| Introduction

Tropical deforestation is the primary driver of the worldwide loss of forests, occurring at an alarming rate of 5.5 million hectares per year (Veldkamp et al. 2020). Implementing effective restoration strategies in the tropics is a global priority (Strassburg et al. 2020) that would have a positive impact on preserving biodiversity and recovering ecosystem functionality (Derhé et al. 2016;Nanni et al. 2019). One limiting factor in restoration seems to be the availability of seeds (Holl 1999;Melo et al. 2009;Shaw et al. 2020;Wijdeven and Kuzee 2000). One strategy to increase seed availability in deforested areas is to promote the activity of seed-dispersing animals (Freeman et al. 2021;Silva et al. 2020). In neotropical ecosystems, fruit-eating bats and birds play a fundamental role in dispersing many pioneer trees and shrubs (Galindo-González, Guevara, and Sosa 2000), promoting and sustaining forest diversity (De la Llata Quiroga, Ruedas, and Mora 2022; Muscarella and Fleming 2007;Ripperger et al. 2019). In comparison to birds that tend to deposit seeds under trees, bats are more likely to fly over open areas and defecate seeds during flight (De la Llata Quiroga, Ruedas, and Mora 2022; Medellin and Gaona 1999). Therefore, fruit-eating bats are ideal candidates for facilitating forest recovery (Parolin et al. 2021).

Several strategies have been implemented to attract bats to deforested areas. As fruit bat foraging is predominantly guided by olfaction (Ripperger et al. 2019;Santana et al. 2021;Thies, Kalko, and Schnitzler 1998), some strategies rely on deploying scent lures. Examples of scent lures include the deployment of commercial fruits (Preciado-Benítez et al. 2015), and essential oils extracted from bat-dispersed plants (Bianconi et al. 2007(Bianconi et al. , 2012;;Cabral et al. 2016;Mikich et al. 2003). Lured bats are assumed to be attracted to the essential oils, inspecting these as potential new food sources. During the process, bats would likely defecate, increasing the seed availability in the area (Bianconi et al. 2012;Parolin et al. 2021). While the efficacy of essential oils extracted directly from ripe fruits as bat attractants has been documented (reviewed in Parolin et al. 2021), it remains unknown whether attracting bats to open areas can effectively increase seed rain, which might contribute to forest regeneration (Camargo et al. 2020;Reid and Holl 2013; but see Holl, Joyce, and Reid 2022). Although effective at attracting bats, essential oil production requires large-scale collections of wild ripe fruits, which are often difficult to locate, followed by expensive extraction processes. A more feasible approach would involve designing lures using synthetic compounds, informed by a fundamental understanding of the chemical mechanisms involved in fruit-frugivore interactions. Chemical ecology studies have identified the main VOCs (volatile organic compounds) in the scent of bat-dispersed fruit (Hodgkison et al. 2013;Leiser-Miller et al. 2020;Ripperger et al. 2019) and linked some of these compounds to bat attraction (Santana et al. 2021), providing valuable insights for the development of synthetic lures.

In this study, we tested the effectiveness of synthetic chemical lures to attract fruit bats and increase seed rain in a lowland rainforest in northeastern Costa Rica. The lures tested are composed of a few commonly found volatiles in bat-dispersed ripe fruits. Specifically, our study aims to assess the impact of chemical lures on fruit bat captures and community composition (objective 1). Additionally, we aim to describe the seed rain (i.e., the collection of dispersed seeds found in traps) in the study area and assess the impact of chemical lures on seed collection and seed rain composition (objective 2). We hypothesize that fruit bats will be attracted to the chemical lures, increasing bat capture success and seed rain in the study area. Although additional research would be required to assess the broader ecological consequences and long-term effectiveness of lures, identifying specific volatile compounds that attract seed-dispersing animals lays the foundation for developing novel forest regeneration strategies.

| Methods

| Study Site and Study Organisms

This study was conducted at La Selva Biological Station (hereafter La Selva), province of Heredia, Costa Rica. La Selva is a 1536ha protected area managed by the Organization for Tropical Studies (OTS) and connected to Braulio Carrillo National Park.

