Biotic and abiotic factors in promoting the starting point of hybridization in the Neotropical flora: implications for conservation in a changing world

Abstract

Coevolution between floral traits and specific pollination behaviour is a significant evolutionary force in angiosperm diversification. However, hybridization is also reported to occur between plants with specialist pollination syndromes. Understanding the role of pollinators in plant diversification is crucial, especially in megadiverse regions, such as the Neotropics. In this review, we examine plant hybridization studies in the Neotropics with the aim of providing a perspective on biotic and abiotic factors starting hybrid zone formation. The Pleistocene was the most widely cited time for the occurrence of hybridization facilitated by geographical range shifts, but time-calibrated analyses are needed to recover a more realistic scenario. Our synthesis of hybridization and pollination systems suggested that specialist and generalist pollinators were associated with the starting point of hybridization in the Neotropical flora. Bees and hummingbirds are most likely to be the primary vectors of interspecific gene flow, and even sporadic visits by bees or other generalist pollinators could allow the formation of a new hybrid zone. We highlight that seed and pollen dispersal vectors should be included in an integrative discussion on hybridization in the Neotropical flora. We also provide a preliminary map of hybrid zones in the Neotropics, including Brazilian vegetation cover and losses in the last 30 years, with the aim of encouraging research into human-driven anthropogenic changes and formation and/or shift of hybrid zones through time.

INTRODUCTION

The dual role of hybridization can be easily understood when plants are considered, as it may drive diversification or promote genetic homogenization. Also, hybridization and diversification can be driven by other factors and, alternatively, there may be no relationship between the two (Mitchell & Whitney, 2021). The relationship between hybridization and increased speciation rate has a limited positive correlation. However, the causation directionality is challenging to establish. There is empirical evidence that hybridization may enable rapid diversification (Rieseberg, 2006; Mallet, 2007; Soltis & Soltis, 2009; Arnold, Ballerini & Brothers, 2012; Abbott et al., 2013; Mitchell et al., 2019). On the other hand, in some cases, it can also erode divergence by mixing adaptive alleles, promoting homogenization and resulting in the extinction of species (Rhymer & Simberloff, 1996; Wolf, Takebayashi & Rieseberg, 2001; Buggs & Pannell, 2006; Todesco et al., 2016; Campbell, Blanchette & Small, 2019). The ecological or evolutionary factors leading to one outcome or the other are unknown.

The occurrence and after-effects of hybridization in Neotropical plant species were recently reviewed by Schley, Twyford & Pennington (2022). Their results reinforce hybridization as an important evolutionary force in the region. However, despite the different outcomes of species hybridization in the Neotropics, little is known about biotic and abiotic characteristics and the occurrence and evolutionary role of hybridization. According to Schley et al. (2022), a positive effect of hybridization, increasing genetic diversity and promoting adaptation is more frequent in species-rich montane radiations. In addition, there is an assumption that plant migration is often associated with climate change that could facilitate secondary contact hybridization. The effects of Pleistocene climatic changes are correlated with rapid species diversification and the positive impact of hybridization in the Neotropics (Schmidt-Lebuhn, Kessler & Kumar, 2006; Dušková et al., 2017; Vargas, Ortiz & Simpson, 2017; Morales-Briones, Liston & Tank, 2018; Nevado et al., 2018; Flantua et al., 2019; Pouchon et al., 2021). Other factors, such as pollinator behaviour, could also affect the occurrence and outcomes of hybridization. However, homoploid hybridization is difficult to detect as it can be confused with other populational and stochastic phenomena such as incomplete lineage sorting (Chase, Paun & Fay, 2010), and estimates of hybridization may be lower than the actual values. Gene flow could also mask longer divergence times and inflate population size estimates (Leaché et al., 2014). Furthermore, when hybridization is a consequence of rapid diversification, it is difficult to establish the true lineages, leading to an erroneous estimate of the diversification rate (Mitchell & Whitney, 2021).

In part, floral diversification in angiosperms has been attributed to selection pressure exerted by animal pollinators (Stebbins, 1970; Pellmyr, 1986), promoting speciation. However, for hybridization to occur, some biotic or abiotic mechanisms must promote pollen exchange among species. Pollen grains of c. 87.5% of angiosperms are transported by various animal pollinators (Ollerton, Winfree & Tarrant, 2011). A wide range of studies of plants and their floral visitors indicate the prevalence of generalist relationships (Waser & Ollerton, 2006; Gomez, 2002), even though those visitors may not effectively pollinate (King, Ballantyne & Willmer, 2013). Some authors concluded that pollination mediated by animals is too non-specific to contribute to reproductive isolation (Waser et al., 1996; Waser, 2001; Fenster et al., 2004; Ollerton et al., 2007). On the other hand, there are highly specialized pollination systems such as that of Ficus L. (figs) and agaonid wasp pollinators (Souto-Vilarós et al., 2018; Wang et al., 2021). In between these extremes, there are species pollinated by a group of animals, such as large bees (André et al., 2022), bats (Arakaki et al., 2021) or butterflies (Leal et al., 2020). Although a relationship between specialist or generalist pollination systems and the presence of hybridization has not been established, a pollinator shift between hybrids and progenitors was reported in most of the studies reviewed by Rezende et al. (2020). The most frequent floral traits affected by hybridization were floral morphology, colour and phenology. It is expected that interspecific pollen transfer will reduce fitness, and angiosperms have evolved a wide range of strategies to reduce the impact (Moreira-Hernández & Muchhala, 2019). However, there is a lack of isolating barriers between closely related species, and little is known about the role of pollinators in preventing or promoting gene flow. Finally, according to our knowledge, whether different floral syndromes or pollinators are related to a higher rate of hybridization in the Neotropics has still not been addressed.

