http://orcid.org/0000-0001-7572-532X
Figures
Abstract
Arthropod community composition in agricultural landscapes is dependent on habitat characteristics, such as plant composition, landscape homogeneity and the presence of key resources, which are usually absent in monocultures. Manipulating agroecosystems through the insertion of in-field floral resources is a useful technique to reduce the deleterious effects of habitat simplification. Food web analysis can clarify how the community reacts to the presence of floral resources which favour ecosystem services such as biological control of pest species. Here, we reported quantitative and qualitative alterations in arthropod food web complexity due to the presence of floral resources from the Mexican marigold (Tagetes erecta L.) in a field scale lettuce community network. The presence of marigold flowers in the field successfully increased richness, body size, and the numerical and biomass abundance of natural enemies in the lettuce arthropod community, which affected the number of links, vulnerability, generality, omnivory rate and food chain length in the community, which are key factors for the stability of relationships between species. Our results reinforce the notion that diversification through insertion of floral resources may assist in preventing pest outbreaks in agroecosystems. This community approach to arthropod interactions in agricultural landscapes can be used in the future to predict the effect of different management practices in the food web to contribute with a more sustainable management of arthropod pest species.
Citation: Haro MM, Silveira LCP, Wilby A (2018) Stability lies in flowers: Plant diversification mediating shifts in arthropod food webs. PLoS ONE 13(2): e0193045. https://doi.org/10.1371/journal.pone.0193045
Editor: Raul Narciso Carvalho Guedes, Universidade Federal de Vicosa, BRAZIL
Received: July 5, 2017; Accepted: February 2, 2018; Published: February 16, 2018
Copyright: © 2018 Haro et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Grant Nº: 1122/31-2 (www.capes.gov.br/) to MMH and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (cnpq.br) to LCPS, MMH and Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) (www.fapemig.br) to LCPS, MMH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
In agricultural landscapes richness and abundance of arthropods are dependent on local habitat characteristics, which are determined by the different strategies used in the management of the agroecosystem [1, 2]. Several pieces of evidence indicate that arthropod communities are directly affected by field simplification and loss of non-crop habitats in monocultures, causing distortion in species relative abundance, removing natural enemies from the crop may increase the likelihood for outbreaks [3–5]. The absence of key resources, such as shelter, alternative food (prey, hosts and nectar) and alternative habitat, is the most important factor limiting the effectiveness of ecosystem services, such as the biological control of arthropod pest species in monocultures [6]. Habitat manipulation through the insertion of flower strips adjacent to or within crops is an useful technique to counteract landscape simplification, and to provide resources for service-providers [7].
Crop diversification increases species richness, abundance and fitness of natural enemies, which can minimize pest damage and increases yield in commercial crops [8–10]. However, resource diversification studies usually focus only on natural enemies and how species abundance and richness are affected by attractiveness, phenology, floral accessibility, seeds, pollen and nectar production of non-crop plants [11–16]. Notwithstanding, diversity indices, richness and abundance are not sufficient to describe changes in community structure, mainly by ignoring interaction between species [17]. Moreover, these descriptors are superficial and may be of little use describing plant/crop outcome, ecosystem functioning and provision of essential services such as the biological control of arthropod pest species [18, 19].
Food web analysis provides an alternative approach and constitutes a powerful mechanism for quantifying species interactions at different trophic levels in a community and revealing functional biodiversity components in an ecological network [20, 21]. This representative approach allows analysis of similarities between complex systems originated from different environments [22, 23], providing robust information about community structure and insights into the dynamic processes associated with their structuring, including robustness to species loss, stability of herbivore species population and susceptibility to species invasion [21, 24–27].
Furthermore, fine and more comprehensive details about ecological community descriptors can be visualized using a food web trivariate approach, combining data from numerical abundance, body size and biomass abundance [28]. Overviews of literature reports dealing with flower resources and biological control enhancements do not describe or confirm alteration in species interaction, energy flow and changes in trophic webs levels. The quantification of relationships between plants and arthropod species through food webs can clarify the ecological mechanisms favouring the biological control of pest species in diversified environments. Here, in addition to resulting in higher abundance and richness, it was hypothesized that presence of floral resources also modifies community dynamics in a horticultural commercial sized crop, based on a smallholder organic model farm, optimizing provision of essential ecosystem services such as biological control, which can result in a more stable agroecosystem. Thus, the objective of this study was to evaluate how food web complexity is affected by the presence of additional floral resources in agroecosystem community networks.
Material and methods
Ethic statement
This study did not involve any endangered or protected species and no specific permission was required.
Experimental site
The study was carried out at the Horticultural Experimental Station of the Federal University of Lavras (Lavras, Minas Gerais, Brazil; 21°13'51.06"S, 44°58'34.36” W) from September/2012 to January/2013, representing a smallholder organic model farm, with 2 ha area with cultivated with diversified vegetables crops (e.g., tomato, cabbage, kale, aromatic plants). The field experiments were performed in a 0.1 ha area (20 m width and 50 m length), representing a commercial sized crop, isolated from the native vegetation by a 2 m border maintained without plants. The cultivated area was divided in beds (1.2 m width, 45 m length) spaced 0.5 m from one another composed by approximately 30 grids (1.7 m width, 1.2 m length) (Fig 1).
