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. Although string connections did increase ant resource recruitment efficiency and pest removal rates compared to the control treatments, A. sericeasur exhibited a clear preference for natural vegetation over string connections.
Interestingly, while distance from the A. sericeasur nest tree did negatively impact ant activity, recruitment to baits, and ant-mediated CBB removal on the control and string treatments, drainage collection pot distance did not affect these response variables on the vegetation treatment plants. Vegetation connectivity influences the distribution, diversity, and interspecific competition of arboreal ant species by affecting the availability of nesting habitats, foraging ranges, and resource availability. Higher degrees of vegetation connectivity provide a range of food resources to arboreal ants, including access to honeydew-producing insects, extrafloral nectaries, and other insects. Arboreal ants can take advantage of these resources more efficiently when connected vegetation provides a network of foraging opportunities, which increases access to patchy resources while enhancing predator avoidance capability. In contrast, disconnected vegetation may limit ant distribution to isolated tree patches.Observed increases in ant activity on plants with vegetation connectivity suggest that structural connectivity facilitates ant mobility and movement efficiency on foraging paths. Between the control, string, and vegetation treatments, ant activity was highest on vegetatively connected plants . After string placement, ant activity did increase on the string plants, indicating that A. sericeasur learned to use strings as foraging paths over time; however, the overall ant activity levels on the string treatment plants were not significantly different from the control treatment. The significant positive interaction between the string treatment and time explains the observed increase in ant activity on the string treatment plants. On the control and vegetation treatment plants, there was no change in ant activity throughout the 5-week duration of the experiment. Consistent with Jiménez-Soto et al., ant activity decreased on coffee plants with increasing distance from A. sericeasur nest trees on the control and string plants. Notably, our additional treatment of naturally occurring vegetation connectivity overrode the effect of distance from ant nest trees, with no impact of distance on the amount of ant activity on the vegetation treatment plants. This important finding suggests that vegetation pathways can facilitate A. sericeasur foraging activity on coffee plants that are farther away from their nest tree.
Allowing a longer network of vegetation connectivity could even increase A. sericeasur foraging ranges. Additionally, increasing their foraging range may help the ants to avoid parasitic phorid flies in the genus Pseudacteon, which parasitize A. sericeasur and decrease in density with increasing distance from the A. sericeasur nests. A. sericeasur activity may be highest on coffee plants with naturally occurring vegetation connections because vegetation connections are generally larger and more structurally complex. Additionally, on the existing vegetation pathways, the ants had more time to establish foraging trails and chemical cues as compared to the string connections. Furthermore, in addition to providing linear foraging trails, vegetation bridges may also contain useful resources including extrafloral nectaries and plant fluids that the strings do not provide. Vegetation pathways can also offer protection from phorid flies beneath the leaves, whereas strings are open and unprotected foraging paths. Studies have also suggested that ants have preferences for foraging on certain surfaces, and that surface characteristics impact foraging speed and chemical communication on the ants’ trails. The A. sericeasur preference for vegetation surfaces may therefore result from texture-based foraging efficiency differences between vegetation and string. Yanoviak et al. studied ant recruitment to baits on bare vs. moss-covered tree trunk surfaces and observed the Azteca spp. actively avoiding baits on moss-covered trunks, indicating a clear surface preference for smoother pathways. In our study, we observed A. sericeasur walking around stray threads on the jute strings , decreasing their foraging efficiency compared to smoother thread-free vine surfaces. In some instances, we observed A. sericeasur “cleaning” the string pathways by biting off jute string threads from the connections to minimize obstacles and enhance their efficiency on these pathways. Another explanation for higher A. sericeasur activity on the vegetation treatment coffee plants is that A. sericeasur may already be tending established green scale colonies on vegetatively connected plants, drawing their activity to these plants over the string treatment plants. C. viridis, a sessile coffee scale insect, has been linked to increased A. sericeasur activity. In a mutualistic relationship, A. sericeasur protect C. viridis from predation in exchange for the honeydew that these scales produce. Increased connectivity, by increasing ant mobility, may also increase the scale tending activity by A. sericeasur. Notably, interactions between A. sericeasur and CBB on coffee plants occur more frequently with higher densities of C. viridis, indicating a relationship between scale tending activity and CBB control services.A. sericeasur recruited most efficiently to tuna baits on the vegetation treatment plants, and significantly more ants recruited to tuna baits on the string treatment plants than on the control plants . The significant difference between control and string treatments found in our resource recruitment results, but not seen in our ant activity results, may have occurred because the ants did not have an incentive to travel to the string treatment plants in the absence of the tuna baits.
