SEARCH

SEARCH BY CITATION

Keywords:

  • biodiversity;
  • Cirsium arvense;
  • herbivory;
  • landscape context;
  • organic farming;
  • weeds

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    The expansion of simplified ecosystems such as intensively managed annual crops plays a big part in driving the global biodiversity crisis. Field-scale diversification, for example leaving weeds to grow in crops, is one way in which diversity in agro-ecosystems can be restored. However, little is known about the determinants of the non-crop plant-based insect communities within arable fields at local and larger spatial scales, an essential component in extrapolating plant diversity benefits to higher trophic levels.
  • 2
    We investigated how diversification of agro-ecosystems at the field and landscape levels affects the insect community of the creeping thistle Cirsium arvense. Artificial plots of the host-plant were established in three regions of Germany in 48 paired organic (diverse, weeds not controlled with herbicides) and conventional (simplified, very low weed density and species richness) wheat fields across a gradient of landscape heterogeneity, from simple arable-dominated to heterogeneous, diverse landscapes.
  • 3
    Leaf-feeding herbivores were monitored directly, while stem-boring herbivores and their parasitoids were quantified by dissecting the stems of the thistles. Land-use types and naturally occurring thistle stands were mapped within a radius of 1 km around each thistle plot.
  • 4
    Herbivore species richness was enhanced by both organic farming and landscape heterogeneity but not by higher densities of thistles in the landscape. For most of the species, host-plant plots in organic fields were more likely to be colonized than those in the conventional fields. The enhancement of diversity in organic fields is probably the result of a slightly higher natural cover of the host-plant Cirsium arvense.
  • 5
    Synthesis and applications. Both diversification of landscape (fewer arable crops, more perennial habitats) and extensification through organic management are effective measures of enhancing arthropod diversity on weeds. The impact of field-scale agri-environment schemes on biodiversity should be supplemented by including landscape-scale diversification programmes to include a minimum level of perennial habitat cover. Biodiversity benefits of organic agriculture rely for a large part on non-crop plants. Weed populations should be allowed to coexist with the crop to maintain these benefits, which are threatened by more intensive ‘organic’ management, such as heavy mechanical weed control.

Introduction

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

Relatively little is known about the determinants of species diversity of the insect fauna on non-crop plants within arable fields compared with natural or near-natural habitat types. By providing resources for pollinators and herbivorous insects, which in turn provide food for predatory birds and mammals, non-crop plants are key players in the above-ground food chain within agro-ecosystems (Swift & Anderson 1994; Marshall et al. 2003). As such, they can be considered as ecological goods that may be produced alongside harvestable crops in the framework of multifunctional agriculture (Gerowitt et al. 2003). Complex food web interactions are known to exist between different arable weed species and insects using their resources (Gibson et al. 2006). However, many important questions remain unanswered, for example the relative accessibility of arable weeds to insects. Non-crop plants within annual crops support less herbivorous insects than plants of the same species growing in perennial habitats (Kruess 2003). In Europe, organic agriculture has greatly increased weed cover and diversity by not using synthetic fertilizers and pesticides. This provides an opportunity to test the effects of field-scale vegetation diversification on the insect community of weeds.

Many important population processes take place at larger scales than the habitat patch itself (Roland & Taylor 1997; Tscharntke et al. 2005). Arable field ecosystems rely on their surroundings more than other types of habitats because they are cleared at harvest and are usually ploughed every year, so much of their biodiversity depends on recolonization from surrounding perennial habitats (Wissinger 1997; Tscharntke & Kruess 1999; Roschewitz et al. 2005), which serve as overwintering habitats and contain alternative resources for arthropods (Russell 1989; Kruess 2003; Pfiffner & Wyss 2004).

In the present study we examined how diversifying arable fields through organic management affects the insect community of a model non-crop plant species in landscapes with different amounts of perennial habitats. To do so we established plots of the creeping thistle Cirsium arvense (L.) Scop. (Asteraceae) as an indicator plant species, or phytometer, in paired organic and conventional winter wheat fields across a landscape gradient in three regions of Germany, and surveyed leaf-feeding and stem-boring herbivores, as well as their parasitoids.