La Selva encompasses diverse landscapes, including primary forests, secondary forests, and abandoned plantation areas (McDade et al. 1994), characterized by high abundances of plant and animal species. Specifically, in La Selva-Braulio Carrillo, more than 80 bat species have been reported in the area (Timm et al. 1989).

| Chemical Lures

| Volatile Organic Compound (VOC) Selection

We selected the VOCs primarily based on studies that explored the chemical ecology of the interaction between Carollia bats (Phyllostomidae) and Piper plants (Piperaceae), one of the most ubiquitous seed dispersal interactions in the Neotropics (Fleming 1988;Florez-Montero et al. 2022). Due to logistical limitations in purchasing and transporting the synthetic volatiles to the study site, we were unable to test lures with identical chemical compositions in both objectives. Recent studies have elucidated the VOC composition of ripe Piper infructescences, and have established the role of some of them in attracting bats (Leiser-Miller et al. 2020;Santana et al. 2021). For objective 1, the chemical lures tested contained 2-heptanol, an aliphatic alcohol, and β-caryophyllene, a sesquiterpene. Both VOCs are present in the ripe scent of the common bat-dispersed fruit, Piper sancti-felicis (Piperaceae) (Leiser-Miller et al. 2020), and several Ficus spp. (Moraceae) (Borges, Bessière, and Hossaert-McKey 2008;Hodgkison et al. 2007). Captive experiments have shown that these VOCs elicit positive responses in the fruit-eating bat Carollia castanea (Santana et al. 2021). For objective 2, ocimene, together with 2-heptanol and β-caryophyllene, was incorporated into the lures. Ocimene is one of the most common monoterpenes in nature (Mostafa et al. 2022) and it is present in the ripe scent of several neotropical bat-dispersed species, including Piper spp. (Da Silva et al. 2017;Leiser-Miller et al. 2020) and Ficus spp. (Hodgkison et al. 2013). The selected compounds are likely to elicit a positive response, are commercially available, inexpensive (Table S1), and not restricted to Piper plants, which could potentially attract several bat species.

| Chemical Lure Design

For objective 1, two synthetic VOCs were delivered through a system engineered from 2 mL autosampler glass vials and paper filter strips. We combined 1 mL of 2-heptanol (Sigma-Aldrich, H3003, 98%) and 1 mL of β-caryophyllene (Sigma-Aldrich, W225207, ≥ 80%) in a 2 mL vial and sealed with a septum. We inserted three 15 × 0.5 cm filter paper strips (Whatman No. 3) into the septum to enable a slow and constant release of volatile compounds. Finally, we attached two 2 mL vials to poles at approximately 2 m, positioning them upside down to facilitate the release of volatiles into the environment. These lures were prepared and deployed daily. For objective 2, we delivered three volatiles in 20 mL vials, as these chemical lures were used for an extended period. We placed 6 mL of 2-heptanol (Sigma-Aldrich, H3003, 98%), 6 mL of β-caryophyllene (Sigma-Aldrich, W225207, ≥ 80%), and 6 mL of ocimene (Sigma-Aldrich, W353901, a mixture of isomers, stabilized, based on the QC data, the (Z)-beta isomer (cis) comprises 26.5%, and the (E)-beta ocimene (trans) comprises 69.0%, ≥ 90%) in the 20 mL vial. Then, we sealed the vial and inserted three 15 × 0.5 cm filter paper strips into the septum for volatile release. Finally, we attached two 20 mL vials, positioned upside down, to poles at approximately two meters. These lures were prepared and deployed every 15 days. All volatiles were used as the pure form purchased from the provider, without dilution.

| Assess the Impact of Chemical Lures on Fruit Bat Captures and Community Composition (Objective 1)

This experiment was conducted in March 2023, throughout 12 non-consecutive nights. We selected six different open and semiopen areas (Figure S1). We considered each site and each sampling night a replicate (Figure S3; N Control = 28, N Chemical lures = 27). The sites were situated in areas with considerable human activity, including artificial lights and noise. Nightly, we deployed the chemical lures in randomly selected treatment sites, while the remaining sites served as controls (Figure S3). To monitor bat activity, we placed one 12-m mist net (Avinet Polyester 38 mm mesh) per site, with the chemical lures positioned approximately half a meter from the mist net, and roughly 10-20 m from the forest edge. We set up a minimum of two and a maximum of six mist nets each night, one per site (Figure S3). We deployed the chemical lures at 5:00 pm, and captured bats from 6:00 pm to 11:00 pm, except for nights with heavy rain when we temporarily closed the mist nets until weather conditions improved. To prevent bats from learning to avoid mist nests, we slightly shifted the position of each mist net every 2 or 3 days. After capturing individuals, we used a taxonomic key to identify fruit bat genera (York et al. 2019), and when possible, we recorded the sex, forearm length, and reproductive status. Bats belonging to different guilds were classified as nectarivorous, insectivorous, or hematophagous. After the data collection process, each individual was released in proximity to the capture location.

