Frugivory and seed dispersal in a hyperdiverse plant clade and its role as a keystone resource for the Neotropical fauna

Abstract

Background and Aims

Much of our understanding of the ecology and evolution of seed dispersal in the Neotropics is founded on studies involving the animal-dispersed, hyperdiverse plant clade Miconia (Melastomataceae). Nonetheless, no formal attempt has been made to establish its relevance as a model system or indeed provide evidence of the role of frugivores as Miconia seed dispersers.

Methods

We built three Miconia databases (fruit phenology/diaspore traits, fruit–frugivore interactions and effects on seed germination after gut passage) to determine how Miconia fruiting phenology and fruit traits for >350 species interact with and shape patterns of frugivore selection. In addition, we conducted a meta-analysis evaluating the effects of animal gut passage/seed handling on Miconia germination.

Key Results

Miconia produce numerous small berries that enclose numerous tiny seeds within water- and sugar-rich pulps. In addition, coexisting species provide sequential, year long availability of fruits within communities, with many species producing fruits in periods of resource scarcity. From 2396 pairwise interactions, we identified 646 animal frugivore species in five classes, 22 orders and 60 families, including birds, mammals, reptiles, fish and ants that consume Miconia fruits. Endozoochory is the main dispersal mechanism, but gut passage effects on germination were specific to animal clades; birds, monkeys and ants reduced seed germination percentages, while opossums increased it.

Conclusions

The sequential fruiting phenologies and wide taxonomic and functional diversity of animal vectors associated with Miconia fruits underscore the likely keystone role that this plant clade plays in the Neotropics. By producing fruits morphologically and chemically accessible to a variety of animals, Miconia species ensure short- and long-distance seed dispersal and constitute reliable resources that sustain entire frugivore assemblages.

‘… I made systematic observations for two years on the various species of melastomaceous trees and shrubs of the genus Miconia, which are a conspicuous element in the secondary vegetation and whose fruits bulk large in the diet of the smaller frugivorous birds such as manakins and tanagers [...] as well as some that are not normally considered to be fruit-eaters’.

D. W. Snow, 1965

INTRODUCTION

Frugivore-mediated seed dispersal constitutes ancient interactions driving plant evolution and diversification worldwide (Lomáscolo et al., 2010; Eriksson, 2016; Brodie, 2017; Onstein et al., 2017; Valenta and Nevo, 2020). These mutualistic interactions directly affect the dynamics of plant communities and are a vital ecosystem function for biodiversity maintenance (García and Martínez, 2012; Valiente-Banuet et al., 2015; Morán-López et al., 2018). The consequences of seed dispersal range from seedling recruitment to carbon storage–release balance in tropical forests (Bello et al., 2015; Peres et al., 2016; Culot et al., 2017). Moreover, fruit–frugivore interactions have direct applications in ecosystem management, and support more effective restoration practices (Cole et al., 2010; Zahawi et al., 2013; Peters et al., 2016; González-Castro et al., 2018). Frugivory and seed dispersal are recognized as important issues spanning ecology, evolution and conservation (Jordano et al., 2011). Nevertheless, the seminal studies that have shaped the theoretical development of fruit–frugivore interactions date back to five decades ago, and were mostly catalysed by the groundbreaking studies of David W. Snow (e.g. Snow, 1965, 1971).

Snow’s landmark studies paved the way for the novel research field of ecology and evolution of fruit–frugivore interactions. His observations on sympatric Miconia species in the tropical forests of Trinidad showed little overlap in fruiting phenology, leading him to propose that the staggered fruiting pattern could have evolved as a strategy to reduce competition for avian seed dispersers (Snow, 1965). Staggered fruiting not only provides a competitive advantage for plants, but is also beneficial for frugivores, as the year-round fruit availability would allow maintenance of their populations (Snow, 1971). The validity of Snow’s proposition regarding staggered phenologies was subject to later debate by some who disputed his ideas because of failure to detect a non-random uniform fruiting sequence in other systems (e.g. Rathcke and Lacey, 1985; Wheelwright, 1985a; van Schaik et al., 1993). However, his pioneering and influential studies highlighted the coevolutionary processes shaping the traits of both fruiting plants and their seed dispersers.

Snow’s ideas also stimulated the publication of seminal papers (e.g. McKey, 1975; Howe and Estabrook, 1977) which set the theoretical background for exploration of this topic in the next decades (Howe, 1993). McKey (1975) proposed that the evolution of plant strategies related to attraction and reward of seed dispersers would fall along an ecological gradient of seed dispersal specialization, with specialist and generalist species at the opposite ends of an adaptive continuum. Specialized plants such as Virola (Myristicaceae), Casearia (Salicaceae) and mistletoes, which produce large, lipid-rich rewards and one or a few large seeds, and have low fecundity and an extended fruiting season, are expected to rely on a small number of more specialized frugivore species acting as highly effective seed dispersers. At the other end of the continuum, Ficus (Moraceae), Cecropia (Urticaceae) and Miconia (Melastomataceae) species are examples of generalized seed dispersal systems, characterized by the production of fruit which are small in size with water- and sugar-rich fruit pulp enclosing many small seeds. These species produce much larger crops, albeit over a shorter fruiting season, and attract large assemblages of less specialized frugivores, with consequential lower seed dispersal effectiveness. The early postulates of the ‘specialized vs. generalized paradigm’ were targets of much criticism due to conceptual problems and limited capacity to explain the observed patterns (Wheelwright and Orians, 1982; Jordano, 1987; Fleming et al., 1993). However, the ‘paradigm’ was important for recognizing that fruit traits and plant phenology may determine the identity of the interacting partners. Why do some plants have a small set of frugivores dispersing their seeds, whereas fruits of other species are consumed by much broader animal assemblages within the same community? This question would lead to the subsequent quest for keystone mutualists in tropical forests (Howe, 1993).

Terborgh (1986) was the first to suggest that a relatively small number of plant species have a disproportionally large role in the maintenance of frugivore communities in tropical forests. Species providing fruit resources during periods of scarcity, especially at the wet and dry season transitions, may sustain a higher diversity of seed dispersers and structure species-rich communities. Therefore, the notion of keystone plant species is linked to widespread resource use by fauna and their significance for the stability of trophic interactions within communities. Since Terborgh’s proposal, the concept of keystone plant resources (KPRs) has been widely discussed, especially regarding their underlying ecological attributes (Lambert and Marshall, 1991; Peres, 2000; Escribano-Avila et al., 2018). Traits supporting identification of KPRs include (1) copious fruit production to sustain a large assemblage of frugivorous animals; (2) fruiting phenology extended through periods of relative food scarcity; (3) low redundancy with other food sources; and (4) small to medium fruit size enabling consumption by a wide diversity of frugivorous animals. The KPR concept is currently largely accepted, and many taxa have been recognized as keystones based on their widespread consumption by fauna; examples include palms (340 frugivore species associated with 126 palm species; Terborgh, 1986; Zona and Henderson, 1989; Muñoz et al., 2019), Ficus trees (1274 bird and mammal frugivores associated with 260 Ficus species worldwide; Lambert and Marshall, 1991; Shanahan et al., 2001), Parkia pods (Peres, 2000), mistletoes (associated frugivores include 66 families within 12 orders of birds and 30 families within ten orders of mammals worldwide; Watson, 2001), some Myrtaceae species (Staggemeier et al., 2017; Escribano-Avila et al., 2018) and melastome berries (Galetti and Stotz, 1996; Maruyama et al., 2013; Escribano-Avila et al., 2018; Messeder et al., 2020).