In this sense, to better understand hybridization and its consequences in the Neotropics, it is crucial to investigate the biotic and abiotic factors leading to hybrid formation. The central questions of this systematic review are as follows. (1) Can hybridization events be associated with certain geological times and climate events in the history of the Neotropics? (2) What is the role of generalist and specialist pollinators in the evolution of the Neotropical plants, and which functional groups of pollinators are more frequently cited as being involved in hybrid formation? (3) What is the location of hybrid zones for plants in the Neotropical area? (4) What are the influences of Anthropocene climate change and environmental degradation on the occurrence and consequences of hybridization? We conducted a systematic search in the literature databases to address these questions. First, we consulted the systematic review articles in this subject area to help construct a script with the keywords and determine the inclusion/exclusion criteria. Then we used the script to search the literature in the databases. We discuss the knowledge synthesis in this area for the Neotropics and provide some perspective for future studies.

SEARCHING FOR HYBRIDIZATION STUDIES IN THE NEOTROPICAL FLORA

To gain insights into the process driving hybridization in the Neotropical flora, we reviewed articles focusing on this subject. First, we retrieved the articles from the recent review by Schley et al. (2022) from the Google Scholar database. Then, we used the exact keywords they used to search for studies from July 2021 to March 2022 (scholar.google.com, accessed March 2022). Additionally, we searched articles on the Web of Science (Institute of Scientific Information, Thomson Scientific - https://apps.webofknowledge.com/ accessed March 2022) using the following key phrase ‘ts=(hybridization OR hybrid OR “hybrid zone” OR reticulation OR introgression OR “phylogenetic discordance” OR “gene flow”) AND ts=(Neotropic* OR Brazil OR Belize OR “Costa Rica” OR “El Salvador” OR Guatemala OR Honduras OR Mexico OR Nicaragua OR Panama OR Bolivia OR Colombia OR Ecuador OR “French Guiana” OR Guyana OR Peru OR Suriname OR Venezuela OR Amazonia OR Llanos OR Páramo OR Cerrado OR Caatinga OR Pampa OR Pampas OR Chaco OR Andes OR Andean OR “Atlantic Forest”) AND ts=(Plant) NOT ts=(breeding OR crop OR mammalian OR fish OR reptiles OR Mammalia OR frog OR mtDNA)’.

In total, we reviewed 90 empirical studies that were published between 1995 and 2021 (Supporting Information 1). For each paper, we recorded the following information: (1) sample taxa; (2) study type (cytogenetics, population genetics, phylogeography, phylogenetics, reproductive biology and morphology); (3) molecular markers; (4) main biome/ecoregion; (5) hybrid zone or sympatric sites; (6) reproductive aspects of hybridizing species (pollination syndrome, pollinators and pollinator vectors); (7) edaphoclimatic and ecological aspects of species distribution and (8) age estimation of hybridization. Hybridization was most addressed in studies using population genetics and phylogenetics, encompassing 52.22 and 25.55% of studies, respectively (Supporting Information 2). These two areas can reveal recent and ancient hybridization because they are primarily based on molecular markers, which allow for testing for ancestral introgression and the recent occurrence of hybridization (see Schley et al., 2022). Most studies focused on herbs (42.22%) and trees (30.0%) (Supporting Information 3). Overall, 33 families, corresponding 10.3% of flowering plant families recognized in the Neotropics (Neotropikey. Interactive key and data resources for Latin American plants) and 64 genera were studied (Supporting Information 1). Bromeliaceae and Orchidaceae were the most studied families, accounting for 14.3 and 10% of hybridization studies, respectively (Supporting Information 4). Fabaceae, Fagaceae and Asteraceae accounted for 7.7% each, and Cactaceae and Solanaceae for 4.4% each. Other families were less represented in these studies (Supporting Information 4). The most studied families are among the largest flowering plant families. This bias could be explained by the fact that hybridization could be more common in some families than others, as reported by Whitney et al. (2010). Nonetheless, this study did not evaluate the Neotropical plants, and we noticed other species-rich families, such as Lauraceae, that were not considered in any hybridization studies we reviewed. Melastomataceae and Myrtaceae, also species-rich families with many taxa distributed in the Neotropics, were less studied considering hybridization. Thus, more studies of hybridization in the Neotropics, including these and other families, will shed light on this pattern in the Neotropical flora.

In the following sections, we discuss abiotic and biotic factors on the origin of hybridization. For this, we selected key studies to analyse the time of plant hybridization in the Neotropics and the plant-pollinator role in hybridization (Supporting Information 1). To unveil the process leading to hybridization and map the putative hybrid zones in the Neotropical region, we compiled the geographical coordinates of sites cited as hybrid zone/contact zone/sympatric populations (Supporting Information 1). Some papers make the georeference explicit in the text or table whereas others mention the specific location. For this last case, we search for coordinates on the Google Earth platform. We used the software QGis to plot the points under the Neotropical regionalization map (Morrone et al., 2022).

ABIOTIC FACTORS IN HYBRIDIZATION IN THE NEOTROPICS

Geological and climatic fluctuation drove the diversification of the Neotropical flora, and hybrid zone formation was widely observed during climatic oscillation in the Quaternary period (Hewitt, 2011). We searched for the chronology of hybrid events, and when this information was not present, we considered the hybridizing species divergence to attempt to reveal the abiotic forces behind hybridization. Few studies reported the times of hybridization events and/or lineage divergence (c. 14%; Fig. 1). Six studies used approaches which allowed them to detect shallow/recent hybridization, and seven carried out methods to detect deep/ancient hybridization (see fig. 1 of Schley et al., 2022, for details concerning study types). Most of them showed that hybridization events would have occurred in the Pleistocene, and some would have occurred in the Miocene.