| Assess the Impact of Chemical Lures on Seed Collection and Seed Rain Composition (Objective 2)

This experiment was conducted between April and July 2023. We selected four open and semi-open areas, distinct from those used in objective 1 (Figure S2), and estimated seed rain over 104 days (Figure S3). We considered each site and each 15-day collection event as a replicate (Figure 3; N Baseline = 12, N Chemical lures = 8, N Treatment = 8). We constructed seed traps to deploy at each site using four polyvinyl chloride (PVC) pipes measuring 0.5 × 0.5 m and a fine plastic mesh. These seed traps capture fecal samples from flying vertebrates that defecate on top of them. At each site, we positioned an array of eight seed traps, totaling 32 seed traps. To collect seeds, we pooled all the samples from the eight traps at each site. During the first 48 days (baseline period), we collected seeds to determine an initial assessment of seed trap effectiveness and identification of the common seeds obtained in each study site. However, due to the differences in length period collection, the baseline period was not included in the statistical comparison between controls and treatments. Then, during the next 56 days, we placed the chemical lures in the center of the eight seed traps in half of the sites, while the other half remained as control sites. At 15-day intervals, we inspected and collected defecated seeds, switched treatment and control sites, and replenished the chemical lures (Figure S3). To identify the collected seeds, we used a publicly available seed reference collection (ECO project, created by Ricardo Bedoya) in the herbarium at La Selva (detailed information about seed collection is located in Table S2, Figures S4 and S5).

| Statistical Analysis

We performed all statistical analyses in R v.4.2.1 (R Core Team 2021). Generalized linear mixed models (GLMMs) described were performed using the glmmTMB package (Magnusson et al. 2017), and multivariate analyses described below were performed using the vegan package (Oksanen et al. 2020).

For objective 1, we assessed the overall impact of chemical lures on fruit bat activity using GLMMs. Total fruit bat count was included as the response variable, treatment (control or chemical lure) as a fixed effect, and site and capture date were included as random intercepts coded as (1|site) + (1|date). Initially, count data were modeled using the Poisson distribution. Then, we examined the assumptions of overdispersion and zero inflation of the best model using the "check_overdispersion()" and "check_ zeroinflation()" functions from the performance package (Lüdecke et al. 2021). When zero inflation or overdispersion was detected, we added a zero inflation term in our model, coded as zi = ~site. If overdispersion or zero inflation persisted, data was modeled using a negative binomial distribution (Brooks et al. 2017). Then, we constructed a set of GLMMs to assess the impact of chemical lures on the count of each bat species following the same approach. We reported effect sizes based on estimated marginal means calculated using the "emmeans()" function from the emmeans package (Lenth et al. 2021). Model parameters and the results of overdispersion and zero inflation tests are presented in Table S3. To address whether olfactory cues from previous nights influenced the capture success, we checked for temporal autocorrelation in model residuals, using an autocorrelation function (ACF) analysis. Then, we used a Durbin-Watson test to test for statistically supported temporal autocorrelation, using the "dwtest()" function from the lmtest packages (Hothorn et al. 2015). We did not find evidence of temporal autocorrelation (Figure S6). Finally, we explored differences in fruit bat communities across sites and treatments. Given that the most common fruit bat species within the same genus often consume similar diets (i.e., Maynard et al. 2019), we visualized the fruit bat communities at the genus level (except for Ectophylla alba, which is a monotypic genus) across sites and treatment using non-metric multidimensional scaling (NMDS) analyses, using the Bray-Curtis distance. To test for statistically supported differences among different sites and treatments, we first checked for homogeneity of variances using the function 'betadisper()' with the centroid type. Then, we conducted a permutational multivariate analysis of variance (PERMANOVA) using the function 'adonis2()' with 999 permutations, and when evaluating the treatment effect, we constrain permutations by site (strata = site).

