Introduction
Endozoochory, or seed dispersal after ingestion by animals, is an integral process shaping plant reproductive success and the spatial structure and species composition of plant communities (Nathan and Muller-Landau, Reference Nathan and Muller-Landau2000; Levine and Murrell, Reference Levine and Murrell2003; Snell et al., Reference Snell, Beckman, Fricke, Loiselle, Carvalho, Jones, Lichti, Lustenhouwer, Schreiber, Strickland, Sullivan, Cavazos, Giladi, Hastings, Holbrook, Jongejans, Kogan, Montaño-Centellas, Rudolph, Rogers, Zwolok and Schupp2019). Approximately half of all fruit-producing, flowering plants are animal-dispersed (Aslan et al., Reference Aslan, Zavaleta, Tershy and Croll2013) and adaptations to facilitate dispersal may be found in up to 80% of tree species in tropical forests (Howe and Smallwood, Reference Howe and Smallwood1982). How animals facilitate seed dispersal can be critical to many ecological questions, including mutualism, plant abundance, seed competition, and coexistence (Levine and Murrell, Reference Levine and Murrell2003; Snell et al., Reference Snell, Beckman, Fricke, Loiselle, Carvalho, Jones, Lichti, Lustenhouwer, Schreiber, Strickland, Sullivan, Cavazos, Giladi, Hastings, Holbrook, Jongejans, Kogan, Montaño-Centellas, Rudolph, Rogers, Zwolok and Schupp2019). The benefit an individual plant receives from an animal disperser can be quantified in terms of seed dispersal effectiveness, specifically as quantity × quality (Schupp et al., Reference Schupp, Jordano and Gómez2010). Within this framework, quantity is the number of seeds dispersed, and quality is the probability of a seed producing a new adult plant (Schupp et al., Reference Schupp, Jordano and Gómez2010).
An important aspect of seed dispersal quality is the effect of gut treatment of frugivores on seed germination (Bewleyl, Reference Bewleyl1997; Schupp et al., Reference Schupp, Jordano and Gómez2010). Animals considered high-quality seed dispersers tend to enhance germination success (Traveset et al., Reference Travaset, Robertson, Rodriguez-Perez, Dennis, Schupp, Green and Westcott2007; Fuzessy et al., Reference Fuzessy, Conelissen, Janson and Silveira2016; Fricke et al., Reference Fricke, Bender, Rehm and Rogers2019) by swallowing fruits whole without damaging seeds, facilitating fruit pulp removal, and providing gentle gut treatment of seed tissue (i.e., in the form of mechanical and chemical scarring), before regurgitating or defecating seeds intact (Levey, Reference Levey1987; Schupp et al., Reference Schupp, Jordano and Gómez2010). Germination success post-gut treatment can be measured as (1) a greater percentage of seeds that germinate and/or (2) a lower time to germination (Traveset et al., Reference Travaset, Robertson, Rodriguez-Perez, Dennis, Schupp, Green and Westcott2007). Across animal taxa, the effect of frugivore ingestion on seed germination is generally positive, though the magnitude of the effect varies. For example, birds, primates, and bats tend to have a greater positive effect on proportion and speed of seed germination compared to other mammals or reptiles (Barnea et al., Reference Barnea, Yom-Tov and Friedman1991; Traveset, Reference Traveset1998; Verdú and Traveset, Reference Verdú and Traveset2004; Traveset and Verdú, Reference Travaset, Verdú, Levey, Silva and Galetti2009; Fuzessy et al., Reference Fuzessy, Conelissen, Janson and Silveira2016).
