Introduction
As postulated by Elton (Reference Elton1958), disturbed, less biodiverse habitats are the most vulnerable to biological invasions. Consequently, the impoverishment of native biological communities driven by anthropogenic impacts, intensified since the twentieth century (Williams et al. Reference Williams, Leinfelder, Barnosky, Head, McCarthy, Cearreta, Himson, Holmes, Waters, Zalasiewicz, Turner, McGann, Hadly, Stegner, Pilkington, Kaiser, Berrio, Wilkinson, Zinke and Delong2022), has favoured the spread and establishment of invasive exotic species.
Exotic invasive plants typically exhibit high fecundity, producing numerous fruits and seeds that are frequently consumed and dispersed by the local fauna (Richardson et al. Reference Richardson, Pyšek, Rejmánek, Barbour, Panetta and West2000). In such cases, their impact extends beyond competition with local plant species, altering ecological networks in which they participate (Genes and Dirzo Reference Genes and Dirzo2022). One of these plants is the jackfruit Artocarpus heterophyllus Lamk. (Moraceae) in Brazil. Originating from the tropical rainforests of India and Bangladesh, the species has its sweet, fleshy fruits and seeds consumed by at least 11 species of reptiles, birds and mammals within their native distribution, and seed dispersal has been reported to bears and elephants. Outside its native range, at least 32 vertebrate species in 14 countries consumed their fruits and seeds (Taconi and Pires Reference Taconi and Pires2021). The fruits of A. heterophyllus are massive infructescences that can exceed 35 kg, and an adult tree is capable of producing more than 100 fruits annually, each fruit containing more than 500 seeds (Elevitch and Manner Reference Elevitch, Manner and Elevitch2006). Seeds vary from 30 to 40 mm in length and 10–20 mm in width, with a mass ranging approximately between 1.5 and 14 g (Kham Reference Kham2004). Regarding dissemination, the species exhibits primary dispersal syndromes by barochory and secondary dispersal by zoochory. The rich nutritional content of the fruits, combined with their high productivity and availability throughout the year, can lead native fauna to preferentially consume their fruits and alter the dispersal of native seeds (Raíces et al. Reference Raíces, Ferreira, Mello and Bergallo2017). In the Brazilian Atlantic Forest, jackfruit is recognised as an aggressive invader (Bergallo et al. Reference Bergallo, Rocha and Rocha2016), reaching high local densities in diverse localities (e.g. Abreu and Rodrigues Reference Abreu and Rodrigues2010; Bergallo et al. Reference Bergallo, Rocha and Rocha2016; Mileri et al. Reference Mileri, Passamani, Eutrópio and Oliveira2012). One of these areas is Tijuca National Park (TNP; 3953 hectares; 22° 57′ 47″ S and 43° 14′ 40″ W), located within Rio de Janeiro city, southeastern Brazil. The area was deforested for sugarcane and coffee production up to the nineteenth century, when a massive reforestation effort was undertaken to maintain the city’s water supply. Besides native species, exotic trees with commercial and/or ornamental value were also planted, including the jackfruit tree (Freitas et al. Reference Freitas, Neves and Chernicharo2006), which has been spreading throughout the park since then.
Nowadays, despite the successful structural recovery of the vegetation, TNP exhibits signs of the ‘Empty Forest Syndrome’ (Redford Reference Redford1992), including recruitment failures in large-seeded tree species. To address this functional erosion, agoutis (Dasyprocta leporina) were reintroduced to the TNP in 2010, as part of a trophic rewilding programme (Fernandez et al. Reference Fernandez, Rheingantz, Genes, Kenup, Galliez, Cezimbra, Cid, Macedo, Araujo, Moraes, Monjeau and Pires2017). The goal was to establish a self-sustaining population and to restore ecological interactions and processes (Cid et al. Reference Cid, Figueira, Flora, Pires and Fernandez2014). Although the exact cause of the local extinction of agoutis at TNP is unknown, previous deforestation, coupled with hunting and attack by domestic dogs – which was common in the past – are likely contributing factors. Agoutis play a crucial role as large seed dispersers, promoting its recruitment through scatter-hoarding (Mittelman et al. Reference Mittelman, Dracxler, Santos-Coutinho and Pires2021). Thus, it was expected that they would favour the seed dispersal and seedling recruitment of some TNP large-seeded tree species, as already reported for the palm Astrocaryum aculeatissimum (Zucaratto and Pires Reference Zucaratto and Pires2015) and for a tree in the Euphorbiaceae family, Joannesia princeps (Mittelman et al Reference Mittelman, Kreischer, Pires and Fernandez2020).
