The Anoa, including the Mountain Anoa (Bubalus quarlesi) and the Lowland Anoa (Bubalus depressicornis) according to current taxonomy, are endangered bovid species endemic to the tropical island of Sulawesi. They are grouped in their own subgenus Anoa. The historical biogeography and phylogeny of this subgenus have not been well characterized. There are two hypotheses describing the colonization routes of large mammals in Southeast Asia: land-bridges and insular route. The present study aimed to understand the molecular phylogeny and test the historical colonization route of Anoa. A total of 71 mitochondrial DNA (cyt b gene and control region) sequence datasets from Anoa and other related species were analyzed. Molecular phylogeny reconstruction was done using a Bayesian inference model. Calibration points were used to estimate divergence time in Anoa based on a Bayesian phylogeny tree. Estimation of ancestral areas in Anoa ancestors using Reconstruct Ancestral State in Phylogenies was carried out to test the Anoa colonization route. The phylogenetic tree revealed the existence of two groups of Anoa, lowland and mountain Anoa, which diverged during the Middle Pleistocene 1.42 Mya (highest posterior density interval: 1.14–1.71 Mya). The Sundaland area is the most probable (P > 0.7) ancestral area for the Bubalus spp. in the Indonesian archipelago. The most probable Anoa colonization route appears to be through land-bridges, which geological records indicate formed between Sundaland and Sulawesi during the Pleistocene.

Anoa, historical biogeography, insular dwarfisme, Land-bridge hypothesis, mtDNA phylogeny

The genus Bubalus includes several species of wild and domesticated buffaloes, primarily found in Asia and Africa. The most notable species within this genus are the water buffalo (Bubalus bubalis), the Anoas (Bubalus depressicornis and Bubalus quarlesi), and the African buffalo (Syncerus caffer), which is sometimes classified under the same genus due to historical taxonomic confusion (Bibi and Vrba 2010). The Anoas, specifically, are endemic to Indonesia and are recognized for their smaller size compared to their relatives, which is a result of insular dwarfism, a phenomenon where species evolve to be smaller in size on islands due to limited resources and space (Benítez‐López et al. 2020). The taxonomy of Bubalus has undergone revisions, with ongoing debates regarding the classification and relationships among species, particularly between domesticated and wild forms (Bibi and Vrba 2010).

The Anoa, comprising Bubalus quarlesi and Bubalus depressicornis in the subgenus Anoa, have declined by 90% over the past 16 years, and they are even believed to be extinct in some locations. In 2016, the International Union for the Conservation of Nature (IUCN) Red List categorized both the Mountain Anoa (Bubalus quarlesi) and the Lowland Anoa (Bubalus depressicornis) as endangered species, with estimated populations of fewer than 2,500 individuals for each species (Burton et al. 2005). The Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) included it in Appendix I in 2010, which means that Anoa are protected and cannot be traded. Anoa are grouped into lowland and mountain Anoa based on morphology and habitat. The taxonomic ambiguity surrounding the subgenus Anoa, is a significant issue in the field of wildlife conservation and biology (Burton et al. 2005). This ongoing debate highlights the necessity for comprehensive genetic and morphological assessments to clarify the taxonomic status of these endemic species and ensure effective conservation strategies (Tanaka et al. 1996; Pitra et al. 1997; Priyono et al. 2018, 2020, 2022).

Bubalus, a genus of Asiatic bovines within the family Bovidae, includes species that vary in body size. In descending order, from largest to smallest, these species are: Bubalus arnee (Kerr 1792) also known as the wild water buffalo, and its domesticated form, Bubalus bubalis (Linnaeus, 1758), from mainland Asia; tamaraw (Bubalus mindorensis Heude, 1888) from Mindoro Island, Philippines; Bubalus cebuensis (Croft, Heaney, Flynn & Bautista, 2006; extinct) from Cebu Island; Bubalus depressicornis (Smith, 1872), known as the lowland Anoa, from lowland areas of Sulawesi Island; and Bubalus quarlesi (Ouwens, 1910), known as the mountain Anoa (Fig. 1). Average body weight for each species, are provided in Fig. 1 to facilitate comparison, based on previous studies (Rozzi 2017, 2018; Shrestha et al. 2019). This size reduction is often attributed to limited resources, reduced predation pressures, and ecological niches that favor smaller body sizes (Benítez-López et al. 2021). For example, studies have shown that large mammals, such as elephants and hippos, significantly reduce size when isolated on islands (Benítez-López et al. 2021). Other mammals in Sulawesi, such as Stegodon sompoensis (Falconer, 1847) and Elephas celebensis (Hooijer, 1949), have also evolved according to insular dwarfism (Hooijer 1949). In the case of the Anoas, their smaller stature compared to other buffalo species is a direct result of these insular pressures, allowing them to thrive in the unique ecological conditions of Sulawesi (Benítez-López et al. 2021). Tougard (2001) stated that body size gradation, especially in large mammals, correlated with the possibility of their colonization of Sulawesi via Taiwan and the Philippine archipelago.