Materials and methods

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

study area

We selected 48 paired organic and conventional wheat fields in three different regions of Germany: seven pairs in the Soester Boerde, 10 pairs in the Leine Bergland and seven pairs in the Lahn-Dill Bergland (details about the regions can be found in Clough et al. 2005). In each region, the paired fields were situated in landscapes with varying composition, along a gradient from simple arable-dominated to more complex landscapes. Distance between fields in a pair was never more than 600 m. Even though fields were selected to be as similar as possible, conventional fields tended to be larger than organic fields (mean ± SE; conventional 5·06 ± 0·76, organic 2·98 ± 0·65). Field boundaries in the research regions were narrow (< 3 m) and grassy. One-third of the conventional farmers sprayed insecticides (lambda-cyhalothrin or dimethoate) in a single application. Organic fields were fertilized with manure only and were not sprayed with any pesticides. Despite mechanical weed control (comb harrowing), weed cover and diversity were higher in the organic fields. Total floral cover (percentage soil cover of all flowering species) was close to 0 for conventional fields (always below 0·005%) and for organic fields the first quantile, median and third quantile were 0·006%, 0·053% and 0·652%. Natural creeping thistle cover outside the experimental plots was higher in organic fields than in conventional fields (mean ± SE; organic 0·92 ± 0·44%, conventional 0·06 ± 0·04% soil cover).

phytometer species

Cirsium arvense, known as creeping or Canada thistle, was chosen as a model system because it is extremely common and very characteristic of arable habitats, although it also occurs in other disturbed ruderal environments. Creeping thistle may also harbour a large number of different species of insect herbivores and their associated parasitoids (Zwölfer 1965; Redfern 1995). Biology, control and detrimental effects of Cirsium arvense have been reviewed in detail by Donald (1990, 1994).

host-plant plots

A row of three Cirsium arvense plots with 12 plantules each was established in the centre (distance from edge between 35 and 55 m) of every study wheat field in the first week of May 2003. Each plot had an area of 1 m2 and plots were 0·5 m apart. All data from the plots were summed at the field level. Plots were hand-weeded on every monitoring occasion. The thistles were harvested between 23 and 28 July. The shoots were dried at 70 °C for 48 h and the biomass recorded. Number of stems (53 ± 4) and amount of dry biomass (144 ± 14 g) did not differ between the Cirsium arvense plots within pairs of organic and conventional fields or between landscapes of different compositions.

herbivores and parasitoids

Leaf-feeding herbivores were surveyed by examining the thistle plants on four occasions during spring and summer (end of May, mid-June, end of June, mid-July). In addition to species richness and presence–absence data, the following specific measures of herbivore abundance were included in the analyses (Fig. 1): incidence per field (number of plants affected/total number of plants) of the shield beetle Cassida rubiginosa Müll. (Coleoptera: Chrysomelidae); cumulative counts of caterpillars of the butterfly Vanessa cardui L. (Lepidoptera: Nymphalidae); number of leaf mines (Diptera: Agromyzidae); number of larvae and pupae of Melanagromyza aenoventris Fall. (Diptera: Agromyzidae); number of larvae of unidentified species of weevils (Coleoptera: Apionidae); number of Auchenorrhyncha (Hemiptera) eggs. Cumulative counts for Vanessa cardui caterpillars may have overestimated the number of caterpillars but because the lag between survey occasions approximately equalled the larval development time (12–18 days; Ekkehard 1983) and most caterpillars were found at intermediate larval stages, results are unlikely to have been biased. For the calculation of species richness we considered higher taxa that were not identified to species level as single species. Sample size was reduced by two field pairs for species richness and endophagous insect analyses because plants either died as a result of herbivory or could not be recovered before harvest. Parasitoids were collected from the plant material after stem dissection. Parasitism rates were calculated for Melanagromyza aenoventris as number of parasitoids divided by the sum of the numbers of larvae, pupae and parasitoids found in the plots. If the host was absent from a field pair it was not included in the parasitoid analysis.

image

Figure 1. Herbivore insect taxa commonly found in experimental Cirsium arvense plots in German winter wheat fields in 2003. (a) Vanessa cardui L. (Lepidoptera: Nymphalidae); (b) Cassida rubiginosa Müll. (Coleoptera: Chrysomelidae); (c) Apionidae (Coleoptera); (d) Melanagromyza aenoventris Fall. (Diptera: Agromyzidae); (e) leafminers (Diptera: Agromyzidae); drawings not to scale.