Three of four studies that dated the time of hybridization events placed them in the Pleistocene (Eaton et al., 2015; Baena-Díaz, Ramírez-Barahona & Ornelas, 2018; Turchetto, Schnitzler & Freitas, 2019a; Möbus et al., 2021). One that was based on the time of splitting of a pair of hybridizing species indicated that hybridization had probably occurred since species divergence in the Pliocene (Bedoya, Leaché & Olmstead, 2021). Through a systematic review of phylogeographic studies in South America, Turchetto-Zolet et al. (2013) observed that most plant lineages split during the Pleistocene period, suggesting the role of climatic cycles in lineage divergence within species and probably also in hybridization between pairs of species brought in contact due to their change in distribution range driven by climate fluctuations.

Turchetto et al. (2019b), using a population sampling approach and nuclear simple sequence repeat (SSRs) markers, found that, despite hybridization, the species boundary between two annual herbs from the Pampa grasslands in southern Brazil is maintained. The species [Petunia axillaris (Lam.) Britton, Sterns & Poggenb. and P. exserta Stehmann (Solanaceae)], have a suite of traits related to the main pollination systems (nocturnal hawkmoths and hummingbirds, respectively; Lorenz-Lemke et al., 2006; Segatto et al., 2014). The study pointed out that hybridization occurred c. 27 kya and resulted in a new morphologically variable genetic lineage (Teixeira et al., 2020). Turchetto et al. (2019a), also studying the same system, investigated the current pollen flow between these two species through paternity analyses and found a percentage of hybrids in the progeny of both species, suggesting that hybridization could occur each generation. Hybridization probably resulted from secondary contact (Caballero-Villalobos et al., 2021) and would be related to changes in the geographical distribution of P. axillaris during the Pleistocene. The site of sympatry between P. exserta and P. axillaris is at the edge of the distribution of the widespread P. axillaris (Turchetto et al., 2014); this was affected by past climatic changes as shown by dated demographic changes using molecular data (Turchetto et al., 2014) and projections of past habitat suitability (Giudicelli et al., 2019a). These population expansions and retractions probably affected the divergence and secondary contact within the evolutionary lineage of P. axillaris (Giudicelli et al., 2019b) and the contact with P. exserta. The migration events were reported as one primary force linked to the evolution of Petunia Juss., an endemic genus of Solanaceae that originated and diversified in open fields in southern South America (Freitas, 2022).

In another study, using a population sampling method and SSR data, Baena-Díaz et al. (2018) investigated the timing of hybridization events between two species of Psittacanthus (Loranthaceae) occurring in sympatry in montane forest in the Central Valleys of Oaxaca, Mexico. The hybrid lineage arose c. 99.1 kya through secondary contact, a period of overlap between the two parental species estimated by projections of past habitat suitability and historical demography assessed with molecular markers (Ornelas et al., 2016). Birds have been cited as seed dispersers in these species, and contrary to what was observed for other bird-dispersed plant species, the Isthmus of Tehuantepec may not act as a barrier for dispersal. The authors mentioned that the result also suggests contemporary hybridization derived from the dispersal of seeds by birds (Baena-Díaz et al., 2018). In addition, several hybrid zones have been reported in Mexican Quercus L. (Schley et al., 2022), having started in the Pleistocene. One study assessed the age estimation of lineage divergence through population sampling and restriction site associated DNA sequencing (RADseq) using single nucleotide polymorphisms (SNPs) data (Eaton et al., 2015; divergence of Q. sagraeana Nutt. occurred c. 0.119 Mya). Epidendrum L. (Orchidaceae) has also been the focus of several studies, with ten studies reporting hybridization (Supporting Information 1). A recently published study by Pessoa et al. (2022) used a species sampling strategy and Sanger sequencing (nuclear and plastid sequences) to estimate the time of divergence in a group of five Brazilian Epidendrum spp. from Atlantic Forest (dated age c. 4.37 Mya; Cardoso-Gustavson et al., 2018). The study reported the divergence time of Brazilian species related to E. secundum Jacq. (1.2 Mya), and past hybridization and introgression before lineage divergence.

Möbus et al. (2021) used a species-level approach and a time-calibrated phylogenetic tree based on plastomes to estimate the age of hybrids with Tillandsia marconae W.Till & Vitek (Bratzel et al., 2020) and T. landbeckii Phil. or T. purpurea W.Till & Vitek from Atacama Desert. They also found the age of origin of a hybrid species in Pleistocene (c. 0.119 Mya), younger than the estimated crown age of T. landbeckii. The time of hybrid species origin was associated with the rainy period in the Last Interglacial (LIG), which facilitated the migration of T. purpurea through southern Peru, bringing it into contact with T. landbeckii. The author also discussed the crown age of the parental species with El Niño-Southern Oscillation cycles (Guerrero et al., 2013) that affect rainfall and the distribution of fog in the Neotropics.

Landscape change during the Miocene and the Pliocene shaped the evolution of some species studied, as suggested by the divergence time of hybridizing species pairs (e.g. Scotti-Saintagne et al., 2013; Vargas et al., 2017; Pouchon et al., 2018; Nicola, Johnson & Poznera, 2019; Schley et al., 2020; Bedoya et al., 2021; André et al., 2022; Loiseau et al., 2021). Bedoya et al. (2021), through population sampling strategies and genomic SNPs obtained by RADseq, and target capture, investigated the hybridization history of two aquatic species in the genus Marathrum Bonpl. (Podostemaceae). The authors found that the divergence time fits with the formation of drainage basin associated with landscape evolution by Andean uplift and the Sierra Santa Marta (Villagomez et al., 2011; Parra et al., 2019) that was changing during the time (from c. 11 Mya until present; Latrubesse et al., 2010; Hoorn et al., 2010) giving rise to multiple rivers. The study suggested the origin of a hybrid population (Boyacá drainage basis) that probably arose c. 4.1 Mya through a lowland inter-Andean fluvial portal (c. 13–4 Mya; Montes et al., 2021). Furthermore, although rare, long-distance dispersal events by birds probably promote contact between species making hybridization possible (Philbrick, Bove & Stevens, 2010).