Decades before the proposal of the KPR concept, Land (1963) reported his pioneer observations on frugivory in the tropics and recognized the importance of Miconia trinervia fruits as a resource for birds in a Guatemalan rain forest. However, Galetti and Stotz (1996) were the first to suggest a Miconia species as a KPR, mostly based on observations of copious fruit production and widespread consumption by a primate and avian frugivores during periods of fruit scarcity in the Atlantic Forest. Later, Maruyama et al. (2013) also highlighted M. chamissois as a KPR in a savannah–swamp–forest habitat mosaic in Central Brazil, as it produced fruits consistently across years during periods of scarcity when most avian frugivores consumed its fruits. More recently, Escribano-Avila et al. (2018) emphasized the role of Miconia species as KPRs in Neotropical fruit–frugivore networks. Although experimental removal of plant species to test their role as keystones are scarce (e.g. Watson and Herring, 2012), Messeder et al. (2020) recently developed an interaction network-based approach to identify KPRs over a broader ecological scale. By recognizing the most central plant species in 38 Neotropical fruit–frugivore interaction networks and by simulating their removal, they were able to compare distinct plant taxa for the effects of their removal on community stability descriptors derived from networks. They concluded that Miconia species were – by far – the top candidates as KPRs for Neotropical avian frugivores. However, no formal attempt has yet been made to address the importance of these plants as keystone resources for frugivores across animal clades.

Miconia belongs to the Melastomataceae, a pantropical family with >5000 species in nearly 150 genera (www.melastomataceae.net), with the highest diversity in the Neotropics (approx. 3000 species; Renner et al., 2001; Michelangeli et al., 2013). More than a third of its diversity is within the tribe Miconieae, the largest clade with >1900 species (Michelangeli et al., 2004; Goldenberg et al., 2008). Due to paraphyly in many lineages within the Miconieae (Michelangeli et al., 2019), recent taxonomic and nomenclatural changes recognized Miconia as the only genus within the tribe, making this the most diverse flowering plant genus exclusively distributed in the Neotropics (Michelangeli et al., 2019). Miconia is not only a taxonomically hyperdiverse clade, but it is also a functionally diverse group with a variety of life forms, including lianas, herbs, epiphytes, shrubs, treelets and trees. Miconia is also recognized for producing berries with a variety of shapes, sizes and colours (Fig. 1). The geographic distribution of Miconia spans the entire Neotropical realm, with species ranging from microendemics to those with widespread distribution. Miconia species are usually pioneers and can be abundant in several plant physiognomies, and occur from sea level to highlands mostly in the mainland and even on distant islands such as the Galapagos (Michelangeli et al., 2013; Silveira et al., 2013a). Despite the historical importance of Miconia to our current understanding of the ecology of fruit–frugivore mutualisms (Land, 1963; Snow, 1965), no study to date has summarized the major findings on the seed dispersal ecology of these plants.

Fig. 1.

Diversity in shapes, sizes and colours of Miconia fruits. (A) M. capilliflora; (B) M. subciliata; (C) M. blepharodes; (D) M. macrosperma; (E) M. loligomorpha; (F) Miconia sp.; (G) M. calvescens; (H) M. albicans; (I) M. leacordifolia; (J) M. affinis; (K) M. gratissima; (L) M. dolichostachya; (M) M. sellowiana; (N) M. ampla; (O) Miconia sp.; (P) M. candelabriforme; (Q) M. heliotropoides; (R) M. impetiolaris; (S) M. rimalis; (T) M. minutiflora; (U) M. dasytricha; (V) M. oxymeris; (W) M. neourceolata; (X) Miconia sp. (All pictures were kindly provided by Renato Goldenberg.)

In this study, we synthesize the published information on frugivory and seed dispersal of Miconia species in order to produce an integrative review that highlights the role of its fruits as KPRs for Neotropical fauna. Our study had three major goals. First, we aimed at compiling data on fruit traits and fruiting phenology of Miconia species relevant to their keystone role. Second, we integrated the scattered published information on frugivory of Miconia species to quantitatively assess the number of animal taxa consuming their fruits. The effectiveness of animals as seed dispersers is determined by their handling behaviour, ingestion and deposition patterns which directly affect seedling establishment (Schupp et al., 2010). Because the quality of seed dispersal services is intimately associated with the capacity of handled or ingested seeds to germinate (Traveset et al., 2007), our third goal was to quantitatively evaluate the effects of consumption by different animal clades on the germination of Miconia species. Through a meta-analytical approach, we compared the consequences for Miconia seed germination of fruit manipulation by ants and vertebrate gut passage of seeds across a broad range of frugivore groups.

MATERIALS AND METHODS

Phenology and fruit and seed traits

To produce a comprehensive database of dispersal-relevant fruit and seed traits and phenological patterns of Miconia species, we performed a systematic literature review of papers published from 1945 to 2017 in the Web of Science, Scopus and SciELO databases, using the following terms and their combinations in the title, abstract and keywords: ‘frugivory’, ‘fruit chemical content’, ‘fruit dispersal syndromes’, ‘fruit trait’, ‘fruiting phenology’, ‘Neotropics’ and ‘seed dispersal’. In addition, we checked for fruit/seed trait information on the Kew Seed Information Database (http://data.kew.org/sid/), TRY Plant Trait Database (Kattge et al., 2011; https://www.try-db.org/TryWeb/Home.php), FRUBASE data set (Jordano, 1995) and the ATLANTIC-FRUGIVORY data set (Bello et al., 2017) to complement our database.

To be included in our database, studies must have reported at least one of the following traits – life form, duration of fruiting season (months with availability of ripe fruit), ripe fruit colour, fruit size (diameter in centimetres) and weight (fresh fruit mass in milligrams), fruit crop size (average number of fruit estimated for the individual), seed size (largest dimension in centimetres) and weight (in milligrams), number of seeds per fruit and fruit nutritional content (percentage of water, sugar, lipids and proteins on a centesimal composition basis) – for any species currently recognized as Miconia (sensuMichelangeli et al., 2019). This included species previously distributed in 17 genera within the Miconieae (i.e. Anaectocalyx, Calycogonium, Catocoryne, Charianthus, Clidemia, Conostegia, Killipia, Leandra, Maieta, Mecranium, Necramium, Ossaea, Pachyanthus, Pleiochiton, Sagraea, Tetrazygia and Tococa; all species names were checked and updated using the MEL names database at http://www.melastomataceae.net/ and Michelangeli et al., 2019). We also gathered information on IUCN conservation status, if available (the target year for IUCN listing was 2017; http://www.iucnredlist.org).

To envisage fruiting phenology patterns, we constructed phenological line plots for each sampled community where interactions among at least three sympatric fruiting Miconia species and frugivores were recorded. Production of berries is costly, requiring allocation of high amounts of nutrients and water for their development and maintenance (Coombe, 1976; Fenner, 1998; Jordano, 2014). Consequently, water availability is regarded as one of the main factors constraining the number of fruiting species and the length of the fruiting period in a community (van Schaik et al., 1993; Fenner, 1998; Mendonza et al., 2017). For these reasons, the dry season is a period of food scarcity for frugivores (Terborgh, 1986; Peres, 2000; Jordano, 2014), when a lower proportion of fleshy fruit is available to sustain frugivores across different ecosystems (Hilty, 1980; Peres, 2000; Batalha and Martins, 2004; Jordano, 2014; Brito et al., 2017a; Staggemeier et al., 2017; Maruyama et al., 2019). To investigate whether Miconia species have phenological patterns related to seasonality, for each community we also retrieved its geographical location and average monthly historical precipitation (mm) from 1901 to 2018 (data obtained from Harris et al., 2020), which was plotted within the community phenology plots to evaluate the relationship between precipitation and phenology. For each community, we also calculated the seasonality index (SI), as the sum of the absolute deviations of mean monthly rainfall from the overall monthly mean divided by the mean annual rainfall (for detailed information on the calculation, see Walsh and Lawler, 1981). In theory, the SI can vary from 0 (precipitation equally distributed throughout the year) to 1.83 (precipitation restricted to a single month; Walsh and Lawler, 1981). We classified each community as seasonal or non-seasonal, and evaluated the number of fruiting species and their fruiting patterns.