Download figure to PowerPoint

field and landscape data

Weed cover and species richness as well as Cirsium arvense cover (in both cases percentage soil cover) were established visually for 42 of the 48 fields. We collected data on the number of insecticide applications and amount of nitrogen applied on the fields. At the landscape scale, land use and naturally occurring thistle populations were mapped based on a field survey in a radius of 1000 m around each field. From the land-use maps we calculated the percentage non-crop area in the landscape sectors (range 17–83%). Thistle abundance in the landscapes was estimated by quantifying area and density of each natural patch and calculating total shoot number. We also used isolation indices following Hanski (1994) using either only the nearest neighbouring patch or all patches within the 1000-m radius, both unweighted and weighted by the number of stems, but analyses using these had a similar outcome to those using thistle abundance in the landscapes (data not shown).

statistical analysis

We used generalized linear mixed models (GLMM), with Poisson error distribution (link = log) for species richness and abundance data and with a quasibinomial error distribution (link = logit) for incidence and presence–absence data. In both cases two error levels were added for field pair (random factor) and management type (fixed factor) nested within pair. We thus took into account the spatial autocorrelation between fields from the same pair. To test for significance of the fixed effects we used a type I sum of squares with the following sequence: region, percentage non-crop area, number of shoots in the landscape, management and number of thistle shoots in the experimental plot. With this sequence we could take into account the spatial nestedness of these factors. We did not assume any special form for the random effects variance–covariance matrix. To test for local fixed effects (weed cover, weed diversity, thistle cover in the field, kg nitrogen fertilizer, insecticide application) we used separate generalized linear models (GLM) with the appropriate error distribution for organic and conventional fields, given the difference in range of the values for these explanatory variables between the two management types. The same variable sequence as in the GLMM was used. To obtain best-fit models, non-significant effects and interactions (P > 0·05) were discarded using a manual stepwise backwards selection procedure. Region remained in the models to avoid spurious significances as a result of spatial autocorrelation. Differences between regions, whenever significant in the mixed model, were investigated further by comparing the confidence intervals of the estimators for the levels of the factor. All analyses were performed using R (version 2·1·1; R Development Core Team 2005) and the R package MASS. In the text, data in parentheses are mean ± 1 SE.

Results

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

species richness

Overall species richness of herbivores was correlated positively with percentage non-crop area and was significantly higher in organic fields than in conventional fields (Fig. 2 and Table 1). On average, plots in organic fields were colonized by four herbivore species, compared with three in the conventional fields, without this effect being attributable to a particular species. Species richness was not related to the amount of thistles in the landscape (denominator degrees of freedom (den.d.f.) = 19, F= 1·30, P > 0·26) and did not differ between regions (den.d.f. = 19, F= 0·92, P > 0·42). Species richness decreased significantly with weed cover for organic fields (den.d.f. = 13, F= 7·12, P= 0·02), but not for conventional fields (den.d.f. = 13, F= 2·38, P > 0·15; Fig. 3). Other local variables did not contribute significantly to the models.

image

Figure 2. Species richness of herbivores of Cirsium arvense in wheat fields in relation to field management and percentage non-crop area; circles indicate the values per field and lines are the fitted linear mixed model for organic (filled circles, solid line) and conventional fields (empty circles, broken line).

Download figure to PowerPoint

Table 1.  Wald tests for best-fit generalized linear mixed models explaining species richness and incidence of Cassida rubiginosa (Coleoptera: Chrysomelidae), cumulative counts of Vanessa cardui larvae (Lepidoptera: Nymphalidae), number of larvae/pupae of Melanagromyza aenoventris (Diptera: Agromyzidae) and number of larvae of pear-shaped weevils (Coleoptera: Apionidae) on Cirsium arvense in organic and conventional wheat fields, in landscapes differing in percentage non-crop cover in three different regions of Germany
 Factord.f.*FEffectP
  • *

    Denominator degrees of freedom.

  • a, Leine Bergland; b, Soester Boerde; c, Lahn-Dill Bergland.

