INTRODUCTION

The Cervidae is one of the most diverse family of large mammals, containing more than 60 species that currently inhabit nearly all conti nents (^{Wilson & Mittermeier 2011}; ^{Heckeberg & Wörheide 2019}). Their origin and evolutionary history date back to almost 20 million years ago (MYA) to the Miocene and Early Pliocene in Eurasia (^{Webb 2000}). By the early Pliocene, true cervid morphotypes became identifiable in North America; nevertheless, the evolutionary history of deer in North America and the Neotropics still remains somewhat obscure (^{Stehli & Webb 1985}).

Approximately 2.5-3 MYA, during the late Pliocene, the uplift of the Panamanian land bridge allowed deer to spread southwards, as part of the “Great American Biotic Interchange” between North and South America (^{Stehli & Webb 1985}).

The Neotropical is considered one of the richest biogeographical regions for the diversity of deer. It contains several biodiversity hotspots and encompasses an impressive range of biomes, reflecting high gamma diversity, including both dry and moist tropical forest, woodlands, savannahs, mangroves, and montane habitats (^{Myers et al. 2000}). The success of these taxa in South America may be attributed in part to the absence of other ruminants occupying the same ecological niche (^{Webb 2000}). However, the evolution and taxonomy of the Cervidae in this region still remain unclear (^{Webb 2000}; ^{Duarte et al. 2008}; ^{González et al. 2017}).

The origin and evolution of Neotropical deer species has been a matter of much speculation and debate, as is reflected in the uncertainties about their evolutionary relationships at differ ent taxonomic levels (^{Gilbert et al. 2006}; ^{Duarte et al. 2008}; ^{Merino & Rossi 2010}). Until recently, several questions regarding the origin of American deer and the timing of their colonization of South America had not been addressed (^{Gilbert et al. 2006}; ^{Duarte et al. 2008}).

In this article we review the taxonomy, evolution, and patterns of speciation of Neotropical Cervidae, and update the current conservation situation of these taxa.

NEOTROPICAL DEER TAXONOMY

Seventeen species of Neotropical cervids are currently recognized and are grouped into six genera: Blastocerus, Hippocamelus, Mazama, Odocoileus, Ozotoceros, and Pudu. Two major morphological forms that are differentiated mainly by size are often recognized (^{Eisenberg 1987}, ²⁰⁰⁰; ^{Weber & González 2003}; ^{Merino et al. 2005}; ^{Merino & Rossi 2010}; ^{González et al. 2017}): Small deer less than 60 cm high at the shoulder (usually < 25 kg), where males have unbranched spike antlers, with morphological and ecological adaptations to forest and other closed vegetation habitats (Mazama and Pudu). Medium to large deer species (> 25 kg), typical of more open vegetation types, where males have branched antlers (Odocoileus, Hippocamelus, Ozotoceros, and Blastocerus).

The phylogeny of the Cervidae based on morphological characters has been repeat edly questioned because of high levels of homoplasy (^{Merino et al. 2005}; ^{Duarte et al. 2008}; ^{Merino & Rossi 2010}). There are several examples of species that were considered synonyms (e.g., Mazama nemorivaga and M. gouazoubira, which were considered as valid different species after proper review, additional cases of misidentification involved several red brocket deer species, such as M. americana, M. temama, M. rufa, and M. bororo (^{Duarte 1996}).

Between 6 and 18 species of brocket deer (genus Mazama) have been identified based on morphological characteristics (^{Allen 1915}; ^{Eisenberg 2000}; ^{Merino et al. 2005}; ^{Merino & Rossi 2010}). They are adapted to live in dense vegetation habitats and show high levels of con vergence. Brocket deer have been traditionally divided into two groups, red and gray, using pelage coloration patterns and body size and shape (^{Allen 1915}; ^{Duarte et al. 2008}).

Red Brockets: This group of ecological specialist species inhabits closed forests in montane areas in the tropical Andes, and in lowlands from Argentina north to tropical Mexico. Taxonomic revisions based on cytogenetic data combined with morphological approaches have proven useful to recognize new cryptic species of red brockets (^{Groves & Grubb 1990}). Studies using cytogenetics led to the identification of another species from Mexico, the temama brocket (M. temama) as distinct from M. americana (^{Taylor et al. 1969}; ^{Jorge & Benirschke 1977}; ^{Neitzel 1987}; ^{Sandoval 2019}). Other karyotypical studies reveal sharp divergences between several red brocket species (^{Duarte 1992}; ^{Duarte & Jorge 2003}; ^{Abril et al. 2010}; ^{Cursino et al. 2014}; ^{Rincón 2016}; ^{Salviano et al. 2017}; ^{Luduvério 2018}; ^{Sandoval 2019}).