In addition, to evaluate fruiting phenological patterns in Miconia at a large spatial scale, we performed circular analysis with the phenological data using the R package ‘Circular’ (Lund et al., 2017). We then calculated the circular standard deviation, the circular mean (µ) and the length of the mean vector (r) which informs how the data are clustered around the mean (with 0 meaning uniformly distributed and 1 meaning perfectly clustered). We then performed Rayleigh’s test of uniformity (at α = 0.05), which indicates unimodal distribution and significant seasonality in fruiting patterns (Morellato et al., 2010).

Fruit consumption by fauna

To produce a comprehensive database of frugivore interactions with Miconia fruits, we performed a systematic literature review using ‘ecological networks’, ‘fruit diet’, ‘fruit–frugivore’, ‘frugivore community’, ‘frugivory’, ‘Neotropics’, ‘plant seed disperser interaction’ and ‘seed dispersal’ as keywords in electronic databases. We also checked the Interaction Web Database (www.nceas.ucsb.edu/interactionweb/), the Web of Life Ecological Networks Database (http://www.web-of-life.es/) and the ATLANTIC-FRUGIVORY data set (Bello et al., 2017) to complement our database. Finally, we searched the references listed in the studies surveyed to ensure that we did not miss any interactions between frugivores and Miconia fruits.

To be included in our database, studies must have reported at least one event of frugivory of any Miconia species. Our database consists of data generated by different methods to record the consumption of Miconia fruits, including focal plant observations, diet studies reporting seed presence in faecal samples, regurgitation samples and stomach contents, and direct foraging observations. We recorded taxonomy and IUCN status for all animal species interacting with Miconia. For each interaction reported, we recorded geographical co-ordinates and elevation, and classified the ecoregion according to Olson et al. (2001). To estimate the richness of frugivores consuming Miconia fruits, we constructed rarefaction curves for animal species belonging to distinct taxonomic groups using Miconia species as samples in the R package ‘vegan’ (Oksanen et al., 2019). To determine the average number of seed dispersers of Miconia species, we considered only focal plant studies (in our database this method was used only by studies dealing with avian seed dispersers). The same approach was used to determine the average number of ant species attracted to Miconia fruits in studies of fruit removal by ants. We also investigated whether different animal taxonomic groups have fruit colour preferences by partitioning their consumption records according to different fruit colours.

Meta-analysis assessing effects of gut passage on Miconia seed germination

Investigation of how fruit handling (i.e. manipulation by invertebrates and vertebrate gut passage) by a wide range of taxa affects seed germination patterns is especially relevant for comparing the effect of services provided by different groups on the fate of dispersed seeds. We used a meta-analytical approach (Hillebrand and Gurevitch, 2014) to determine the overall magnitude of the effect of animal manipulation/gut passage on the percentage and speed of seed germination. We then partitioned the magnitude of these effects by animal taxonomic group.

To build a germination database, we searched for the following terms and combinations: ‘ant seed germination’, ‘frugivore gut effect’, ‘gut passage effect’, ‘Neotropics’, ‘vertebrate gut passage’, ‘seed fate’ and ‘seed germination’. We also checked the reference list of the studies surveyed to supplement our literature survey. To be included in our database, a study must have met the following criteria: (1) reported data as germination proportion (or percentage) of manipulated/gut-passed and control treatments (manually extracted seeds from fruits); and (2) reported a measure of germination kinetics (speed of or time to germination, which we converted to days whenever needed) for both treatment and control seeds. Whenever the data were unclear, or not clearly reported in the results or supplementary materials, we contacted the corresponding author to ask for permission to use the original data. In some cases, we were not able to determine the exact sample sizes (the number of replicates and seeds set to germinate in each replicate), so we instead used the values given for total seed sample size. When data were available only in figures, we digitized them and extracted the data using ImageJ software (Abramoff et al., 2004).

First, we provided information on the overall percentage of seed germinated in different experimental groups (Samuels and Levey, 2005). We plotted the germination percentage of each plant species reported in each study according to control I (seeds germinating within intact fruits), control II (manually extracted seeds from fruits) and treatment, according to handling by each animal group. For the four vertebrate orders (Aves, Primates, Didelphimorphia and Rodentia) included in our review, we considered gut passage as the treatment, while for ants (Formicidae) we considered seed cleaning (depulping) as the treatment. We then conducted separate meta-analyses for different aspects of germination performance: germination percentage, days to germination of the first seed and mean germination time for total germination (Traveset et al., 2007). As most of the studies did not consider seed germination within fruits (control I) as an additional control treatment, we only included manually extracted seeds from fruits (control II) as the control group (see Samuels and Levey, 2005). We investigated the total effect of animal fruit handling/gut passage on germination performance and moderated further analyses by taxonomic categories (class, order and family) to explore specific germination outcomes.

We used the unweighted log response ratio (Hedges et al., 1999) to summarize gut passage effects of vertebrate and ant manipulation on seed germination of Miconia species. We chose this metric because most data from the literature lacked the information needed to calculate standard deviations of individual effect size estimates. The 95 % confidence intervals (CIs) around the effect size were calculated, and estimates of the effect sizes were considered significant only when the CIs did not overlap zero (Hedges et al., 1999).

The response ratio is the ratio of some measured quantity in experimental vs. control treatments and is commonly used as a measure of experimental effects because it quantifies the proportional change that results from experimental manipulation (Hedges et al., 1999). We calculated the natural log of the response ratio for each effect studied (Hedges et al., 1999) as

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{L}\mathrm{R}=\mathrm{l}\mathrm{n}({\displaystyle \overline{X}}\cdot treatment)-\mathrm{l}\mathrm{n}({\displaystyle \overline{X}}\cdot control).}\end{array}}$$

Effects were reported as the proportional change from control groups (seeds manually extracted from fruits). Negative percentage changes indicate decreases in seed germination percentage and/or germination time of ingested/manipulated seeds compared with control groups, and positive values indicate an increase in the effect measured due to passage through vertebrate gut or ant manipulation (Rosenberg et al., 2000). To estimate the cumulative effect size (E++) for a sample of studies addressing the same effect, effect sizes were combined across studies using an unweighted randomization test (Rosenberg et al., 2000). All analyses were conducted using Open MEE (Wallace et al., 2017).

By using meta-analysis, it is possible to split the variance within groups and to evaluate whether or not categorical groups (in our study taxonomic groups) are homogeneous by using heterogeneity analysis (Q; Gurevitch and Hedges, 1999). We calculated total heterogeneity (QT) and heterogeneity within (QW) and between groups (QB). We used a χ ² distribution to assess whether or not the value of Q was significant. Fail-safe numbers were calculated for each effect tested, indicating how many non-significant, unpublished or missing studies would need to be added to the sample to change the results from significant to non-significant (α = 0.05; Rosenberg et al., 2000). Studies that show large and significant effects are more likely to be published than studies that show weak or no effects (Rosenthal, 1979). As a rule of thumb, results are considered robust when the fail-safe number exceeds 5n + 10, where n is the number of comparisons (Møller and Jennions, 2001). To evaluate publication bias, we used funnel plots as a graphical method, where a symmetrical funnel shape is obtained in the absence of bias when the effect size of each study is plotted against sample size.

RESULTS

Traits of Miconia diaspores

Our trait database included at least one trait datum for 357 species derived from 126 papers (Supplementary data Appendix 1). Data on life form were available for 68.3 % of the species, followed by fruit diameter (60.7 %), ripe fruit colour (51.8 %), fruiting season (46.7 %), seed size (35 %), seeds per fruit (24.3 %) and seed weight (20.7 %). Other traits were less commonly reported (<12 %). The IUCN conservation status was available only for five species: M. robinsoniana and M. quadrangularis (both endangered) and M. amoena, M. paucidens and M. rimalis (least concern).