Lettuce (Lactuca sativa var. Solaris) was used as a model vegetable crop, due to its low defense against herbivores during commercial stage. Seeds were obtained from Seminis© (Curitiba, PR, Brazil) and seedlings were cultivated in nursery for 30 days before transplantation to the field, in which lettuce plants were cultivated in four lines per bed using 0.3 x 0.35 m spacing.
Plants of the Mexican marigold Tagetes erecta L. were used as a model floral resource due to its capacity to attract and sustain natural enemies in agroecosystems [29]. Mexican marigold seeds were obtained from ISLA© Sementes (Porto Alegre, RS, Brazil) and seedlings were cultivated in nursery for 20 days before transplant to the field. They were cultivated as a floral resource in the border and between the beds of the vegetable crop field in small linear plots with 1.7 m and plant spacing of 0.15 m; within each bed, the plots were 5.1 m apart from each other (Fig 1).
Lettuce plants were cultivated in three consecutive commercial cycles (35 days each) alongside floral resource plots with T. erecta plants (which were not changed during the three lettuce cycles) allowing evaluation of possible alterations in food web composition with development stage of the T. erecta, which was classified as either: 1) pre-flowering, encompassing the period between seedling transplant into the field, 30 days after germination, until the appearing of the floral buds; 2) complete flowering, containing all plants with flowers, immature seeds and few floral buds; and 3) late-flowering, with the majority of the plants exhibiting flowers with mature seeds, few flowers with immature seeds and almost no floral buds.
Arthropod and plant sampling
Arthropod and lettuce plant sampling were performed 15, 21, 28 and 35 days after seedling transplant to the crop field in all the three crop cycles, in six randomly selected grids per week. The average of three plants per grid constituted a replicate, totalling 18 lettuce per week. During the process, the entire plants were covered by a translucent plastic bag and totally removed from the field. Care was taken to avoid contact with the plant before covering to allow collection of fast moving arthropods.
Fresh biomasses of lettuce plants were determined with an analytical balance (Shimadzu AUW220D, Kyoto, Japan), aiming to measure the relationship between floral resources and biological control in the agroecosystem. The arthropods from each sampled lettuce were subsequently identified and counted to obtain species richness (S) and numerical abundance (N; individuals/m2). Specific body mass (M; mg) was obtained by the average of weighing 10 arthropods from each species on an electronic scale (model XS3DU, Mettler Toledo, Columbus, OH, EUA). Aiming to analyze energy flow patterns in the community [28], biomass abundance (B; mg/m2) was calculated, by multiplying numerical abundance per plant by specific body mass (i.e., B = N x M).
The collected specimens were maintained in vials with 70% alcohol until the identification process, which was performed by trained personnel using available taxonomic keys and reference collections from the Entomology Museum of the Federal University of Lavras (S1 Table).
The average number of T. erecta flowers per plot/week was determined by randomly sampling 8 floral resource plots per week.
Food web connections
Interaction among species was first recognized by in-field visual observation. Specimens of arthropods were also collected and reared in laboratory, which permited the assessment of herbivory, parasitism, hyperparasitism and predatory behaviours [30, 31]. This step allowed to recognize approximately 82% of the trophic interaction observed (S2 Table). All relationships among the different arthropod species were confirmed by comparison with existing literature [32, 33]. Complementary interactions (19%) were computed based exclusively in the literature register. In a food web with cannibalism cycles (A eats A) and mutual predation loops (A eats B eats A, or longer) a species may have different trophic positions, depending on which food chain is specified. As stated in the literature, cannibalism cycles and mutual predation loops were ignored in computing trophic levels, to assure that they are all finite, since loops makes it impossible to delimit the food web [28].
Data analyses
Data from species richness (S), numerical abundance (N), body mass (M), biomass abundance (B) and all connections registered between arthropods (S2 and S3 Tables) were used for the food web parameters computation (Table 1) [28, 34, 35]. All parameters were calculated in R [36] using the cheddar package [37].
Trophic webs from pre-flowering, flowering and late flowering period were plotted using the trivariate pattern N, M and B, which were log-transformed with base 10 to reduce the dimensionality of the data [28]. Trends of N, M and B distribution over time for each trophic level were analyzed using multiple linear regressions. The curve-fitting software TableCurve 3D (Systat, San Jose, CA, USA) was used to estimate the best fit of the equation. Low error, high F-value and R2 were the parameters used for model selection and validation. In addition, linear regression was used to assess the effect of the number of T. erecta flowers on species richness, numerical abundance and biomass abundance for each trophic level.
Results
Species richness
A total of 57 arthropod taxa were collected throughout the three lettuce cultivation cycles and assembled in food webs (Figs 2, 3 and 4; Table 2). Remarkable differences in species richness were registered during different floral resource developmental stages. In the pre-flowering cultivation, the entire food web was composed by 27 taxa, increasing to 49 and 56 during flowering and late flowering respectively (Table 2). Linear regression indicated a direct relationship between alterations in richness and the increasing number of T. erecta flowers in the field (y = 0.94 + 0.16x; R2 = 0.39; F = 46.09; P < 0.001). Although the difference in richness was significant, the arthropods collected were distributed across the 5 trophic levels with at least one representative of each level in each the three cultivations (Figs 2, 3 and 4).