André et al. (2022) also used population sampling strategies and target capture to investigate the relationship between the Chamaecostus subsessilis (Nees & Mart.) C.D.Specht & D.W.Stev. species complex in Cerrado and southern Amazonia. They found that C. subsessilis and C. acaulis (S.Moore) T.André & C.D.Specht and populations from Cerrado and Amazonia of C. acaulis diverged during the Miocene. These two species are restricted to the Tocantins Basins, and the authors found one population of C. acaulis clustering with C. subsessilis, suggesting possible hybridization between them. The hybrid population is placed on the west side of the river and at the edge of the Amazon and Cerrado transition. This species occurs preferentially through gallery forest, and ants disperse the seeds. The ecotone area between Cerrado and the adjacent ecosystem, as Atlantic Forest was also cited as a hybridization zone between two evolutionary lineages in a tree species of Fabaceae (Plathymenia reticulata Benth.; Muniz et al., 2022).

Other studies have also reported species divergence, which began in the Miocene, suggesting that hybridization could have occurred in this period. For example, Scotti-Saintagne et al. (2013) studied the tree species Carapa surinamensis Miq. and C. guianensis Aubl. (Meliaceae) through a population sampling strategy, SSR markers and Sanger sequencing, and they suggested that divergence between these species began during the Miocene: however,  SSR data suggested recent hybridization between them. Loiseau et al. (2021) studied species of Vriesea Hassk. (Bromeliaceae) from Brazil using a species sample strategy and genome skimming data and found that diversification of the group started during the Miocene, but most diversification occurred in the Pleistocene. They found evidence of hybridization in several species. Vargas et al. (2017), using a species-level sampling strategy and RADseq data, found ancient hybridization in a Northern Andean clade of Diplostephium Kunth that diverged c. 10 Mya. Pouchon et al. (2018), also using a species sample strategy and genome skimming data, found a diversification at c. 2.3 Mya for species of Espeletiinae (Asteraceae) in the high Andean region driven by ecological and morphological adaptations and possible hybridization as a role in speciation of Espeletia spectabilis Cuatrec. during the Pleistocene, as discussed by Schley et al. (2022).

In summary, the divergence of the dated lineage/hybrid origin results indicated that the Pleistocene period is the most probable time of hybridization events related to a secondary contact between divergent species through a shift of its range distribution. However, taking together the results from the divergence time of hybridizing species pairs is unclear whether hybridization could occur during the divergence period of some species (Miocene or Pliocene) or as a secondary contact that may also be in the Pleistocene. Thus, in addition to range shift driven by Quaternary climate fluctuations, landscape changes, such as river basis formation/shift, could also be considered a driver of ancient hybridization in the Neotropical flora. We argue the need for more studies on age estimation to understand the most abiotic drive to hybridization in the Neotropical flora. The estimation of divergence time and demography population analyses would provide information concerning the spatial mode of hybrid zones as to whether species diverged in a continuum environmental gradient or were a result of secondary contact. Few papers discuss the causal role of biological traits, e.g. as seed could be dispersed and bring species in contact. In this sense, we suggest future studies should investigate the role of seed dispersal mode on hybridization rates. Some studies have pointed out that rivers could also serve as dispersers of plant seeds, at least in part for some plant groups, even though the fauna associated with the riverine location also act as occasional dispersers. These findings reinforce the need for integrative studies on ecological aspects of hybridization.

POLLINATORS AND HYBRIDIZATION IN THE NEOTROPICS

Hybridization is linked to angiosperm diversity in the Neotropics and was reported mainly as having a positive than a negative effect, increasing genetic diversity (Schley et al., 2022). Pollinators are the main vectors of pollen transfer and play a key role in angiosperm diversification by connecting with a suit of floral traits and serving as a vector of hybridization. Here, we look for an integrative view considering morphology, pollination strategies and hybridization. We selected 30 studies using this framework to assess the drivers of hybridization. Although we cannot directly quantify the contribution of pollinator types to hybridization, we could provide a perspective based on hybridization in plant groups and its biological information. To do this, in some cases, we also search for information in the literature regarding the pollination system in the genera and species discussed here.

We note in the papers analysed that hybridization could occur in specialist and generalist pollination systems. In specialized pollination systems, such as birds and moth flowers, many plant species present a secondary pollinator, generally bees, that play a role in reproduction (Rosas-Guerrero et al., 2014). Thus, we argue that secondary pollinators or sporadic visits could lead to hybrid formation. We will present and discuss these findings next, considering the studies by plant family.