Miconia species showed a variety of life forms, with shrubs (160 species) dominating, followed by trees (49), shrubs/trees (16; dual life form depending on the habitat), treelets (16), lianas (4) and herbs (4). Most fruits were small berries [diameter 0.57 ± 0.01 cm (mean ± s.e.), range 0.2–1.76 cm] with many tiny seeds (seed size: 0.23 ± 0.02 cm, range 0.002–1.3 cm; seeds/fruit: 106.92 ± 19.01, range 1.44–1005.2; Table 1; Supplementary data Appendix 5, Fig. S1A–C). Water was the predominant component in fruit pulp, with sugars (mainly glucose and fructose) figuring as the most important reward, followed by smaller amounts of lipids and proteins (Table 1). Ripe fruit colour in Miconia varied, but 72.3 % of the species exhibited black and purple berries (Fig. 2A). Other colours, in decreasing order of frequency, were dark blue, blue, white, red, orange, green and yellow (Fig. 2A).

Table 1.

Fruit and seed traits and fleshy pulp nutritional contents of Miconia species.

	Mean	Median	s.e.	Range	n
Morphological
Fruit diameter (cm)	0.57	0.54	0.01	0.2–1.76	217
Fruit weight (mg)	166.15	100	35.36	10–1300	39
Seed size (cm)	0.23	0.1	0.02	0.002–1.3	125
Seed weight (mg)	4.89	0.093	2.37	0.002–150	74
Seeds per fruit	106.92	33.4	19.01	1.44–1005.2	87
Fruit crop size	11 817.09	1358	5099.24	74.3–168 696	41
Fleshy pulp chemical composition (%)
Water	80.93	79.86	2.01	70.05–93.7	12
Lipids	1.43	0.65	0.46	0.04–5.98	14
Protein	2.34	1.55	0.59	0.045–8	14
Total non-structural carbohydrates	13.04	8.74	2.58	0.67–59.22	28
Sugar composition
Fructose (%)	44.25	46	4.43	1–62	12
Glucose (%)	48.54	48	1.97	37–56	11
Sucrose (%)	7.25	3	4.01	1–51	12

Values for chemical composition were obtained in a centesimal composition basis (s.e. = standard error, n = sample size).

Fig. 2.

Percentage of Miconia species with corresponding ripe fruit colour (n = 185 species) (A) and partitioned consumption by different animal groups according to Miconia ripe fruit colours; animal icons corresponds to the animal clades Aves (n = 90 Miconia species), Mammalia (n = 34), Reptilia (n = 3) and Insecta (Formicidae; n = 17) (B).

Fruit production and phenology

Average crop size ranged from 74.3 fruits per plant in M. auricoma to 168 696 fruits per plant in M. rubiginosa. However, 56 % of the 41 species with estimated average crop size produced fruit very copiously, usually exceeding 1000 fruits per plant (Table 1; Supplementary data Appendix 5, Fig. S1D). Data on fruiting phenology were available for 167 Miconia species, and showed a large diversity of fruiting patterns. Most species fruited once a year, 11 species twice a year and M. rufescens three times a year (Supplementary data Appendix 4, Table S1). Average duration of the fruiting period was 6.5 ± 0.27 months, ranging from a single month to year-round fruit production (29 species) (Supplementary data, Appendix 4, Table S1).

At the community level, sympatric Miconia species provided food for frugivores during virtually the whole year across several ecoregions, regardless of the rainfall regime, and exhibited a highly staggered fruiting pattern (Supplementary data Appendix 5, Fig. S2). According to the SI, 54 % of the communities (n = 35) were classified as seasonal (SI ≥0.4) (Supplementary data Appendix 4, Table S2). The seasonal communities had on average 4.5 ± 0.9 fruiting species, with around 50 % of these (2.33 ± 0.28 species) fruiting during or comprising the dry season, and 14 communities had on average 1.5 ± 0.13 species with long fruiting periods (≥7 months). Only three communities had one species each with year-round fruit production (Supplementary data Appendix 4, Table S2). Non-seasonal communities, on the other hand, had on average 6.6 ± 1.1 fruiting species, and all communities, except one, had on average 2.86 ± 0.57 species with long fruiting periods, and around half of them (1.45 ± 0.28 species) with year-round fruit production (Supplementary data Appendix 4, Table S2). Overall, the large-scale pattern of the Miconia fruiting season indicated no phenological peak during the year, with Miconia species fruiting every month in many ecoregions across the Neotropics (Supplementarydata Appendix 5, Fig. S3).

Fruit consumption by fauna

Our literature survey found 182 studies reporting Miconia fruit consumption and handling by animals, including both vertebrates and ants (Supplementary data Appendix 2), that yielded a total of 2396 pairwise interactions. A third of the pairwise interaction records came from community-wide studies, 29 % from species-focused studies and 28 % from dietary studies (other type of studies accounted for <7 % of the recorded interactions).We recorded frugivory events for 243 Miconia species, 78 (32 %) of which were not fully identified to the species level. Fruit–frugivore interactions were found in 47 ecoregions across latitudinal and elevational gradients, with elevation ranging from zero to 4200 m a.s.l., and extending from 19°31′N to 26°15′S (Supplementary data Appendix 5, Fig. S4; Appendix 4, Table S3).

Pairwise interactions were recorded between 243 Miconia species and 646 animal species distributed across five taxonomic classes: Aves, Mammalia, Actinopterygii, Reptilia and Insecta (Fig. 3; Supplementary data Appendix 4, Table S4). The shape and steepness of the rarefaction curves indicated that the diversity of frugivores associated with Miconia fruits was underestimated (Fig. 3). Birds were the most diverse group consuming Miconia fruits, representing 73 % of total animal interactions. Within Aves, frugivore diversity was distributed across ten orders and 37 families, with the Passeriformes standing out as the most diverse order (384 species representing >80 % of the within-Aves diversity) and the Thraupidae as the most diverse family (117 species) (Supplementary data Appendix 4, Table S4). Ants were ranked second for species interacting with Miconia fruits, with 99 species unevenly distributed amongst six subfamilies, with the Myrmicinae representing 75 % of total diversity (Supplementary dataAppendix 4, Table S4). Sixty-one species of mammals consumed Miconia fruits; these were distributed in seven orders and 13 families, including Primates (41.3 %), Chiroptera (17.5 %), Didelphimorphia (17.5 %), Rodentia (14.3 %), Carnivora (4.8 %), Perissodactyla (3.17 %) and Cetartiodactyla (1.6 %) (Supplementary data Appendix 4, Table S4).

Fig. 3.

Species rarefaction curves describing the number of animal species of different clades increasing in function of Miconia species with reports of fruit consumption. Lines represent means and shaded areas the 95 % confidence interval. Blue = Aves; green = Formicidae; pink = Actinopterygii; red = Mammalia; yellow = Reptilia.

We also found reports of 13 fish species in five families consuming Miconia fruits (Supplementary data Appendix 4, Table S4). Reptiles were represented by only five species from the families Iguanidae and Teiidae (Squamata) and Geoemydidae and Testudinidae (Testudines) (Supplementary data Appendix 4, Table S4). IUCN conservation status information was available for nearly 78 % of animal species consuming Miconia fruits, with 36 species (distributed within three classes, 11 orders and 20 families) listed as threatened with extinction to some extent (Supplementary data Appendix 4, Table S5).

According to 21 plant-focal studies assessing frugivory, each Miconia species had an average of 16.5 ± 1.4 avian seed dispersers (n = 37). According to ant fruit removal studies (14 studies), Miconia fruits attracted on average 6 ± 1 ant species (n = 28). Miconia albicans attracted the most diverse assemblage of frugivores, with records of consumption by at least 122 animal species (Fig. 4A). The passerine bird Catharus ustulatus (Turdidae) consumed the highest number of Miconia species (32 species; Fig. 4B). In the case of plant life form, shrubs were associated with the highest number of frugivore species (401 species), followed by trees (328).

Fig. 4.

Top 15 Miconia species consumed by Neotropical frugivores (A) and top 15 animal species consuming Miconia fruits (B).