The width of the black, gray and white horizontal bars shows log10 body mass (mg), log10 numerical abundance (individuals/m2) and log10 biomass abundance (mg/m2). Solid lines represent direct consumption interaction. Dotted line represents presence, but without direct consumption of living tissues. Species numbers are identified in Table 2. The vertical position indicates trophic level. Horizontal position is arbitrary. Isolated species, cannibalism or loops are ignored.
The width of the black, gray and white horizontal bars shows log10 body mass (mg), log10 numerical abundance (individuals/m2) and log10 biomass abundance (mg/m2). Solid lines represent direct consumption interaction. Dotted line represents presence, but without direct consumption of living tissues. Species numbers are identified in Table 2. The vertical position indicates trophic level. Horizontal position is arbitrary. Isolated species, cannibalism or loops are ignored.
The width of the black, gray and white horizontal bars shows log10 body mass (mg), log10 numerical abundance (individuals/m2) and log10 biomass abundance (mg/m2). Solid lines represent direct consumption interaction. Dotted line represents presence, but without direct consumption of living tissues. Species numbers are identified in Table 2. The vertical position indicates trophic level. Horizontal position is arbitrary. Isolated species, cannibalism or loops are ignored.
Food web properties
The increasing richness, probably triggered by the presence of floral resources, also affected the number of trophic links during the flower and late flowering cultivation periods, which were two and three-times greater, respectively, when compared with pre-flowering. Link density, which is directly dependent on link number and richness, also followed the same ascending pattern as the Mexican marigold developed (2.778, 3.878 and 4.578 in pre-flowering, flowering and late flowering stages respectively), (Table 3).
The proportions of taxa in each food web status (producer, consumer) varied during experiment (Table 3). Producer (lettuce plants) represented 7% of all taxa the in the pre-flowering cultivation, decreasing to 4 and 3% in flowering and late flowering cultivation periods respectively. In contrast, the proportion of intermediate taxa rose from 70% in pre-flowering to 73% in late flowering, due to additions of parasitoids and predators in this category, which were also responsible for lower prey-to-consumer ratios in flowering and late flowering treatments (Table 3). The insertion of intermediate taxa affected the probability of any two species interacting with each other (see connectance in Table 3) shortening the average distance from resource to consumers (see average link length Table 3). However, this compartmentalization resulted in an ascending average chain length from pre-flowering to late flowering cultivation (Table 3).
The additions of detritivores, herbivores, predators and parasitoids changed not only community composition, but also the asymmetries between prey and consumers during the experiment (Figs 2, 3 and 4). As an example, generality was affected by the development of flowers in the marigold T. erecta, increasing from 3.00 during the pre-flowering cultivation, to 4.04 during flowering and 4.75 in late flowering (Table 3). The same pattern was registered for the food web vulnerability, which responded to augmentative number of flowers in the field increasing by 60% its value between the pre-flowering (3.57) and late flowering cultivation (5.93) (Table 3).
Numerical abundance
The total food web numerical abundance, including all arthropods collected, was also altered by the floral resource developmental stages, responding positively to the number of flowers per plot (y = 4.05 + 1.53x; R2 = 0.51; F = 73.08; P < 0.001). However, the same pattern was not registered in all trophic levels. The abundance of natural enemies located on the third (y = 5.35 + 1.24x; R2 = 0.63; F = 120.37; P < 0.001) and fourth trophic levels (y = 4.21 + 0.30x; R2 = 0.24; F = 17.96; P < 0.001), in addition to detritivores (trophic level 0) (y = 20.14 + 0.95x; R2 = 0.57; F = 96.15; P < 0.001), positively responded to the presence of flowers in the field. Inversely to the natural enemies, herbivores located in the second trophic level significantly decreased in abundance as marigold flowering advanced (y = 140.41–1.5x; R2 = 0.13; F = 9.83; P = 0.002). The responses of some specific arthropod groups deserve attention, such as the aphids, whose abundance also responded negatively to the number of T. erecta flowers in the field (y = 1.33–0.51x; R2 = 0.29; F = 29.21; P < 0.001), decreasing during flowering and late flowering periods, probably due to the increasing number of natural enemies. Another interesting response was registered for thrips (Thysanoptera), encompassing two of the agricultural pest species collected during the study, Caliothrips phaseoli Hood, 1912 and Frankliniella schultzei schultzei (Trybom, 1910) (Thripidae), which significantly decreased with the presence of flowers in the field (y = 0.6–0.22x; R2 = 0.26; F = 25.46; P < 0.001). Conversely, harmless thrips species such as Echinothrips mexicanus Moulton, 1911 and Neohydatothrips gracilipes Hood, 1924 (Thripidae), which were sampled in lettuce plants, increased their numerical abundance during the experiment (y = - 0.005–0.102x; R2 = 0.30; F = 31.24; P < 0.001), denoting a specific substitution triggered by floral resources.