Figure 1.

Map of Southeast Asia depicting conditions around the Pleistocene period. The solid line shows the colonization route of large mammals according to the land-bridge hypothesis; dotted lines interpret the colonization route according to the insular hypothesis. This study aims to test these routes using mitochondrial DNA phylogeny to better understand the biogeographic history and evolutionary relationships of the Anoa (adapted from Tougard 2001).

Von Koenigswald (1935, 1939) proposed that mammals migrated during the Pleistocene era to Indonesia through the Philippines; his hypothesis was based on affinities between animals found in China and Java. An alternative theory postulates that the most probable colonization route for mammals to Sulawesi was through ancient corridors or land-bridges via Myanmar, Thailand, or Indochina (Bird et al. 2005; De Vos et al. 2007). At present, the relationship between body size gradation and insularism in Bubalus and its connection to potential colonization routes, as suggested by Tougard (2001), remains unknown. However, the specific colonization route of the Anoa to Sulawesi remains uncertain.

Tracing the biogeographic history of the Anoa group and related species is necessary to determine a possible colonization route for Anoa. Many studies have addressed this question using phylogenetic analysis, molecular dating, and reconstruction of ancestral geographic ranges. However, most of these studies focus predominantly on the Bovidae family in general such as Pseudois (Tan et al., 2017). Tan et al. (2017) effectively demonstrated the utility of Reconstruct Ancestral State in Phylogenies (RASP) in reconstructing the biogeographic history of the Pseudois genus within the Bovidae family. By integrating molecular data with advanced biogeographic modeling, the study provided valuable insights into the evolutionary dynamics and historical factors influencing the distribution of these species.

Molecular and particularly mitochondrial DNA data on Bubalus is abundant in GenBank. Mitochondrial DNA is inherited maternally and does not undergo recombination, which allows for clearer lineage tracing and a more straightforward interpretation of phylogenetic relationships. This characteristic makes mtDNA particularly useful for reconstructing colonization histories, as it can serve as a “molecular clock” that reflects historical population dynamics and colonization events more directly than nuclear DNA, which is subject to recombination and more complex inheritance patterns (Hassan et al. 2009; Hassanin et al. 2012; Budakva 2022). Statistical analysis of mitochondrial DNA cytochrome b (cyt b) gene and control region (CR) using Reconstruct Ancestral State in Phylogenies (RASP) has solved numerous biogeographic and historical questions (Banguera-Hinestroza et al. 2014; Lauron et al. 2014; Song et al. 2014; Clementz et al. 2014; Mahmoudi et al. 2017; Tan et al. 2017). The use of RASP is a powerful approach for tracing the biogeographic history of species. RASP allows researchers to reconstruct ancestral geographic ranges based on phylogenetic trees derived from molecular data, particularly mitochondrial DNA, which is advantageous due to its relatively high mutation rate and maternal inheritance patterns Song et al. (2016). This method facilitates the identification of historical colonization routes and the ecological factors influencing species distribution. RASP provides a comprehensive framework for understanding the evolutionary processes that have shaped the distribution of the Anoa and its relatives, thereby contributing valuable insights into the dynamics of insular dwarfism and the ecological adaptations of these unique mammals. The present study aims to test two hypothesized colonization routes (land-bridge and insular routes) and trace the biogeographic history of the Anoa using mitochondrial DNA phylogeny. This includes examining whether the two Anoa species are monophyletic, understanding their evolutionary relationship, and determining the significance of colonization patterns. Investigating these questions provides insight into the evolutionary mechanisms shaping biodiversity on islands and informs conservation strategies for these endangered species.

Studies on insular dwarfism, such as those by Meiri (2007) and Lomolino (2005), emphasize the significance of ecological factors influencing body size evolution in island environments. RASP’s ability to integrate phylogenetic uncertainty into biogeographical reconstructions allows researchers to understand better the dynamics of size evolution in isolated populations (Nylander et al. 2008). This integration is crucial for elucidating the mechanisms behind insular dwarfism, as demonstrated in analyses of various taxa, including lizards and mammals (Meiri 2007; Geer and Lyras 2016).

