Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2022 Jun 22;32(7):e2653. doi: 10.1002/eap.2653

An aggressive nonconsumptive effect mediates pest control and multipredator interactions in a coffee agroecosystem

Jonathan R Morris 1,✉, Ivette Perfecto 1
PMCID: PMC9788006  PMID: 35543106

Abstract

Natural pest control is an alternative to pesticide use in agriculture, and may help to curb insect declines and promote crop production. Nonconsumptive interactions in natural pest control that historically have received far less attention than consumptive interactions, may have distinct impacts on pest damage suppression and may also mediate positive multipredator interactions. Additionally, when nonconsumptive effects are driven by natural enemy aggression, variation in alternative resources for enemies may impact the strength of pest control. Here we study control of the coffee berry borer (CBB), Hypothenemus hampei, by a keystone arboreal ant species, Azteca sericeasur, which exhibits a nonconsumptive effect on CBB by throwing them off coffee plants. We conducted two experiments to investigate: (1) if the strength of this behavior is driven by spatial or temporal variability in scale insect density (an alternative resource that Azteca tends for honeydew), (2) if this behavior mediates positive interactions between Azteca and other ground‐foraging ants, and (3) the effect this behavior has on the overall suppression of CBB damage in multipredator scenarios. Our behavioral experiment showed that nearly all interactions between Azteca and CBB are nonconsumptive and that this behavior occurs more frequently in the dry season and with higher densities of scale insects on coffee branches. Our multipredator experiment revealed that borers thrown off coffee plants by Azteca can survive and potentially damage other nearby plants but may be suppressed by ground‐foraging ants. Although we found no non‐additive effects between Azteca and ground‐foraging ants on overall CBB damage, together, both species resulted in the lowest level of plant damage with the subsequent reduction in “spillover” damage caused by thrown CBB, indicating spatial complementarity between predators. These results present a unique case of natural pest control, in which damage suppression is driven almost exclusively by nonconsumptive natural enemy aggression, as opposed to consumption or prey behavioral changes. Furthermore, our results demonstrate the variability that may occur in nonconsumptive pest control interactions when natural enemy aggressive behavior is impacted by alternative resources, and also show how these nonconsumptive effects can mediate positive interactions between natural enemies to enhance overall crop damage reduction.

Keywords: ants, coffee, coffee berry borer, crop damage, multipredator interactions, natural enemy aggression, natural pest control, nonconsumptive effects

INTRODUCTION

With the increasing alarm surrounding global insect declines (Hallmann et al., 2017; van Klink et al., 2020; Wagner, 2020), a prescription that continues to emerge is the need for drastic reductions in pesticide and insecticide use (Harvey et al., 2020; Kremen & Merenlender, 2018), which appears to be one of the major drivers of the observed declines (Sánchez‐Bayo & Wyckhuys, 2019; Wagner, 2020). Natural pest control, through the conservation of natural enemy habitat in agricultural landscapes, can serve as an alternative to promote the production of crops, the regulation of pests, and the conservation of biodiversity (Bianchi et al., 2006; Dainese et al., 2019; Karp et al., 2013). One of the challenges to this approach is understanding how diverse communities of natural enemies impact the overall functioning of pest regulation (Casula et al., 2006; Griffin et al., 2013; Letourneau et al., 2009; Straub et al., 2008). However, in focusing primarily on the relationship between predator richness and pest control, this research often overlooks the complexity of interactions between predators and pests that serve as the ultimate mechanisms for pest suppression (Crowder & Jabbour, 2014). When interactions are studied in pest control, most often, direct, consumptive interactions receive the bulk of the attention (Eubanks & Finke, 2014). This is despite the growing awareness of the ubiquity of trait‐mediated and nonconsumptive interactions in communities, which, more generally, have been shown to have impacts of equivalent magnitudes on prey regulation (Preisser et al., 2005; Werner & Peacor, 2003) and plant communities through trophic cascades (Schmitz et al., 2004).

In pest control, nonconsumptive effects of natural enemies can increase pest risk and reduce pest damage by magnitudes comparable with consumptive interactions (Eubanks & Finke, 2014; Hermann & Landis, 2017; Thaler & Griffin, 2008). These interactions can have effects on pest populations that are disproportionate to the density of natural enemies, potentially serving as mechanisms for the influence of keystone biocontrol agents (Meadows et al., 2017). This may occur when the mere presence of natural enemies changes the behavior of prey to reduce pest feeding rates, although not necessarily impacting pest densities themselves (Eubanks & Finke, 2014). Importantly, nonconsumptive effects can also mediate the interactions between natural enemies (Davenport & Chalcraft, 2013) and may potentially influence the impact of natural enemy diversity on pest control (Ingerslew & Finke, 2018; Meadows et al., 2017). When enemies compete directly over shared prey resources or space, multipredator interactions often have negative impacts on prey regulation, but when predators are spatially separated or functionally distinct, nonconsumptive effects can result in positive synergistic pest regulation (Ingerslew & Finke, 2018; Meadows et al., 2017).

A particularly interesting example of this was documented by Losey and Denno (1998), in which they showed that the presence of a Coccinellid predator on plants caused aphids to drop to the ground, making them more available to a ground‐foraging beetle (Losey & Denno, 1998). This study demonstrated how the spatial separation of predators on plants and the ground, along with the nonconsumptive effect of one of the predators on the pest, enhanced overall control and resulted in positive synergistic multipredator effects (Losey & Denno, 1998). Other research on aphid dropping in multiple wasp enemy communities has found more conflicting results, in which consumptive effects between enemies resulted in interference and reduced overall prey suppression, but nonconsumptive effects yielded positive additive prey suppression (Ingerslew & Finke, 2018). Despite this research, few studies have explored the importance of nonconsumptive interactions in multienemy pest control scenarios, in which impacts on crop damage are explicitly tested (Hermann & Landis, 2017). Furthermore, the dynamics of nonconsumptive interactions are not well understood in pest control, and few studies have tested how nonconsumptive enemy behavior changes across space or time (Hermann & Landis, 2017; Sheriff et al., 2018), despite the long history of dynamical research on consumptive predator–prey interactions. This may be particularly important when nonconsumptive effects are driven by predator aggression, as opposed to prey behavioral responses, and spatial or temporal variation in alternative resources for predators impacts the strength of nonconsumptive effects, potentially resulting in variable or inconsistent pest damage suppression.