Gray Brockets: The gray brocket group is similarly widespread. Among the gray brocket deer species, ^{Cabrera (1960)} considered the Amazonian gray brocket deer (M. nemorivaga) to be a synonym of the gray brocket deer (M. gouazoubira). However, the two species can be distinguished morphologically (^{González et al. 2018}) and genetically (^{Borges 2017}; ^{Donoso 2017}). Additional examples of species that were not recognized in the 1960s include species with M. pandora, in the Yucatán of Mexico (^Allen1915; ^{Medellín et al. 1998}). restricted ranges, such as M. chunyi, an Andean species found in southern Peru, and M. pandora, in the Yucatán of Mexico (^{Allen 1915}; ^{Medellín et al. 1998}).

Morphological approaches have not proven to be efficient tools to discriminate cryptic species in sympatry. The taxonomy of brocket deer has, however, proven challenging due in large part to morphological differentiation not clearly associated with the wide karyotypic diversification among the species in this genus (^{Duarte & Merino 1997}; ^{Duarte & Jorge 1998}; ^{Duarte et al. 2008}). Diploid chromosome numbers vary in brocket from 32 to 70 (^{Duarte & Jorge 2003}). Compelling arguments further exist to split the genera Pudu and Hippocamelus, with the latter including also two cryptic species (^{Duarte et al. 2008}).

NEOTROPICAL DEER GENETIC AND PHYLOGENETIC RELATIONSHIPS

The evolutionary history of South American deer has been studied using a variety of genetic markers: isozymes (^{Smith et al. 1986}), cytogenetics (^{Neitzel 1987}; ^{Spotorno et al. 1987}; ^{Duarte & Merino 1997}), and mitochondrial and nuclear markers (^{Gilbert et al. 2006}; ^{Duarte et al. 2008}; ^{González et al. 2017}).

The development of powerful molecular techniques has improved the approaches that can be used to estimate the genetic diversity in Neotropical deer species. It is now possible to quantify the genetic variability of historical populations and their living descendants, and thus assess the rate at which genetic variation is being lost and restructured in fragmented populations (^{González et al. 1998}; ^{Márquez et al. 2006}, ^{Mantellatto et al. 2017}).

Analysis of mitochondrial cytochrome b sequences showed an astounding and complex evolutionary pattern and phylogenetic relationships among Neotropical deer, particularly in brocket deer. Phylogenetic analysis showed two clades with different evolutionary histories, as well as hybridizations episodes (^{Duarte et al. 2008}; ^{González et al. 2010}; Fig. 1). The mitochondrial DNA phylogeny shows a clade that includes gray brockets (M. gouazoubira and M. nemorivaga), marsh deer (B. dichotomus), huemul (H. bisulcus), and pampas deer (O. bezoarticus) whose ancestor was estimated to have lived approximately 5 MYA. A second clade includes the red brocket deer group of species (M. bororo, M. nana, M. americana, and M. temama), white tail deer (O. virginianus) and mule deer (O. hemiomus), whose common ancestor was estimated to have lived approximately 2 MYA ago. These phylogenetic analyses suggest that what is now considered the genus Mazama actually corresponds toa polyphyletic arrangement, with one cladeincluding only the red brocket and the otherthe gray brocket species. Consequently, many of the morphological traits used to identify Mazama (e.g., unbranched antlers) are evolutionary convergent features, likely associated with adaptations to similar environments.

SPECIATION AND REPRODUCTIVE ISOLATION

The colonization of South America by deer appears to have occurred by at least 6 distinct forms: (1) the ancestor of M. gouazoubira and H. bisulcus; (2) the ancestor of B. dichotomus; (3) the ancestor of O. bezoarticus, H. antisensis, and M. nemorivaga; 4) the ancestor of P. puda; and 5) an ancestor that gave rise the red brocket deer M. americana, M. nana, M. bororo, and 6) O. virginianus (^{Duarte et al. 2008}, Fig. 1). Apparently, not all invasions occurred during the formation of the Panama Isthmus at the end of the Pliocene. In particular, the invasion of O. virginianus likely occurred more recently. Its southern population has changed little in relation to the North American populations and has not yet been able to cross the barrier of the Amazon rainforest southwards.