The observed data were subsequently subjected to multiple regression analysis and a polynomial model equation (z = a + bx + cy + dy2 + ey3 + fy4 + gy5) was obtained to explain the dispersion of numerical abundance in the different trophic levels mediated by the increasing number of marigold flowers (R2 = 0.60; F6, 425 = 107.25; P < 0.001) (Fig 5A) and used to generate the simulated distribution (Fig 5B). Numerical abundance increased in response to a higher number of flowers (Fig 5A).
The color-graded scale at the right of each plot represents the level of either abundance (a,b), body size (c,d), or biomass (e,f) distribution of the lettuce-associated arthropods. The lowest to the highest levels are represented by the transition from dark blue to dark red respectively.
Body mass
Total body mass was not affected by the addition of floral resources in the field. However, specific effects were noticed in particular trophic levels. Average body mass of herbivores decreased during the marigold flowering (y = 10.71–6.14x; R2 = 0.23; F = 21.99; P < 0.001). In contrast, specialist natural enemies located in third trophic level increased their average body mass with the presence of floral resources in the field (y = - 6.14x + 10.71; R2 = 0.23; F = 21.99; P < 0.001). A more complex X-Y polynomial model (z = a + bx + cx2 + dx3 + ey + fy2 + gy3 + hy4 + iy5) better fitted the body mass distribution in different trophic levels during the marigold flowering (R2 = 0.91; F8, 423 = 541.35; P < 0.001) (Fig 5C). The model was used to simulate the distribution pattern (Fig 5D). In both graphics were registered increasing body sizes in higher trophic levels.
Biomass abundance
Total biomass increased due to presence of flower resources in the field (y = 0.58 + 0.14x; R2 = 0.44; F = 55.67; p > 0.001). However, change in total biomass reflected the increasing biomass in specific trophic levels, particularly of detritivores (y = -0.13 + 0.51x; R2 = 0.70; F = 168.82; P < 0.001) and natural enemies located in third trophic level (y = 0.37 + 0.92x; R2 = 0.45; F = 58.5; P < 0.001). Following the same pattern registered for numerical abundance, aphid biomass decreased with the presence of higher number of flowers per plot (y = 0.86–0.44x; R2 = 0.50; F = 71.98; P < 0.001). Data was assessed by multiple regression analysis and the non-linear extreme value model equation (z = a + bx + cexp(−exp(−(y−d)/e)−(y−d)/e + 1)) was selected to explain the dispersion of biomass abundance in the five trophic levels mediated by increasing flower number during the three field cultivations (R2 = 0.85; F4, 427 = 604.28; P < 0.001) (Fig 5E). The model was used to access the simulated distribution (Fig 5F) showing translocation of biomass from basal to higher trophic levels responding to a high number of flowers in field.
Discussion
Planting of T. erecta within the crop modified the arthropod community associated with lettuce plants, changing the structure, biomass distribution, assumed species interactions and consumer/prey properties in the arthropod food webs. However, identifying the mechanisms involved in these changes is complex and requires a comprehensive approach [38]. In the first scenario, the presence of non-crop plants in the agroecosystem likely diversified microhabitats, allows the recruitment and conservation of species with different dietary, ecological and habitat requirements mainly due to the increasing availability of floral resources, which results in a diversified community [39–41]. The increasing complexity in the lettuce crop, arising from flowers, nectar and shelter, partially explains the higher richness and abundance in lettuce recorded during the marigold flowering stages [42]. The main trophic levels responsible for species richness increase were detritivores (ranked as trophic level 0), composed principally of springtails; and natural enemies, either specialists (third trophic level) or generalists (fourth trophic level), whose positive responses to the increasing complexity of the environment have been previously reported [43, 44].
Secondly, the availability of alternative food resources might also explain the observed changes in arthropod community. Increased organic matter arising from plant diversification probably benefited detritivorous arthropods [45]. As a third scenario, parasitoids and predators in the third trophic level were likely assisted directly or indirectly by the presence of non-prey food items in the marigold plants. Specialist natural enemies, mainly parasitoids, depend on these alternative food resources at least during part of their life cycle, and the absence of these resources, as during the marigold vegetative stage, has been shown to directly affect the abundance and richness of this group in the field [46, 47]. Finally, generalist natural enemies, especially spiders, are in a unique position in that they can potentially benefit from the three previous scenarios feeding on non-prey food from floral resources [12, 47], on specialist natural enemies, and on detritivores [45].
Higher richness of natural enemies consequently leads to higher number of trophic links in the food web, modifying the entire community structure and composition through a combination of bottom-up and top-down effects. The increasing number of links, sometimes with more than one prey per natural enemy species, may potentially reduce the abundance of harmful herbivores and their damage to the crop field [48, 49]. As an example, in this study there was a significant reduction in the abundance of herbivores, including arthropod pest species, during the marigold flowering stage. This reduction may be linked to the population increase of a second species (a detritivore, or herbivore present in the crop, or non-crop plant itself) mediating numerical increase of a third species located in a higher trophic level (generalist or specialist natural enemy). This process is known as "apparent competition" in which natural enemies shared by more than one prey species result in a negative interaction among the prey species [50].