Germination success can also vary with seed size (Traveset and Verdú, Reference Travaset, Verdú, Levey, Silva and Galetti2009). In tropical tree species, if the resources are available, it is generally advantageous to produce larger seeds, as a greater number of internal resources can decrease seedling competition for limited external resources (Murali, Reference Murali1997; Deb and Sundriyal, Reference Deb and Sundriyal2017). Larger seeds in tropical forests tend to have a greater proportion of seeds germinating and faster germination rates (Daws et al., Reference Daws, Garwood and Pritchard2005, Reference Daws, Crabtree, Dalling, Mullins and Burslem2008; Fuzessy et al., Reference Fuzessy, Conelissen, Janson and Silveira2016; Deb and Sundriyal, Reference Deb and Sundriyal2017, but see Fricke et al., Reference Fricke, Bender, Rehm and Rogers2019). Seed size can also interact with seed retention time in the frugivore gut to affect germination. For example, larger seeds of terrestrial plants tend to spend less time in the gut and are more likely to be regurgitated than defecated because they limit nutrition intake (Levey, Reference Levey1987; Fukui, Reference Fukui2003). Shorter retention times potentially reduce the likelihood that seeds may be excessively abraded, augmenting germination success (Traveset and Verdú, Reference Travaset, Verdú, Levey, Silva and Galetti2009). In birds, larger species are generally perceived as being more effective seed dispersers, as they consume more seeds, disperse them longer distances, and have longer gut retention times; however, this effectiveness is often dependent on seed sizes consumed (Godínez-Alvarez et al., Reference Godínez-Alvarez, Ríos-Casanova and Peco2020). Thus, quantifying the general contribution of various birds to seed success in relation to seed size can give context to avian benefits on local plant populations and greater plant meta-population dynamics (Godínez-Alvarez et al., Reference Godínez-Alvarez, Ríos-Casanova and Peco2020).
Eugenia uniflora is a fleshy fruit-producing tropical tree native to South America (Morton, Reference Morton1987), that is largely vertebrate-dispersed (Stricker and Stiling, Reference Stricker and Stiling2013). Although fruits are similar in size, E. uniflora seeds vary considerably in size among fruits on the same individual tree, which can have a maximum length ranging from 11.0 to 15.5 mm (Smiderle et al., Reference Smiderle, Souza and Souza2016). Toucans (Ramphastidae) are large-bodied, highly frugivorous birds that consume fruits from a wide variety of plants in the Neotropics and are generally considered high-quality seed dispersers as their large gape permits them to swallow fruits whole without damaging seeds (Short and Horne, Reference Short, Horne, Perrins, Bock and Kikkawa2002). For instance, they have been found to be effective whole fruit-swallowing dispersers for Virola trees, removing and dispersing approximately 57% of seeds (Holbrook and Loiselle, Reference Holbrook and Loiselle2009). Here, we assessed toucan seed regurgitation, seed size, and seed germination relationships for E. uniflora. Compared to control seeds that experienced no toucan gut passage, we predicted that E. uniflora seeds regurgitated by toucans (1) would have improved germination success both in terms of the number of seeds germinated and the speed of germination and (2) gut treatment effects would be more prevalent for larger seeds. Finally, we predicted that (3) larger seeds would have a higher number of germinated seeds and would germinate faster, regardless of gut treatment.
Methods
In Costa Rica, several toucan species co-occur with E. uniflora trees. Within this region, we observed collared aracaris (Pteroglossus torquatus), a medium-sized toucan species, eating fruits of E. uniflora. We also found E. uniflora seeds underneath their roost and nest sites. As a result of observed toucan E. uniflora foraging, we picked ripe E. uniflora fruits from two trees at El Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), located 3 km east of Turrialba, Costa Rica (Figure 1) in mid-March 2013. The day after picking, we brought fruits to the Toucan Rescue Ranch in San Isidro de Heredia, Costa Rica (Figure 1) and hand fed them to four yellow-throated toucans (Ramphastos ambiguus) and five keel-billed toucans (Ramphastos sulfuratus). Both Ramphastos species in our captive population readily consumed E. uniflora and regurgitated fruit seeds. We fed five randomly selected E. uniflora fruits per bird within a 1-minute period and collected all seeds that could be located after regurgitation by the toucans. We repeated this procedure for a total of 28 hand-feeding events across the nine toucans (N = 140 seeds). None of the seeds were defecated. We brought all regurgitated seeds to CATIE the same evening in which they were collected and stored them in a sealed container to prevent desiccation.