The ancestral forms that invaded South America quickly differentiated and established, with some species having a clear taxonomic identity, such as B. dichotomus (^{Márquez et al. 2006}), O. bezoarticus (^{González et al. 1998}, ²⁰⁰², ²⁰¹⁰), H. bisulcus (^{Vila et al. 2010}), and H. antisensis (^{Barrio 2010}). The species currently recognized in the Mazama genus underwent a more complex evolutionary process linked to chromosomal rearrangements that produced a very rapid speciation (^{Abril et al. 2010a}; ^{Potter et al. 2017}).

Brocket deer (Mazama) have impressive chromosomal variation, with diploid numbers ranging from 2n = 32 to 2n = 70 (^{Duarte & Jorge 1996}, ²⁰⁰³; ^{Abril & Duarte 2008}; ^{Abril et al. 2010b}). This wide variation also occurs in Muntiacus and seems to be explained to chromosome fragility (^{Yang et al. 1995}). Several studies performed to evaluate chromosome stability applying mutagenic agents showed that, particularly in the case of M. gouazoubira, chromosomes show great breakability (^{Vargas-Munar et al. 2010}; ^{Tomazella et al. 2017}). This presents a high degree of intra-population polymorphism. For instance, three different chromosomal rearrangements were found in individuals from a restricted area (2000 ha) in the Brazilian Pantanal (^{Valeri et al. 2018}). The centric fusions (Robertsonian translocations) found in this population have the capacity to generate post-zygotic isolation, which can produce sympatric speciation.

Phylogenetic tree showing the relationships among 59 deer haplotypes derived from a 934 base pair fragment of the mitochondrial cytochrome b compiled with the computer program MEGA 4 (^{Tamura et al. 2007}) after performing a branch length test (^{Takezaki et al. 1995}) to test for differences in base substitution rates. Bootstrap values (1000 replicates) and Bayesian posterior probabilities (> 50%) are denoted above nodes. The geographic location is denoted with abbreviations for each of the following Brazilian states: PA, Pará; RO, Rondônia; GO, Goiás; PR, Paraná; AM, Amazonas; AC, Acre. (gb), GenBank. Yellow bar refers to the timing of the uplift of the formation of the land bridge and timing of the entry of cervids into South America. The scale on top corresponds to time whereas the scale below corresponds to the observed mean sequence divergence using the substitution model K2P (^{Kimura 1980}). Tree modified from ^{Duarte et al. (2008)} with permission of Duarte & González (2010).

Fig. 1. 

The Amazonian gray brocket deer species (M. nemorivaga) also need deeper assessment. Morphological analyses allowed to discriminate the gray brocket deer (M. gouazoubira) from the Amazonian gray brocket deer (M. nemorivaga), but no intraspecific differences could be identified (^{González et al. 2018}; ^{Rossi 2000}). Previous cytogenetic and molecular data suggested, however, the existence of at least three clearly distinct species within this taxon (^{Fiorillo et al. 2013}; ^{Figueiredo 2014}; ^{Donoso 2017}).

The most impressive example of chromosomal polymorphism (from 2n = 42 to 54) occurs in the red brocket deer complex (M. americana), which includes cryptic and sympatric species (^{Abril et al. 2010b}; Fig. 2). The karyotypical analysis discriminated among the red brocket species (M. americana) from the northern Neotropics (Mexico) and identified another taxon from Mexico, the temama brocket (M. temama) (^{Taylor et al. 1969}; ^{Jorge & Benirschke 1977}; ^{Neitzel 1987}; ^{Sandoval 2019}). These karyotypical divergences between red brocket species have now been confirmed (^{Duarte 1992}; ^{Duarte & Jorge 2003}; ^{Rincón 2016}; ^{Borges 2017}; ^{Donoso 2017}; ^{Luduvério 2018}; ^{Sandoval 2019}).

^{Rincón (2016)} proposed a neotype for the red brocket deer (M. americana) collected at the type locality in French Guiana. The neotype was then compared with other M. americana cytotypes from several Brazilian locations showing that they are undoubtedly genetically distinct, thus indicating that a detailed taxonomic revision is necessary. The application of cytogenetic and molecular analyses has already allowed revalidation of M. rufa, which inhabits the interior of the Atlantic Forest, one of the most threatened areas of this biome (^{Luduverio 2018}). This species was previously described by ^{Azara (1802)} and nominated by ^{Illiger (1811)}, but was later considered a subspecies (^Cabrera1960) or a junior synonym of M. americana (^{Rossi 2000}).