In shaded coffee agroforests, communities of natural enemies can be quite diverse, leading to a host of potential multipredator interactions and nonconsumptive effects (Perfecto et al., 2014; Vandermeer et al., 2010, 2019). Ants have received much attention in the coffee pest control literature, particularly in Latin America, where some species are known natural enemies of the coffee berry borer (CBB), Hypothenemus hampei (Ferrari 1867), (Morris et al., 2018; Philpott & Armbrecht, 2006), a major global pest of coffee, which bores into fruits and significantly reduces yield. In southern Mexico, the ecology of the arboreal ant, Azteca sericeasur (Longino 2007), has been investigated extensively for its apparent keystone role in the interaction networks of coffee farms (Vandermeer et al., 2010, 2019). It is well documented that this ant suppresses the damage of CBB, in both the laboratory (Pardee & Philpott, 2011; Philpott et al., 2012), and in the field (Gonthier et al., 2013; Jiménez‐Soto et al., 2013; Morris et al., 2015). This species is commonly observed foraging on coffee bushes in this region where it tends hemipteran insects (scale) for honeydew. When Azteca ants participate in these mutualistic relationships with scale, they can indirectly benefit coffee by patrolling plants and preying on other herbivores that might threaten the supply of honeydew resources from their scale partners (Morris et al., 2018; Perfecto & Vandermeer, 2006). However, this behavior may also drive nonconsumptive interactions with coffee herbivores, including CBB. This has been observed previously, when Azteca ants will attack CBB individuals during their colonization of coffee plants, often throwing or pushing them off plants to the ground (Jiménez‐Soto et al., 2013). Unlike many other cases of nonconsumptive enemy–prey interactions (Hermann & Landis, 2017), this effect appears to be driven by the enemy's antagonistic, but nonconsumptive impact, and not by a behavioral response of the prey.

Despite this aggressive behavior, it is still not well understood what the overall impact of this nonconsumptive interaction is on coffee pest control. Previous work in this system has reported conflicting frequencies of this behavior compared with direct consumption of CBB (Jiménez‐Soto et al., 2013; Perfecto & Vandermeer, 2006). Variability in the intensity of this interaction or in the proportion of nonconsumptive to consumptive interactions by ants may have important impacts on the dynamics and efficacy of pest control in this system. Interestingly, because this nonconsumptive interaction is driven by enemy aggression, rather than prey defense, these dynamics may be governed by the availability of honeydew resources for ants from scale insects on coffee, which previous research suggests may influence Azteca's reduction of CBB damage (Rivera‐Salinas et al., 2018). Although it is not always clear how honeydew availability (Clark & Singer, 2018) or hemipteran insect density (Kaplan & Eubanks, 2005) influences ant–plant defensive interactions more generally, research on other species of Azteca ants in this region has demonstrated that seasonal variability in plant carbon pools may impact scale insect honeydew and the strength of ant–plant defense (Pringle et al., 2013). Furthermore, it is not clear what impact this nonconsumptive effect has on overall CBB damage and the broader control of coffee pests in the community, where the flux of resources from coffee plants to the ground may mediate interactions between Azteca and other ground‐foraging predators, potentially resulting in enhanced control of pests and positive multipredator effects.

To assess these questions and gain an understanding into the natural history of this complex pest control interaction, we conducted both a field behavioral experiment and a laboratory multipredator experiment. With our behavior experiment, we aimed to test (1) if the strength of Azteca's aggressive nonconsumptive effect on CBB varies positively across space (on coffee) or time (between seasons) with the density of scale insects (an alternative resource). With our laboratory experiment we aimed to understand (2) how this behavior impacts overall CBB damage in multipredator communities and (3) if CBB throwing results in positive interactions or spatial complementarity between Azteca and ground‐foraging predators. Ultimately, we aimed to illuminate how this unique nonconsumptive effect, driven by the aggressive behavior of a dominant keystone consumer, influences the overall regulation of this important agricultural pest.

METHODS

To understand the impact of the nonconsumptive interaction between Azteca sericeasur and CBB on pest control we conducted two experiments. First, we performed a behavioral experiment in the field to investigate variability in the strength of this nonconsumptive interaction, regarding the availability of scale insect resources for Azteca across space on coffee bushes and time between seasons. Second, we conducted a laboratory experiment to assess the impact of this behavior on borer damage in multipredator scenarios and to understand how this nonconsumptive effect mediates interactions between Azteca and ground‐foraging predators. All experiments were conducted at Finca Irlanda in Chiapas, Mexico. The farm is a 300‐hectare, certified organic, shaded coffee polyculture situated at roughly 1000 m elevation at 15°11′ N, 92°20′ W. Data for the behavioral experiment were collected during July 2019 and February 2020, whereas data for the multipredator experiment were collected during October 2016.

Field behavior experiment

To assess the variability of this nonconsumptive behavior in the field we conducted a behavioral assay. We haphazardly selected 20 sites where Azteca sericeasur ants were active on coffee bushes. We chose sites that were separated by a minimum of 6.13 m (between ant nest host trees) to increase the likelihood that ants from each site represented different colonies from different shade trees, however the average distance between sites was roughly 43 m. We measured the activity of ants on coffee by counting the number of individuals crossing a fixed point on the central trunk of the coffee bushes for 1 min. We selected one coffee bush at each site with at least five individual ants observed during the 1‐min period. Then, we selected one branch on each coffee plant and measured the branch‐level ant activity by visually scanning the branch and counting all individuals. We chose branches that had a minimum of three ant individuals for our experiment. Finally, we estimated the availability of resources for Azteca ants from scale insects by counting the number of adult scale insects on branches of a few common species, including the green coffee scale, Coccus viridis. On one coffee bush replicate some aphid individuals were also being tended by ants and were included in our count of scale, although this did not statistically alter our results.

To assess the variability of borer throwing by Azteca, we placed individual live adult borers on coffee plants and recorded the resulting behavior of Azteca ants. Adult borers were extracted from bored green coffee fruits collected in the field. For each behavioral trial, borer individuals were placed on a leaf of the chosen branch of coffee plant replicates and observed for up to 3 min. We recorded three possible outcomes from these trials. First, we recorded “consumption events” when Azteca ants encountered borers, captured them, and returned to the ant foraging trail on the trunk of the coffee plants (which we assumed indicated that the ants were bringing borers back to their nest). Second, we recorded the nonconsumptive dropping behavior when ants interacted with borers by capturing them and dropping them from leaves or by pushing them off leaves, in either case resulting in the removal of the borer from the coffee plant. Finally, if ants encountered borers but did not remove them or if ants failed to encounter borers, we recorded the result as a non‐interaction. Because we were only interested in the overall rate of borer throwing in this study, and other studies have reported more on the details of ant–borer behavioral interactions (Jiménez‐Soto et al., 2013), we chose to simplify our interaction outcomes into these three principal categories. The time of all interactions was recorded and if no interaction occurred, we recorded the end time of the trial at 3 min. We repeated behavioral trials five times per bush on the same branches to estimate the frequency of borer throwing per each site using new CBB individuals for each replicate.

To test for variability in the proportion of borers thrown by ants due to resource variability by season, we conducted this experiment in two different seasons. First, during the rainy season in July 2019, when scales insects are typically considered to be more abundant and then in February 2020, during the dry season, when there are typically fewer arthropods, including hemipterans (Williams‐Guillén et al., 2008). Local precipitation at the farm varied significantly between these sampling points, with 249 mm rainfall measured during July 2019 and 43 mm during February 2020, indicating that these sampling periods represent distinct seasons. In most cases, the trials were repeated at the same sites for both sampling periods and on the same coffee bushes. When this was not possible due to low ant activity during the second sampling in the dry season, we substituted another nearby bush with sufficient Azteca activity at the same site, or in four cases a new site was chosen. Only 19 site replicates were used during the dry season.