Birds and mammals consumed Miconia fruits regardless of their colour, except for yellow fruits, which were consumed exclusively by ants (Fig. 2B). Reptiles only consumed purple fruits, although this group accounted for a small proportion of total fruit consumption (Fig. 2B).

Effects of vertebrate gut passage and ant manipulation on seed germination

Our literature survey found 25 published studies addressing the effect of vertebrate gut passage or ant manipulation on seed germination that met our inclusion criteria (Supplementary data Appendix 3). Germination data were reported for 35 Miconia species and generated 207 outcomes: 121 for germination percentage, 78 for germination time and 8 for days until first germination. With the exception of ants, most animal groups seem to substantially increase germination percentage relative to seeds germinating within fruits (control I), with no apparent differences compared with manually extracted seeds (control II; Fig. 5).

Fig. 5.

Percentage of seed germination in Miconia species across different experimental treatments. Control I, seeds within fruits; control II, manually extracted seeds; Aves; Primates; Didelphimorphia; Rodentia; Formicidae. Dots are the independent germination percentage of each plant species reported in each study according to the experimental group. Dashed lines represent means, and whiskers the standard error. n is the number of observations for each experimental group.

Overall, we found a non-significant effect of gut passage on germination percentage (E++ = –0.06, 95 % CI = –0.1435 to 0.0217), days to first seed germination (E++ = 0.10, 95 % CI = –0.2270 to 0.4198) and mean germination time (E++ = 0.05, 95 % CI = –0.0071 to 0.1169; Fig. 6). However, when effects were moderated by taxonomic class (Q_B = 60.63, P ≤ 0.001), gut passage had significant effects on Miconia seed germination.

Fig. 6.

Effect sizes of vertebrate gut passage and fruit handling by ants on germination performance. Overall effects are for germination percentage (circle) and mean germination time (square). Animal groups mediating germination are Formicidae, Rodentia, Didelphimorphia, Primates and Aves (from top to bottom). The cumulative effect size is reported with its 95 % confidence interval. Numbers in parentheses indicate the number of outcomes for each effect. Effects are significant if the confidence intervals do not overlap zero.

Unexpectedly, passage through the gut of bird species significantly reduced germination percentage by 15 % (E++ = –0.15, 95 % CI = –0.2618 to –0.0504) and diaspore handling by ants reduced seed germination by 90 % (E++ = –0.90, 95 % CI = –1.2369 to –0.5669; Fig. 6). Conversely, gut passage in mammals significantly increased germination percentage by 26 % (E++ = 0.26, 95 % CI = 0.1288 to 0.3956; Fig. 6). Birds (E++ = 0.07, 95 % CI = –0.0017 to 0.1542), mammals (E++ = 0.07, 95 % CI = –0.0450 to 0.1917) and ants (E++ = –0.34, 95 % CI = –0.7899 to 0.1084) had no significant effect on germination time. We could not compare the effects between different animal classes on days to first seed germination due to small sample sizes.

Among birds, gut passage in the Passeriformes resulted in a significant reduction in germination percentage by 20 % (E++ = –0.20, 95 % CI = –0.2761 to –0.1279), whereas it was non-significant for the Galliformes (E++ = 1.32, 95 % CI = –1.6024 to 4.2404). To evaluate the gut passage effect of passerine birds on germination percentage, we partitioned the variance among bird families (Supplementary data Appendix 4, Table S5). The overall trend was a reduction in germination percentage. Species within the Turdidae and Thraupidae caused a significant reduction of 27 % (E++ = –0.27, 95 % CI = –0.4209 to –0.1173) and 17 % (E++ = –0.17, 95 % CI = –0.2895 to –0.0494), respectively (Supplementary data Appendix 4, Table S6).

For mammals, we found contrasting within-class effects of gut passage on seed germination. Primates significantly reduced germination percentage by 28 % (E++ = –0.28, 95 % CI = –0.4879 to –0.0799), whereas other non-flying mammals (i.e. Didelphimorphia and Rodentia grouped together) significantly increased it by 45 % (E++ = 0.45, 95 % CI = 0.3406 to 0.5595). Both primates (E++ = 0.19, 95 % CI = –0.0321 to 0.4103) and non-flying mammals (E++ = –0.00, 95 % CI = –0.1659 to 0.1648) had no significant effect on mean germination time. Gut passage in didelphimorphs increased germination percentage by 50 % (E++ = 0.50, 95 % CI = 0.3889 to 0.6102), but the effects were not significant for rodents (E++ = –0.02, 95 % CI = –0.7512 to 0.6939; Fig. 6).

Fail-safe numbers for the effects of vertebrate gut passage and ant fruit manipulation on seed germination percentage and mean germination time were small relative to the number of outcomes included in the meta-analysis, indicating that our results should be interpreted with caution and we should avoid generalizations. The scatter plots of effect size against sample sizes of pooled data showed a classical funnel shape (Supplementary data Appendix 5, Fig. S5), indicating that studies with small sample sizes had a larger dispersion of effect size around the true effect, whereas studies with large sample sizes had an effect size closer to the true value.

DISCUSSION

Our results support the idea of Miconia species as KPRs (Escribano-Avila et al., 2018; Messeder et al., 2020) for most fruit-eating vertebrates and ant lineages. We showed that Miconia frugivory and seed dispersal involve a functionally diverse range of animals, including mostly terrestrial but also aquatic species, varying in morphology, body size, behaviour and habitat use. As we reported, Miconia produces large fruit crops year round, especially at times of resource scarcity, and sustain frugivore assemblages, including threatened species, in both forest and non-forest ecosystems.

Diaspore traits and fruiting phenology in Miconia

Most Miconia fruits are blackish/purplish small-sized berries, enclosing numerous small seeds within water- and sugar-rich fleshy pulp – traits typical of ornithochoric species (Snow, 1981; Wheelwright et al., 1984; Stiles and Rosselli, 1993). However, we showed that a much broader spectrum of vertebrates acts as primary seed dispersers of Miconia. The suite of traits of Miconia fruits may largely explain the widespread consumption by Neotropical fauna. Small fruits and seeds are key traits that drive patterns of interaction in fruit–frugivore networks (Burns, 2013; González-Castro et al., 2015; Sebastián-González et al., 2017). Large fruits containing large seeds constrain interactions with small frugivores, whereas plants producing small-sized fruits and seeds allow consumption by animal species with a much wider range of gape sizes, encompassing frugivores across different sizes and taxonomic groups (Wheelwright, 1985b; Levey, 1987; Jordano, 2014; Fuzessy et al., 2018).

Likewise, the high sugar content of the fruit pulp of Miconia berries also contributes to consumption by fauna (Wheelwright et al., 1984; Jordano, 1995; Baker et al., 1998; Galetti et al., 2011). Although lipid-rich fruits are supposedly more rewarding in energetic terms, their digestion is costly, with longer retention times, because lipids need to be emulsified, hydrolysed, absorbed and then metabolized (Witmer and van Soest, 1998; Levey and Martinez-del-Rio, 2001). In contrast, digestion of soluble monosaccharides (Wheelwright et al., 1984; Baker et al., 1998) is much simpler and entails faster absorption favouring greater net energy gains (Karasov and Diamond, 1988; Worthington, 1989; Witmer and van Soest, 1998; Lepczyk et al., 2000). Therefore, it is reasonable to expect that most frugivores are able to digest fruits, such as Miconia, primarily containing non-structural carbohydrates (Baker et al., 1998; Maruyama et al., 2019).