Fig. 2. Male specimens of red brocket deer M. americana and their respective karyotypes showing the corresponding karyotypes, 2n = 45 and 53, respectively. Hybrids between these two cytotypes are sterile (^{Cursino et al. 2014}; ^{Salviano et al. 2017}). 

Another example of two cryptic and sympatric species were the small red brocket deer (M. bororo) and the red brocket deer (M. americana), that can only be distinguished by their remarkably different karyotypes (^{Duarte 1996}; ^{Duarte & Jorge 2003}).

Experimental crosses of individuals with different karyotypes within the red brocket complex resulted in hybrid infertility, indicating post zygotic isolation (^{Cursino et al. 2014}; ^{Salviano et al. 2017}).

CONSERVATION IMPLICATIONS

Based on the latest Red List™, 10 (59%) of the 17 species of Neotropical deer are listed in a threatened category, two are Data Deficient (DD), two are listed as Near Threatened (NT), and two as Least Concern (LC) (^{IUCN 2019}; Table 1). The conservation situation of the Neotropical deer species is considerably worse when compared with all the mammal species assessed, of which 22% are threatened and 15% data deficient (^{Schipper et al. 2008}). The number of threatened species will certainly increase once the taxonomy has been updated, especially with the revision of the Mazama species. An easy distinction of the species is crucial for monitoring population trends, designing effective management plans to protect and recover endangered species, and performing a sustainable use of Neotropical deer species in rural communities. As most of them inhabit forested and forestry areas, developing non-invasive sampling techniques will be necessary to diagnose species and determine occurrence areas.

Several Neotropical deer species have broad geographic ranges in vulnerable Latin American ecosystems. Most threats for the species are directly linked to human activities, such as habitat loss, degradation, fragmentation and overhunting, which have impacted and limited them to a small portion of their former range. There are also indirect threats as a result of human activities that have affected the conservation status of Neotropical deer, such as urbanization, development, the use of dogs in rural areas and globalization trends (^{González et al. 2017}).

The trophic competition and pathogen transmission with exotic ungulate species (Sus scrofa, Bos taurus) and introduced deer (Cervus elaphus, Axis and Dama dama) also affect the native populations of Latin American deer and lead to their decline (^{Patz et al. 2004}).

The deer species that shown a decline trend are: marsh deer, pampas deer, taruca, huemul, northern pudu, southern pudu, Brazilian dwarf brocket deer, small red brocket deer, merida brocket deer, and Peruvian dwarf brocket deer. Reductions of species distributions are estimated to range from 40 to 90% (^{Weber & González 2003}). If the goal is to maintain long-term population stability and preserve genetic variation, conservation efforts should focus on the restoration of deer habitats in Latin America.

Deer were included in the Convention on International Trade in Endangered Species of Wild Fauna and Flora, with the aim to protect them and avoid illegal traffic (^{CITES 2017}).

Table 1 Detailed geographic range, habitat and Red List™ (www.iucnredlist.org) category of Neotropical deer species. 

FINAL REMARKS AND PERSPECTIVES

We recommend continuing the detailed taxonomic revision of the genus Mazama. Sampling cryptic species in more detail would allow defining their geographic distribution patterns. Field studies on the behavior, home range, feed ing ecology, demography, and geographic range of small forest species (e.g., Mazama) are urgently needed. Brocket deer are elusive animals, and intensive efforts should include sampling to obtain karyotypes and DNA sequence data. We have designed a set of primers that amplify informative sequence regions of mitochondrial DNA from feces that allow discriminating between species and analysing intraspecific variability (^{González et al. 2009}). Additional samples can be effectively secured with the help of trained scat detection dogs (^{Oliveira et al. 2012}; ^{Duarte et al. 2016}). Additionally, the isolation of DNA from museum collections and archaeological sites are a promising tool for surveying deer and reconstruct their past geographic ranges. Next-generation sequencing technologies (NGS) have boosted the field of ancient archaeological DNA (^{Moreno et al. 2016}). These techniques promise to revolutionize the field of molecular genetics with applications in evolution, ecology, medicine, and ancient DNA (^{Davey et al. 2011}). In addition, these novel methodologies will be useful to recognize new species and evolutionary significant units and achieve a better resolution of the molecular phylogeny and evolution for managing populations (^{Moritz 1994}).

The contribution of genetic and biogeographic data will be useful for solving the uncertain taxonomy and update the assessment of the conservation status of Neotropical deer species. This key information will provide data to perform models for testing management and conservation policies. Finally, if the goal is to maintain long-term population stability and preserve genetic variation, conservation efforts should focus on the restoration of deer habitats in Latin America.