To test if season or scale insect density on branches drove differences in the frequency of nonconsumptive behavior by Azteca, we conducted a generalized linear mixed model (GLMM). We included season as a categorical fixed effect, branch‐level scale density as a continuous fixed effect, and their interaction as a fixed effect, as well as site as a random effect (to control for spatial non‐independence between repeated replicates on the same plant). We modeled behavioral outcomes using a binomial error distribution and logit link function (outcomes were reduced to two possible types: nonconsumptive throwing or no interaction given the lack of observed consumption; please refer to the section  Results ). To assess whether there were differences in scale insect abundance on coffee branches between seasons we conducted a generalized linear model (GLM). We included the sampling time (season) as a fixed effect and ran the GLM using a Poisson error distribution with a log link function, to account for count data.

Multipredator interaction experiment

To better understand the impact of Azteca's nonconsumptive effect on CBB damage reduction in multipredator communities we conducted an additive, fully factorial laboratory experiment with two predators. We designed mesocosms in the laboratory using coffee plant saplings (Figure 1). Mesocosms included four different treatments: a control with no ants, an Azteca‐only treatment, a ground‐foraging ant treatment, and a treatment with both ant species. We used the ant species Wasmannia auropunctata (Roger 1863) as the ground‐foraging species, because it is a known predator of CBB (Gonthier et al., 2013; Newson et al., 2021) and can be easily collected and manipulated in the laboratory. Coffee plants were acquired from the nursery at Finca Irlanda and were all Coffea arabica individuals of the same variety and age. All plants were roughly 70 cm tall and were watered every 2–3 days in the laboratory. Plants were placed in 70‐cm diameter plastic washtubs and were kept in plastic containers to avoid water or dirt from spilling into the mesocosms. A coating of fluon was applied to the sides of tubs to keep ants and borers from escaping mesocosms (although some borers could potentially fly out). Additionally, fluon was applied to the outside of plant containers and tanglefoot to the base of coffee plants to avoid the direct interaction of the different ant species. We chose to limit direct ant interactions because of the artificial proximity of the ants in our mesocosms (on small coffee saplings) and the potential for this to unnaturally amplify ant aggression by reducing the amount of territory that multiple ants would typically share in the field. In total, 10 tubs and 20 coffee saplings were used throughout the experiment. For treatments with Wasmannia ants, ants were placed on the floor of washtubs and kept in open plastic containers along with pieces of moss, plants, and twigs collected during ant collection in the field to provide temporary shelter and suitable microclimatic conditions (Figure 1). For treatments with Azteca ants, ants were placed directly on coffee bushes. A small drop of honey was placed on five leaves of each coffee plant to simulate the honey dew resources provided by scale insects for Azteca ants. Roughly 40 individuals of Azteca were placed on bushes. For Wasmannia treatments, we filled ant containers with a minimum of 100 individuals, although this number is likely to vary substantially between replicates due to the difficulty in counting such small ants. These densities of ants were chosen given our observations of what was typical for these species in the field. We used the same number of ants for the treatment with both ant species as for individual treatments to conduct an additive experiment. We used this design to directly test for non‐additive effects from the interaction of ants on borer control, assuming that the resulting borer damage measured from the treatment with both ants would differ from the sum of that of the individual ant treatments, if a synergistic or facilitative interaction occurs (Cardinale et al., 2003).

FIGURE 1.

FIGURE 1

Open in a new tab

Experimental setup of multipredator interaction mesocosms.

All ants were collected in Finca Irlanda and stored in plastic containers with perforated lids between trials. Azteca ants were collected from different nest trees separated by a minimum of 5 m for different replicates. Both majors and minors of Azteca were collected along with fragments of carton nest material. Wasmannia ants and brood were collected by scraping epiphytes and bark from trees and from hollowed out branches collected from the ground. Although Wasmannia ants in the area of the study may exist in large “supercolonies” spread out across farms (Yitbarek et al., 2017), we attempted to collect from different areas separated by at least 2 m for each replicate. All ants were kept in containers in the laboratory until the morning of experimental trials, but no longer than 1 week. In two instances some ants were reused to supplement Wasmannia colonies with low activity.

Because coffee saplings were too young to produce fruits, we collected fruits from plants in the field to add to our mesocosms. We added branches with 20 unbored fruits to plants in the laboratory by positioning them across branches of the saplings and added 20 unbored individual green coffee fruits on the floor of mesocosms (Figure 1). Fruits were placed both on coffee plants and on the ground to track individual borers and borer damage in both places. Branches that were added to coffee plants had all leaves removed and any additional fruits until each branch had exactly 20 unbored fruits. Fruits and coffee plants were arranged in experimental mesocosms at least 2 h before beginning trials with ants and borers to allow plant volatiles to dissipate.

At the start of each experimental trial 40 individual CBBs were placed on the coffee plants in our mesocosms. We chose this density to more easily track CBB damage given the possibility that some borers may die or escape in the laboratory, but this density has been also frequently observed on individual plants in the field (Barrera, 2008 and unpublished data). All borers were collected from bored fruits in the field by dissecting them. To prevent falling or thrown borers from being lost in plant containers, we fitted a small plastic skirt around each plant to deflect borers onto the ground arena of mesocosms (Figure 1). 24 h after placing borers in mesocosms we checked all coffee fruits for evidence of borer holes and counted the total number of CBB in fruits on plants and on the ground. Tracking fruit damage as a measure of pest suppression was preferred over measuring borer mortality, because it was sometimes difficult to assess when individual borers were dead and if the cause of death was due to ant attack or environmental exposure in mesocosms. This also allowed for a more direct assessment of the impact of this nonconsumptive effect on crop damage suppression. Additionally, we counted all CBB individuals that were found outside fruits at the end of trials to track CBB movement and ensure that few borers were escaping the mesocosms (Appendix S1: Figure S1). To minimize the impact of residual ant pheromones or plant volatile chemicals, all plastic tubs were cleaned with alcohol in between trials. Coffee plants were alternated such that at least 48 h passed before being used again in experiments. To control for differences between mesocosms or coffee plants, treatments were assigned to each mesocosm randomly. In total, 59 trials were conducted (N = 15 control, N = 13 Wasmannia only, N = 17 Azteca only, N = 14 both ants) in blocks during the first 2 weeks of October 2016.