Species richnessRegion19 0·612   0·552
% non-crop20 4·989+  0·037
Management2016·613+< 0·001
Cassida rubiginosaRegion21 1·414   0·265
% non-crop22 7·685+  0·011
Management22 7·997+< 0·001
Vanessa carduiRegion19 4·528a, c > b  0·023
% non-crop2310·104  0·004
Melanagromyza aenoventrisRegion19 3·099   0·068
Stems per field21 8·238+  0·009
ApionidaeRegion19 0·048   0·953
Management20 0·041+  0·841
Stems per field2010·565+  0·004
image

Figure 3. Relationship between the species richness of Cirsium arvense herbivores and percentage weed cover in (a) organic fields (inline image = 0·30) and (b) conventional fields.

Download figure to PowerPoint

plot colonization by herbivores

Analysis of the presence–absence data for individual species (Table 2) showed that plots in organic fields were significantly more likely to be colonized by Cassida rubiginosa, Melanagromyza aenoventris and stem-boring weevils and not by leafminers and Vanessa cardui. There were no significant landscape effects on the colonization probability of individual species. All thistle plots were colonized by Vanessa cardui and all but three by leafminers. Of the endophagous herbivores, only the presence of Melanagromyza aenoventris was more likely in plots containing a greater number of thistle stems. Other local variables were not significant.

Table 2.  Percentage presence of five herbivorous insect taxa in plots of Cirsium arvense in organic and conventional wheat fields, with significance levels from Wald-tests for linear mixed models
 Management
ConventionalOrganicSignificance
  • ***

    P < 0·001;

  • **

    P < 0·01;

  • *

    P < 0·05; NS, not significant.

  • Sample sizes: n= 22 field pairs for Apionidae and Melanagromyza aenoventris, n= 24 field pairs for Cassida rubiginosa, Vanessa cardui and leafminers.

Apionidae 45·5 81·8**
Cassida rubiginosa (Coleoptera: Chrysomelidae) 62·5 83·3*
Melanagromyza aenoventris (Diptera: Agromyzidae) 68·2 90·9*
Vanessa cardui (Lepidoptera: Nymphalidae)100100NS
Leafminers (Diptera: Agromyzidae) 79·2 87·5NS

herbivore abundance measures

Analysis of best-fit models for individual species data are shown in Table 1. The incidence of Cassida rubiginosa was higher in the organic fields than in the conventional fields and increased with percentage non-crop area in the landscape (Fig. 4). Numbers of immature stages of Apionidae and Melanagromyza aenoventris did not differ between organic and conventional fields. Furthermore, the densities of both stem-boring herbivores depended on the number of stems per plot. Cumulative densities of Vanessa cardui larvae were significantly higher in the Lahn-Dill Bergland and Leine Bergland than in the Soester Boerde, and decreased with increasing percentage of non-crop area in the landscape. The number of leaf mines (median 10 field−1, or approximately 0·3 thistle shoot−1) could not be explained by any of the factors included in the model. Local variables were not significant for any of the species.

image

Figure 4. The percentage of Cirsium arvense plants damaged by Cassida rubiginosa (Coleoptera: Chrysomelidae) per field in relation to management and percentage non-crop area in the landscape; circles indicate the values per field and lines the fitted linear mixed model for organic (filled circles, solid line) and conventional fields (empty circles, broken line).

Download figure to PowerPoint

parasitism

Because parasitoids of the apionids had emerged before the thistle harvest, the parasitoids found in the stems were all from Melanagromyza aenoventris and of the genera Stenomalina and Chlorocytus (Hymenoptera: Pteromalidae). The average parasitism rate was 41 ± 6%. There were no relationships between local and larger scale variables and parasitism rates or colonization success by parasitoids (presence–absence data). Parasitism of Vanessa cardui by an unidentified braconid wasp (Hymenoptera: Braconidae) was also observed, but only in four of the study fields.