Our results indicate vegetation connectivity, and, to a lesser extent, artificial connectivity increases the discovering and recruiting to resources. In other systems, the ants discover resources more A. sericeasur efficiency at quickly by walking on fallen branches than by traveling through leaf litter; in this case, the fallen branches are a form of vegetation connectivity that advantageously increase colony resource acquisition by reducing the searching effort. Consistent with the ant activity results, the number of ants recruiting to baits on control and string plants declined with distance from the ant nest tree but remained consistent over all distances for vegetation treatment plants. These results confirm the A. sericeasur preference for vegetation foraging paths over artificial ones, as explained in Section 4.1. Between treatments, 10 liter pot the control treatment plants had the lowest ant recruitment to baits. Other ants in tropical systems similarly prefer vegetation pathways over ground and leaf litter, optimizing networks of vines, leaves, and branches in their foraging trails. Clay et al. suggest that ants may even favor vines over bark or moss because the linear nature of vines reduces the necessity for intensive chemical trail maintenance. Strings might similarly provide this linear path advantage, which reduces chemical trail maintenance and opportunities for path confusion compared to ground trails. Because ants account for energy efficiency when deciding between foraging paths, the control plant baits were likely the least attractive because they required the highest energy expenditure due to traveling over ground and leaf litter. Because none of the tuna baits were depleted within the 20 min observation period, recruiting to control baits while more energy-efficient paths were present is an inefficient use of ant workforce.Over time, 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 that recruited to the baits on the string treatment. Decreases in ant recruitment rates on the control and vegetation treatment plants may have resulted from the presence of phorid flies, which greatly reduce A. sericeasur activity. Phorid fly attacks may have curtailed ant recruitment along the pre-existing foraging routes, as phorids are likely more abundant in leaf litter along popular foraging routes. Because the strings were a novel foraging route, it is likely that fewer phorids frequent those routes and interfere less with ant recruitment to the baits.Between treatments, A. sericeasur removed the most CBB from vegetation treatment plants and removed more CBB on the string plants as compared to the control plants . Decreases in CBB removal may similarly have resulted from phorid fly attacks inhibiting pest removal activity, as occurred in Philpott et al. and Pardee and Philpott. Consistent with the results of our ant activity and resource recruitment experiments, the number of CBB removed on the control and string plants declined with distance from the nest tree, but remained consistent with distance from the nest on the vegetation treatment plants. Interestingly, Jiménez-Soto et al. did not find any effect of distance from the nest tree on CBB removal for the control or string plants. Our contrasting results may be the result of the A. sericeasur preference for the vegetatively connected plants in our study; in the absence of vegetation pathways, ants may forage more on artificial connections. Our results reinforce how habitat complexity in the form of vegetation connectivity impacts interspecific interactions, specifically ant-mediated CBB removal at the local scale.Our results confirm the importance of naturally occurring vegetation connectivity and habitat complexity in facilitating arboreal ant mobility and ant-mediated CBB removal. Our findings have important implications for the practical application of ant-provided pest removal in coffee systems, indicating that A. sericeasur may most effectively control CBB on coffee plants with natural vegetation connectivity connected to their nest trees. In the absence of vegetation connectivity, implementing artificial connections between ant nests and coffee plants can increase CBB removal by A. sericeasur; however, with increasing distance from the ant nest tree, the strength of this pest removal service decreases on artificially connected plants. The observed preference of A. sericeasur for vegetation pathways underscores the importance of maintaining or promoting vegetation connectivity via habitat complexity and structural diversity within coffee agroecosystems. In managing agroecosystems in support of ant-mediated ecosystem services, artificial connectivity does not provide an equal substitute for the naturally occurring vegetation connectivity provided through forest conservation and structural complexity. Consistent with studies affirming the influence of vegetation connectivity on predatory arthropod movement and predation range, our results illustrate how vegetation connectivity facilitates A. sericeasur foraging mobility and pest removal. In coffee systems, higher degrees of vegetation connectivity are associated with shade trees, as well as more heterogeneous habitat complexity and variability in plant structure. In other studies, ants generally increase predation services in shaded systems as compared to monocultures and, in coffee plants, more effectively remove CBB in shaded coffee systems as compared to sun monoculture systems. Interestingly, most studies find the opposite effect of structural complexity on parasitoid behavior, with higher degrees of plant structural complexity leading to decreased parasitoid foraging efficiency. This negative relationship between parasitism and habitat complexity transfers to coffee systems, where the parasitic phorid flies exert a greater inhibiting effect on Azteca ants in simple, low-shade farms thanin complex, high-shade farms. Together with the aforementioned study, our combined results illustrate how habitat complexity at the landscape scale and vegetation connectivity at the plot scale dually facilitate A. sericeasur-mediated pest removal: by facilitating ant mobility and by reducing the efficiency of the parasitoid that interferes with their pest removal ability. In order for A. sericeasur to provide ant-mediated pest removal services, coffee agroforests must include enough shade trees to provide sufficient habitats for ant nests. Planting coffee plants close enough to shade trees to allow for direct connectivity and leaving some vegetation connections between coffee plants and shade trees rather than chopping them or relying on herbicides can facilitate ant-provided ecosystem services by providing foraging paths through naturally occurring structural connectivity.