The natural history, non-invasive sampling, activity patterns and population genetic structure of the Bornean banteng (Bos javanicus lowi) in Sabah, Malaysian Borneo


Counterpart: Dr. L. Ambu/Sabah Wildlife Department

Collaborators: Datuk S. Mannan (Sabah Forestry Department), Dr. M. Ancrenaz (HUTAN), MaluaBioBank & New Forests, FieldSkills.

Status: Completed


The Bornean banteng is an endangered wild bovid that is endemic to the island of Borneo. Following the extinction of the Sumatran rhino in the wild, the Bornean banteng is now the rarest large mammal in Sabah, which is believed to be their last stronghold and may support a declining population of about 500 individuals. Their decline has been brought about over the past few years by deforestation and habitat loss, conversion to oil palm and timber plantations, and widespread hunting using firearms in forest areas that were previously difficult to access. Due to such detrimental habitat disturbances, this shy cattle species is highly elusive and rarely seen; they prefer to spend their time in secluded forest and grassland habitats where disturbance is minimal. Due to this behaviour, they are a highly challenging species to study, and traditional survey techniques are largely unsuitable or yield poor results. This is coupled with a lack of baseline data or an awareness of their historical distribution. Prior to 2011, no extensive quantitative surveys of the Bornean banteng had been undertaken in Sabah, and little information was available on their ecology to underpin their conservation and management.

Research Objectives

  1. To describe the natural history of banteng.
  2. Identify suitable and effective non-invasive survey methods that are appropriate for long-term monitoring and estimating population parameters.
  3. To characterise activity patterns and identify the effect of habitat disturbance upon activity budgets and site use.
  4. To investigate the population genetic structure of banteng


A comprehensive account of the natural history of all three subspecies of banteng was compiled through collaborations with stakeholders and researchers to help identify the gaps in knowledge and to highlight future research topics.

A comparison of non-invasive survey methods (signs and camera traps) for locating and monitoring banteng subpopulations was conducted in two forests: Tabin Wildlife Reserve and Malua Forest Reserve. Individual bantengs were identified using naturally-evolving tags such as scars because it was impossible to observe or capture/tag individuals directly. As this method can incur errors, three levels of identification were created to simulate the impact of this error on population size estimates. A comparison of survey methods was conducted using Generalised Estimation Equations fitted using a Generalised Linear Model (GEEGLM) due to the nature of the data. The population size was estimated using species accumulation curves and non-parametric estimators because the data was unsuitable for capture-recapture modelling.

Activity budgets were estimated from photos of banteng captured using remote infrared camera traps. Habitat type, canopy cover and site type were classified in the field based on observations and measures of percentage leaf cover. The ambient temperatures of each forest were recorded using the camera traps when the banteng were captured. Activity patterns were bootstrapped due to the limited data available on this low-density species, and environmental effects upon activity were estimated using correlations.

Markers of mitochondria (mtDNA) were designed and mtDNA was amplified to identify the haplotype diversity of subpopulations, the distribution of haplotypes and potential division of management units. Samples were sexed using markers of the Y-chromosome. Microsatellite markers designed by the FAO guidelines were used to amplify nuclear DNA to establish the genotyping errors that are incurred when using poor quality faecal DNA. Neighbour-joining analysis and Bayesian likelihood analysis were used to create phylogenies and estimate the ancestral lineage of the Bornean banteng in relation to other banteng subspecies and to other ungulates. Non-invasively collected DNA was used due to the inability to obtain tissue samples of banteng.


The natural history account combined data from multiple researchers in Thailand, Cambodia, Vietnam, Laos, Indonesia (Java, Kalimantan), Malaysia (Sabah and Sarawak), and Brunei Darussalam. This information-sharing encouraged the networking of banteng researchers and the exchange of research ideas. This account was published in a scientific textbook by Cambridge University Press in 2014 entitled ‘Ecology, Evolution and Behaviour of Wild Cattle: implications for Conservation’ and was edited by M. Melletti and J. Burton. It will increase the awareness and understanding of the banteng within the scientific community, draw attention to the research topics needed to fill the gaps in our knowledge, and provide a textbook for biology students and cattle enthusiasts.

We now understand the banteng herd locations within two forest reserves, the population sizes within restricted study areas inside each forest, and the bantengs’ use of the habitat. The most efficient non-invasive survey method was camera traps, which were not unduly affected by the environmental conditions. Banteng were captured primarily during the first 40 survey days using 60 camera trap stations, suggesting surveys can be shorter. The population size within our study areas within Tabin was 18-53 individuals in an area of 30.5 km2 and 16-20 individuals in an area of 27 km2 in Malua.

Activity patterns of banteng were found to vary according to the structure of the forest (i.e., the regeneration time since the last logging episode), whereby they adapt the duration of their behaviours and the type of behaviour they perform according to the ambient temperature. When temperatures are high, the banteng switches to a forest habitat with a dense canopy and performs less energetic activities to mitigate thermal stress, and during cool hours, they take advantage of secluded grassland areas and perform more energy-demanding activities like foraging. Illegal activity occurred within key banteng areas, therefore increasing the potential risk of human-wildlife conflict. In addition to this, illegal activity was found to be prevalent along old logging roads. Therefore, timber harvesting facilitates hunting even years after the cessation of harvesting activities. This information has helped to inform wildlife patrols by the local government agencies and will, in turn, help the conservation of the banteng populations.

A total of six haplotypes were identified, with two common haplotypes found across four forest reserves and four rare haplotypes restricted to two forest reserves. The rare haplotypes were located in a habitat where hunting is commonplace. Therefore, these locations should be priority areas for conservation, patrols, and enforcement. A median-joining network revealed a recent population expansion and divergence from the gaur (Bos Taurus) between 217 to 634 thousand years ago when sea levels were low during the last glacial maximum.

The PhD thesis concluded with recommendations for mitigating population decline and loss of genetic diversity, which include better protection of herds that are vulnerable to eradication by hunting through increasing wildlife security. The data arising from this PhD have informed an update of the IUCN Red List banteng account (currently in review) and will also inform the first Population and Habitat Viability Analysis and the drafting of the first Action Plan for banteng in Sabah, which will take place in 2017.