For objective 2, we followed a similar approach. We assessed the overall impact of chemical lures on seed rain using GLMMs. Total seed count was included as a response variable, treatment (control or chemical lure) as a fixed effect, and site and collection date were included as random intercepts, coded as (1|site) + (1|date). We examined assumptions of overdispersion and zero inflation of the best model using the 'check_overdispersion()' and 'check_zeroinflation()' functions from the performance package (Lüdecke et al. 2021). As zero inflation and overdispersion was detected, we added a zero inflation term in our model, coded as zi = ~site. Model parameters and the results of overdispersion and zero inflation tests are presented in Table S3. We checked for temporal autocorrelation in model residuals using an autocorrelation function (ACF) analysis. Then, to test for statistically supported temporal autocorrelation we used a Durbin-Watson test, using the 'dwtest()' function from the lmtest packages (Hothorn et al. 2015). We did not find evidence of temporal autocorrelation (Figure S7). We did not fit individual GLMM to the plant families due to the high abundance of zeros in the data. Finally, we explored differences in seed rain communities across sites and treatments. We visualized the seed communities at the family level across sites and treatment using NMDS analyses, using the Bray-Curtis distance. We log-transformed the seed abundances before analysis. To test for statistically supported differences among different sites and treatments, we first checked for homogeneity of variances using the function "betadisper()" with the centroid type, then conducted a PERMANOVA using the function "adonis2()" with 999 permutations, and when evaluating the treatment effect, we constrain permutations by the site (strata = site).

| Results

| Assess the Impact of Chemical Lures on Fruit Bat Captures and Community Composition (Objective 1)

During the 12 non-consecutive nights, we captured 124 fruit bats in mist nets without lures, and 150 bats with lures in mist nests with chemical lures nearby (Table 1). We did not find significant differences in the total number of fruit bats captured with and without chemical lures (p = 0.811; Figure 1; Table S3). However, when we evaluated the effect of the treatment in each bat species, we found that the deployment of chemical lures resulted in a daily ID Total number of bats captured without lures Total number of bats captured with lures Total frugivorous bats 124 150 Artibeus sp. 1 0 Carollia spp. 13 29 Ectophylla alba 60 86 Sturnira spp. 5 2 Uroderma spp. 45 33 Total insectivorous bats 6 5 Total nectarivorous bats 27 14 Total hematophagous bats 1 0 average 2.1-fold increase (107%) in the capture success of Carollia spp. compared to sites without lures (p = 0.048; Figure 1; Table S3).

We did not find a significant effect of the lures on the other two most abundant fruit bat species captured (E. alba, p = 0.287 and Uroderma spp., p = 0.610; Figure 1; Table S3). We found that the presence of chemical lures did not change the composition of fruit bat communities (p = 0.471; Figure 2). Instead, we found that the community was determined mostly by the site (p < 0.001; Figure 2). TABLE 2 | Comparison of seed rain (average number of seeds per collection period) collected with and without chemical lures in four semi-open and open areas. Collections were performed every 15 days from May to July 2023. Total number of seeds is presented in Table S2. Plant family Average number of seeds collected without chemical lures (mean ± standard deviation) Average number of seeds collected with chemical lures (mean ± standard deviation) Piperaceae 2.38 ± 6.32 23.5 ± 56.7 Urticaceae 115.75 ± 271.49 92.75 ± 220.7 Poaceae 1.12 ± 2.42 2 ± 5.66 Moraceae 0.62 ± 1.77 12.12 ± 34.29 Solanaceae 1.88 ± 4.91 0.75 ± 2.12 Unknown 2.62 ± 2.62 1.5 ± 3.21 Total 124.38 ± 269.14 132.62 ± 212.43

| Assess the Impact of Chemical Lures on Seed Collection and Seed Rain Composition (Objective 2)

Between May and July 2023, we collected 995 seeds in seed traps without chemical lures and 1061 seeds in seed traps with chemical lures nearby, belonging to a minimum of five plant families (Table 2 and Table S2; Figures S4 and S5). Our GLMM predicted that the deployment of chemical lures resulted in an average 2.3fold increase (131%) of seed collected every 15 days, compared to seed traps without lures (p < 0.001; Figure 3; Table S3). Similar to the results obtained for the bat community, we found that the presence of chemical lures did not alter the composition of the seed rain communities (p = 0.547; Figure 4). Instead, we found that the seed rain community was determined mostly by the site (p = 0.001; Figure 4).

| Discussion

For almost two decades, using essential oils as bat attractants has been proposed as a strategy to accelerate forest recovery in tropical ecosystems (reviewed in Parolin et al. 2021). While there is compelling evidence demonstrating that different essential oils attract fruit bats (Bianconi et al. 2007(Bianconi et al. , 2012;;Cabral et al. 2016;Mikich et al. 2003), translating this knowledge to application has been hindered by a limited understanding of specific compounds in the oils that could be synthesized and developed into commercially available products. Here, we simplified the use of essential oils extracted from fruits or leaves of bat-dispersed plants by using synthetic chemical lures composed of just a few volatiles that are likely involved in bat attraction. Additionally, the effectiveness of the chemical lures to augment seed rain remained unstudied.