In the case of jackfruit, conflicting traits present an ecological puzzle. On the one hand, the large seed size and pulp availability suggest potential for dispersal, matching size selection criteria by large rodents (Jansen et al. Reference Jansen, Bongers and Hemerik2004). On the other hand, the lack of mechanical protection (no hard endocarp) and the recalcitrant nature of jackfruit seeds (Warrier et al. Reference Warrier, Singh, Ravichand, Sivakumar, Geetha, Kumar and Hegde2009) are traits that typically favour immediate consumption over long-term hoarding (Preston and Jacobs Reference Preston and Jacobs2009). Given this duality, the specific nature of the interaction between D. leporina and A. heterophyllus remains unknown since its reintroduction. It remains uncertain whether agoutis act primarily as dispersers (facilitating invasion) or predators (controlling it) for this exotic species. Considering this, our objective was to evaluate the role of D. leporina in the seed fate of A. heterophyllus by comparing seed predation and removal rates in areas with and without agoutis in TNP. We hypothesised that D. leporina would act primarily as a seed predator rather than a disperser for this exotic species, given that the rapid germination and/or the high perishability of its seeds makes it unsuitable for hoarding. Thus, we predict that the presence of agoutis would lead to higher seed predation rates, thereby enhancing biotic resistance against invasion, with negligible effective dispersal.
Methods
Seed removal experiments were carried out in 10 sites where agoutis had established home ranges after release (Cid et al. Reference Cid, Figueira, Flora, Pires and Fernandez2014) and 10 sites without the presence of these animals; the minimum distance among points was 200 m. Agoutis’ absence was confirmed using camera trap surveys, using jackfruit seeds as bait, carried out immediately before the beginning of the experiments. Experiments took place in 2014 during the peak (February) and low (July) fruiting seasons of Artocarpus heterophyllus (Abreu and Rodrigues Reference Abreu and Rodrigues2010).
We applied two treatments to assess the role of different seed predators, open control and selective exclusion. In the open control treatment, seeds were placed directly on the forest floor, accessible to all terrestrial fauna. The selective exclusion treatment, adapted from Mileri et al. (Reference Mileri, Passamani, Eutrópio and Oliveira2012), excluded medium and large mammals (such as D. leporina), allowing access only to small rodents. In this treatment, seeds were placed inside PVC pipes (45 cm length, 7.5 cm diameter) fixed to the ground. Each site received one station of each treatment, spaced 5 m apart, with five seeds per station. To track seed fate, each seed was attached to a 30-m thread bobbin, which was placed inside a small container attached to a nearby tree. To facilitate identification in case of line breakage, each seed had a numbered tag matching its numbered container. Experiments were monitored weekly for one month. Seeds were categorised as either ‘predated by vertebrates’ or intact (not damaged by vertebrates). Seeds removed from experimental stations but not found were considered as preyed upon. Animals that interacted with seeds in the experiments were identified through camera traps, placed in a subset of sampling points, totalling 380 camera-days. For the same species, only images obtained after a one-hour interval were considered as independent records.
To test if treatments (open vs. exclusion) and areas (with vs. without agoutis) influenced seed survival, we fitted a Cox proportional hazards model using the ‘survival’ package in R 4.0.1 (R Core Team Reference Team2020; Therneau Reference Therneau2020). We used clustering with robust standard errors at the levels of location and time period to account for the correlation between seeds located in the same plot and sampled in the same season.
The data and code that support the findings of this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.19679214.
Results
Of the 400 monitored seeds, 292 (73%) were predated by vertebrates. Camera-trapping monitoring registered four seed consumers at the experiments: the agouti D. leporina (n = 9 independent records), followed by the black rat Rattus rattus (n = 8), the spotted paca Cuniculus paca (n = 4) and the southern black-eared opossum Didelphis aurita (n = 3). No seed was dispersed or scatter-hoarded. During the low fruiting season (July), which corresponds to winter in TNP, seed deterioration due to fungal attack was observed for remaining seeds at the end of monitoring. Fungi affected 11 seeds in the open control (5.5% of 200) and 34 seeds in the exclusion treatment (17% of 200).
Seed survival probability was significantly lower in areas with agoutis when seeds were exposed (Figure 1, n = 200, z = −3.74, P < 0.001). However, in the selective exclusion treatments, there was no statistically significant difference between areas (Figure 1, n = 200, z = 1.77, P = 0.07). The selective exclusion treatment increased seed survival compared to open controls in areas with agoutis (Figure 1, n = 200, z = −4.79, P < 0.001). However, in areas without agoutis, there was no significant difference between exclusion and control (Figure 1, n = 200, z = 0.61, P = 0.54). The occurrence of other seed consumers (Cuniculus paca, Rattus rattus, Didelphis aurita) did not affect removal rates between areas without agoutis and those where they were present. Overall, limiting agouti access – either spatially (areas without agoutis) or experimentally (exclusion treatments) – significantly increased seed survival (Figure 1), highlighting the determinant role of D. leporina in A. heterophyllus seed predation at TNP.