To assess whether the number of borers observed inside coffee fruits differed between treatments after 24 h we conducted GLMMs. We ran individual GLMMs on the number of borers in fruits on plants and the ground separately, and on the combined outcome. We included Azteca and Wasmannia presence or absence as fixed effects in the models. Their interaction was also included as a fixed effect to determine statistical significance of the both‐ant treatment, which would indicate non‐additive predator effects (for the combined data). Due to the heavily nonconsumptive nature of the Azteca–CBB interaction (please refer to the section  Results ) we used additive models, as opposed to a multiplicative risk model (Sih et al., 1998), which is based on prey depletion effects through consumption (Mccoy et al., 2012). To account for count data, the models were run using a Poisson error distribution with a log link function. We corrected for observed overdispersion in our plant level model by running a Poisson‐lognormal error distribution using an observation‐level random effect (Elston et al., 2001). In all models, mesocosm number and trial date (block) were added as random effects to account for any impacts of inconsistency in our laboratory environment. For all GLMMs, fixed effect parameters and the variance of random effects were estimated by maximum likelihood with Laplace approximation. All GLMMs were run using the “glmer” function from the lme4 package, whereas GLMs were run using the “glm” function, both in R version 4.0.2 (R Core Team, 2021).

RESULTS

Does the nonconsumptive interaction vary across time or space with scale insect density?

Interestingly, in nearly 200 behavioral trials we observed only one occurrence of what appeared to be consumptive behavior by Azteca on the CBB. Overall, nearly 43% of cases resulted in the nonconsumptive effect of Azteca throwing or dropping borers from plants. Alternatively, borers were either not removed by Azteca or not found during the 3‐min trials (this includes five cases in which borers fell or flew off plants on their own). Unexpectedly, we found no difference in the amount of adult scale on coffee branches between seasons (Figure 2 and Table 1), indicating that scale density available to Azteca did not vary across time. However, we did find a significant difference in the proportion of dropped borers by Azteca between the two sampling times (Figure 3 and Table 1), when more borers were thrown off plants during the dry season (removing the one case of consumption from the analysis). The proportion of borers that were thrown by Azteca was also positively influenced by the amount of scale that were present on individual coffee branches (Figure 4 and Table 1). Additionally, the results from our GLMM showed a significant interaction between branch scale and season on the nonconsumptive behavior, in which CBB throwing was more consistent across a range of scale densities during the dry season (Figure 4 and Table 1).

FIGURE 2.

FIGURE 2

Open in a new tab

Number of adult scale insects on coffee branches by season. Raw data and means (±standard error) are shown (means shown in blue). Data from the rainy season were sampled in July 2019, whereas data from the dry season were sampled in February 2020. Data are pooled across several species of scale that are typically tended by Azteca sericeasur in this system.

TABLE 1.

Summary of statistical model results for the field behavioral experiment.

Parameter Estimate (±SE) z‐value p (>|z|)
GLMM: proportion of CBB thrown ~
Intercept −1.544 ± 0.452 −3.413 <0.001
Season 1.113 ± 0.462 2.409 0.016
Scale 0.023 ± 0.008 3.026 0.002
Season × Scale −0.017 ± 0.009 −1.973 0.048
GLM: branch scale ~
Intercept 3.615 ± 0.037 98.537 <0.001
Season <0.001 ± 0.053 0.004 0.997
Open in a new tab

Notes: The generalized linear mixed model (GLMM) was run using a binomial error distribution (logit link). The generalized linear model (GLM) was run using a Poisson error distribution (log link). Parameter estimates (±SE), z‐values, and p‐values are provided. SE, standard error.

FIGURE 3.

FIGURE 3

Open in a new tab

Number of coffee berry borers (CBB) thrown off plants by Azteca sericeasur by season. Total counts from behavioral trials are tallied for the different sampling times. Data from the rainy season were sampled in July 2019, whereas data from the dry season were sampled in February 2020. The overall proportion of CBB thrown by season is shown in bold at the top of columns.

FIGURE 4.

FIGURE 4

Open in a new tab

Proportion of coffee berry borers (CBB) thrown per plant across scale insect density. The relationship between the number of adult scale insects on coffee branches and the proportional result of Azteca behavioral trials for each coffee plant is shown. Proportions are calculated from the five behavioral trial replicates conducted per coffee plant. Trend lines show the relative effect of season by scale density.

What is the impact of the nonconsumptive interaction on pest damage and multienemy interactions?

Overall, in the plant and ground combined data, we did not observe a difference in the number of borers found in fruits between our control treatments and treatments with Azteca only. However, we did observe a significant decrease in borers in the Wasmannia and both‐ant treatments (Appendix S1: Figure S2; Table 2). We did not find a significant interaction between Azteca and Wasmannia presence in our treatments, indicating that non‐additive multipredator effects were not observed in the combined data set (Appendix S1: Figure S2; Table 2).

TABLE 2.

Summary of generalized linear mixed model results for the multipredator experiment

Parameter Estimate (±SE) z‐value p (>|z|)
Overall CBB damage~
Intercept 2.944 ± 0.063 46.477 <0.001
Azteca −0.043 ± 0.087 −0.495 0.620
Wasmannia −0.316 ± 0.099 −3.198 0.001
Azteca × Wasmannia −0.019 ± 0.140 −0.136 0.891
Plant CBB damage~
Intercept 2.182 ± 0.177 12.327 <0.001
Azteca −0.950 ± 0.213 −4.456 <0.001
Wasmannia 0.045 ± 0.217 0.207 0.836
Azteca × Wasmannia −0.025 ± 0.322 −0.076 0.939
Ground CBB damage~
Intercept 2.213 ± 0.098 22.502 <0.001
Azteca 0.436 ± 0.111 3.924 <0.001
Wasmannia −0.779 ± 0.167 −4.664 <0.001
Azteca × Wasmannia 0.352 ± 0.200 1.761 0.078
Open in a new tab

Notes: All models were run using Poisson error distributions (log link). Parameter estimates (±SE), z‐values, and p‐values are provided. The overall category shows the statistical results for the combined plant and ground CBB fruit damage. SE, standard error.

Separating plant and ground borer data, we observed opposite impacts of Azteca presence on borer infestation levels, which accounted for the lack of an effect for Azteca in the overall data. On the coffee plant, Azteca ants lowered the number of borers found in fruits by roughly 57% compared with controls (Figure 5a and Table 2), in line with previous evidence demonstrating the effectiveness of Azteca at reducing borer damage (Gonthier et al., 2013; Jiménez‐Soto et al., 2013; Morris et al., 2015). However, because Wasmannia were restricted to the ground in this experiment, they had no effect on borer damage on the plant (Figure 5a and Table 2). The treatment with both ants also resulted in significantly fewer CBB in fruits on plants than in the control, reflecting the positive effect of Azteca ants on coffee plants (Figure 5a and Table 2).

FIGURE 5.

FIGURE 5

Open in a new tab

Number of coffee berry borers (CBB) in fruits on the plant and ground in mesocosms. The total number of CBB found bored into fruits on the plant (a) and the ground (b) at the end of the laboratory experiment after 24 h is shown. Raw data are shown for each treatment along with the mean (±standard error). Significant differences in means (from the generalized linear mixed models with Poisson error distribution and log link) for each figure are indicated with different letters.