Discussion

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

In this study we were able to separate the effect of enhanced within-field diversity through organic farming and the effect of the landscape-scale complexity around the field on species richness and abundance of thistle herbivores in arable fields. At the local scale, plots of the host-plant situated in organic fields were more likely to be colonized by Cassida rubiginosa, Melanagromyza aenoventris and apionid weevils than plots situated in conventional fields. Species richness was also higher in the organic fields. Host-plant acceptance by foraging herbivores is unlikely to have been different in organic and conventional fields. A small-scale pot experiment showed no difference in colonization of Melanagromyza aenoventris between plants grown in organic soil or conventional soil but in otherwise similar conditions (Y. Clough, unpublished data). Insecticide applications seem to have had negligible effects on thistle herbivores, as treated fields did not have lower herbivore species richness or densities than untreated conventional fields. The enhanced diversity in organic fields might therefore be the result of the higher Cirsium arvense cover in the organic fields, although there was no significant correlation between herbivore parameters and percentage thistle cover. Overall weed cover was significantly higher in the organic fields than in the conventional fields (Gabriel et al. 2006). Herbivore species richness was lower at high levels of weed cover (above 35–40%) in organic fields. As none of the other weed species recorded in the fields were within the host-plant spectrum of the herbivores that were recorded, weed cover could have been expected to have little or no impact. Our results suggest that the composition and diversity of the vegetation matrix (determined by the weeds) may play a role in host-plant colonization by arable weed herbivores. More generally we conclude that organic management does succeed in reversing some of the negative effects of local agricultural intensification on weed herbivore diversity. Surprisingly, such an effect was not found for parasitism of Melanagromyza aenoventris. One might have expected the nectar-feeding parasitoids to benefit from the greater floral resources in organic fields (A. Holzschuh, unpublished data).

While the abundance of thistles in the landscape had no measurable impact on the insect community of our indicator plots, landscape heterogeneity was related to both species richness and individual species densities of thistle herbivores. In homogeneous arable-dominated landscapes, less species colonized the thistle plots than in landscapes with a higher percentage of perennial habitats. The latter provides overwintering areas for insect colonizers. For example, for species such as Melanagromyza aenoventris, which overwinter in the pupal stage within the dry stems of the host-plant, thistle stands on arable fields represent a population sink because of high mortality caused by tillage long before the emergence of the adults. Other species overwinter away from the thistles but still rely on non-tilled habitat for survival. One herbivore species, Cassida rubiginosa, which overwinters as an adult in litter, was more frequent in landscapes with a high percentage of non-crop habitat. Thus our results show that the landscape determines potential insect diversity that may be realized or not depending on local conditions. The fact that the percentage non-crop area at the landscape scale had no effect on the parasitism rate of Melanagromyza aenoventris does not support the results of Kruess & Tscharntke (2000), who found parasitism rates to increase with percentage cover of perennial habitats. However, parasitism is generally lower in annual crops than in less-disturbed habitats (Kruess 2003) so disturbance and isolation within the crop as well as small plot size may be the main limiting factors, leaving other effects undetectable.

In conclusion, both local management and the landscape context affected the thistle herbivores. This is the first time this has been shown for such a community, although similar results have been reported for other systems, for example the parasitism of aphids (Östman, Ekbom & Bengtsson 2001). Based on our results we make a number of suggestions.

(i) The biodiversity decline in agricultural landscapes should be tackled at several scales. While organic agriculture results in more species-rich arthropod communities on weeds, arthropods depend on undisturbed habitat for part of their life cycle, and we would recommend further incentives to diversify agricultural landscapes to ensure a minimum level of cover of perennial habitat.

(ii) Weeds (and the arthropods depending on them) can only contribute to agricultural biodiversity when they are allowed to co-exist with the crop hence further intensification of organic agriculture through the use of repeated mechanical weeding would greatly reduce biodiversity. If biodiversity benefits are to be an integral part of the advantages of organic agriculture, restricting mechanical control might be appropriate.

(iii) From a methodological point of view, we suggest that the use of phytometers to survey arthropods may not only provide a complementary approach to more traditional methods, such as pitfall and pan trapping, with respect to the species caught, but also provides direct evidence of species’ interactions, including herbivory and parasitism of herbivores. This provides much-needed insight into the links between trophic levels, an essential component of biodiversity.

