INTRODUCTION
The middle ’90s saw the beginning of Mastozoología Neotropical (Ojeda 1994), now the leading journal for Neotropical mammalogy, and also the rise of a new field called macro- ecology (Brown & Maurer 1989; Brown 1995). The macroecological approach focuses on large spatial and temporal scales to find and explain recurrent patterns that emerge consistently enough to suggest that general mechanism are in operation (Brown & Maurer 1989; Brown 1995). Such mechanisms can take place at the assemblage and ecosystem levels, as well as at the individual and population levels (Brown 1995; Brown et al. 2004; Smith & Lyons 2011), helping to distinguish macroecology from biogeography (the latter usually centered on species-level processes such as speciation, extinction and range expansion).
Nevertheless, macroecology and biogeography are not always easily distinguished, and with the omnipresence of molecular phylogenetics, both are now closely connected to macroevolution. Common questions in these fields involve the search for processes behind patterns of species richness (e.g., Pereira & Palmeirim 2013), phylogenetic lineage distribution (e.g., Duarte et al. 2014), body size variation (e.g., Martinez et al. 2013), and range size distribution (e.g., Arita et al. 2005), among others. The merger of macroecology and biogeography should be encouraged and embraced, since using a common language and concepts can only help us to understand patterns and processes at regional and larger scales (Jenkins & Ricklefs 2011).
While the overall number of macroecological studies have been rising fast in the last decades (Smith et al. 2008; Smith & Lyons 2011; Weber 2018), studies considering only Neotropical rodents are scarce. In order to contextualize the current state of macroecological/biogeographical research for this group, I conducted a search for studies of macroecology specifically focused on Neotropical rodents. This search returned only 10 articles in the last 25 years (searched in the database of Web of Science using the terms: [macroecol* AND (neotrop* OR South America OR Central America) AND (rodent* OR sigmodont* OR caviomorp* OR hystrico*)], including mentions in the title, abstract and key words). Replacing macroecology with biogeography (term: [biogeogr*], other terms unchanged) returned 258 articles for the same time window. While some of the biogeography articles may have used an ecological approach considering a relationship between organisms and environment, still a macroecological em phasis on these rodents seems to be scarce.
In the search for general regularities, macroecology has focused on patterns characterizing broad taxonomic groups, such as mammals (Safi et al. 2011) and birds (Jetz et al. 2012), which explains the small number of articles mentioned previously. Indeed, such a broad focus can help uncover patterns sustained by deep biological functioning (e.g., a pattern repeatedly found, regardless of the taxonomic group analysed), such as metabolic scaling theory (Brown et al. 2004; West & Brown 2005). However, this approach can also mask some interesting patterns that occur in particular clades. For example, bats have an unusual biology, and Neotropical clades of bats have been targets of interesting discoveries using macroecological investigations (Stevens 2005; López-Aguirre et al. 2018). Monophyletic clades with singular histories of colonization in a given region can illuminate particular macroecological patterns, which are obscured when more comprehensive groups with multiple independent histories are mixed. Focusing on monophyletic groups as the hierarchical level of analyses (Eldredge 1985) offers interesting perspectives into ecological patterns and processes.
Neotropical rodents comprise the largest component of mammalian diversity in the Neotropics, including more than half of all mammal species in this region (Patterson 2000). Two major radiations contain most of the ~650 rodent species in the New World (Patton et al. 2015): the sigmodontines and the caviomorph rodents. Caviomorphs have a ~50 million-year history in the Neotropics, which began after transoceanic dispersal from Africa, and the group has about 250 living species today (Rowe et al. 2010; Upham & Patterson 2015; Vucetich et al. 2015). Sigmodontines have a shorter history of diversification, dating back ~10 million years (Steppan et al. 2004; Leite et al. 2014), and the colonization of South America was performed via dispersal from North America (Parada et al. 2013; Vilela et al. 2014). Nevertheless, sigmodontines underwent a rapid radiation resulting in over 400 living species (D’Elía & Pardiñas 2015). The contrasting histories of colonization of caviomorphs and sigmodontines in the Neotropics pose a number of questions that call for a macroecological perspective; few of these questions have been answered. In this Perspective, I offer a brief standpoint on the current research on the most basic subjects of macroecological research—patterns and processes of species and lineage diversity, body size, and range size distribution—focusing on Neotropical rodents.