On the ground, we observed a base level of damage in fruits in our control treatments from borers that had either fallen during trials on their own or flew off plants to the ground (Figure 5b and Table 2). Wasmannia‐only ant treatments significantly reduced borer damage from these levels, however Azteca‐only treatments significantly increased borer damage on the ground compared with the control (Figure 5b and Table 2), reflecting the nonconsumptive throwing behavior of the ants. Although significantly less than the Azteca‐only treatment, ground borer damage in the treatment with both ants was not different from the control (Figure 5b and Table 2).

DISCUSSION

Collectively, our results document a unique case study in natural pest control in which crop damage suppression is driven almost exclusively by the nonconsumptive attack of a dominant natural enemy. Our behavioral data show that the strength of this aggressive nonconsumptive interaction—throwing CBBs off plants by Azteca sericeasur—may be driven by variability in alternative resources for Azteca, potentially resulting in variable pest control efficiency. Our multipredator laboratory experiment demonstrated how this nonconsumptive effect mediated spatial complementarity between arboreal and ground‐foraging natural enemies and may result in enhanced reduction of borer damage, when ground foragers are present on farms.

Surprisingly, from our behavioral experiment, we show that this pest control interaction is almost exclusively nonconsumptive, which is interesting considering the previously documented efficiency of Azteca in reducing borer damage on plants (Gonthier et al., 2013; Jiménez‐Soto et al., 2013; Morris et al., 2015). Although we set out to test the hypothesis that this behavior may vary due to seasonal variation in resources from scale insects, we did not find a significant difference in scale abundance on coffee plants between seasons. However, we did still find a significant effect of season on borer throwing, in which slightly more borers were thrown during the dry season compared with the rainy season. We also found an overall significant positive effect of the density of scale insects on coffee branches on the frequency of Azteca's throwing behavior, regardless of season, which aligns with related research showing a negative relationship between scale density and CBB damage on coffee with Azteca (Perfecto & Vandermeer, 2006; Rivera‐Salinas et al., 2018). This suggests that the tendency for Azteca to exhibit this nonconsumptive behavior may be explained both by spatial variation in scale resources at different sites across coffee farms and by temporal variation across seasons.

Despite not finding a difference in scale abundance between seasons, it is possible that the effect of season on borer throwing that we observed was driven by the quality of honeydew resources from scale insects that may vary seasonally in the region of this study (Pringle et al., 2013). Research on ant–scale–plant interactions in Central America and Mexico has shown that water stress during dry seasons changes the concentration of carbohydrates within plants, which may cascade upward to affect scale honeydew quality and hemipteran‐tending ant activity, ultimately impacting the ant defense of plants (Pringle et al., 2013). A previous study in our system used exclosure experiments on coffee bushes to show that Azteca's suppression of borer damage also varied by time of year (Rivera‐Salinas et al., 2018), which may be explained by variation in honeydew quality. Although we did not test honeydew composition in this experiment, our results aligned with the findings from Pringle et al. (2013) in that ant defense of plants is more consistent during the dry season than the wet season, potentially implying that seasonal differences in honeydew quality are at play. This would also explain the greater consistency in throwing we observed across a range of scale densities during the dry season. Additionally, variation in ant throwing behavior may be driven by seasonal differences in the nutritional needs of ant colonies across time (Cook et al., 2011). An important caveat is that, although we found these differences between two different sampling times with distinct precipitation levels, we were limited in our ability to infer long‐term seasonal trends in nonconsumptive effects from these data. Regardless of the precise mechanism, nonconsumptive effects driven by predator aggression as opposed to prey behavior may result in distinct dynamics when driven by variation in alternate resources for enemies, which may ultimately result in variable or unreliable pest suppression. To our knowledge, this phenomenon has not been explored previously in natural pest control, and should be investigated further to understand the general impact of spatial and long‐term seasonal (Hermann & Landis, 2017) variation in nonconsumptive enemy aggression on pest damage suppression.

In addition to our behavioral experiment, our multipredator experiment helped to illuminate some important questions about the community ecology of this interaction. First, it appears that when there are no other predators in the system, many of the borers that are thrown off coffee plants by Azteca survive those attacks, potentially remaining in the borer population pool as reproductive individuals. Furthermore, thrown CBB individuals that survived attacks could relocate to old fruits on the ground or other coffee bushes and damage new fruits if Azteca is not foraging on those bushes. Despite the potential negative consequences of this “spillover” effect of thrown CBB, we also confirmed the results of previous laboratory (Pardee & Philpott, 2011; Philpott et al., 2012) and field experiments (Gonthier et al., 2013; Jiménez‐Soto et al., 2013; Morris et al., 2015) that showed that Azteca is a highly efficient pest control agent in reducing borer damage on coffee bushes where they forage. Additionally, we demonstrated that a ground predator, Wasmannia auropunctata, suppressed borer individuals while foraging on the ground. Although it is already known that these ants, which nest both on plants and the ground, are important predators of borers on coffee plants (Gonthier et al., 2013; Newson et al., 2021), less research has focused on their potential to consume borers on the ground. This is despite their ability to enter borer holes where they potentially predate CBB larvae and pupae in old fruits that fall to the ground (Morris & Perfecto, 2016).

Regarding the impact of this nonconsumptive effect on multipredator interactions, the results from both experiments suggest that Azteca may significantly increase resource availability on the ground for ground‐foraging predators such as other ants, which could potentially result in facilitation or synergistic predator effects (Morris et al., 2018). However, despite this clear potential benefit for borer control, we did not find any evidence of non‐additive effects from our multipredator experiment. Although Wasmannia presence consistently reduced borer damage, and in treatments with both ants more borers were made available to Wasmannia on the ground by Azteca, their rate of damage reduction did not appear to increase under these circumstances. In fact, Wasmannia ants reduced borer damage in fruits on the ground by roughly the same quantity in the both‐ant treatment as in the Wasmannia‐only treatment. One potential limitation is that we did not actually know from our data what quantity of borers were being directly consumed by Wasmannia, because we were only measuring the reduction in berry damage and not directly tracking ant behavior. However, we did observe parts of CBB individuals near Wasmannia containers in several replicates, and ants carrying CBB in their mandibles, suggesting that consumption did occur. Further research should investigate interactions between ants, ground‐foraging predators, and herbivores under more realistic conditions in the field (Hermann & Landis, 2017) on mature coffee plants where predators can interact freely.