Fruit production and phenology also help to explain the widespread consumption of Miconia fruits. First, we found that even small shrubs or treelets usually produce copious fruits, with crop sizes often exceeding thousands of fruits per plant (Levey, 1990; Blendinger et al., 2008; Christianini and Oliveira, 2009; Kessler-Rios and Kattan, 2012; Guerra et al., 2017; Santos et al., 2017), but large trees produce up to 160 000 fruits (Christianini and Oliveira, 2009). Second, Miconia species are abundant in plant communities, leading to high fruit density in habitat patches where they occur (Levey, 1990; Blendinger et al., 2008; Christianini and Oliveira, 2009; Maruyama et al., 2013; Guerra et al., 2017). Third, as we have shown, the fruiting phenology of most Miconia species extends on average over 6 months, with year-round fruiting patterns common in species in non-seasonal habitats (Levey, 1988; Poulin et al., 1999) where water is less limited (Morellato et al., 2000).

Our investigation shows that Miconia species commonly produce fruits during the dry season in several seasonal habitats, making them a reliable food source during periods of food scarcity (Rathcke and Lacey, 1985; Terborgh, 1986; Peres, 2000; Jordano, 2014; Mendonza et al., 2017; Maruyama et al., 2019). Year-round fruit availability at the community level entails a higher probability of interaction (Jordano et al., 2003; Olesen et al., 2011), for both resident and migratory species, ensures seed dispersal services and helps to maintain biodiversity (Messeder et al., 2020).

Current evidence also indicates that Miconia species produce fruits consistently between years (Christianini and Oliveira, 2009; Kessler-Rios and Kattan, 2012; Maruyama et al., 2013; Brito et al., 2017a). With the phenological patterns identified in our study, we emphasize the role of Miconia in providing consistent and reliable food resources to frugivores over the years, as well as being one of the plant taxa most likely to sustain many animal species during periods of food scarcity in seasonal ecosystems throughout the Neotropics. The combination of fruit traits and fruiting phenology is similar to those identified for other KPRs such as Ficus and Cecropia (Terborgh, 1986; Lambert and Marshall, 1991; Shanahan et al., 2001), and largely explains the relevance of Miconia species to Neotropical frugivores (Messeder et al., 2020).

Frugivores associated with Miconia fruits

Miconia fruits are consumed by taxonomically and functionally highly diverse vertebrates across all lineages, except amphibians (Table 2; Supplementary data Appendix 4, Table S4). Remarkably, bird species accounted for >70 % of recorded interactions, with >460 species, encompassing a wide range of lineages, body sizes and functional groups. Although our study represents <10 % of Miconia diversity, we have shown that Miconia fruits are indeed a significant food source for 13 % of all Neotropical bird species and for >70 % of the species classified as primarily frugivorous (Kissling et al., 2009, 2012).

Table 2.

A summary of the main seed dispersal agent clades of Miconia species (based on the number of interaction records and species diversity) of each animal class and their preferential fruit traits

Frugivore clade	Species	Interactions	Fruit colour (%)	Crop size	Fruit size	Seeds/fruit	%Water	%Carbohydrates	%Lipids	%Protein	Key references
Aves Passeriformes Thraupidae	117	744	Purple (44 %); black (25.23 %); dark blue (13.88 %); green (7.09 %); blue (6.15 %); orange (1.57 %); red (1.26 %); white (0.78 %)	16 450.4 (26)	0.62 (45)	105.97 (35)	80.06 (8)	16.04 (16)	1.27 (11)	2.88 (11)	Silva et al. (2002); Parrini and Pacheco (2011); Vidal et al. (2014); Palacio et al. (2016)
Aves Passeriformes Tyrannidae	76	235	Purple (50.47 %); green (17.61 %); black (16.66 %); dark blue (12.38 %); orange (1.90 %); red (0.47 %); blue (0.47 %)	24 486.75 (16)	0.64 (28)	36.46 (22)	80.06 (8)	12.90 (15)	1.04 (9)	2.21 (8)	Luck and Daily (2003); Maruyama et al. (2013); Camargo et al. (2015)
Aves Passeriformes Pipridae	26	237	Purple (38.70 %); black (27.95 %); dark blue (43.11 %); blue (3.76 %); orange (3.22 %); green (1.61 %); white (1.61 %)	24 061.86 (18)	0.58 (39)	132.57 (31)	78.64 (6)	15.89 (17)	1.24 (8)	2.66 (8)	Snow (1965); Loiselle and Blake (1990); Gorchov et al. (1995);
Aves Passeriformes Turdidae	26	228	Purple (44.14 %); black (23.40 %); dark blue (17.55 %); blue (6.91 %); green (3.19 %); orange (2.65 %); white (1.59 %); red (0.53 %)	14 789.39 (24)	0.63 (41)	100.15 (28)	80.06 (8)	11.99 (19)	1.23 (10)	3.11 (10)	Blake and Loiselle (1992); Kessler-Rios and Kattan (2012)
Aves Passeriformes Fringilidae	16	61	Purple (38.18 %); black (30.90 %); dark blue (12.72 %); blue (7.27 %); orange (7.27 %); green (3.63 %)	12 825.17 (13)	0.67 (24)	139.23 (13)	80.94 (4)	11.58 (10)	1.37 (6)	3.18 (6)	Silva et al. (2002); Vidal et al. (2014)
Mammalia Primates Callitrichidae	12	31	Purple (54.16 %); black (20.83 %); orange (12.5 %); dark blue (8.33 %); green (4.16 %)	31 921.94 (6)	0.42 (10)	74.45 (14)	77.46 (3)	30.44 (4)	1.63 (4)	2.46 (4)	Dietz et al. (1997); Miranda and Faria (2000); Lapenta et al. (2008)
Mammalia Didelphimorphia Didelphidae	11	54	Black (64.58 %); purple (29.16 %); green (6.25 %)	2239.66 (5)	0.60 (12)	133.32 (11)	77.19 (2)	12.73 (2)	1.32 (2)	2.88 (2)	Lessa and Costa (2010); de Camargo et al. (2011); Lessa et al. (2013)
Mammalia Chiroptera Phyllostomidae	11	18	Purple (36.36 %); black (27.27 %); dark blue (27.27 %); white (9.09 %)	3656.66	0.7 (6)	181 (3)	74.76	12.08	2.56	3.6	Hernández-Conrique et al. (1997); Hernández-Montero et al. (2015); Lima et al. (2016)
Mammalia Perissodactyla Tapiridae	2	21	Dark blue (66.66 %); black (33.34 %)	NA	0.60 (3)	360	NA	NA	NA	NA	Lizcano and Cavelier (2004); Tobler et al. (2010); Barcelos et al. (2013)
Mammalia Primates Cebidae	4	15	Purple (54.54 %); black (18.18 %); orange (18.18 %); blue (9.09 %)	12 135.62 (6)	0.47 (8)	84.2 (6)	78	15.52 (3)	2.60 (2)	4.48 (2)	Wehncke and Dalling (2005); Canale et al. (2016)
Insecta Hymenoptera Formicidae Myrmicinae	74	126	Purple (52.41 %); green (23.38 %); dark blue (12.09 %); black (9.67 %); yellow (2.61 %); orange (0.8 %)	25 790.45 (8)	0.58 (14)	38.59 (15)	76.27 (4)	14.16 (8)	0.99 (5)	2.47 (5)	Kaspari (1993); Leal and Oliveira (1998); Christianini et al. (2012)
Insecta Hymenoptera Formicidae Formicinae	10	21	Green (38.09 %); purple (33.33 %); black (28.57 %)	2609.5 (2)	0.65 (5)	33.56 (6)	79.63	25.24 (2)	1.14 (2)	2.83 (2)	Leite et al. (2013); Lima et al. (2013)
Insecta Hymenoptera Formicidae Ponerinae	7	10	Purple (66.66 %); green (33.34 %)	85 895.5 (2)	0.61 (4)	44.53 (3)	74.84 (2)	22.95 (3)	0.79 (3)	2.43 (3)	Christianini et al. (2012); Santana et al. (2013)
Actinopterygii Characiformes Serrasalmidae	6	6	NA	NA	NA	NA	NA	NA	NA	NA	Hawes and Peres (2014a)
Actinopterygii Characiformes Bryconidae	3	3	NA	NA	NA	NA	NA	NA	NA	NA	Correa et al. (2016)
Reptilia Testudines Geoemydidae	2	2	Purple (100 %)	NA	0.55	40	NA	7.65	NA	NA	Moll and Jansen (1995)

Morphological and chemical fruit traits are reported by the mean value and sample size in parentheses.