Comparative analysis on the food web properties resulting from the different lettuce cultivations also demonstrates the community impact arising from the alteration of natural enemy numerical abundance. The susceptibility to predation in a food web can be measured using two parameters: generality (i.e. the average number of prey for natural enemy species), and vulnerability (i.e. the average number of consumers per prey). Increases of these two parameters as observed in these study, especially vulnerability, reflects changes in top-down control and predation strength, which results in large changes in biomass registered in the study, potentially reflecting the degree of energy propagation in the food web [51–54]. In addition to the presence/absence of natural enemies, another factor affecting the generality and vulnerability of the trophic network is the body size of consumers or prey, which may limit consumption in terrestrial environments [51]. In this study, during the Mexican marigold flowering stage, both the higher richness and abundance of natural enemies and the reduction of prey body size potentially were responsible for increasing community vulnerability and generality.
Curiously, connectance (L/S2) declined with increasing species richness, usually, food webs with low connectance are highly sensitive to species loss and this sensitivity tends to decrease with increasing connectance [55]. Theoretical research hypothesized that natural food webs in more complex environments are characterized by having a large number of weak interactions and a small numbers of strong interactions [56]. Therefore, rich food webs are more connected with weaker interactions between species on average, and this low strength is probably responsible for high stability even at low connectances [57, 58]. Thus, rich environments as observed during T. erecta flowering stages tend to be more robust to the loss of abundant less connected species and sensitive to loss of the rare more connected ones [22].
Omnivorism is another factor that can contribute to stability in diversified communities. Theoretical studies suggest that higher omnivory rates, as were observed during the flowering stage of marigold, might favour the stabilization of food webs with frequent, but weak, interactions [59, 60]. Empirical studies also suggest the stabilizing effect of omnivory in communities [61–63]. In the present study, increasing omnivorism was associated with the insertion of natural enemies at higher trophic levels of the food web. In general, environmental degradation causes a reduction in ecosystem size, especially through the loss of top predators, subsequently shortening the food chain length, as reported in the marigold vegetative stage [64, 65]. Food chain length is one of the community properties most affected by human activities [66, 67], and comparative study of observed food chain length originating from different lettuce cultivation allows us to hypothesize about the impact of the floral resources in decreasing impacts of agriculture management in the arthropod food web.
Earlier theoretical studies have demonstrated that longer food chain length tends to be less stable than shorter chain length [68–70]. However, recent studies have suggested that systems with long food chain length exhibit more stable population dynamics. The presence or absence of predatory species influences food web stabilization through its effects on herbivore populations [71], which are most affected by the addition of omnivorous species in the highest trophic levels [72]. In this study, the deleterious effect of diversification on the herbivore population was demonstrated by the systematic biomass loss as floral resources increased, which coincides with the longer food chain length and the higher rate of omnivorism registered. Lower stability between herbivore populations caused by natural enemy pressure was determined by the decline of dominant species, particularly important agricultural pests such as aphids. This disruption in dominance may also have allowed the recruitment and entrance of harmless herbivore species in the crop field.
The analyses of food web parameters allows us to hypothesize that agroecosystem diversification deeply affects the arthropod community dynamics, redirecting numerical abundance and biomass from herbivores to higher trophic levels, inducing a decrease in average body size of herbivores/natural enemies and potentially resulting in a more stable food web. Knowing the effect of additive floral resource on properties of crop food webs may allows optimization of floral resource use in integrated pest management programs by allowing the identification and matching of higher abundance and richness of beneficial arthropods with critical stages of the crop. As suggested in this study, the use of T. erecta plants as a floral resource during their flowering stage may prevent asynchronous colonization between natural enemies and pest population in field, considered the biggest challenge to effective natural biological control [73]. One limitation of this approach is related with the recognition of species feeding interaction in field. Next-generation sequencing is probably an important tool for these studies in the future, which is able to identify all DNA in digestive system of natural enemies and confirm interactions.
Our results suggested the addition of floral resources performs an effective and environmentally friendly strategy for farmers in the pest management. Food web complexity registered during flowering stages provided higher richness and abundance of specialists and generalists natural enemies in the crop, which increased community stability and functioning, benefiting biological control and avoiding outbreaks of herbivore population, such as aphids. Future studies should be conducted using other floral resources or different cultures to clarify the ecological relationships determining natural biological control. The effect of different external factors, such management practices, in the food web should also be tested in the future in order to contribute with a more sustainable management of harmful arthropods in the field.
Supporting information
S1 Table. Species and functional groups of arthropods associated with lettuce plants during the flowering stages of T. erecta.
https://doi.org/10.1371/journal.pone.0193045.s001
(PDF)
S2 Table. Trophic links registered for the arthropods registered in the lettuce plants during the flowering stages of T. erecta.
https://doi.org/10.1371/journal.pone.0193045.s002
(PDF)
S3 Table. Raw number of flowers, trophic level and number of individuals (m2) associated with lettuce plants during the experiment.
https://doi.org/10.1371/journal.pone.0193045.s003
(PDF)
Acknowledgments
We thank R. N. C. Guedes for comments on the manuscript and statistical analysis. We thank V. Tomazella and C. F. Haro for providing technical assistance.
References
- 1. Clough Y, Kruess A, Tscharntke T. Local and landscape factors in differently managed arable fields affect the insect herbivore community of a non-crop plant species. Journal of Applied Ecology. 2007;44(1):22–8.
