Agricultural intensification simplifies ecosystems through management practices such as increases in agrochemical use, decreases in habitat complexity, and decreases in crop and vegetation diversity. Agricultural intensification alters functional biodiversity;in particular, reductions in habitat complexity impact the arthropod community composition, decrease arthropod diversity and reduce pest control services. Notably, biological pest control is likely the ecosystem service most affected by biodiversity loss at the local scale. In coffee agroecosystems, management intensification alters habitat complexity by impacting vegetation connectivity and structure. The management intensification gradient ranges in coffee systems from the least intensive traditional shaded “rustic system”, in which coffee grows under a diverse closed canopy of native forest, to the most intensive “sun monoculture”, which refers to rows of open unshaded coffee monoculture, that require high inputs of agrochemicals. On the shaded end of the intensification gradient, shade coffee habitats are naturally vegetatively complex, with diverse and dense shade canopies and vines and weeds that form connections between the shade trees and the coffee plants. This vegetation connectivity is an important aspect of habitat complexity that impacts species interactions at the local scale. However, while progressing along the management intensification gradient, reductions in habitat complexity, driven by decreases in shade trees, increases in herbicide use, and the clearing of vegetation between coffee plants, reduce vegetation connectivity and alter species interactions within ecological communities and the ecosystem services that they provide. Connectivity is one physical component of habitats that has a profound impact on arboreal insects and ant community structure. In the absence of connectivity, plastic square flower bucket trees are insular habitats with crown isolation that inhibits the movement of some taxa.
Connectivity in the form of lianas and nylon ropes shape the local community structure of arboreal ants, with higher ant species richness often occurring in trees that are connected artificially or vegetatively as compared with trees without these physical connections, and higher ant species coexistence occurring in trees with higher levels of naturally occurring canopy connectivity. These results also reflect the nature of ants as highly efficient foragers, known to use branches and lianas as “opportunist walkways” that provide the quickest foraging routes by allowing for faster traveling speeds through avoiding obstacles on the ground, even if these routes are not necessarily the shortest distance. The variation in texture of natural walkways, characterized as “surface roughness”, further impacts both arboreal and ground ant running speeds and foraging efficiency. Physical connections between trees are thus important structures that facilitate not only arboreal ant mobility but also their foraging success, resource recruitment efficiency, and ant-provided ecosystem services, including pest removal. Ants play an important role in the control of the coffee berry borer , the most damaging insect pest of coffee. In particular, the aggressive arboreal ant Azteca sericeasur nests in shade trees, forages on coffee shrubs, and is a keystone predator that controls the CBB. Like many arboreal ants, A. sericeasur prefers walking on branches and vegetation to avoid traveling on the ground. Given the role of A. sericeasur as a biological control agent, understanding how connectivity at the local scale impacts these ants has potential implications for coffee agroecosystem management. In Chiapas, Mexico, Jiménez-Soto et al. found that artificially increasing connectivity between A. sericeasur nests and coffee plants by tying jute string between ant nest trees and coffee plants increased the capacity for A. sericeasur to remove the CBB by throwing them off the coffee plants.
These results suggest that naturally occurring vegetation connectivity might have a similar effect as that of artificial string connectivity on A. sericeasur activity and their associated pest removal services. Our study tests and expands on this hypothesis by examining the impact of both artificial connectivity and naturally occurring vegetation connectivity on A. sericeasur activity, its ability to recruit to resources, and its removal of the CBB with a manipulative experiment. Specifically, we tested the following hypotheses: We predicted that the coffee plants with vegetation or artificial connections to the ant nest tree have higher A. sericeasur activity than that of the isolated control plants; A. sericeasur ants recruit to resources more efficiently on coffee plants with vegetation or artificial connections to the nest tree; coffee plants with vegetation or artificial connections to the nest tree have greater CBB removal rates by A. sericeasur ants; and A. sericeasur activity, resource recruitment rates, and CBB removal rates decrease with increased distance from A. sericeasur nests. This study was conducted in the Soconusco region of Chiapas, Mexico at Finca Irlanda, a shaded, 300-hectare commercial polyculture coffee plantation. The plantation is located in the Sierra Madre de Chiapas Mountains at an elevation of 1100 m.a.s.l. The average canopy cover throughout the farm is 75 percent and the majority of the plantation shade trees are of the genus Inga. The climate is semi-tropical, with the rainy season occurring between May and October. Vegetation management at Finca Irlanda frequently includes “chaporreo”, in which farmworkers periodically use machetes to clear the weeds and epiphytes that grow between coffee plants. This management practice facilitates farmworker movement between coffee plants and reduces competition between weeds and coffee plants, but in the process inadvertently eliminates vegetation connections between the coffee plants and A. sericeasur nest trees.We collected data between June and August in the summer of 2022.