We began germination trials 2 days after collection from toucans in an indoor laboratory that was subject to temperature variation from outside daily temperatures, which ranged from 17° to 27° C during germination trials (temperatures which are well above the 12° C recommended lower limit for cultivation; World Agroforesty Centre). The morning of our trials, we collected E. uniflora fruits as controls from the same two trees harvested for regurgitation trials and mechanically removed the seeds from fruit pulp (Traveset et al., Reference Traveset, Riera and Mas2001). Each treatment and control seed was wrapped in a moist piece of paper towel approximately 6 × 6 cm and placed in separate petri dishes (Young and Evans, Reference Young and Evans1977). We randomly assigned all control and toucan-regurgitated petri dishes to spatial positions in a grid dispersed across counter space in the laboratory to minimize any potential bias in environmental effects. We determined that successful germination occurred if a radicle emerged from the seed (Barnea et al., Reference Barnea, Yom-Tov and Friedman1991; Bewleyl, Reference Bewleyl1997). Germination trials ran for 30 days, after which we considered all seeds that did not have radicle emergence as not germinated. Time to germination was measured in days, where day 0 was the first day of germination trials. All procedures were approved by El Ministerio del Ambiente y Energía (MINAE, permit # ACCVC-R-INV-025-2011) in Costa Rica and the Institutional Animal Care and Use Committee of the University of Louisiana at Lafayette (permit # 2012-8717-059).
For analysis, we considered all seeds regurgitated together as the same treatment because we could not reliably assign seeds to either species or specific individuals in gut retention trials. We examined the presence–absence of germination using logistic regression, with toucan treatment and seed size included as main and interactive effects. We then examined the time to germination using a cox proportional hazards model, also with toucan treatment and seed size included as main and interactive effects. We considered both significant (P < 0.05) and marginally significant (P < 0.10) effects and present beta-coefficients ±SE as measures of effect size (Nakagawa and Cuthill, Reference Nakagawa and Cuthill2007). For cox proportional hazard models, we present odds ratios (OR) as measures of effect size. Analyses were conducted in R v. 4.1 (R Core Team, 2021).
Results
We collected a total of 151 control seeds and 137 seeds regurgitated from our nine toucans. The size of E. uniflora seeds ranged from 3.7 to 14.3 mm (Figure 2a). Although toucan regurgitation did not influence the proportion of germinated seeds (β ± SE = 2.27 ± 1.92, z = 1.19, P = 0.24), 93.4% of toucan-regurgitated seeds germinated (128 total) compared to 76.8% for 116 control seeds. There was, however, an effect of seed size on the proportion of germinated seeds (β ± SE = 0.72 ± 0.12, z = 5.97, P < 0.001), with an increase in the germination proportion as seed size increased (Figure 2b). The interaction of regurgitation and seed size on the number of seeds germinated was not significant (β ± SE = −0.11 ± 0.22, z = −0.49, P = 0.62).
Whereas toucan regurgitation (OR ± SE = 0.27 ± 0.85, z = −1.52, P = 0.13) and seed size (OR ± SE = 1.07 ± 0.06, z = 1.08, P = 0.28) alone did not affect the speed of germination, the interaction between regurgitation and seed size did affect germination speed, although this was a marginal effect (OR ± SE = 1.16 ± 0.08, z = 1.82, P = 0.07). Mean speed of germination for toucan-regurgitated seeds was 10.7 ± 0.4 days compared to 11.1 ± 0.5 for control seeds (Figure 2c). Mean speed of germination was 12.3 ± 0.7, 11.7 ± 0.7, 10.2 ± 0.5 and 9.8 ± 0.5 days for seeds <10, 10–11, 11–12, and >12 mm in size, respectively (Figure 2d), indicating germination time declined as seed size increased. Overall, germination rates were 3.6× faster when seeds were both large and regurgitated by toucans.