Agoutis reduce jackfruit seed survival. Lines represent the seed survival probability of Artocarpus heterophyllus seeds over time under different animal exclusion conditions in areas with and without agoutis. Solid lines represent seeds exposed to all animals, while dashed lines represent seeds accessible only to small rodents. Colors distinguish study areas: red tones indicate areas with agoutis, and blue tones indicate areas without agoutis. Shaded regions around each curve represent 95% confidence intervals of the estimated survival probabilities. Estimates and confidence intervals were derived using a Cox proportional hazards model.
Figure 1. Long description
A line graph showing seed survival probability over time in different areas and treatments. The horizontal axis represents time in weeks, ranging from 0 to 4 weeks. The vertical axis represents seed survival probability, ranging from 0.00 to 1.00. The graph includes four lines representing different areas and treatments: Area With Agoutis - Open to all animals, Area With Agoutis - Small rodents only, Area Without Agoutis - Open to all animals, and Area Without Agoutis - Small rodents only. The lines are color-coded and styled differently: solid red for Area With Agoutis - Open to all animals, dashed red for Area With Agoutis - Small rodents only, solid teal for Area Without Agoutis - Open to all animals, and dashed teal for Area Without Agoutis - Small rodents only. The shaded areas around the lines represent the confidence intervals. The seed survival probability decreases over time for all treatments. The Area With Agoutis - Open to all animals shows the steepest decline, while the Area Without Agoutis - Small rodents only shows the least decline. The other two treatments fall in between these two extremes.
Discussion
The enemy release hypothesis (Keane and Crawley Reference Keane and Crawley2002) postulates that exotic species may thrive in certain habitats due to the absence of natural enemies. Our results identify D. leporina as an important jackfruit seed predator in TNP, emphasising the importance of functionally intact communities in preventing invasions by exotic plants. Increased predation rates were strictly confined to areas with agouti presence. Consequently, by exerting such top-down control, agoutis can limit the recruitment and expansion of A. heterophyllus at TNP.
Although rare, some studies have reported jackfruit seed burial by agoutis (Cid et al. Reference Cid, Figueira, Flora, Pires and Fernandez2014; Galetti et al. Reference Galetti, Donatti, Steffler, Genini, Bovendorp and Fleury2010), which was not observed in this study. Jackfruit seeds are a highly perishable food source rather than a storable resource. Recalcitrant jackfruit seeds require specific humidity and temperature conditions for their rapid germination and quickly lose viability and rot when germination fails (Neves Reference Neves1994; Warrier et al. Reference Warrier, Singh, Ravichand, Sivakumar, Geetha, Kumar and Hegde2009), making them an unreliable reserve for the future. In a short-term basis, by its turn, the great abundance of seeds could prevent agoutis from caching seeds to keep them safe from competitors. Thus, while D. leporina acts as an effective disperser for native species with storable seeds at TNP, such as Astrocaryum aculeatissimum (Zucaratto and Pires Reference Zucaratto and Pires2015) and Joannesia princeps (Mittelman et al Reference Mittelman, Kreischer, Pires and Fernandez2020), it functions exclusively as a seed predator for A. heterophyllus.
Asynchronous fruiting ensures fruit presence in July, but seed survival during this low-fruiting period is further compromised by fungal attacks. Although the Park management plan (ICMBio 2008) mentions this pathology, our study quantifies its specific impact as a supplementary biotic filter during the winter. During this season, reduced light incidence, combined with consistently higher humidity, can foster the proliferation of fungi (Tapia et al. Reference Tapia, Castillo-Novales, Riquelme, Valencia, Larach, Cautín and Besoain2025).
Our study underscores the potential of trophic rewilding not merely as a strategy for fauna recovery but as a cost-effective tool for invasive species management. The high predation rates suggest that D. leporina acts as an effective biological filter against A. heterophyllus, a species with otherwise high initial viability (Haq Reference Haq2006). We propose a synergistic management framework: by spatially expanding the reintroduced agouti population, managers can simultaneously restore lost ecological functions and maximise predation pressure on exotic flora in new areas of the park. In the context of novel ecosystems (Hobbs et al. Reference Hobbs, Arico, Aronson, Baron, Bridgewater, Cramer, Epstein, Ewel, Klink, Lugo, Norton, Ojima, Richardson, Sanderson, Valladares, Vila, Zamora and Zobel2006), where traditional eradication of invasive flora is often unfeasible from a logistical or financial standpoint, the use of reintroduced native antagonists offers a cost-effective and self-sustaining management strategy (Derham et al. Reference Derham, Duncan, Johnson and Jones2018).
Acknowledgements
The authors thank Tijuca National Park staff and Refauna team for logistical support. RCM received a scholar grant from CAPES, and ASP received research grants and financial support from CNPq (Proc. no. 311791/2022-0) and FAPERJ (Proc. no. 202.650/2019). Two anonymous reviewers and Emerson M. Vieira provided valuable suggestions that significantly improved earlier versions of the manuscript.
Financial support
This work was supported by CAPES (R.C.M), CNPq (A.S.P, grant number 311791/2022-0) and FAPERJ (A.S.P., grant number 202.650/2019).
Competing interests
The authors declare no conflicts of interest.