NA = information not available.

Frugivory has emerged independently several times in the evolutionary history of birds (Fleming and Kress, 2011), with the Passeriformes standing out as the most species-rich order with the highest proportion of frugivorous species (Kissling et al., 2009). Accordingly, we have demonstrated the relevance of Miconia fruits to small-sized passerines, especially tanagers (Thraupidae), flycatchers (Tyrannidae), thrushes (Turdidae) and manakins (Pipridae; Table 2; Supplementary data Appendix 4, Table S4). Thus, coevolution with highly frugivorous passerine bird clades may be a major force driving changes in Miconia fruit traits towards smaller size and purplish pulp rich in water and sugar (Table 2). However, studies on coevolution between fruiting plants and frugivores remain scarce and inconclusive, and it is a topic worthy of future investigation (Valenta and Nevo, 2020).

Although mostly overlooked, the reports gathered here also emphasize the importance of Miconia fruits for a functionally diverse group of mammals, ranging from large terrestrial to small volant and non-volant species. Frugivory is well distributed along mammalian lineages, having evolved independently numerous times (Fleming and Kress, 2011). The major families of Neotropical frugivorous mammals are the Atelidae, Cebidae, Pitheciidae and Aotidae (Primates), Phyllostomidae (Chiroptera), and Echimyidae and Dasyproctidae (Rodentia; Fleming and Kress, 2011). However, we document a much broader diversity of mammals relying on Miconia fruits as a food resource, including some threatened species (Table 2; Supplementary data Appendix 4, Tables S4 and S5).

Primates are by far the most frugivorous clade of mammals, and accordingly are the most diverse group of mammals consuming Miconia. Mutualistic interactions with fruits have played an important role in promoting primate diversification, especially in the Neotropics (Gómez and Verdú, 2012), where most species depend greatly on fruit as part of their overall diet (Hawes and Peres, 2014b). These large-bodied vertebrates are considered generalist frugivores, due to their capacity for consuming a wide variety of fruit, ranging from very small to very large seeded (Fuzessy et al., 2018). We emphasize the importance of the small-seeded Miconia fruits as food resources for many primate species, which rely heavily on fruits of several different Miconia species, especially during periods of resource scarcity (Pavelka and Knopff, 2004; Lapenta et al., 2008; Canale et al., 2016).

Although less common, Miconia fruit consumption by fishes and reptiles is underestimated. We found at least 13 species of fish that consume Miconia fruits (Kubitzki and Ziburski, 1994; Hawes and Peres, 2014a; Correa et al., 2016; Weiss et al., 2016) in the Amazon forest (Supplementary data 4, Table S4) where, once a year, seasonal flooding allows aquatic fauna to access forest interiors and consume diaspores that fall into the water (Kubitzki and Ziburski, 1994). Miconia fruits are also consumed by reptiles, including tortoises (Moll and Jansen, 1995; Guzmán and Stevenson, 2008) and lizards (Swanson, 1950; Magnusson and Sanaiotti, 1987; Guerra et al., 2018). Although documented infrequently, consumption of Miconia fruits by fishes and reptiles is not necessarily rare, simply because diet studies sampling stomach contents rarely provide precise taxonomic identification of plant items.

Despite the scarcity of studies (17 studies) providing data on interactions with Miconia (19 species), ants were ranked as the second most diverse group of consumers of the genus (Supplementary data Appendix 4, Table S4). In fact, the shape of the species accumulation curve indicates that the diversity of ants interacting with Miconia fruits could match that of birds. Miconia species are primarily vertebrate dispersed and lack specialized structures, such as elaiosomes, to mediate ant interactions. Ants are, therefore, important secondary seed dispersers that use both fruit pulp and seeds as food resources (Kaspari, 1993; Christianini et al., 2007; Christianini and Oliveira, 2009; Lima et al., 2013; Guerra et al., 2018). Miconia fruits are consumed by fungi-growing ants (Leal and Oliveira, 1998; Dalling and Wirth, 1998), opportunistic exploiters of fruit pulp (Christianini et al., 2007; Christianini and Oliveira, 2009; Lima et al., 2013) and granivores (Kaspari, 1993; Levey and Byrne, 1993; Guerra et al., 2018). Therefore, our results highlight the overlooked relevance of Miconia fruits and seeds as food resources for Neotropical ants.

Effects of gut passage on germination of Miconia seeds

Endozoochory is the main mechanism of Miconia seed dispersal. Germination of Miconia seeds is light dependent and rarely occurs within fleshy pulp (Silveira et al., 2013a; Santos et al., 2017). Miconia fruits have germination inhibitors; therefore, release of seeds from fleshy pulp is a key service provided by dispersers (Silveira et al., 2013a, b; Ribeiro et al., 2016). Although our results showed that some animal groups caused an overall decrease in seed germination when compared with hand extraction, fruit consumption and seed handling allow a much higher likelihood of germination than situations when seed cleaning does not take place.

Our meta-analysis highlighted that vertebrates and ants may affect Miconia seed germination, but also revealed that the outcomes of gut passage are clade specific. Fruit-eating birds are recognized to enhance seed germination, as passage through their guts generally increases germination by up to 40 % and reduces germination time (Traveset et al., 2001; Traveset and Verdú, 2002; Lehouck et al., 2011). Contrary to this expectation, we found that birds negatively affected seed germination. Passage through the avian gut may reduce resistance to breakage (Traveset et al., 2008) in some Miconia species (Ribeiro et al., 2016). The reduced germination of gut-passed seeds may be related to their small size and thinner coats, which could be damaged during passage through muscular gizzards, or because some species produce physiologically dormant seeds (Silveira et al., 2012a).

Seed germination outcomes were far from general, showing very specific results strongly dependent on frugivore identity. Studies with specialized frugivores have shown positive or neutral influences on Miconia seed germination (Ellison et al., 1993; Gomes et al., 2008; Acosta-Rojas et al., 2012; Silveira et al., 2012b). However, most of the studies in our database evaluated omnivorous birds that included both fruit and insects in their diets (Alves et al., 2008; Gomes et al., 2008; Silveira et al., 2012b). Generalist birds often possess more muscular gizzards, leading to longer retention times and mechanical and chemical abrasion of seeds, consequently decreasing the likelihood of germination (Traveset et al., 2008; Jordano, 2014). However, the frequency of Miconia seeds in faecal samples of some passerines may reach >90 % (Stiles and Rosselli, 1993), whereas the average number of seeds in bird droppings can reach hundreds in some cases (Loiselle and Blake, 1999). Therefore, even with <50 % of seeds germinating after gut passage, birds could still effectively disperse many viable seeds, successfully contributing to Miconia seedling establishment (Ellison et al., 1993; Krijger et al., 1997; Blendinger et al., 2011).

Our meta-analysis showed that diaspore handling by ants considerably reduced germination of Miconia seeds. However, it is important to acknowledge that seeds handled by ants are those that have fallen under mother plants, after being mashed or defecated by birds (Kaspari, 1993; Christianini and Oliveira, 2009; Lima et al., 2013; Guerra et al., 2018). Therefore, most seeds collected and transported by ants have already suffered – to some degree – from pre-dispersal predation and mandibulation, and/or ingestion by primary dispersers. Despite their apparent negative effects on seed germination, several ant species present behaviours that allow their recognition as legitimate dispersers of Miconia seeds, including depulping and transporting intact seeds to sites favourable to establishment (Dalling and Wirth, 1998; Christianini and Oliveira, 2009; Lima et al., 2013). Some ants seem to be attracted mostly to fruit pulp rewards, and transport seeds unintentionally while carrying fruit pulp (Christianini and Oliveira, 2009; Lima et al., 2013; Guerra et al., 2018). Although ants may harm Miconia seeds during handling and reduce their germination potential, their contribution to short-distance seed dispersal must be acknowledged (Christianini and Oliveira, 2009; Lima et al., 2013; Guerra et al., 2018).