These contrasting results revealed the complexity of pest control interactions when nonconsumptive interactions are involved. Although there may be instances when Azteca ants do consume borers in the field, our data suggested that this occurred very infrequently, despite the well known importance of this ant for reducing borer damage. Interestingly, this behavior may actually explain the efficiency of Azteca as a keystone pest control agent. A previous field experiment conducted in this study system manipulated the densities of borers on coffee plants to which Azteca were exposed and found the same level of borer damage reduction across a range of pest densities (Morris et al., 2015). Importantly, this study failed to find a satiating effect at high densities of borers, which, given our results here, is probably explained by the fact that Azteca are simply not consuming CBB. This behavior may then ultimately result in the most efficient reduction of borers on coffee plants with high levels of Azteca activity. Subsequently, in farms with sufficient ground‐foraging predator abundance, the “spillover” of these borer individuals from bushes with Azteca can be regulated by other predators, such as Wasmannia, reducing their survival and colonization of other coffee bushes. The both‐ant treatment in our laboratory experiment resulted in the best overall control of CBB by first reducing borer damage on plants to its lowest levels and then minimizing the “spillover” from Azteca's nonconsumptive effect of throwing borers to the ground, demonstrating the potential for spatial complementarity between arboreal and ground‐foraging natural enemies. Beyond ants, other predators, such as certain species of web‐building spiders that form associative relationships with A. sericeasur in this region (Marín et al., 2015), may act as filters to collect resources thrown by Azteca and potentially buffer the effect of thrown borers and other pests. Additionally, thrown borers may also experience increased mortality risk from the loss of energy spent searching for refuges in old fruits on the ground, from potentially shifting their diets to these inferior resources, or from searching for new coffee plants to colonize.

Although nonconsumptive effects have long been studied in community ecology (Preisser et al., 2005; Schmitz et al., 2004; Werner & Peacor, 2003), there have been fewer examples of nonconsumptive mediated pest control (Hermann & Landis, 2017; Walzer & Schausberger, 2009). This study provides an unique addition to the existing literature, in which nonconsumptive natural enemy aggression, rather than prey defensive behavior (or direct predation), dominates pest damage suppression and drives spatial complementarity between natural enemies. Although this interaction may enhance coffee pest suppression in diverse communities when ground‐foraging predators are conserved, variation in other resources that mediate Azteca's aggressive behavior, such as scale insect abundance and potentially honeydew quality, could result in variable pest control efficacy. Future research exploring the impact of similar nonconsumptive interactions on pest populations (Sheriff et al., 2018) in diverse agroecological communities and under more realistic field conditions across growing seasons (Hermann & Landis, 2017) will help to illuminate the broader importance of these interactions for natural pest control. Ultimately, this case study demonstrates the complexity of natural pest control ecology and highlights the need to consider specific interaction mechanisms and spatial and temporal variability in those interactions for the management of this important ecosystem service (Vandermeer & Perfecto, 2017).

AUTHOR CONTRIBUTIONS

Jonathan R. Morris and Ivette Perfecto conceived of the project and experiments; Jonathan R. Morris designed experiments and collected data; Jonathan R. Morris analyzed the data; Jonathan R. Morris wrote the first draft of the manuscript; Jonathan R. Morris and Ivette Perfecto edited the manuscript together.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Supporting information

Appendix S1

Click here for additional data file. (124.2KB, pdf)

ACKNOWLEDGMENTS

We thank Finca Irlanda, Walter Peters, and the farm workers for allowing us to be part of their community and to conduct this research. We thank Miguel Crisóstomo for providing coffee saplings from the farm nursery. We thank H. Uciel Vazquez Perez, Braulio E. Chilel, Aldo de la Mora, Gustavo López Bautista, and Cruz Elena Gomez Vazquez for assistance with field work and experiments. We thank Kevin Li and Pranav Yajnik for assistance with data analysis. We thank Iris Saraeny Rivera‐Salinas, Gordon Fitch, Nicholas Medina, Chatura Vaidya, and Zachary Hajian‐Forooshani for helpful feedback on the manuscript. Funding was provided by the Tinker Field Research Grant & Individual Fellowship, from the International Institute at the University of Michigan, the Rackham Graduate Student Research Grant from the University of Michigan, and National Science Foundation Grant (DEB 1853261).

Morris, Jonathan R. , and Perfecto Ivette. 2022. “An Aggressive Nonconsumptive Effect Mediates Pest Control and Multipredator Interactions in a Coffee Agroecosystem.” Ecological Applications 32(7): e2653. 10.1002/eap.2653

Handling Editor: Juan C. Corley

Funding information Division of Environmental Biology, Grant/Award Number: 1853261

DATA AVAILABILITY STATEMENT

Data (Morris, 2022) are available in Dryad at https://doi.org/10.5061/dryad.hmgqnk9k0.