- 2. Rusch A, Valantin-Morison M, Sarthou JP, Roger-Estrade J. Effect of crop management and landscape context on insect pest populations and crop damage. Agriculture, Ecosystems & Environment. 2013;166(0):118–25. http://dx.doi.org/10.1016/j.agee.2011.05.004.
- 3. Chapin FS III, Zavaleta ES, Eviner VT, Naylor RL, Vitousek PM, Reynolds HL, et al. Consequences of changing biodiversity. Nature. 2000;405(6783):234–42. pmid:10821284
- 4. Schmidt MH, Thies C, Nentwig W, Tscharntke T. Contrasting responses of arable spiders to the landscape matrix at different spatial scales. Journal of Biogeography. 2008;35(1):157–66.
- 5. Perović DJ, Gurr GM, Raman A, Nicol HI. Effect of landscape composition and arrangement on biological control agents in a simplified agricultural system: A cost–distance approach. Biological Control. 2010;52(3):263–70. http://dx.doi.org/10.1016/j.biocontrol.2009.09.014.
- 6. Landis DA, Wratten SD, Gurr GM. Habitat Management to Conserve Natural Enemies of Arthropod Pests in Agriculture. Annual Review of Entomology. 2000;45(1):175–201. pmid:10761575.
- 7. Jonsson M, Wratten SD, Landis DA, Gurr GM. Recent advances in conservation biological control of arthropods by arthropods. Biological Control. 2008;45(2):172–5. http://dx.doi.org/10.1016/j.biocontrol.2008.01.006.
- 8.
Heimpel GE, Jervis MA. Does floral nectar improve biological control by parasitoids? Plant-Provided Food for Carnivorous Insects. In: Wäckers F, van Rijn PCJ, Bruin J, editors. Plant-provided Food for Carnivorous Insects: A Protective Mutualism and its Applications. Cambridge: Cambridge University Press; 2005. p. 267–304.
- 9. Cullen R, Warner KD, Jonsson M, Wratten SD. Economics and adoption of conservation biological control. Biological Control. 2008;45(2):272–80. http://dx.doi.org/10.1016/j.biocontrol.2008.01.016.
- 10. Griffiths GJK, Holland JM, Bailey A, Thomas MB. Efficacy and economics of shelter habitats for conservation biological control. Biological Control. 2008;45(2):200–9. http://dx.doi.org/10.1016/j.biocontrol.2007.09.002.
- 11. Baggen LR, Gurr GM, Meats A. Flowers in tri-trophic systems: mechanisms allowing selective exploitation by insect natural enemies for conservation biological control. Entomologia Experimentalis et Applicata. 1999;91(1):155–61.
- 12. Wäckers FL. Assessing the suitability of flowering herbs as parasitoid food sources: flower attractiveness and nectar accessibility. Biological Control. 2004;29(3):307–14. http://dx.doi.org/10.1016/j.biocontrol.2003.08.005.
- 13. Rebek EJ, Sadof CS, Hanks LM. Manipulating the abundance of natural enemies in ornamental landscapes with floral resource plants. Biological Control. 2005;33(2):203–16. http://dx.doi.org/10.1016/j.biocontrol.2005.02.011.
- 14. Hickman JM, Wratten SD. Use of Phacelia tanacetifolia strips to enhance biological control of aphids by hoverfly larvae in cereal fields. Journal of Economic Entomology. 1996;89(4):832–40.
- 15. Nicholls CI, Parrella MP, Altieri MA. Reducing the abundance of leafhoppers and thrips in a northern California organic vineyard through maintenance of full season floral diversity with summer cover crops. Agricultural and Forest Entomology. 2000;2(2):107–13.
- 16. Ambrosino MD, Luna JM, Jepson PC, Wratten SD. Relative Frequencies of Visits to Selected Insectary Plants by Predatory Hoverflies (Diptera: Syrphidae), Other Beneficial Insects, and Herbivores. Environmental Entomology. 2006;35(2):394–400.
- 17. Tylianakis JM, Tscharntke T, Lewis OT. Habitat modification alters the structure of tropical host-parasitoid food webs. Nature. 2007;445(7124):202–5. http://www.nature.com/nature/journal/v445/n7124/suppinfo/nature05429_S1.html. pmid:17215842
- 18. Kremen C. Managing ecosystem services: what do we need to know about their ecology? Ecology Letters. 2005;8(5):468–79. pmid:21352450
- 19. Tscharntke T, Klein AM, Kruess A, Steffan-Dewenter I, Thies C. Landscape perspectives on agricultural intensification and biodiversity–ecosystem service management. Ecology Letters. 2005;8(8):857–74.
- 20. McCann K. Protecting biostructure. Nature. 2007;446(7131):29–. pmid:17330028
- 21. Morris RJ, Lewis OT, Godfray HCJ. Experimental evidence for apparent competition in a tropical forest food web. Nature. 2004;428(6980):310–3. pmid:15029194
- 22. Strogatz SH. Exploring complex networks. Nature. 2001;410(6825):268–76. pmid:11258382
- 23. Newman MEJ. The Structure and Function of Complex Networks. SIAM Review. 2003;45(2):167. pmid:9924428.
- 24. Müller CB, Adriaanse ICT, Belshaw R, Godfray HCJ. The structure of an aphid–parasitoid community. Journal of Animal Ecology. 1999;68(2):346–70.