This study included Boselaphus as an outgroup in phylogenetic tree to clarify the evolutionary relationships within the Bubalus genus and its subgenus Anoa. Including Boselaphus as an outgroup helps better understand the divergence between Anoa and other species of the Bovidae family. Previous studies, such as those by Pitra et al., (1997), suggested a closer relationship between the lowland Anoa and Boselaphus using noncoding regions of aromatase cytochrome P450 and lactoferrin. However, these studies were limited by small sample sizes (only 1–2 Anoa samples). In contrast, our phylogenetic analysis demonstrates that Boselaphus is more distantly related to Bubalus, reinforcing the classification of Anoa as a subgenus within Bubalus, rather than suggesting a closer relationship with Boselaphus. Most morphological and genetic studies seem to agree with our findings, grouping Anoa into the Bubalus genus rather than the Boselaphus one (Schreiber et al. 1993; Kakoi et al. 1994; Kikkawa et al. 1997; Burton et al. 2005; Groves and Grubb 2011; Rozzi 2017). Recent genetic studies, including those by subgenus Kakoi et al. (1994) and Tanaka et al. (1996), have suggested that Anoa is a subgenus of Bubalus.

The phylogenetic tree of Bubalus, based on this mitochondrial DNA, revealed two distinct clades: Clade I, which includes Bubalus arnee and Bubalus cebuensis, and Clade II, which includes Bubalus bubalis and Bubalus mindorensis. This topology supports the classification of Anoa species (Bubalus quarlesi and Bubalus depressicornis) as part of the Bubalus genus, specifically within the subgenus Anoa. The separation of Anoa species in Clade I from B. bubalis and B. mindorensis in Clade II highlights the distinct evolutionary trajectory of Anoa, suggesting a unique biogeographic history compared to other Bubalus species.

The designation of specimens as B. d. quarlesi and B. d. fergusoni across all phylogenetic trees in this study follows the naming conventions used in the GenBank sequence database, which stem from historical inconsistencies in the nomenclature of Anoa. This taxonomic ambiguity, resulting from the morphological similarities and overlapping ecological niches of Bubalus depressions and Bubalus Charles, complicates their classification. The Biological Species Concept, which focuses on reproductive isolation, may not fully apply to these species due to their restricted geographical ranges and potential hybridization in overlapping habitats (Burton et al. 2005; Frantz et al. 2018; Priyono et al. 2018, 2022).

The RASP analysis conducted in this study suggests that the colonization route of Anoa in Sulawesi likely occurred through Sundaland, rather than via the insular route through the Philippines. Previous molecular studies on the Bubalus distribution in Indonesia have also suggested that Bubalus migrated from mainland Asia to Sundaland and then onward to the western part of Indonesia (Sumatra and Java) (Barker et al. 1997; Lau et al. 1998). These studies are in agreement with the fossil record, whereby the oldest water buffalo and other large herbivores in the Southeast Asia archipelago were found within the Sundaland region during the Pleistocene (Pitra et al. 1997; Price and Webb 2006; Louys et al. 2007; Hassanin et al. 2012). During prehistoric time, dry spells were common and they formed a sort of land corridor that connected mainland Asia with the Indonesian Archipelago (Cockrill and Mahadevan 1974; Nachtsheim and Stengel 1977). Our result suggests that Bubalus used a land-bridge corridor from mainland Asia to the Indonesian Archipelago. This result complies with the fossil record, which indicates that the oldest modern water buffalo ancestor, Bubalus palaeokerabau (Dubois, 1908), existed in Cisaat, Java, Sundaland, during ancient times (Louys et al. 2007). At that time, the sea level was lower by 40–50 m, which allowed the Sunda Shelf to connect the Malay Peninsula to Java and Borneo; whereas a decrease of the sea level of 120 m extended the dry land mass to include the Philippines and Sulawesi (Voris 2000; Tougard 2001; Woodruff and Turner 2009). We believe that B. mindorensis, which is endemic to the Philippines, migrated from Sundaland to the Mindoro Island by way of Borneo and Palawan, Philippine Archipelago. The high similarity between mammals in mainland Asia and Palawan indicates that Palawan is an extension of Sundaland (Esselstyn et al. 2010) and supports the colonization of an ancestor of B. mindorensis through the same route.

