.: Tenebrio molitor; R.p.: Rhopalosiphum padi; D.r. Eggs: eggs of Deroceras reticulatum; and H.a. Eggs: eggs of Helix aspersa.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Partitioning the effects of ground beetle size diversity in our study revealed that prey suppression by ground beetles depends on their functional identity rather than on functional diversity, with one group of species (large beetles) contributing more to prey suppression than the other groups. Indeed, in the assemblages containing at least one large ground beetle species, prey suppression was significantly greater than in assemblages without large ground beetles. Such relationships between predator size diversity and prey suppression is common in a variety of systems and taxa (Radloff & Du Toit, 2004; Bruno et al., 2005, 2006; Arenas et al., 2006; Philpott et al., 2009; Valdivia et al., 2009; Rudolf, 2012; Toscano & Griffen, 2012). An identity effect for large ground beetles found in our study is in agreement with the hypothesis given by Rudolf (2012) to explain the ecological effects of variation in predator body size on prey suppression, and according to which this variation can increase prey suppression through increases in the average size that increases the per capita consumption rates of predators. As for other organisms (Reiss et al., 2011), large ground beetles most likely have higher mass-specific metabolic rates, higher energy demands, and higher ingestion rates than small species.

The neutral effect of increasing ground beetle size diversity was not surprising given that there was no partitioning of resource between small, medium, and large species. Although we showed distinct patterns in prey consumption, the different-sized ground beetles did not occupy distinct trophic niches. The results revealed an increase in the amount and range of prey consumed with increasing ground beetle body size; small ground beetles showed low prey consumption and did not attack adult slugs, whereas large ground beetles showed maximum prey suppression and demonstrated an ability to consume all of the prey species. Neutral effects of predators' diversity on prey suppression are often expected to occur when increasing the diversity within a predator population increases the potential for intra-guild predation as well as for negative behavioural interference (Letourneau et al., 2009). For example, in the presence of a large predator, small ones often alter their activity rates or habitat use, which also reduces their foraging rates that indirectly alter prey suppression (Prasad & Snyder, 2006; Crumrine, 2010; Rudolf, 2012). In our case, such negative interactions could not be responsible for the neutral effect of the increasing ground beetle size diversity on prey suppression. All of the ground beetles were alive at the end of the experiment, indicating that there was no intra-guild predation. Additionally, we expect that large ground beetles in the more diverse assemblages did not reduce or only slightly reduced the foraging rates of the small ground beetle species, as the prey suppression was low even when these small species were present alone (Assemblage 1, Fig. 1).

Excluding negative interactions to explain the neutral effect of increasing the ground beetle size diversity effect on prey suppression does not exclude that it is the absence of intra-guild predation among the large ground beetle species (best performing predators in our study) that may leads to the evident identity effect of this species group rather than ground beetles size diversity per se. Indeed, in the absence of negative interferences among species that overlap in resource use, there would be no advantage of resource use differentiation and no significant difference between predator diversity treatments (Snyder & Ives, 2003). At least two explanations can account for the absence of intra-guild predation and ground beetles' functional diversity effect in our study.

Predator to prey ratio

It may be possible that the ratio of predators to prey in our experiment was too low (i.e. the prey offered were sufficient enough) for antagonistic interactions to occur among the ground beetles. Most of the studies that have examined the effects of prey density on the outcome of predator–predator interactions have found that lower prey density increases intra-guild predation, whereas higher prey density results in synergetic interactions (Lucas et al., 1998; Kajita et al., 2000; Hindayana et al., 2001; Burgio et al., 2002; de Clercq et al., 2003; Nóia et al., 2008). Werling et al. (2012) showed that combining two functionally distinct predators increased the predation of Colorado potato beetle only when the larval density of this insect pest was low. However, Griffin et al. (2008) found that functionally diverse communities of crab species showed greater prey consumption only when the predator density was high. This observation has also been reported in other previous studies (Griffin et al., 2008; Takizawa & Snyder, 2011), in which reducing the intra-specific density of predators reduced the resource overlap and interference and, therefore, had a greater effect on prey suppression.