PATTERNS
Advances in taxonomy and field studies have led to refined data on the distribution of rodent species over the years (Wilson & Reeder 2005; Patton et al. 2015). This effort allowed the depiction of patterns of rodent richness over the Neotropics (Amori et al. 2013), revealing a latitudinal gradient with peaks of richness in tropical montane regions. Concatenating data on biogeographical ranges for each major clade separately reveals that sigmodontines and caviomorphs share similar patterns of richness, although sigmodontine richness is largely associated with mountainous regions in the Andes and in the Atlantic forest, while caviomorphs better reflect a latitudinal gradient with richness positively associated with temperature (Maestri & Patterson 2016) and peaks in richness even at low elevations in Amazonia (Upham & Patterson 2012). The Andes harbor/support regions of high richness for both groups, as well as regions of endemism (Ferro 2013; Upham et al. 2013; Prado et al. 2015), and are crucial to sustaining rodent diversity in South America (Patterson et al. 2012; Novillo & Ojeda 2014). Unsurprisingly, spatial patterns of species beta-diversity uncover the highest values of rodent turnover along the Andes chain (Maestri & Patterson 2016), a pattern shared by both caviomorphs and sigmodontines.
Patterns of species richness and diversification dynamics can vary according to the phylogenetic scale used (Morlon et al. 2011; Graham et al. 2018), making it important to evaluate clade-specific patterns. For example, variation in species richness by elevation varies by family of anurans (Hutter et al. 2017), which is similar to the strong effect of elevation found on sigmodontine richness than on caviomorph richness (Maestri & Patterson 2016). Therefore, by pulling contrasting clades together in a single analysis, dominant groups can mask richness patterns occurring in less species rich groups, and interesting patterns can vanish. Given the different histories of diversification of sigmodontines and caviomorphs—as is also true for many other large groups—it is important that both rodent clades are treated separately in macroecological studies. The resulting distinct patterns/processes can then be compared to achieve synthesis on the factors affecting biodiversity.
The uneven distribution of lineages inside big clades (e.g., sigmodontines and caviomorphs) can also generate distinct patterns of lineage diversity and distribution (Heard & Cox 2007), calling for a phylogenetic perspective. A number of articles on biogeography of Neotropical rodents confirm that a historical approach is important to understand current patterns (Leite et al. 2014; Upham & Patterson 2015; Prado & Percequillo 2018; Gonçalves et al. 2018; Machado et al. 2018). Moreover, using phylogenetic metrics of diversity can be a way to approximate historical biogeography and macroecology (Davies & Buckley 2011; Cisneros et al. 2014; Stevens & Gavilanez 2015). Few studies have considered phylogenetic metrics of diversity for Neotropical rodents at a continental scale. At least one study with caviomorphs has shown regions of high phylogenetic diversity (controlled for richness effects) associated with open areas in the Neotropics (Fergnani & Ruggiero 2015), and one study with sigmodontines revealed lower average phylogenetic relatedness for assemblages in the Amazon basin and higher values for assemblages in Central America and northern South America (Maestri et al. 2019). Patterns of phylogenetic diversity and phylogenetic turnover are still to be entirely explored for Neotropical rodents and its sub-clades at a macroecological scale.
It is important to stress that well-resolved phylogenetic trees are essential to reliably interpret phylogenetic metrics in macroecology. Despite the existence of many phylogenetic hypotheses for Neotropical rodents (e.g., Upham & Patterson 2015; Maestri et al. 2017), better phylogenies are still needed to resolve the relationships among species and subclades within caviomorphs and sigmodontines. For example, tribal-level relationships among sigmodontine rodents are still poorly resolved (Steppan et al. 2004; Parada et al. 2013; Leite et al. 2014). Moreover, the scarcity of fossils for sigmodontines cast doubts on the chronology of diversification (Barbière et al. 2019). A broader taxonomic and gene coverage is needed to increase support for phylogenetic patterns in macroecology.