Seven studies reported hybridization in species of Asteraceae in Mexican grasslands and the Andes (Supporting Information 1). Asteraceae mostly show generalist pollination, because the inflorescence does not address assortative specialization to pollinators and most species are pollinated by a wide range of pollinators (Hymenoptera, Diptera, Lepidoptera and Coleoptera; Torres & Galetto, 2008; Vogel, 2015). However, ornithophily was also suggested in Asteraceae on the basis of predictions of floral syndromes probably occurring in shrubs and trees, preferentially at high elevations (Vogel, 2015). The annual herbs Tithonia tubaeformis (Jacq.) Cass. and T. rotundifolia (Mill.) S.F.Blake from grasslands and savannas in Mexico (Tovar-Sánchez et al., 2012; López-Caamal et al., 2013) present morphological and ecological differences. The first occurs mainly in temperate zones > 1,000 m a.s.l., whereas the second occurs preferentially in arid zones < 1,000 m a.s.l. Five hybrid zones have been studied in these species; most occur at > 1,000 m (López-Caamal et al., 2013). Besides maintaining parental phenotypes, the study found a mosaic of morphology in the hybrid zones with transgressive characters. These parental species are self-incompatible and need a pollinator to reproduce. The authors suggested that a wide range of pollinators could visit the flowers based on knowledge of Asteraceae, but no field observations were conducted. Other four studies on Asteraceae were from the high Andean region (Dušková et al., 2010, 2017; Vargas et al., 2017; Pouchon et al., 2018; Nicola et al., 2019). These studies used a species sampling strategy in a phylogenetic context and detected hybridization; however, they did not provide detailed information regarding morphology or pollination systems. Besides that, they also suggested transgressive and intermediate morphology in Senecio L. (Dušková et al., 2017) and possibly a generalist pollination system in Diplostephium (Vargas et al., 2017). Because of the ecological trait of the high Andean region (elevational gradient), the diversity of pollinators is low (Berry & Calvo, 1989), resulting in the same pollinators visiting different plant species. Hybridization is common in Asteraceae and is probably involved in the success of this angiosperm group. Hybridization in Asteraceae was also reported to increase the invasiveness of foreign species by minimizing founder effect and increasing genetic diversity, e.g. in Senecio madagascariensis Poir. in Australia (Dormontt et al., 2017). This species is also invasive in Pampa grasslands in southern Brazil, showing high genetic diversity (Mäder et al., 2016), but hybridization with native species has not yet been investigated. Senecio L. is a species-rich genus occurring worldwide, and it is evident that hybridization is a positive driver for diversification in this group. However, the impact of hybridization in species invasiveness on native community composition driven by human changes in the landscape in the Anthropocene needs to be investigated.

In Orchidaceae, hybridization could be common due to the large number of sympatric species, the lack of reproductive barriers and a food-deceptive pollination system in at least one-third of species. Epidendrum is the largest genus of orchids in the Neotropics, showing morphological and chromosome number diversity and occurring in a broad range of habitats (Pinheiro & Cozzolino, 2013); it was the most represented genus in hybridization studies in this review (Supporting Information 1). In general, these studies cited that Epidendrum does not have a specific pollination system, and many pairs of hybridizing species share the same pollinator, especially Lepidoptera (diurnal butterflies; Pansarin & Amaral, 2008, Pinheiro et al., 2010, 2015, 2016; Pinheiro & Cozzolino, 2013; Vega et al., 2013; Marques et al., 2014; Arida et al., 2021). Many species are food-deceptive (they do not provide nectar as a reward for pollinators), except in one study that included a rewarding species (specialist nocturnal pollination, with a strong odour at night; Arida et al., 2021). These authors showed that hybrids probably present a mixture of signals related to colour and scent, suggesting that the hybrids attract diurnal and nocturnal pollinators. On the other hand, Szlachetko et al. (2017), through observation of yellow-flowered species of Cypripedium L. from montane forest in Guatemala, suggested that the hybridizing species are preferentially pollinated by small bees (Trigona spp. and Halictidae) in a sympatric area where hybrids are formed. This study also identified an overlap niche estimation coincident with the hybrid zone. Many Epidendrum spp. have been reported to be self-compatible, but they need pollinators to transfer pollen grain. One study modelled the persistence of hybrid zones in the context of future climatic changes in the Andes of Ecuador, and they found a substantial reduction of suitable habitats for the pair of species E. madsenii Hágsater & Dodson and E. rhopalostele Hágsater & Dodson and probable maintenance of a hybrid zone though favourable climatic conditions (Marques et al., 2014). Epidendrum fulgens Brongn., occurring in sand dunes in southern Brazil is involved in hybridization with three different species. For example, Pinheiro et al. (2010) reported hybridization of E. fulgens and E. puniceoluteum F.Pinheiro & F.Barros, that differ in floral colour (orange with yellow labellum and red with yellow labellum callus, respectively). These species are self-compatible, but a pollinator vector is necessary following a model of pollination by deceit because they do not produce nectar. These species co-occur, and hybrids with intermediate morphology have been found. The hybrids present physiological adaptations that might confer ecological tolerance (Leal et al., 2020). Pansarim & Amaral (2008) reported hybridization in sympatric population of E. fulgens and E. secundum, a pink-flowered species and probably anthropogenic perturbation could put the widespread E. secundum in contact with E. fulgens.