Acknowledgements

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

We thank the farmers for participating in the project and are grateful to Doreen Gabriel and Indra Roschewitz for the vegetation mapping, to Stéphanie Domptail for the landscape digitization and to Tobias Purtauf for his participation in the field selection. Comments by Tibor Bukovinszky, Sabine Eber, Heiri Wandeler, two anonymous referees and the associate editor greatly improved the manuscript. This research was carried out within the framework of the EU-funded project ‘EASY’ (QLK5-CT-2002-01495), co-ordinated by David Kleijn.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Clough, Y., Kruess, A., Kleijn, D. & Tscharntke, T. (2005) Spider diversity in cereal fields: comparing factors at local, landscape and regional scales. Journal of Biogeography, 32, 20072014.
  • Donald, W.W. (1990) Management and control of Canada thistle (Cirsium arvense). Reviews of Weed Science, 5, 193250.
  • Donald, W.W. (1994) The biology of Canada thistle (Cirsium arvense). Reviews of Weed Science, 6, 77101.
  • Ekkehard, F. (1983) Handbuch der Schmetterlingszucht. Franckh, Stuttgart, Germany.
  • Gabriel, D., Roschewitz, I., Tscharntke, T. & Thies, C. (2006) Beta diversity at different spatial scales: plant communities in organic and conventional agriculture. Ecological Applications, 16, 20112021.
  • Gerowitt, B., Bertke, E., Hespelt, S.-K. & Tute, C. (2003) Towards multifunctional agriculture: weeds as ecological goods? Weed Research, 43, 227235.
  • Gibson, R.H., Nelson, I.L., Hopkins, G.W., Hamlett, B.J. & Memmott, J. (2006) Pollinator webs, plant communities and the conservation of rare plants: arable weeds as a case study. Journal of Applied Ecology, 43, 246257.
  • Hanski, I. (1994) A practical model of metapopulation dynamics. Journal of Animal Ecology, 63, 151162.
  • Kruess, A. (2003) Effects of landscape structure and habitat type on a plant–herbivore–parasitoid community. Ecography, 26, 283290.
  • Kruess, A. & Tscharntke, T. (2000) Effects of habitat fragmentation on plant–insect communities. Interchanges of Insects Between Agricultural and Surrounding Landscapes (eds B.Ekbom, M.E.Irvin & Y.Robert), pp. 5370. Kluwer, Dordrecht, the Netherlands.
  • Marshall, E.J.P., Brown, V.K., Boatman, N.D., Lutmans, P.J.W., Squire, G.R. & Ward, L.K. (2003) The role of weeds in supporting biological diversity within crop fields. Weed Research, 43, 7789.
  • Östman, Ö., Ekbom, B. & Bengtsson, J. (2001) Farming practice and landscape heterogeneity influence biological control. Basic and Applied Ecology, 2, 365371.
  • Pfiffner, L. & Wyss, E. (2004) Sown wildflower strips to enhance natural enemies of agricultural pests in annual and perennial cropping systems. Ecological Engineering for Pest Management: Advances in Habitat Manipulation for Arthropods (eds G.M.Gurr, S.D.Wratten & M.A.Altieri), pp. 167188. CSIRO Publishing, Melbourne, Australia.
  • R Development Core Team (2005) R: A Language and Environment for Statistical Computing. Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org, accessed 17 July 2006.
  • Redfern, M. (1995) Insects and Thistles. Richmond, Slough, UK.
  • Roland, J. & Taylor, P.D. (1997) Insect parasitoid species respond to forest structure at different scales. Nature, 386, 710713.
  • Roschewitz, I., Gabriel, D., Tscharntke, T. & Thies, C. (2005) The effects of landscape complexity on arable weed species diversity in organic and conventional farming. Journal of Applied Ecology, 42, 873882.
  • Russell, E.P. (1989) Enemies hypothesis: a review of the effect of vegetational diversity on predatory insects and parasitoids. Environmental Entomology, 18, 590599.
  • Swift, M.J. & Anderson, J.M. (1994) Biodiversity and ecosystems in agricultural systems. Biodiversity and Ecosystem Function (eds E.-D.Schulze & H.A.Mooney), pp. 1541. Springer, New York, NY.
  • Tscharntke, T. & Kruess, A. (1999) Habitat fragmentation and biological control. Theoretical Approaches to Biological Control (eds B.A.Hawkins & H.V.Cornell), pp. 190205. Cambridge University Press, Cambridge, UK.
  • Tscharntke, T., Klein, A.-M., Kruess, A., Steffan-Dewenter, I. & Thies, C. (2005) Landscape perspectives on agricultural intensification and biodiversity: ecosystem service management. Ecology Letters, 8, 857874.
  • Wissinger, S.A. (1997) Cyclic colonization in predictably ephemeral habitats: a template for biological control in annual crop systems. Biological Control, 10, 415.
  • Zwölfer, H. (1965) Preliminary list of phytophagous insects attacking wild cynareae (Compositae) in Europe. Technical Bulletin of the Commonwealth Institute for Biological Control, 6, 81154.