Here, we have demonstrated that the daily deployment of synthetic chemical lures results in a significant increase in the average capture success of Carollia spp. and the 15-day deployment of more complex chemical lures results in a significant increase in the average count of seeds collected. Taken together, our findings demonstrate that synthetic chemical lures attract key fruit bats and have the potential to enhance seed rain, offering a promising strategy for accelerating forest regeneration.

The daily deployment of the synthetic lures resulted in a significant increase in the capture success of Carollia spp., presumably because the VOCs released by the chemical lures are naturally found in the primary food source of Carollia spp. The selection of synthetic volatiles for this study was informed by prior research that identified the scent composition of two common and abundant bat-dispersed plant genera, Piper (Piperaceae) and Ficus (Moraceae). Carollia and Artibeus, two of the most abundant neotropical bat genera, are the primary dispersers of these two plant genera (Fleming 2004;Lopez and Vaughan 2007;Pineda-Lizano and Chaverri 2022). Therefore, our findings suggest that Carollia spp. may rely on the most abundant VOCs found in ripe Piper fruit scent as a foraging cue, consistent with previous observations in captivity (Santana et al. 2021). However, the VOCs used here are not exclusively released by Piper spp. β-caryophyllene is perhaps the most common compound found in the scent of a wide range of wild and cultivated flowers and fruits (Gupta and Phulara 2021;Knudsen et al. 2006). Similarly, 2-heptanol is found in several commercial fruits, such as pomegranate (Tripathi et al. 2014), blackberries (Meret et al. 2011), and bananas (Jordán et al. 2001). In contrast, other studies have shown that the use of essential oils does not consistently attract bat species that primarily feed on the plants from which the oils were extracted. Essential oils extracted from Piper gaudichaudianum (Piperaceae) significantly increased the capture success of Artibeus lituratus, although P. gaudichaudianum is not their preferred food source (Bianconi et al. 2007).

Incorporating other commercially available volatile compounds might increase the diversity of bats attracted to the chemical lures. For instance, it has been shown that monoterpenes (e.g., ocimene, α-pinene) and sesquiterpenes (e.g., α-copaene) are abundant in bat-dispersed fruits and captive experiments have demonstrated these terpenes elicit a positive effect on bat foraging (Hodgkison et al. 2013;Parolin et al. 2019). Future advancements in elucidating the chemical composition of wild fruit scents, along with identifying the scent compounds involved in frugivore attraction, can provide valuable insights for improving the formulation of novel and more complex chemical lures.

The 15-day deployment of synthetic lures resulted in an increased average number of seeds collected, many of which are primarily dispersed by bats (Galindo-González, Guevara, and Sosa 2000). However, the treatment effect accounted for < 5% of the total variance (Table S3), indicating that most of the variability in seed rain was due to other unmeasured factors.