Other patterns such as the distribution of range sizes (Ruggiero & Werenkraut 2007; Novillo & Ojeda 2012) and traits like body size (Medina et al. 2007) are still little known for Neotropical rodents. Despite the all-importance of Andes to determine rodent richness in the Neotropics, we lack an understanding of patterns of range-size distribution (Rapoport 1982) and its association with elevation (Stevens 1992). Body size patterns have been explored for a few rodent groups (e.g., Medina et al. 2007; Maestri et al. 2016). Body size variation in caviomorphs is very heterogeneous in its rates of evolution (Álvarez et al. 2017) and seems to be associated with variation in life mode (Upham 2014), although we still lack a clear picture of how size is spatially distributed. For sigmodontines, assemblages of species with larger body sizes seem to be associated with open and warm areas in South America (Maestri et al. 2016), but more studies are needed to refine body size estimates and investigate within-species and cross-species patterns. Skewness of body size and other traits are also relevant and its patterns are still to be explored. Patterns in the distribution of other phenotypic traits are still poorly understood at a macroecological scale, although efforts have been made to understand the evolution of such traits as appendicular morphology, tail and feet length, and molar morphology across species (Morgan & Álvarez 2013; Carrizo et al. 2014; Parada et al. 2015; Tulli et al. 2016; Tavares et al. 2018a). Attempts to depict spatial patterns of shape variation are promising and have shown, for example, an association between relatively larger and rounded tympanic bullas with arid areas in the Neotropics at an assemblage level (Maestri et al. 2018), a pattern recurring generally among rodents (Alhajeri et al. 2015). Evidently, phenotypic traits carry a relation ship with overall body size, and allometry is an important factor in trait evolution (Marroig 2007; Tavares et al. 2018b), which must also be accounted for in future macroecological studies of traits. Another pattern frequently overlooked for these rodents, but little studied, is the frequency of sexual dimorphism across species, which can generate a macroecological pattern for sexual selection (Macías-Ordóñez et al. 2014).
A voluminous number of recent investigations have studied body size and other traits at the macroevolutionary level (Alhajeri et al. 2016; Álvarez et al. 2017; Maestri et al. 2017; Tavares et al. 2018a), in large part enabled by the availability of data in museums and the ability to access DNA sequences from repositories (Lessa et al. 2014; Dunnum et al. 2018). General macroecological patterns for traits can be expected to emerge from spatially explicit approaches to trait evolution (e.g., Polly et al. 2017), and integrating concepts and analysis from related disciplines, such as metacommunity and phylogenetic ecology (Duarte et al. 2018), which can be achieved by mapping the trait’s averages and disparities among point localities or grid cells and associating their spatiotemporal variation with major environmental and geological events.
PROCESSES
Understanding the processes that generate broad-scale patterns of diversity is a long- term goal of macroecology. The task is not straightforward and involves the adoption of mechanistic and process-based models that consist of explicit expectations derived from biological reasoning (Keith et al. 2012; Connolly et al. 2017). A recent advance incorporated estimates for parameters as dispersal, evolutionary rate, time for speciation, and competition into a macroecological model (Rangel et al. 2018), offering a promising approach to the study of processes in macroecology. Simple macroecological models could be improved by incorporating biotic interactions and historical factors, which would allow the development of proper null predictions. To date, progress in the understanding of processes for Neotropical rodents has been large gained through inferences from statistical associations between biodiversity variables and environmental and biogeographical predictors.
How the processes of speciation, extinction, and dispersal unfolded to generate the current diversity patterns are still poorly understood. Integrating information from studies on phylogenetics and biogeography of Neotropical rodents (e.g., Leite et al. 2014; Upham & Patterson 2015) with a spatially explicit macroecological standpoint offers a new perspective on how diversification occurred across space. For example, regions of faster diversification for sigmodontines have been found in southern and northeastern South America, which were among the last regions to be occupied during the group’s biogeographical history (Maestri et al. 2019), suggesting that processes of diversification depend on the sequence of biogeographical occupation. Studies that integrate estimates of diversification and dispersal (e.g., using new promising approaches as the geographical state-dependent diversification—Goldberg et al. 2011) with explicitly spatial and environmental contexts are still lacking, and may offer novel perspectives on the processes behind the observed patterns of species richness and phylogenetic diversity for Neotropical rodents.
The relationship between richness and abundance across space has been the subject of few Neotropical studies (e.g., Novillo & Ojeda 2014). Insights into macroecological patterns of abundance can be gained by exploring new refined data, such as those recently compiled for the Atlantic forest (Figueiredo et al. 2017). Another potentially interesting approach to understand processes is to investigate the role of large-scale biotic interactions on species richness. Caviomorphs had more than 30 million years to colonize the Neotropics (Lessa et al. 2014; Upham & Patterson 2015), and still, incumbency effects seem to be unimportant given the astonishing radiation of sigmodontines (although caviomorphs have an extensive record of extinctions). Yet, a proper assessment of the influence of biotic interactions on richness patterns has not been conducted. A perspective focused on interactions among species may help to elucidate the causes of richness differences among clades (Fig. 1).