Experimental term

Our experiment may not have lasted long enough to detect the positive indirect effects on prey suppression that may result from interferences among ground beetle species of the same group. As shown above, intra-guild predation occurrence usually involves a lower prey density. The short time scales of our experiment could have been insufficient for prey density to be reduced by the ground beetle consumption and then to reach the ‘critical threshold’ that trigger intra-guild predation. Previous studies suggested that intra-guild predation requires prey density changes, which may take some time to manifest (Rosenheim, 2001)

While we found evidence for a functional identity effect, there was no species identity effect, as prey suppression in the assemblages where the three large ground beetles occurred separately was not different. Additionally, the effects of these three ground beetle species on prey suppression seems to be non-additive but redundant in the sense that there was no difference in prey suppression between the assemblages in which two species were present together and those in which each species was present separately. The redundant effects of the large ground beetle species suggest that they occupy similar trophic niches, have overlapping resources, and interfere with each other. Indeed, when such effects occurred, they were often related to competition for food among predators that alter the behaviours of the predators. They were also related to intra-guild predation that often occurs between certain life history stages, such as adults preying on juveniles, or when one of the predators is an intermediate predator (often because of its small size) and is vulnerable to predation by the top predator (Rosenheim, 1998; Finke & Denno, 2002; Crumrine & Crowley, 2003; Lang, 2003; Griffen & Byers, 2006). Because all of the ground beetles were recovered at the end of our experiment, we suggest that inter-specific competition for food, and not intra-guild predation, was responsible for the non-additive effect among the three large ground beetle species. However, we have no evidence that this behavioural interference existed in our experiment. A similar non-additive effect of multiple predators was observed in a system composed of three spider species preying upon the grasshopper Melanoplus femurrubrum (De Geer) (Sokol-hessner & Schmitz, 2002). The spiders had different hunting modes and occupied complementary habitats, but their combined effects were equivalent to the average of the corresponding single-species effect. The authors concluded that the species effects of the three spiders were substitutable and that it is reasonable to aggregate them into a single functional unit. The same results were also found by Woodcock and Heard (2011). Similar effects of the three large ground beetles in our experiment validate the fact that they have been treated as a single functional group and suggest that the effect of ground beetles on prey suppression should depend on the presence of large ground beetles, independent of their species identity.

Although predators of the same guild could have a non-additive (redundant) effect, increasing their number could increases prey suppression through a resource use complementarity (Powell et al., 2006). Moreover, species that appear functionally redundant under some environmental conditions are functionally diverse when environmental conditions change (Naeem & Li, 1997). This effect is encompassed in the ‘insurance hypothesis’, which states that maintaining different predators that perform better or worse in particular environments provides functional compensation and reliable pest suppression in spite of changing conditions (Loreau et al., 2003). Also, it has been suggested that if predator communities contain functionally redundant species, key aspects of community and ecosystem processes may remain unchanged by the changes in species' compositions as long as each broad functional group retains at least one functionally competent species (Morin, 1995). This suggestion is consistent with the results found in the present study. Indeed, there were no differences in prey suppression between the assemblages that included at least one large ground beetle species, except assemblage in which the three large species occurred together (A3) and in which prey suppression was more important than in two other assemblages (A7 and A8).

This study found that increasing the size diversity of ground beetles had no effect on the strength of prey suppression. Instead, prey suppression was strongly strengthened by the presence of large ground beetles (irrespective of their species identity). Based on these results and given that large ground beetles are vulnerable to environmental disturbances (Ribera et al., 2001; Kotze & O'Hara, 2003) and predation (Kaspari & Joern, 1993; Brose, 2003), we suggest that conservation biological control strategies should promote the presence of large carabids. These strategies should, however, be established without being in conflict with the global objectives of biodiversity conservation. Redundancy among species should also be considered in conservation biological control strategies. Indeed, given that the local and global extinctions are more likely for species occupying higher trophic levels than for species at lower trophic levels within food webs (Petchey et al., 1999), it is always important to maintain redundancy both among and within local communities because species that are redundant in their effects will not necessarily have the same responses to environmental change (Wellnitz & Poff, 2001). Thus, this redundancy could provide insurance against loss or degradation of the biological pest suppression (Walker, 1991)

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This research was supported by a PhD fellowship from the INRA/Lorraine Region (France) and by the ‘CASDAR Entomophages’ Project (French Agriculture Ministery). The authors wish to thank Catherine Mignolet who gave permission to conduct experiments on the INRA's Mirecourt Experimental Station. We gratefully acknowledge the logistical help of Claude Gallois and David Marcolet. We would also like thank Simon Taugourdeau and Jean Villerd for their advice on statistical analysis. Two anonymous reviewers provided valuable comments that greatly improved the manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References