REFERENCES

  1. Barrera, J. F. 2008. “Coffee Pests and their Management.” In Encyclopedia of Entomology, edited by Capinera J. L.. Dordrecht: Springer. [Google Scholar]
  2. Bianchi, F. J. J. A. , Booij C. J. H., and Tscharntke T.. 2006. “Sustainable Pest Regulation in Agricultural Landscapes: A Review on Landscape Composition, Biodiversity and Natural Pest Control.” Proceedings of the Royal Society B 273(1595): 1715–27. 10.1098/rspb.2006.3530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Cardinale, B. J. , Harvey C. T., Gross K., and Ives A. R.. 2003. “Biodiversity and Biocontrol: Emergent Impacts of a Multi‐Enemy Assemblage on Pest Suppression and Crop Yield in an Agroecosystem.” Ecology Letters 6(9): 857–65. 10.1046/j.1461-0248.2003.00508.x. [DOI] [Google Scholar]
  4. Casula, P. , Wilby A., and Thomas M. B.. 2006. “Understanding Biodiversity Effects on Prey in Multi‐Enemy Systems.” Ecology Letters 9: 995–1004. 10.1111/j.1461-0248.2006.00945.x. [DOI] [PubMed] [Google Scholar]
  5. Clark, R. E. , and Singer M. S.. 2018. “Differences in Aggressive Behaviors between Two Ant Species Determine the Ecological Consequences of a Facultative Food‐for‐Protection Mutualism.” Journal of Insect Behavior 31(5): 510–22. 10.1007/s10905-018-9695-8. [DOI] [Google Scholar]
  6. Cook, S. C. , Eubanks M. D., Gold R. E., and Behmer S. T.. 2011. “Seasonality Directs Contrasting Food Collection Behavior and Nutrient Regulation Strategies in Ants.” PLoS One 6(9): e25407. 10.1371/journal.pone.0025407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Crowder, D. W. , and Jabbour R.. 2014. “Relationships between Biodiversity and Biological Control in Agroecosystems: Current Status and Future Challenges.” Biological Control 75: 8–17. 10.1016/j.biocontrol.2013.10.010. [DOI] [Google Scholar]
  8. Dainese, M. , Martin E. A., Aizen M. A., Albrecht M., Bartomeus I., Bommarco R., Carvalheiro L. G., et al. 2019. “A Global Synthesis Reveals Biodiversity‐Mediated Benefits for Crop Production.” Science Advances 5: 1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Davenport, J. M. , and Chalcraft D. R.. 2013. “Nonconsumptive Effects in a Multiple Predator System Reduce the Foraging Efficiency of a Keystone Predator.” Ecology and Evolution 3(9): 3063–72. 10.1002/ece3.691. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Elston, D. A. , Moss R., Boulinier T., Arrowsmith C., and Lambin X.. 2001. “Analysis of Aggregation, a Worked Example: Numbers of Ticks on Red Grouse Chicks.” Parasitology 122(5): 563–9. 10.1017/S0031182001007740. [DOI] [PubMed] [Google Scholar]
  11. Eubanks, M. D. , and Finke D. L.. 2014. “Interaction Webs in Agroecosystems: Beyond Who Eats Whom.” Current Opinion in Insect Science 2: 1–6. 10.1016/j.cois.2014.06.005. [DOI] [PubMed] [Google Scholar]
  12. Gonthier, D. J. , Ennis K. K., Philpott S. M., Vandermeer J., and Perfecto I.. 2013. “Ants Defend Coffee from Berry Borer Colonization.” BioControl 58(6): 815–20. 10.1007/s10526-013-9541-z. [DOI] [Google Scholar]
  13. Griffin, J. N. , Byrnes J. E. K., and Cardinale B. J.. 2013. “Effects of Predator Richness on Prey Suppression: A Meta‐Analysis.” Ecology 94(10): 2180–7. [DOI] [PubMed] [Google Scholar]
  14. Hallmann, C. A. , Sorg M., Jongejans E., Siepel H., Hofland N., Schwan H., Stenmans W., et al. 2017. “More than 75 Percent Decline over 27 Years in Total Flying Insect Biomass in Protected Areas.” PLoS One 12(10): e0185809. 10.1371/journal.pone.0185809. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Harvey, J. A. , Heinen R., Armbrecht I., Basset Y., Baxter‐Gilbert J. H., Martijn Bezemer T., Böhm M., et al. 2020. “International Scientists Formulate a Roadmap for Insect Conservation and Recovery.” Nature Ecology and Evolution 4(2): 174–6. 10.1038/s41559-019-1079-8. [DOI] [PubMed] [Google Scholar]
  16. Hermann, S. L. , and Landis D. A.. 2017. “Scaling up our Understanding of Non‐Consumptive Effects in Insect Systems.” Current Opinion in Insect Science 20: 54–60. 10.1016/j.cois.2017.03.010. [DOI] [PubMed] [Google Scholar]
  17. Ingerslew, K. S. , and Finke D. L.. 2018. “Multi‐Species Suppression of Herbivores through Consumptive and Non‐Consumptive Effects.” PLoS One 13(5): 1–17. 10.1371/journal.pone.0197230. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Jiménez‐Soto, E. , Cruz‐Rodríguez J. A., Vandermeer J., and Perfecto I.. 2013. “ Hypothenemus hampei (Coleoptera: Curculionidae) and its Interactions with Azteca Instabilis and Pheidole Synanthropica (Hymenoptera: Formicidae) in a Shade Coffee Agroecosystem.” Environmental Entomology 42(5): 915–24. 10.1603/EN12202. [DOI] [PubMed] [Google Scholar]
  19. Kaplan, I. , and Eubanks M. D.. 2005. “Aphids Alter the Community‐Wide Impact of Fire Ants.” Ecology 86(6): 1640–9. 10.1890/04-0016. [DOI] [Google Scholar]
  20. Karp, D. S. , Mendenhall C. D., Sandí R. F., Chaumont N., Ehrlich P. R., Hadly E. A., and Daily G. C.. 2013. “Forest Bolsters Bird Abundance, Pest Control and Coffee Yield.” Ecology Letters 16(11): 1339–47. 10.1111/ele.12173. [DOI] [PubMed] [Google Scholar]
  21. van Klink, R. , Bowler D. E., Gongalsky K. B., Swengel A. B., Gentile A., and Chase J. M.. 2020. “Meta‐Analysis Reveals Declines in Terrestrial but Increases in Freshwater Insect Abundances.” Science 368(6489): 417–20. 10.1126/science.aax9931. [DOI] [PubMed] [Google Scholar]
  22. Kremen, C. , and Merenlender A. M.. 2018. “Landscapes that Work for Biodiversity and People.” Science 362(6412): eaau6020. 10.1126/science.aau6020. [DOI] [PubMed] [Google Scholar]
  23. Letourneau, D. K. , Jedlicka J. A., Bothwell S. G., and Moreno C. R.. 2009. “Effects of Natural Enemy Biodiversity on the Suppression of Arthropod Herbivores in Terrestrial Ecosystems.” Annual Review of Ecology, Evolution, and Systematics 40(1): 573–92. 10.1146/annurev.ecolsys.110308.120320. [DOI] [Google Scholar]
  24. Losey, J. E. , and Denno R. F.. 1998. “Positive Predator‐Predator Interactions: Enhanced Predation Rates and Synergistic Suppression of Aphid Populations.” Ecology 79(6): 2143–52. 10.1890/0012-9658(1998)079[2143:PPPIEP]2.0.CO;2. [DOI] [Google Scholar]
  25. Marín, L. , Jackson D., and Perfecto I.. 2015. “A Positive Association between Ants and Spiders and Potential Mechanisms Driving the Pattern.” Oikos 124(8): 1078–88. 10.1111/oik.01913. [DOI] [Google Scholar]
  26. Mccoy, M. W., A. C. Stier, and C. W. Osenberg . 2012. “Emergent Effects of Multiple Predators on Prey Survival: The Importance of Depletion and the Functional Response.” Ecology Letters 15(12): 1449–56. 10.1111/ele.12005. [DOI] [PubMed] [Google Scholar]
  27. Meadows, A. J. , Owen J. P., and Snyder W. E.. 2017. “Keystone Nonconsumptive Effects within a Diverse Predator Community.” Ecology and Evolution 7(23): 10315–25. 10.1002/ece3.3392. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Morris, J. 2022. “Data for: An Aggressive Non‐consumptive Effect Mediates Pest Control and Multi‐Predator Interactions in a Coffee Agroecosystem.” Dryad, Dataset. 10.5061/dryad.hmgqnk9k0. [DOI] [PMC free article] [PubMed]
