The orange leaf-nosed bat Rhinonicteris aurantia (Hipposideridae) is a rare Australian endemic with a fragmented distribution, an uncertain taxonomic status of distributional groups and is subject to human disturbance pressure in several areas. Two mitochondrial DNA markers (ND4 and D-loop) were partially sequenced to investigate the similarity of allopatric isolates and the phylogeographic structure of colonies. A clear phylogeographic pattern based on female philopatry emerged from both markers. Limited movement of females within the Pilbara isolate was also inferred, supported by AMOVA, and in agreement with spatial models generated previously using GIS. The position of the small number of Kimberley haplotypes available fitted with the phylogeographic pattern of both markers, but further sampling is required from this population. Two repeat arrays (with 104 and 6 base pair units) were present in the D-loop, which might be useful as markers in future studies. The phylogeographic pattern also coincided with differences between the Pilbara and the more northern group observed previously in the nasal apparatus involved in echolocation, as well as the frequency of the calls themselves. Formal taxonomic reclassification of isolates was not supported by the ND4 fragment marker, but the combination of patterns from all available datasets suggested that the current recognition of the Pilbara population as a separate form for conservation purposes is justified, according to the concept of ‘adaptive evolutionary conservation’. Furthermore, the differences observed in the Pilbara suggested that colonies in the mines of the eastern Pilbara should be managed as a unit distinct from those in the Hamersley Range and Barlee Range according to the precautionary principle until further studies have been conducted.