Elucidating the processes behind range size distribution depends first on documenting the patterns. The general relationship between elevation and range size (Stevens 1992; McCain 2009) may hide interesting explanations considering the particular association of rodents with elevation and the peculiarities of the Andes Mountains. For instance, Steven’s rule (Stevens 1992) predicts larger elevational ranges with increasing elevation, following an increased climatic variability at high elevations (McCain & Knight 2013). However, Andes Mountains in South America have extensive Puna regions, which may have homogeneous climatic conditions, possibly disrupting the expected linear relationship at the highest elevations (Patterson et al. 1998).
Patterns of body size can have a number of ramifications still to be explored. Body size is positively related with total metabolic costs (Kleiber 1932; Brown et al. 2004) and negatively with population density (Damuth 1981), such that an ‘energetic equivalency’ exists where energy use is independent of body size (Nee et al. 1991), but see Marquet et al. (1995). Large species are also thought to utilize larger geographic areas because of energetic necessities (Diniz-Filho & Balestra 1998; Olifiers et al. 2004) and have increased extinction rates (Cardillo et al. 2005) compared to small animals. If so, small rodents can be expected to have smaller range sizes than large rodents (still to be properly investigated) and that can contribute to higher speciation rates (greater opportunity for reproductive isolation due to range fragmentation) and less extinction in such clades owing to higher population density. Therefore, understanding the forces behind body size evolution can explain size variation and also illuminate patterns of diversity for these rodents. We still know little about rates of energy intake and consumption for Neotropical rodent species, as well as the effect of ambient temperature on body size (Naya et al. 2018), and such knowledge is essential to understand the basics of metabolic demands and its differences among rodents. This may be fundamental to understand processes in macroecology (Brown et al. 2004).
Finally, we do not understand how traits are associated with diversification rates for Neotropical rodents. Body size, appendicular and molar morphology, and skull shape and size are a few of the phenotypic characteristics that have been studied (Morgan & Álvarez 2013; Parada et al. 2015; Maestri et al. 2017). Any could have acted to trigger niche occupation and species diversification. Yet, few attempts have been made to connect traits to diversification rates (Parada et al. 2015; Álvarez et al. 2017), which promises to reveal the links among rates of diversity, disparity, and its uneven spatial distribution.
FINAL CONSIDERATIONS
Recent decades have seen a rise in the field of macroecology, which could be attributed to both the development of spatial statistics and the availability of data (Smith et al. 2008). A macroecological perspective is highly dependent on studies that collect biodiversity data and/or document the distribution, phenotypic, and genetic characteristics of organisms (Beck et al. 2012). Improvements in macroecological research will thus depend on basic biological research and the availability of specimens in scientific collections and information on their spatial distribution (De la Sancha et al. 2017). Macroecological models would be greatly improved with enhanced biological knowledge, including stronger data on density and dispersal abilities, biotic interactions, as well as diet, habitat, and behavior.
Remaining questions include intensive investigations of the most basic aspects of species diversity, body size and range size distribution, some presented in this perspective. Certainly, the distinctive histories of Neotropical rodent clades must be considered explicitly, since, for example, regions functioning as ‘cradles’ and ‘museums’ of biodiversity are likely to be clade-specific. By comparing processes responsible for biodiversity distribution in those clades, general regularities may emerge, and knowledge about the evolution and biogeography of these rodents can be achieved.
Decades of efforts by taxonomists, paleontologists, ecologists, phylogeneticists and many others working on Neotropical rodents (Reig 1986; Pardiñas et al. 2002; D’Elía 2003; Weksler et al. 2006; Voss et al. 2013; Luza et al. 2015) have created the foundation to document patterns and understand processes in macroecology. In turn, a macroecological perspective can help to integrate ecology, evolution, and biogeography of Neotropical rodents, and elucidate the biological mechanisms behind large-scale ecological patterns. Much work remains to be done, both to document patterns, and especially to understand underlying processes—this task is still in its infancy for Neotropical rodents.