  29. Morris, J. R. , Estelí Jiménez‐Soto S. M., Philpott S. M., and Perfecto I.. 2018. “Ant‐Mediated (Hymenoptera: Formicidae) Biological Control of the Coffee Berry Borer: Diversity, Ecological Complexity, and Conservation Biocontrol.” Myrmecological News 26: 1–17. [Google Scholar]
  30. Morris, J. R. , and Perfecto I.. 2016. “Testing the Potential for Ant Predation of Immature Coffee Berry Borer (Hypothenemus hampei) Life Stages.” Agriculture, Ecosystems and Environment 233: 224–8. 10.1016/j.agee.2016.09.018. [DOI] [Google Scholar]
  31. Morris, J. R. , Vandermeer J., and Perfecto I.. 2015. “A Keystone Ant Species Provides Robust Biological Control of the Coffee Berry Borer under Varying Pest Densities.” PLoS One 10(11): e0142850. 10.1371/journal.pone.0142850. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Newson, J. , Vandermeer J., and Perfecto I.. 2021. “Differential Effects of Ants as Biological Control of the Coffee Berry Borer in Puerto Rico.” Biological Control 160(5): 104666. 10.1016/j.biocontrol.2021.104666. [DOI] [Google Scholar]
  33. Pardee, G. L. , and Philpott S. M.. 2011. “Cascading Indirect Effects in a Coffee Agroecosystem: Effects of Parasitic Phorid Flies on Ants and the Coffee Berry Borer in a High‐Shade and Low‐Shade Habitat.” Environmental Entomology 40(3): 581–8. 10.1603/EN11015. [DOI] [PubMed] [Google Scholar]
  34. Perfecto, I. , and Vandermeer J.. 2006. “The Effect of an Ant‐Hemipteran Mutualism on the Coffee Berry Borer (Hypothenemus hampei) in Southern Mexico.” Agriculture, Ecosystems & Environment 117(2–3): 218–21. 10.1016/j.agee.2006.04.007. [DOI] [Google Scholar]
  35. Perfecto, I. , Vandermeer J., and Philpott S. M.. 2014. “Complex Ecological Interactions in the Coffee Agroecosystem.” Annual Review of Ecology, Evolution, and Systematics 45(1): 137–58. 10.1146/annurev-ecolsys-120213-091923. [DOI] [Google Scholar]
  36. Philpott, S. M. , and Armbrecht I.. 2006. “Biodiversity in Tropical Agroforests and the Ecological Role of Ants and Ant Diversity in Predatory Function.” Environmental Entomology 31: 369–77. [Google Scholar]
  37. Philpott, S. M. , Pardee G. L., and Gonthier D. J.. 2012. “Cryptic Biodiversity Effects: Importance of Functional Redundancy Revealed through Addition of Food Web Complexity.” Ecology 93(5): 992–1001 http://www.ncbi.nlm.nih.gov/pubmed/22764486. [DOI] [PubMed] [Google Scholar]
  38. Preisser, E. L. , Bolnick D. I., and Benard M. F.. 2005. “Ecology 2005 Scared to Death?” Ecology 86(2): 501–9. http://www.esajournals.org/doi/pdf/10.1890/04-0719%5Cnpapers3://publication/uuid/040F9073-624F-43EA-8E10-9FB1CFAFCF04. [Google Scholar]
  39. Pringle, E. G. , Akçay E., Raab T. K., Dirzo R., and Gordon D. M.. 2013. “Water Stress Strengthens Mutualism among Ants, Trees, and Scale Insects.” PLoS Biology 11(11): e1001705. 10.1371/journal.pbio.1001705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. R Core Team . 2021. R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing; http://www.r-project.org/. [Google Scholar]
  41. Rivera‐Salinas, I. S. , Hajian‐Forooshani Z., Jiménez‐Soto E., Cruz‐Rodríguez J. A., and Philpott S. M.. 2018. “High Intermediary Mutualist Density Provides Consistent Biological Control in a Tripartite Mutualism.” Biological Control 118: 26–31. 10.1016/j.biocontrol.2017.12.002. [DOI] [Google Scholar]
  42. Sánchez‐Bayo, F. , and Wyckhuys K. A. G.. 2019. “Worldwide Decline of the Entomofauna: A Review of its Drivers.” Biological Conservation 232(1): 8–27. 10.1016/j.biocon.2019.01.020. [DOI] [Google Scholar]
  43. Schmitz, O. J. , Krivan V., and Ovadia O.. 2004. “Trophic Cascades: The Primacy of Trait‐Mediated Indirect Interactions.” Ecology Letters 7(2): 153–63. 10.1111/j.1461-0248.2003.00560.x. [DOI] [Google Scholar]
  44. Sheriff, M. J. , Peacor S. D., Hawlena D., and Thaker M.. 2018. “Non‐Consumptive Predator Effects on Prey Population Size: A Dearth of Evidence.” Journal of Animal Ecology 89(6): 1302–16. 10.1111/1365-2656.13213. [DOI] [PubMed] [Google Scholar]
  45. Sih, A., G. Englund, and D. Wooster . 1998. “Emergent Impacts of Multiple Predators on Prey.” Tree 13(9): 350–5. [DOI] [PubMed] [Google Scholar]
  46. Straub, C. S. , Finke D. L., and Snyder W. E.. 2008. “Are the Conservation of Natural Enemy Biodiversity and Biological Control Compatible Goals?” Biological Control 45(2): 225–37. 10.1016/j.biocontrol.2007.05.013. [DOI] [Google Scholar]
  47. Thaler, J. S. , and Griffin C. A. M.. 2008. “Relative Importance of Consumptive and Non‐Consumptive Effects of Predators on Prey and Plant Damage: The Influence of Herbivore Ontogeny.” Entomologia Experimentalis et Applicata 128(1): 34–40. 10.1111/j.1570-7458.2008.00737.x. [DOI] [Google Scholar]
  48. Vandermeer, J. , Armbrecht I., De La Mora A., Ennis K. K., Fitch G., Gonthier D. J., Hajian‐Forooshani Z., et al. 2019. “The Community Ecology of Herbivore Regulation in an Agroecosystem: Lessons from Complex Systems.” Bioscience 69(12): 974–96. 10.1093/biosci/biz127. [DOI] [Google Scholar]
  49. Vandermeer, J. H. , and Perfecto I.. 2017. Ecological Complexity and Agroecology. Milton Park: Routledge. [Google Scholar]
  50. Vandermeer, J. , Perfecto I., and Philpott S.. 2010. “Ecological Complexity and Pest Control in Organic Coffee Production: Uncovering an Autonomous Ecosystem Service.” Bioscience 60(7): 527–37. 10.1525/bio.2010.60.7.8. [DOI] [Google Scholar]
  51. Wagner, D. L. 2020. “Insect Declines in the Anthropocene.” Annual Review of Entomology 65: 457–80. [DOI] [PubMed] [Google Scholar]
  52. Walzer, A. , and Schausberger P.. 2009. “Non‐Consumptive Effects of Predatory Mites on Thrips and its Host Plant.” Oikos 118(6): 934–40. 10.1111/j.1600-0706.2008.17299.x. [DOI] [Google Scholar]
  53. Werner, E. E. , and Peacor S. D.. 2003. “A Review of Trait‐Mediated Indirect Interactions in Ecological Communities.” Ecology 84(5): 1083–100. 10.1890/0012-9658(2003)084[1083:AROTII]2.0.CO;2. [DOI] [Google Scholar]
  54. Williams‐Guillén, K. , Perfecto I., and Vandermeer J.. 2008. “Bats Limit Insects in a Neotropical Agroforestry System.” Science 320(5872): 70. 10.1126/science.1152944. [DOI] [PubMed] [Google Scholar]
  55. Yitbarek, S. , Vandermeer J. H., and Perfecto I.. 2017. “From Insinuator to Dominator: Foraging Switching by an Exotic Ant.” Diversity and Distributions 23(7): 820–7. 10.1111/ddi.12568. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix S1

Click here for additional data file. (124.2KB, pdf)

Data Availability Statement

Data (Morris, 2022) are available in Dryad at https://doi.org/10.5061/dryad.hmgqnk9k0.

Articles from Ecological Applications are provided here courtesy of Wiley

RESOURCES