Given the small sample size, our results should be interpreted cautiously. Further studies with larger datasets are necessary to validate these findings. Seeds from pioneer plants, such as Moraceae, Piperaceae, and Solanaceae, grow in early phases of forest succession and create suitable conditions for the future development of more complex plant communities (Delang and Li 2013). In site A, the deployment of chemical lures increased Piper spp. (Piperaceae) seeds. Notably, Cecropia insignis (Urticaceae) seeds, an important pioneer species in neotropical forests (Martínez-Ramos et al. 2021), dominate the seed rain at study site B, regardless of the treatment. Cecropia insignis trees are typically found in mature forest clearings, and their fruits are consumed by both bats and birds (Condit, Pérez, and Daguerre 2010; Pérez and Condit 2023). Among our study sites, we observed that site C exhibited the highest level of human activity, with daily human presence, noise, and considerable artificial light exposure during nighttime. Seeds were not collected at site C during the baseline period or when the site served as the control, likely due to the high disturbance levels, which reduced bat activity. However, during the deployment of chemical lures, we collected 97 Ficus colubrinae (Moraceae) seeds, suggesting that our chemical lures might incentivize bats to visit previously avoided areas. In this instance, F. colubrinae seeds were collected from various seed traps, indicating that the seeds likely resulted from multiple dispersal events. In the study site, many species of birds, and one species of bat, E. alba, are the primary seed dispersers of F. colubrinae (De la Llata Quiroga, Ruedas, and Mora 2022). Considering that birds are more likely to disperse seeds under trees, while bats tend to do so in open areas (De la Llata Quiroga, Ruedas, and Mora 2022), we hypothesize that the majority of seed rain collected in our seed traps is the result of seed dispersion by bats. Regarding the future full implementation of the lures, the synthetic chemical lures tested in this study offer several advantages in comparison to the essential oils obtained directly from plant tissues. The manufacturing capacity of chemical lures based on the extraction of essential oils is constrained by the time-consuming process of fruit and leaf collection and the resource-intensive extraction procedures that require expensive equipment and trained personnel. Second, the nature of fruit chemistry introduces significant temporal and spatial variations (Ripperger et al. 2019; Soler et al. 2018), limiting the reproducibility of the lures. Furthermore, fleshy fruits contain complex mixtures of chemicals, but not all of them are involved in frugivore attraction (Gelambi, Morales-M, and Whitehead 2024; Gelambi and Whitehead 2024; Maynard et al. 2020; Whitehead et al. 2021). Therefore, synthetic chemical lures, containing a few volatiles involved in frugivore attraction, present a more practical, reproducible, and reasonably priced alternative for full-scale implementation. For future manufacturing, the lure design can optimize the release of VOCs by employing a stable matrix. For instance, SPLAT (ISCA Technologies) is an inert controlled-release polymer matrix that has been successfully used to deliver a variety of compounds as a pest management strategy (Mafra-Neto et al. 2013).

Our study contributes valuable insights to the synthetic lure approach but it is important to acknowledge several considerations before considering its full implementation in forest recovery. First, bats could learn that the olfactory cues in the chemical lures are not associated with a food source, potentially leading to a decrease in their attraction to the lures over time. One solution could be combining the chemical lures with other strategies that provide an incentive for bats, such as deploying commercial (Preciado-Benítez et al. 2015), natural fleshy fruits (Silva et al. 2020) and/or planting bat-dispersed plants (Sansevero et al. 2011). Also, our study was short-term and conducted in an area characterized by a remarkable abundance of bats. Future studies should evaluate the long-term effect of the chemical lures in environments where bats are less abundant. The attraction of bats to the lures may have broader negative ecological implications, potentially leading to increased predation risk and wasted energy expended in futile foraging attempts. Regarding the seed rain, caution is warranted when extrapolating our findings to broader and more direct forest recovery strategies. Despite the potential increase in seed rain observed here, previous research has not consistently identified a direct relationship between augmented seed rain and subsequent seedling recruitment (Holl, Joyce, and Reid 2022;Reid, Holste, and Zahawi 2013;Reid and Holl 2013). Future studies should evaluate these crucial parameters to evaluate the direct impact of the restoration method proposed here.

The chemical lures tested here can be used as short-term lures to maximize bat capture success in ecological research. In the Neotropics, fruit-eating bat communities are typically assessed by using mist nests. The use of the lures increases the probability of the fruit bat flying closer to the net, and therefore being caught. Increased success in capturing bats can lead to more accurate estimations of population size and community composition. In addition, the lures can be employed in studies that require the capture of a specific number of fruit bats for behavioral or physiological experiments, minimizing the sampling time and effort.

In sum, our study demonstrates that synthetic chemical lures attract bats from the genus Carollia, a key seed disperser of Piper spp. and other pioneer plants in tropical ecosystems. In addition, our study provides the first evidence demonstrating that chemical lures have the potential to increase the seed rain of several families of pioneer plant species, many of which have the potential to accelerate forest recovery (Muscarella and Fleming 2007).

While the treatment showed a significant effect, the small proportion of variance explained by our treatment indicates that other factors likely play a larger role in shaping seed rain. Future research should investigate the effectiveness of more complex synthetic chemical lures, delivered through a more stable polymer matrix, in attracting a broader bat community and increasing seed rain and seedling recruitment in deforested areas.