In Bromeliaceae, eight studies discuss the role of pollination mode in hybridization. These studies were mainly from the Brazilian Atlantic Forest Domain; one is from the Andean region in Chile (Wendt et al., 2001; Jabaily & Sytsma, 2010; Palma-Silva et al., 2011; Lexer et al., 2016; Matos et al., 2016; Zanella et al., 2016; Neri, Wendt & Palma-Silva, 2018; Mota et al., 2019). These studies provide information regarding the specialized pollination system in the most studied species. Even in these specialized systems, secondary pollinators are observed and probably play a role in the formation of a hybrid zone. For example, Pitcairnia albiflos Herb. and P. staminea G.Lodd. have floral traits suggestive of specialized pollination by nocturnal hawkmoth and diurnal hummingbirds, respectively (Palma-Silva et al., 2011). These species are self-compatible and present ecological differences concerning habitat, and although the species are pollinated preferentially by these pollinator functional groups, several floral visitors could visit the parental species and hybrids in a hybrid zone of P. albiflos and P. staminea. Also, some specialized species are reported to hybridize and share the same pollinator. In these cases, the main reproductive barrier is probably the geographical occurrence and/or flowering time. For example, Mota et al. (2019) reported that a ‘mistake’ by the pollinator of the parental species could promote hybridization between two red-flowered species of Pitcairnia L’Hér., one widespread and the other rare, that share the same pollinator. Three studies evaluated hybrid zones in Vriesea from the Atlantic Rainforest. These studies suggested that the main barrier of species is the geographical distribution and the differences in flowering time because all the species have a specialized pollination system (ornithophilous syndrome; hummingbirds Phaethornis eurynome and Ramphodon naevious) and share the same pollinator. A slight overlap in the flowering period and the sympatric occurrence allows hybridization. Sharing of the same hummingbird pollinator (Patagonia gigas) also was reported for species of Puya Molina from the Andes of Chile in a study by Jabaily & Sytsma (2010). However, the generalist pollinators and sporadic visits of secondary pollinators to improve the hybrid formation was not evaluated.

Solanaceae have their centre of origin in the Andean region (Dupin et al., 2016), and the interaction with pollinators has had a key role in species diversity (Knapp, 2010), as was reported as a driver of diversification in Petunia (Fregonezi et al., 2013). Nevertheless, in the study by Turchetto et al. (2019a, b) from the Pampa grassland in southern Brazil, the authors also reported hybridization in a pair of species with specialized pollination systems (white-flowered, scented P. axillaris pollinated by hawkmoths; red-flowered, scentless P. exserta pollinated by hummingbirds). Although pollen flow could occur in most generations (Turchetto et al., 2019a), some selective post-zygotic barriers are related to specific habitats (Caballero-Villalobos et al., 2021). No field observation on pollination has been done in the hybrid zone, and hummingbirds were suggested to visit both species (Lorenz-Lemke et al., 2006). Moreover, bees also have been observed sporadically visiting the flower of P. axillaris. Bee pollination was identified as the ancestral state, and bees are the primary pollinators of many species with short corolla tubes in Petunia (Reck-Kortmann et al., 2014) and of P. secreta Stehmann & Semir, the sister species of P. axillaris (Rodrigues et al., 2018). Species with specialized pollination and probable natural hybridization have also been reported for species of Nicotiana L. occurring in open habitats in the southern Atlantic Forest Domain on the basis of geometric morphometric analyses (Teixeira et al., 2022). The species have pink, scentless flowers pollinated mainly by hummingbirds (Nicotiana forgetiana hort. ex Hemsl.) and scented flowers with nocturnal anthesis pollinated by hawkmoth species, (N. alata Goodsp.). Through field pollination experiments, Kaczorowski, Gardener & Holtsford (2005) observed bees visiting flowers of N. forgetiana and our research group also observed native and foreign bees collecting nectar and pollen in N. forgetiana and collecting pollen in N. alata (see the video in Supporting Information 5). The field observation makes clear that the bees carry pollen from the flowers, and the fact that the population of both species occur close to each other suggests the potential role of the bees carrying pollen from one species to another. It is intriguing to note the behaviour of bees (Apis mellifera) on flowers of N. alata. The flowers have nocturnal anthesis and remain close during the day (as shown in the video, Supporting Information 5). The bees approach the flower and open the petals to access the pollen. We do not know what signal attracts the bees to these flowers. These reports and floral morphology suggest hybrids have been observed in a region of intensive land use for agriculture.

The example cited previously in Nicotiana spp. is one example of a potential secondary pollinator probably promoting gene flow between species with specialized pollination systems, consequently leading to formation of a hybrid zone. However, systematic observations in the field in the hybrid zone are needed to investigate the vectors of pollen flow between N. alata and N. forgetiana. In fact, bees were cited as pollinators or secondary pollinators in many studies revised here. In addition to nectar, bees also collect pollen as the primary source of protein (Nicolson, 2011). It is crucial to investigate the flexibility of bees collecting pollen effectively from a range of different floral morphologies and thus its role in starting hybrid zone formation through ‘casual’ interspecific pollen flow. This question needs future studies to be tested, and research evaluating the chemistry of pollen and its functional significance on foragers changes could be useful in discussing this question. For example, pollen taste influenced the visit of bumblebees, subsequently influencing visits to visually similar flowers (Muth, Francis & Leonard, 2016).

Other plant families were represented in this review with only a few studies, but also suggested specialized pollinators and visits of secondary pollinators, with bees and hummingbirds being the most cited pollinators. For example, in Cactaceae the two studies reported a specialized system for species of Melocactus Boehm. and also cited bees as pollinators (hummingbirds as the main pollinator; Lambert et al., 2006) and bees or hummingbirds for species Opuntia Desv. (Granados-Aguilar et al., 2020) from the desert and dry forests in Brazil and Mexico. In Costaceae, the two studied species of Costus L. from rainforest in Central America (Kay, 2006; Surget-Groba et al., 2013) were associated with pollination by long-billed hermit hummingbirds (Phaethornis longirostris). Rollo et al. (2016) studied tree species of Inga Mill. (Fabaceae) from Peruvian Amazonia and reported occasional visits by bats and hummingbirds during the day, despite the specialized pollination by hawkmoths. Larson et al. (2021) detected deep hybridization in tree species of Eschweilera Mart. ex DC. from Amazonia and reported that bumblebees (Bombus spp.) visit the flowers of both species that have nectar guides and are probably the pollen transfer vectors. Many other studies suggested insect-mediated pollination and generalist pollination, and bees are cited many times as pollination in the hybridizing species (Nettel et al., 2008; Scotti-Saintagne et al., 2013; Luebert et al., 2014; Mori, Zucchi & Souza, 2015; Baena-Díaz et al., 2018; Nevado et al., 2018; Tapia-Pastrana, 2020). For example, André et al. (2022) for species of Chamaecostus C.D.Specht & D.W.Stev. (Costaceae) from Amazonia and Cerrado and Caddah et al. (2013) for species of Kielmeyera Mart. & Zucc. from Cerrado reported pollination mainly by Xylocopa bees. Roberts & Roalson (2018) studied species of Gesneriaceae and reported bees as pollinators in addition to hummingbirds and butterflies and suggested bees and butterflies as pollen vectors between species. Other studies also reported generalist pollinators, as in species of Begonia L. (Begoniaceae; Twyford, Kidner & Ennos, 2015); and species of Brahea Mart. ex Endl. (Arecaceae; Ramírez-Rodríguez et al., 2011). Some studies, especially for tree species, also cited wind pollination, e.g. in Quercus (Castillo-Mendoza et al., 2019), Rhizophora Pers. (Cerón-Souza et al., 2010) and Polylepis Ruiz & Pav. (Schmidt-Lebuhn et al., 2006).