- 25. Dunne JA, Williams RJ, Martinez ND. Network structure and biodiversity loss in food webs: robustness increases with connectance. Ecology Letters. 2002;5(4):558–67.
- 26. Melián CJ, Bascompte J. Food Web Cohesion. Ecology. 2004;85(2):352–8.
- 27. Ives AR, Cardinale BJ. Food-web interactions govern the resistance of communities after non-random extinctions. Nature. 2004;429(6988):174–7. http://www.nature.com/nature/journal/v429/n6988/suppinfo/nature02515_S1.html. pmid:15141210
- 28. Cohen JE, Jonsson T, Carpenter SR. Ecological community description using the food web, species abundance, and body size. Proceedings of the National Academy of Sciences. 2003;100(4):1781–6. pmid:12547915
- 29. Silveira LCP, Berti Filho E, Pierre LSR, Peres FSC, Louzada JNC. Marigold (Tagetes erecta L.) as an attractive crop to natural enemies in onion fields. Scientia Agricola. 2009;66:780–7.
- 30.
Memmott J, Alonso D, Berlow E, Dobson A, Dunne JA, Solé R, et al. Biodiversity loss and ecological network structure. In: Pascual M, Dunne JA, editors. Ecological Networks: Linking structure to dynamics in food webs. Oxford: Oxford University Press; 2006. p. 325–47.
- 31. Eveleigh ES, McCann KS, McCarthy PC, Pollock SJ, Lucarotti CJ, Morin B, et al. Fluctuations in density of an outbreak species drive diversity cascades in food webs. Proceedings of the National Academy of Sciences. 2007;104(43):16976–81. pmid:17940003
- 32.
Harper-Smith S, Berlow EL, Knapp RA, Williams RJ, Martinez ND. 7.5—Communicating ecology through food webs: Visualizing and quantifying the effects of stocking alpine lakes with trout. In: Melville-Smith PdRWCM, editor. Dynamic Food Webs. 3. Burlington: Academic Press; 2006. p. 407–23.
- 33. O'Gorman EJ, Emmerson M, O'Gorman EJ, Emmerson M. Manipulating Interaction Strengths and the Consequences for Trivariate Patterns in a Marine Food Web. Advances in Ecological Research: Ecological Networks. 2010;42:301–419.
- 34. Bersier L-F, Banašek-Richter C, Cattin M-F. Quantitative descriptors of food-web matrices. Ecology. 2002;83(9):2394–407.
- 35. Cohen JE, Schittler DN, Raffaelli DG, Reuman DC. Food webs are more than the sum of their tritrophic parts. Proceedings of the National Academy of Sciences. 2009;106(52):22335–40. pmid:20018774
- 36.
R Development Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013.
- 37. Hudson LN, Emerson R, Jenkins GB, Layer K, Ledger ME, Pichler DE, et al. Cheddar: analysis and visualisation of ecological communities in R. Methods in Ecology and Evolution. 2013;4(1):99–104.
- 38. Kremen C, Miles A. Ecosystem Services in Biologically Diversified versus Conventional Farming Systems: Benefits, Externalities, and Trade-Offs. Ecology and Society. 2012;17(4).
- 39. Rusek J. Microhabitats of Collembola (Insecta: Entognatha) in beech and spruce forests and their influence on biodiversity. European Journal of Soil Biology. 2001;37(4):237–44. http://dx.doi.org/10.1016/S1164-5563(01)01090-1.
- 40.
Rosenzweig ML. Species diversity in space and time. Cambridge: Cambridge University Press; 1995.
- 41. Stephan JG, Albertsson J, Wang L, Porcel M. Weeds within willow short-rotation coppices alter the arthropod community and improve biological control of the blue willow beetle. BioControl. 2016;61(1):103–14.
- 42. Schneider G, Krauss J, Riedinger V, Holzschuh A, Steffan-Dewenter I. Biological pest control and yields depend on spatial and temporal crop cover dynamics. Journal of Applied Ecology. 2015;52(5):1283–92.
- 43. Sousa JP, Bolger T, Da Gama MM, Lukkari T, Ponge JF, Simon C, et al. Changes in Collembola richness and diversity along a gradient of land-use intensity: a pan European study. Pedobiologia. 2006;50(2):147–56. PubMed PMID: CC:0002386629–0009.
- 44. Chaplin-Kramer R, O’Rourke ME, Blitzer EJ, Kremen C. A meta-analysis of crop pest and natural enemy response to landscape complexity. Ecology Letters. 2011;14(9):922–32. pmid:21707902
- 45. Sunderland K, Samu F. Effects of agricultural diversification on the abundance, distribution, and pest control potential of spiders: a review. Entomologia Experimentalis et Applicata. 2000;95(1):1–13.
- 46. Coll M, Guershon M. Omnivory in terrestrial arthropods: Mixing Plant and Prey Diets. Annual Review of Entomology. 2002;47(1):267–97. pmid:11729076.
- 47.
Wäckers FL, van Rijn PCJ, Bruin J. Plant-provided food for carnivorous insects: a protective mutualism and its applications. Cambridge: Cambridge University Press Cambridge; 2005.