This shows that the Anoa’s ancestor evolved as an endemic species after colonizing the Sulawesi Island from Sundaland rather than from the Philippines. A confirmation of the land-bridges colonization from the nearest parts of Sundaland to Sulawesi (especially from the west or east) is also found in modern mammals of Sulawesi, such as the babirusa (Thenius 1970; Van Den Bergh 1999; Bergh et al. 2001; De Vos et al. 2007), macaque (Abegg 2002; Blancher et al. 2008), and tarsier (Merker et al. 2009). Moreover, during the latest glacial phase, the Makassar strait separating South Sulawesi and the eastern coast of Borneo was only some 50 km wide (Bergh et al. 2001), allowing corridors from Sundaland to the southern or western part of Sulawesi. Additional fossil evidence includes a prehistoric painting (Rozzi 2017) of what appears to be Anoa, and which was found in Gua Batti, South Sulawesi. In summary, this and previous findings point to the land-bridges hypothesis as the most likely Sundaland-to-Sulawesi colonization route.

Previous studies on genomic age estimates have proposed the divergence time for Bubalus to be between 2 and 1 Mya (Tanaka et al. 1996; Hassanin and Douzery 1999; Hassanin et al. 2012; Bibi 2013; Bibi and Métais 2016), which implies Early Pleistocene as the time for the first island colonization by mainland water buffalo. This hypothesis agrees with the fossil record’s earliest indication of water buffalo and other large herbivores from the Southeast Asia Archipelago on Java during the Pleistocene (Pitra et al. 2004; Louys et al. 2007; Ropiquet et al. 2012). Fossil records indicate as well that the ancestor of the Anoa appeared in Sundaland between Early to Middle Pleistocene (Bergh et al. 2001; Tougard 2001). Other fossil records have revealed that the Anoa is closely related to hemibos from Tatrot and the Pinjor formation of the Siwaliks. Magnetostratigraphic analysis assigned the specimens to a time during the Middle-Late Pleistocene (Azzaroli 1983). Calibration of this time interval was dated to between 3.20 Mya and 1.78 Mya (Spell and McDougall 1992), which is similar to our molecular dating analysis. However, other studies suggest an even longer divergence time for the Anoa, for example 12,4 Mya according to Pitra et al. (1997) or 4,3 Mya according to (Stelbrink et al. 2012). The exact time may be influenced by the distant phylogenetic relationship between Bubalus species and the close relationship to Boselaphus (Pitra et al. 1997) discussed before. In conclusion, our result is supported by previous research which indicates that the Anoa is a plesiomorphic member of the Bovini clade and supports the placement of Anoa as a subgenus within Bubalus, whose ancestor it diverged from around 1.42 Mya. Although our knowledge of the endemic Sulawesi buffalo is hindered by limited sample size, the present study provides important information on the biogeographic history of the Anoa, especially its colonization of Sulawesi.

While mitochondrial DNA (mtDNA) analysis provides valuable insights due to its low mutation rate, this method may limit the resolution of more recent evolutionary changes. Incorporating genomic approaches and expanding the sample size can provide a higher resolution of evolutionary history, enabling more comprehensive genetic diversity and population structure analyses. Integrating genomic data and fossil evidence can refine interpretations of phylogenetic relationships and historical colonizations. In this present study, we tested existing hypotheses for large colonization routes to the Indonesian archipelago using molecular data, with a particular focus on the case of Anoa. Our results support the previous hypothesis that the migration of Anoa followed one of the major migration routes to the Indonesian archipelago, specifically through Sundaland. In the evolutionary history of Anoa, habitat conditions have played a crucial role in preserving their genetic diversity. Mountains and lowland Anoa were separated 1.42 million years ago. They are separated due to fragmentation, which is exacerbated by deforestation pressures, leading to a reduction in genetic diversity (Aninta et al. 2024). Connectivity refers specifically to the ability of Anoa populations to move between fragmented habitats, maintaining gene flow between isolated populations. Sulawesi is an island with a high level of endemism, so by protecting the two Anoa species in Sulawesi, it will also help conserve many other species. Future studies should incorporate fossil records alongside genomic data to build a more complete evolutionary framework for Anoa, providing deeper insights into their historical biogeography.

We would like to thank the Indonesia Secretariat of Scientific Authority for Biodiversity (SKIKH) BRIN for the research recommendation and the Directorate General of Nature Resources and Ecosystem (KSDAE) Indonesia for granting permission for this research. We also thank Arin from Anoa Breeding Center, Manado, Indonesia, and Bontomarannu Education Park, Gowa, Indonesia. Finally, we extend our gratitude to Montana Stone and Peter Keaney (Cornell University) and Gita Alvernita (IPB University) for their valuable assistance in reviewing and proofreading this manuscript.