Taken together, these results reveal some aspects of biotic vectors in hybrid zone formation in the Neotropics. Hybridization has occurred in both generalist and specialist systems, and bees and hummingbirds were often cited as pollinator vectors mediating pollen flow between species. For hybridization to occur, first interspecific pollen flow needs to occur. We hypothesize that this is possible even in specialized systems, and sporadic visits by secondary pollinators, especially bees, could promote the start of a hybrid zone formation in the Neotropical flora (Fig. 2). Hybridization in specialized pollination systems mainly results in a mosaic of floral morphology in hybrids, including transgressive characters that could attract different pollinators than the parental species, both specialists and generalists (Fig. 2) and possibly new species formation and/or adaptative introgression. We stress that future studies should consider plant–animal interaction in hybrid zones and floral traits to expand our understanding of the mechanisms of hybrid formation. The regional differences in pollinators and plant communities should also provide key information regarding hybridization and its impact on plant diversity, especially in a changing world.

MAPPING PLANT HYBRIDIZATION IN THE NEOTROPICS

The formation and establishment of hybrid zones are intriguing processes, but understanding them is challenging. A hybrid zone occurs before the species are fully reproductively isolated from each other, with an area of contact (Endler, 1977) or where populations occur in sympatry. Contact and reproduction between genetically distinct groups of individuals resulting in some mixed ancestry offspring define a hybrid zone (Harrison, 1993). The nature, types and structure of hybrid zones have already been well discussed on a worldwide basis (see review by Abbott, 2017). In addition, the recognition that hybrid zones are ‘natural laboratories for evolutionary studies’ (Hewitt, 1988) and ‘windows on the evolutionary process’ (Harrison, 1990) makes their identification widely relevant to understanding the occurrence and evolution of reproductive isolation between species. Population studies based on molecular markers reveal the impact of selection on hybrid zones and its interaction with historical and contemporary gene flow, helping our understanding of genetic diversity distribution (Hamilton & Aitken, 2013; Ma et al., 2019). The balance of dispersal into the centre of the contact zone and selection against hybrids is important to maintaining many hybrid zones in both animals and plants (Barton & Hewitt, 1985). Nonetheless, the factors (biotic and abiotic) that allow the formation and establishment of hybrid zones are still relatively poorly addressed for the Neotropical flora, as discussed in the previous sections. Therefore, the identification and study of hybrid zones are highly relevant both from an evolutionary point of view and for biodiversity conservation.

In this review, we map the hybrid zones reported in hybridization studies and show that most studies (c. 90%) reported hybrid zones (or hybridization among sympatric populations) for the species they analysed. Bromeliaceae, Orchidaceae, Fabaceae, Asteraceae, Solanaceae and Cactaceae present many hybrid zones (Supporting Information 1). This result could be because these are the most studied families regarding hybridization in the Neotropics, but it could also be congruent with the role of hybridization in some taxa in these families. For example, in the largest orchid genus in the Neotropics (Epidendrum), hybridization can be a driving force responsible for the origin of chromosomal and morphological variation (Pinheiro et al., 2009, 2010). The hybrid zones are recorded across most biomes and ecoregions, most of them in the Atlantic Forest Domain (Parana Domain), and mountain in the Andean (Pacific Domain) and Mexican regions (Mesoamerica Domain). For some ecoregions, such as the Amazonian and Cerrado, there are fewer hybrid zone records (Fig. 3). Whether this bias is because hybridization has been poorly tested empirically in such regions remains to be investigated.

Our mapping shows a geographical regionalization of hybrid zones according to families (Fig. 3). Solanaceae, for example, are found forming hybrid zones only in the southern region of the Chacoan domain. The family is one of the most diverse in America, and the centre of diversity is South America (Olmstead, 2013). Peru has the highest number of genera and species and levels of endemism in Solanaceae, but no work reporting hybrid zones was found in the region (Palchetti, Cantero & Barboza, 2020). However, the Andean region has been cited as the centre of origin and diversification of Solanaceae (Dupin et al., 2016). Moreover, some genera, such as Nicotiana, also diversified in the Andean region, with many allopolyploid species of different ages arising there (Clarkson, Dodsworth & Chase, 2017). Thus, hybridization was found more frequently in the centre of origin of these species. Hybrid zones in Bromeliaceae are predominant in the Atlantic Forest, but they are widely distributed in the Neotropics, with three centres of diversity: the Brazilian Atlantic Rainforest; the Andean slopes of Peru, Colombia and Ecuador; and Mexico and adjacent Central America (Zizka et al., 2009; Zanella et al., 2012). The hybrid zones of Asteraceae are concentrated in the Mesoamerican domain, even though many taxa in this family occur in other regions (Zhang et al., 2021). It is difficult to attribute this pattern to any biotic or abiotic factor, and the studies seem to be concentrated in families/genera in which a research group is interested. Nonetheless, the family place of origin and centre of diversification are also possible explanations for different frequencies of hybrid zones in distinct geographical locations (Pennington, Prado & Pendry, 2001; Zanella et al., 2012; Olmstead et al., 2013). More studies of phylogeny and phylogeography are necessary to disentangle the geographical distribution of hybrid zones and remove the bias of the most studied families, especially in regions such as Brazilian Cerrado, Amazonia and Pampa (Schley et al., 2022).