Discussion
Although toucan regurgitation did not increase the prevalence of germination for E. uniflora seeds (prediction 1), the process did enhance germination speed, particularly when seeds were large (prediction 2). The effect of toucan regurgitation on germination speed is not surprising given that seeds ingested by frugivores are more likely to have greater germination rates. For example, a meta-analysis across 351 fruit ingestion experiments shows higher germination rates across 213 tree, shrub, and herbaceous plant species (Traveset and Verdú, Reference Travaset, Verdú, Levey, Silva and Galetti2009). These effects can also be influenced by the sizes of seeds ingested. Larger seeds tend to germinate faster compared to smaller seeds after gut treatment by primates, according to another meta-analysis (Fuzessy et al., Reference Fuzessy, Conelissen, Janson and Silveira2016). Additionally, toucans disperse large seeds of other plants such as palms (e.g., Syagrus romanzoffiana, Vespa et al., Reference Vespa, Zurita, Gatti and Bellocq2018) and nutmeg species (genus Virola, Howe Reference Howe1981, Reference Howe, Fleming and Estrada1993; Holbrook and Loiselle, Reference Holbrook, Loiselle, Dennis, Schupp, Green and Westcott2007; Jones, Reference Jones2017). Further, the functional extinction as the loss of large gape frugivores such as toucans from tropical forests can drive rapid evolutionary reduction in seed sizes (Galetti et al., Reference Galetti, Guevara, Côrtes, Fadini, Von Matter, Leite, Labecca, Ribeiro, Carvalho, Collevatti, Pires, Guimarães, Brancalion, Ribeiro and Jordano2013). Thus, there is precedence for the positive association between toucan regurgitation and seed success for larger seeds of E. uniflora. Although our predictions were only partially supported, our results demonstrate that toucans can provide germination benefits to E. uniflora seeds.
Although our results did not support that toucan regurgitation alone would improve both the number and speed of seeds germinated, and our interaction was marginally significant (i.e. perhaps weaker statistically but not biologically irrelevant; Nakagawa and Cuthill, Reference Nakagawa and Cuthill2007), this could have been due to the initial pulp removal from control seeds in our germination trials. Pulp removal is a key benefit seeds gain from animal gut passage (Levey, Reference Levey1987; Schupp et al. Reference Schupp, Jordano and Gómez2010; Fricke et al., Reference Fricke, Bender, Rehm and Rogers2019); therefore, pulp removal in our study is not truly reflective of natural processes. As a result, germination trials without control seed pulp removal would perhaps be more appropriate in gaging toucan regurgitation benefits (Fricke et al., Reference Fricke, Bender, Rehm and Rogers2019). Beyond pulp removal, there is some chemical scarification that occurs in gut treatment that is also beneficial for seed dispersal quality (Levey, Reference Levey1987; Schupp et al., Reference Schupp, Jordano and Gómez2010). Given that we compared seeds without pulp for both gut-treated and control seeds, we still found a 15% greater germination proportion of seeds with gut treatment, indicating that some scarification likely occurs and provides some benefits to seeds. We would expect that the benefits of toucan regurgitation alone would thus be greater under a depulp-scarification testing scenario (Fricke et al., Reference Fricke, Bender, Rehm and Rogers2019). Finally, gut retention time also plays a large role in germination success (Barnea et al., Reference Barnea, Yom-Tov and Friedman1991; Fukui, Reference Fukui2003; Verdú and Traveset, Reference Verdú and Traveset2004; Traveset and Verdú, Reference Travaset, Verdú, Levey, Silva and Galetti2009; Fricke et al., Reference Fricke, Bender, Rehm and Rogers2019). Although we were unable to disentangle the effect of gut retention time on seed germination rates, future research could investigate the role of retention time on germination of E. uniflora, and the effect of toucan gut retention on other plant species consumed by Ramphastos toucans.