Within the 300 hectares of Finca Irlanda, we selected 17 trees with active A. sericeasur nests as study sites. Each site was located at least 10 m away from any other active A. sericeasur nests to prevent overlapping ant activity, following the methodology used by Jimenez-Soto et al.. We chose six coffee plants within a 5 m radius of each A. sericeasur nesting tree for a total of 102 coffee study plants. At each nesting tree site, we selected two coffee plants for the natural vegetation connectivity treatment, two for the artificial connectivity treatment, and two as isolated control plants . For the vegetation connectivity treatment, we selected two coffee plants with existing vegetation connections. The vegetation connections were either coffee branches directly touching the A. sericeasur nest tree or coffee branches touching a secondary plant, such as a vine or epiphyte that was touching the nest tree. We selected two coffee plants for the artificial connectivity treatment, in which we tied jute strings between the point of the nest tree trunk with the most active ant foraging trail and the central trunk of each coffee plant. We ensured that there were no existing vegetation connections on these plants and that the string was the only point of connection between each coffee plant and the nest tree. For the control treatment, we selected two isolated coffee plants with no connections between the coffee plants and the nest tree. We measured the distance between the central trunk of each study coffee plant and the ant nest tree. At each site, we quantified the ant activity on the coffee plants by counting the number of A. sericeasur that passed a central point on the central trunk of each coffee plant during 1 min . The observations took place between 7:30 AM and 2 PM before the afternoon rainy period. The observations were stopped if it began to rain, as rain significantly reduces ant activity. After setting up the strings, we returned to each site between 7 and 13 days after the initial setup and re-measured ant activity on the coffee plants. To assess the impact of artificial and vegetation connectivity on prey removal by A. sericeasur, plastic plant pot we placed five dead adult female CBB on white index cards on the centraltrunk of each coffee plant . We monitored A. sericeasur interactions on the cards for one hour, ensuring that only A. sericeasur were responsible for removing CBB, and counted the number of CBB removed. Because it has already been well-documented that A. sericeasur remove live CBB from coffee plants, we used dead prey to avoid the possibility of live CBB escaping during a longer observation period. The CBB were collected from infested coffee berries in the field, then frozen for up to 5 days before use. Recruitment is understood to be an integral component of trail-following in which ant workers follow chemical foraging trails to a food source, then re-apply chemical trails until that food source is exhausted. Tuna baiting is an effective and widely used method of assessing ant recruitment in coffee agroecosystems. To assess the impact of connectivity on ant resource recruitment efficiency, we placed 1 g of canned tuna on the central trunk of each coffee plant 1 m above ground and recorded the number of A. sericeasur that recruited to each tuna bait after 20 min. To test for statistical differences in ant activity, resource recruitment efficiency, and CBB removal between the control, string, and vegetation treatment coffee plants over the 5-week experiment, we fit our data with generalized linear mixed models using the lme4 package in R. For each response variable , we included the time , the treatment method , the distance between the coffee plant and the ant nest tree , the interaction between the treatment and the time, and the interaction between the time and the distance as fixed effects.
As random effects, we modeled the coffee plant identity nested within the site to control for site variation and spatial non-independence. To assess count data , we originally fit each model to a Poisson distribution. However, to correct for observed over-dispersion, we instead modified each model to a negative binomial distribution. We observed A. sericeasur using the artificial string connections at 12 of the 17 sites and on 20 of the 33 strings placed . A. sericeasur were the only ants observed using the strings. Out of 33 vegetation treatment plants , 20 plants included primary connections , and 13 plants were connected by secondary connections . We observed A. sericeasur utilizing vegetation connections on every vegetation treatment plant. Ant activity was higher on the vegetation treatment coffee plants than on both the control treatment and the string treatment . There was a significant effect of time on ant activity for the string treatment, indicating an increase in ant activity on the string plants after connecting the strings . There was no effect of time on ant activity for either the control or vegetation treatments, indicating that there was no significant change in ant activity for those treatments over the 5-week experiment . Treatment, time, and distance all impacted the ant recruitment to tuna baits. More ants recruited to the tuna baits on the vegetation treatment plants than on the strings, and more ants recruited to the bait on string plants as compared to the control plants . The overall number of ants that recruited to the baits decreased with time post string placement on both the control and vegetation plants, but there was no significant change in the number of ants recruiting to the baits on the string treatment . The number of ants recruiting to the baits on the control and string plants declined with distance from the nest tree, but was consistent over all distances for vegetation treatment plants .Treatment, time, and distance all impacted the number of CBB removed by ants. Ants removed more CBB from the natural vegetation treatment plants than from the string plants and removed more CBB on the string plants compared to the control plants . The overall number of CBB removed from the vegetation plants decreased with time post-string placement . The number of CBB removed on the control and string plants declined with distance from the nest tree, but was consistent over all distances for vegetation treatment plants .This study asked how connectivity, occurring naturally as vegetation or artificially as string connections, influences A. sericeasur activity, foraging efficiency, and pest removal services in coffee systems. Our research demonstrates that naturally occurring vegetation connectivity and, to a lesser extent, artificial connectivity between A. sericeasur nest trees and coffee plants increased both A. sericeasur mobility and CBB removal on coffee plants. Between the control, string, and vegetation connectivity treatments, all response variables were highest on coffee plants with naturally occurring vegetation bridges between the coffee plants and the ant nest tree.