- 48. Letourneau DK, Armbrecht I, Rivera BS, Lerma JM, Carmona EJ, Daza MC, et al. Does plant diversity benefit agroecosystems? A synthetic review. Ecological Applications. 2010;21(1):9–21.
- 49. Koricheva J, Vehviläinen H, Riihimäki J, Ruohomäki K, Kaitaniemi P, Ranta H. Diversification of tree stands as a means to manage pests and diseases in boreal forests: myth or reality? Canadian Journal of Forest Research. 2006;36(2):324–36.
- 50. Holt RD. Predation, apparent competition, and the structure of prey communities. Theoretical population biology. 1977;12(2):197–29. Epub 1977/10/01. pmid:929457.
- 51. Sinclair ARE, Mduma S, Brashares JS. Patterns of predation in a diverse predator-prey system. Nature. 2003;425(6955):288–90. http://www.nature.com/nature/journal/v425/n6955/suppinfo/nature01934_S1.html. pmid:13679915
- 52. Chase JM. Food Web Effects of Prey Size Refugia: Variable Interactions and Alternative Stable Equilibria. The American naturalist. 1999;154(5):559–70. Epub 1999/11/24. pmid:10561128
- 53. Bishop MJ, Kelaher BP, Alquezar R, York PH, Ralph PJ, Greg Skilbeck C. Trophic cul-de-sac, Pyrazus ebeninus, limits trophic transfer through an estuarine detritus-based food web. Oikos. 2007;116(3):427–38.
- 54. Atlas WI, Palen WJ. Prey Vulnerability Limits Top-Down Control and Alters Reciprocal Feedbacks in a Subsidized Model Food Web. PLoS ONE. 2014;9(1):e85830. pmid:24465732
- 55. Estrada E. Food webs robustness to biodiversity loss: the roles of connectance, expansibility and degree distribution. Journal of theoretical biology. 2007;244(2):296–307. Epub 2006/09/22. pmid:16987531.
- 56. Neutel A-M, Heesterbeek JAP, Ruiter PCd. Stability in Real Food Webs: Weak Links in Long Loops. Science. 2002;296(5570):1120–3. pmid:12004131
- 57. McCann K, Hastings A, Huxel GR. Weak trophic interactions and the balance of nature. Nature. 1998;395(6704):794–8.
- 58. Kokkoris GD, Jansen VAA, Loreau M, Troumbis AY. Variability in interaction strength and implications for biodiversity. Journal of Animal Ecology. 2002;71(2):362–71.
- 59. Borrvall C, Ebenman B, Tomas Jonsson TJ. Biodiversity lessens the risk of cascading extinction in model food webs. Ecology Letters. 2000;3(2):131–6.
- 60.
Emmerson M, Yearsley JM. Weak interactions, omnivory and emergent food-web properties2004 2004-02-22 00:00:00. 397–405 p.
- 61. Lawler SP, Morin PJ. Food web architecture and population dynamics in laboratory microcosms of protists. American Naturalist. 1993;141(5):675–86. Epub 1993/05/01. pmid:19426005.
- 62. Fagan WF. Omnivory as a stabilizing feature of natural communities. American Naturalist Journal. 1997;150(5):554–67. Epub 2008/09/25. pmid:18811300.
- 63. Holyoak M, Sachdev S. Omnivory and the stability of simple food webs. Oecologia. 1998;117(3):413–9. pmid:28307921
- 64. Odum EP. Trends Expected in Stressed Ecosystems. BioScience. 1985;35(7):419–22.
- 65. Petchey OL, Downing AL, Mittelbach GG, Persson L, Steiner CF, Warren PH, et al. Species loss and the structure and functioning of multitrophic aquatic systems. Oikos. 2004;104(3):467–78.
- 66. Pauly D, Christensen V, Dalsgaard J, Froese R, Torres F. Fishing Down Marine Food Webs. Science. 1998;279(5352):860–3. pmid:9452385
- 67. Pauly D, Palomares ML, Froese R, Sa-a P, Vakily M, Preikshot D, et al. Fishing down Canadian aquatic food webs. Canadian Journal of Fisheries & Aquatic Sciences. 2001;58(1):51. pmid:8773515.
- 68. Pimm SL. The complexity and stability of ecosystems. Nature. 1984;307(5949):321–6.
- 69. Pimm SL, Kitching RL. The Determinants of Food Chain Lengths. Oikos. 1987;50(3):302–7.
- 70. Lawton JH, Warren PH. Static and dynamic explanations for patterns in food webs. Trends in Ecology & Evolution. 1988;3(9):242–5. http://dx.doi.org/10.1016/0169-5347(88)90167-X.
- 71. Halpern BS, Borer ET, Seabloom EW, Shurin JB. Predator effects on herbivore and plant stability. Ecology Letters. 2005;8(2):189–94.
- 72. Long ZT, Bruno JF, Duffy JE. Food chain length and omnivory determine the stability of a marine subtidal food web. The Journal of animal ecology. 2011;80(3):586–94. Epub 2011/01/22. pmid:21250990.
- 73. Kean J, Wratten S, Tylianakis J, Barlow N. The population consequences of natural enemy enhancement, and implications for conservation biological control. Ecology Letters. 2003;6(7):604–12.