HYBRIDIZATION IN A CHANGING WORLD

As both laboratories for studying reproductive isolation, evolution and diversification and windows for future studies, hybrid zones are relevant to conservation in many ways. Natural hybrid zones provide a source of novel genetic variation that contributes to the processes involved in speciation and adaptation. In this sense, interspecific gene flow via natural hybrid zones may be a source of genetic variation necessary for adaptation to environmental changes (Janes et al., 2017). In the case, introgression between E. fulgens and E. puniceoluteum would increase the adaptability of both species in natural and human-disturbed habitats (Sujii, Cozzolino & Pinheiro, 2019). Hybrids among two Vriesea spp. are found only in disturbed areas, and the authors regarded the colonization of these areas as a positive result (Matos et al., 2016). On the other site, hybridization can promote genetic uniformity among species, leading to the extinction of rare species or giving rise to a single lineage. In the present day, anthropogenic effects are contributing to changing the habitat rapidly, and the species loss models focus on abiotic factors and rarely incorporate species interactions, such as hybridization or adaptive evolution, making it difficult to predict the future of natural hybrid zones or the emergence of new ones (Taylor, Larson & Harrison, 2015). Habitat changes can lead to hybrid zone movement, changes in range overlap and the origin of new hybrid zones (Chunco et al., 2014; Taylor et al., 2014; Ryan et al., 2018). For example, Zelener et al. (2016) concluded that climatic changes and habitat forest degradation contributed to structuring hybrid zones in the Yungas of north-western Argentina. Wendt et al. (2001) found that hybrids among Pitcairnia albiflos and P. staminea frequently occurred in disturbed places where ecological isolation between these species has broken down, possibly as a result of human activity. Pansarin & Amaral (2008) reported the occurrence of hybridization between two species of Orchidaceae at one studied site where they are found in sympatry. One species is mainly found in Restinga vegetation (Epidendrum fulgens) near the Atlantic shore, whereas the other is found in natural bare places at high elevations in semi-deciduous forests and in disturbed areas along roadsides in Atlantic rain forest (E. secundum), and they are not expected to co-occur. Hybridization can probably be related to anthropogenic disturbance that led to the occurrence of the second species in the Restinga as a result of road construction. Tovar-Sánchez & Oyama (2004) noted that hybrid trees were also more numerous in disturbed habitats. Therefore, the conservation of hybrid populations relies on knowing their origin and the pure populations that remain (Thompson, Gaudeul & Debussche, 2010), and the understanding of hybrid zone dynamics is crucial to helping conservation strategies in the Neotropical regions. In this sense, Figure 4 shows the vegetation loss in Brazil during the last 35 years and the hybrid zones (or hybridization in sympatric populations) we were able to map in the region. The hybrid zones mapped are concentrated in Atlantic Forest, the most threatened Brazilian Biome. This makes it clear that more studies are needed to understand the effects of habitat loss and disturbance in the change and creation of hybrid zones.

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article at the publisher’s website:

Supporting Information 1. List of articles of hybridization in Neotropical flowering plants examined in this review. Columns of the table represent the citation of the study, the biome/ecoregion in which the study was carried out, growth form and taxonomic classification of the study group, the type of study and data used in each article, the number of species examined and hybrid species or pairs of species hybridizing, the information about seed dispersal mode, pollination and morphological characteristics of hybrids, and species hybridizing and hybrid zone sites. Details about pollination and morphological characteristics of hybrids and species hybridizing, as well as divergence time information, are presented in separate sheets.

Supporting Information 2. Number of studies of hybridization in Neotropical flowering plants by research area.

Supporting Information 3. Number of studies of hybridization by growth form.

Supporting Information 4. Number of studies of hybridization by plant families.

Supporting Information 5. Video showing two Brazilian Nicotiana spp. foraged by bees. The two species have different floral specializations related to distinct functional pollinator groups. Nicotiana forgetiana (pink to red flowers and no scent) is pollinated by hummingbirds, and N. alata (white flowers, nocturnal anthesis with strong odour emission at night) is pollinated by hawkmoths. However, during fieldwork, we observed bees foraging flowers of both species. Video and data from C. Turchetto, available at https://youtu.be/noOAUl-eVWM.

ACKNOWLEDGEMENTS

We thank M. Teixeira for his help in figures and video editing. We thank the reviewers for their helpful comments on a previous version of the manuscript. We thank the funding agencies Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPQ: 427575/2018-4; 308135/2020-2), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES, and Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS -ARD/ARC 10/2021).

FINANCIAL SUPPORT

This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq - 427575/2018-4), and the Universidade Federal do Rio Grande do Sul (UFRGS) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). ACTZ were funded by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq; 308135/2020-2).

CONFLICT OF INTERESTS

The authors declare that they have no conflict of interest.

DATA AVAILABILITY

The data used in this article are available in Supplementary files.

REFERENCES

Author notes