We also documented a positive effect of seed size on the proportion of germinated seeds and speed of germination (prediction 3). Our results are similar to other studies on Eugenia species (Amador and Barbedo, Reference Amador and Barbedo2015). Eugenia uniflora seeds have no dormancy period and are sensitive to desiccation (Stricker and Stiling, Reference Stricker and Stiling2013; Pavithra et al., Reference Pavithra, Swamy, Suresh and Ruchita2020; Pirola et al., Reference Pirola, Wagner, Dotto, Cassol, Possenti and Citadin2021). Typically, seeds sensitive to desiccation, which are more common in tropical systems, are larger, have a greater moisture content, and have faster germination rates (Bazzaz and Pickett, Reference Bazzaz and Pickett1980; Vázquez-Yanez and Orozco-Segovia Reference Vázquez-Yanez and Orozco-Segovia1993; Daws et al., Reference Daws, Garwood and Pritchard2005). Additionally, 86.1% of all our E. uniflora seeds germinated within 30 days, irrespective of toucan treatment. In most tropical forest trees, germination is observed and advantageous promptly after dispersal, likely as a means to minimize predation and to capitalize on ephemeral resources, such as moisture, light, and space (Bazzaz and Pickett, Reference Bazzaz and Pickett1980; Vázquez-Yanez and Orozco-Segovia, Reference Vázquez-Yanez and Orozco-Segovia1993; Daws et al., Reference Daws, Garwood and Pritchard2005).
Toucan dispersal effectiveness is likely highly variable and dependent on the plant species. For example, Guettarda viburnoides seeds have greater germination success when eaten by pulp-feeding jays (Cyanocorax cyanomelas) compared to seed-swallowing toucans (Pteroglossus castanotis) because of the differences in the ways they select and handle fruit (Loayza and Rios, Reference Loayza and Rios2014). Additionally, toucans may also be effective dispersers for E. uniflora within other parts of the seed dispersal framework. For example, Ramphastos toucans may be good dispersers in terms of quantity, as they can remove the most seeds from nutmeg trees (Virola sebifera, Howe, Reference Howe1981) and carry and scatter more seeds further from the crown of the parent tree, increasing their survival (dispersal quality, Howe, Reference Howe, Fleming and Estrada1993). Large vertebrates in Neotropical systems, such as toucans, primates, and ungulates, are capable of dispersing seeds too large for smaller animals to consume and tend to disperse seeds over large distances, which makes them influential in maintaining ecosystem diversity and plant meta-population dynamics (Levey, Reference Levey1987; Vidal et al. Reference Vidal, Pires and Guimarães2013; Andresen et al., Reference Andresen, Arroyo-Rodriguez and Ramos-Robles2018; Fuzessy et al., Reference Fuzessy, Janson and Silviera2018). Large frugivorous birds in particular, can be valuable to dispersal as they will cross open areas (Vidal et al., Reference Vidal, Pires and Guimarães2013) and may readily travel across open areas and diverse landscape mosaics (Graham, Reference Graham2001; Moreira et al., Reference Moreira, Riba-Hernández and Lobo2017). As tropical forests are becoming highly disturbed and fragmented by human activities, the ability of large frugivorous birds such as toucans to continue providing seed dispersal services to degraded landscapes may be vital to the persistence of many tropical plants (Holbrook and Loiselle, Reference Holbrook and Loiselle2009; Moreira et al., Reference Moreira, Riba-Hernández and Lobo2017). This is particularly concerning in the face of large vertebrate disperser population declines, which may lead to the collapse of critical plant–animal mutualistic relationships maintaining the integrity of terrestrial ecosystems (Aslan et al., Reference Aslan, Zavaleta, Tershy and Croll2013; Andresen et al., Reference Andresen, Arroyo-Rodriguez and Ramos-Robles2018). Thus, researching seed dispersal services for a variety of plant species, particularly for degraded landscapes and declining tropical communities is of critical conservation importance (McConkey et al., Reference McConkey, Prasad, Corlett, Campos-Arceiz, Brodie, Rogers and Santamaria2012). This work helps contextualize the role large vertebrates, such as toucans, play in the sustainability and resilience of tropical ecosystems.
Acknowledgements
We thank CATIE and MINEA in Costa Rica for enormous logistical support, permitting aid, and permission to conduct the field study. We also thank J. and L. Howle and the Toucan Rescue Ranch for the use of their facility and logistical support. We also thank K. Gibson and several volunteer technicians who conducted germination trials.
Financial support
LJ was supported by a Louisiana Board of Regents Fellowship, U.S. Fulbright Fellowship to Costa Rica for 2011–2012, and the Biology Department at the University of Louisiana at Lafayette.
Conflict of interest
None.