Previous meta-analytical studies investigating the gut passage effect on seed germination across a wide variety of animals and plant taxa (Traveset, 1998; Traveset and Verdú, 2002; Traveset et al., 2008; Torres et al., 2020) showed that non-flying small mammals caused the lowest positive effect on seeds (approx. 5 % increase in germination percentage). However, there is emerging evidence that rodents and opossums can play important roles as seed dispersers, with seeds showing positive germination responses after fruit ingestion, especially in small-seeded species (Grenha et al., 2010; Lessa et al., 2013, 2019; Lessa and Geise, 2014; Sahley et al., 2015; Genrich et al., 2017). Our results support the view that small mammals, especially opossums, are important dispersers of Miconia seeds, considerably increasing germination percentages.

The germination percentage of seeds of Miconia fruit consumed by primates decreased by 28 % when compared with that of hand-extracted seeds. Despite this reduction, germination of primate gut-passed seeds was greater than that of seeds enclosed in fruits, suggesting that seed depulping is an important service also provided by primates. Although primates seem to have a restricted direct role in Miconia seed dispersal, they may complement the mostly short-distance services provided by birds and ants (Christianini and Oliveira, 2009; Guerra et al., 2018). The longer transit times and wider home ranges of primarily frugivorous primates could lead to long-distance dispersal of some Miconia seeds (Fuzessy et al., 2017). Although the results of our meta-analysis were highly informative, they must be interpreted with caution. Future experimental germination studies with Miconia species should investigate gut passage effects across a broader range of frugivores.

CONCLUSIONS AND PERSPECTIVES FOR FUTURE STUDIES

Although studies of Miconia species and their dispersers catalysed the early debate on coevolution between frugivores and their plants (Snow, 1971; Stiles and Rosselli, 1993), the role of fruit–frugivore interactions as drivers of diversification and geographic range in this hyperdiverse genus stands as a major knowledge gap (Reginato et al., 2020). Indeed, the evolution and diversification of Miconia fruits remain an enigma, as fleshy fruits have not been found to be a key innovation within the Melastomataceae (Reginato et al., 2020). Therefore, several aspects of the evolutionary ecology of Miconia fruits remain unexplored. Data gathered in our integrative review strongly support previous studies (Escribano-Avila et al., 2018; Messeder et al., 2020) emphasizing the relevance of Miconia species as keystone resources for other animals in addition to birds. By producing large crops of small fruits year round, especially at times of resource scarcity, Miconia species are able to entirely sustain a wide diversity of frugivore assemblages.

Current evidence indicates that selection by frugivores may drive evolution of fruit and seed traits, including size, colour and nutrient content (Janson, 1983; Jordano, 1995; Galetti et al., 2013; Brodie, 2017; Nevo et al., 2018). Nevertheless, studies addressing phenotypic selection of Miconia fruits and seeds by dispersers remain scarce (Camargo et al., 2015), with available information indicating that avian frugivores may select for individual plants with larger crop sizes (Blendinger et al., 2008; Christianini and Oliveira, 2009; Guerra et al., 2017). We argue that the diverse suite of traits exhibited by Miconia fruits and their exposure to different selective pressures exerted by different frugivore functional groups can be a suitable model system for testing coevolutionary hypotheses (Table 2). Although consumption of Miconia fruits by frugivores seems diffuse with no clearly apparent coevolutionary patterns at this time (Table 2), future studies with proper experimental designs may be valuable. For instance, studies combining observational, experimental and phylogenetic comparative approaches should investigate the role of frugivore clades in shaping the evolution and diversification of Miconia species. Can species dispersed by different frugivore clades evolve specific traits? We hypothesize that, given the germination outcomes specific to animal clades, fruit and seed traits could be positively shaped according to the disperser functional group.

While many frugivory and seed dispersal studies have highlighted the importance of Miconia species for sustaining and structuring frugivore assemblages, their broader role at higher organizational levels (e.g. community, ecosystem and landscape) remains overlooked. Miconia species can be highly diverse and abundant across many communities throughout the Neotropics; thus, their contribution to the structure and function of forest and non-forest ecosystems probably extends far beyond the maintenance of fruit–frugivore interactions. Of note, Miconia flowers are an important resource for pollen-collecting bees, wasps, beetles, flies, moths and hummingbirds (Kriebel and Zumbado, 2014; Brito et al., 2016, 2017b). Moreover, Miconia species are known as hosts of many specialized herbivores, including leaf-chewing lepidopteran larvae (Badenes-Pérez et al., 2010; Scherrer et al., 2010), twig girdler beetles (Paulino-Neto et al., 2005; Paro et al., 2014), seed-feeding carabid beetles (Paarmann et al., 2002), sap-sucking treehoppers (Lopes, 1996; Chacón-Madrigal et al., 2012; Swing, 2012; Alfaro-Alpízar et al., 2020), fruit-galling insects (Centrella and Shaw, 2010, and leaf-galling nematodes (Santos et al., 2012; Viana et al., 2013), among others. Therefore, we highlight the need for further studies addressing their importance for primary consumers other than frugivores, as well as their role in maintaining biodiversity and ecosystem functioning, while facilitating habitat connectivity and providing ecosystem services across Neotropical ecoregions.

Finally, the recognition of Miconia as a keystone plant taxon for Neotropical frugivores is the first step towards applying this knowledge to better conservation and management practices (Peters et al., 2016). Birds are the most common frugivores attending Miconia plants, and also the main group delivering diaspores into sites targeted for restoration (Wunderle, 1997; Cole et al., 2010; Zahawi et al., 2013, Bechara et al., 2016; González-Castro et al., 2018). As a result, planting of Miconia species could intensify the attraction of avian frugivorous species to areas targeted for restoration and increase both the amount and diversity of seed rain. Since we have shown their broad use and dispersal by fauna, we hypothesize that the return of frugivore assemblages and their ecosystem functions can be facilitated by Miconia species. Furthermore, Miconia species are recognized as pioneers present in early successional stages, forest borders and gaps (Levey, 1990; Ellison et al., 1993; Silveira et al., 2013a). Therefore, we suggest that prioritizing the planting of Miconia species selected from regional pools could maximize natural regeneration in increasingly fragmented landscapes.

SUPPLEMENTARY DATA

Supplementary data are available online at www.aob.oxfordjournals.org and consist of the following. Appendix 1: Miconia fruit phenology and trait database reference papers. Appendix 2: Miconia–frugivore interaction database reference papers. Appendix 3: Miconia gut passage seed germination database reference papers. Appendix 4: Table S1: list of Miconia species and their respective fruiting seasons; Table S2: seasonality index, precipitation regime and seasonal/non-seasonal simple classification for each community sampled with at least three sympatric Miconia species with phenological data; Table S3: Neotropical ecoregions encompassing fruit–frugivore interactions with Miconia; Table S4: diversity of frugivores consuming Miconia; Table S5: list of species classified as some level of degree of threat according to the IUCN; Table S6: gut passage effect size on seed germination percentage according to bird families. Appendix 5: Figure S1: data distribution of dispersal-related traits in species of Miconia; Figure S2: fruiting phenology line plots and precipitation data for sampled communities; Figure S3: rose diagram showing the number of fruiting Miconia across the Neotropics; Figure S4: geographical map of the Neotropical realm and its terrestrial ecoregions showing each sampled community; Figure S5: scatter plot of effect size against sample size.

FUNDING

J.V.S.M. was supported with a master student scholarship and T.J.G. with a post-doctoral researcher scholarship funded by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). J.V.S.M. acknowledges his current PhD scholarship funded by the Fulbright Commission and CAPES. F.A.O.S. and T.G.C. acknowledge the financial support from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG).