What did the “Beast of Huli" eat? The diet of this bizarre extinct marsupial, which had similarities to both giant wombats and pandas, has remained a mystery. PhD student Joshua White, with coauthors, examined the teeth of this species to help answer these questions.

More than 30 years ago, the Australian Museum’s Professor Tim Flannery discovered a bizarre extinct marsupial from the remote highlands of Papua New Guinea. He named it Hulitherium, meaning “Beast of Huli” after the indigenous people who lived in the area where the remains of this unusual beast were discovered. Hulitherium belonged to an extinct group of wombat-like marsupials, called Diprotodontidae, which roamed the rainforests of New Guinea during the Pleistocene period1. Like most fossil discoveries, there wasn’t an abundance of remains from this animal, but the fossils described include various teeth, skull fragments, forelimbs and hindlimbs. Fortunately, the remains discovered were enough to identify this fossil as a new and unique genus and species, Hulitherium tomasettii.

Artist’s reconstruction of Hulitherium tomasettii. Figure 8 from Flannery and Plane (1986).

Artist’s reconstruction of Hulitherium tomasettii. Figure 8 from Flannery and Plane (1986).

Image: Peter Schouten
© Peter Schouten

The family Diprotodontidae were large quadrupedal, herbivorous marsupials of Australia and New Guinea that looked perhaps like overgrown or giant wombats. The largest of which was Diprotodon optatum of Australia, weighing in at approximately 3 tonnes and 1.6 metres tall at the shoulders, the largest marsupial that ever lived. The fossil record of New Guinea includes a variety of diprotodontids, classified in the genera Kolopsis, Kolopsoides, Maokopia, and Hulitherium. When originally described, Hulitherium was proposed to be a possible bamboo feeder because it shared similar traits to the giant panda (Ailuropoda melanoleuca), including its post-cranial morphology, weight, and habitat1. Based on its joints, Hulitherium was also capable of rearing up on its hind legs, enabling it to grasp at food that was out of reach for other herbivores.

To test whether Hulitherium was a bamboo specialist feeder, my research included two different methods to evaluate the diet of Hulitherium - and included assessing the fossil molars of the species, in the Australian Museum Palaeontology Collection. One of these methods is called dental complexity analysis, which is a type of topographical analysis used to map out teeth and assess how many features are available to grind plant material2. In our recent study, my co-authors and I applied this principle to the chewing surfaces of molars, and measured regions of low areas and high areas to calculate how ‘complex’ the chewing surfaces are2. The more complex the chewing surfaces of the tooth are, the more likely that the organisms consumed fibrous vegetation because processing this type of food required sophisticated molars. Previous studies that used this same method indicated living species that consumed bamboo have a much higher level of molar complexity than other diet categories (herbivores, omnivores and carnivores)3. Our results show that Hulitherium did not possess the same complexity as modern-day bamboo feeding specialists, such as the giant panda, therefore we concluded that it most likely did not primarily consume tough fibrous vegetation like bamboo.

Dental complexity of the lower third molar of various diprotodontids. Compared to other diprotodontids, Hulitherium has fewer patches indicating that the bilophodont molars did not evolve to break down fibrous vegetation such as bamboo.

Dental complexity of the lower third molar of various diprotodontids. Compared to other diprotodontids, Hulitherium has fewer patches indicating that the bilophodont molars did not evolve to break down fibrous vegetation such as bamboo.

Image: White et al. 2021.
© White et al. 2021.

The second method we applied was dental microwear texture analysis (DMTA), which examines the scratches caused by chewing food and tooth-on-tooth interactions. DMTA has previously been applied to studying other species of diprotodontids, including Diprotodon and Zygomaturus4. We examined the scratches on the molar rows of Hulitherium specimens and compared these to other extinct diprotodontids from previous studies; we also compared them to bamboo feeding taxa, including the giant panda and the Andean bear (Tremarctos ornatus). Our results demonstrated that microwear seen on the teeth of Hulitherium is completely different to that seen in the giant panda.

Lower molar row dental complexity maps

Lower molar row dental complexity maps of living bamboo feeding taxa; greater bamboo lemur (Prolemur simus), giant panda (Ailuropoda melanoleuca), red panda (Ailurus fulgens) and non-bamboo feeding taxa; black lemur (Eulemur macacao), brown bear (Ursus arctos), raccoon (Procyon lotor) and the diprotodontid; Hulitherium. All data from living taxa are from Eronen et al. (2017). This graph demonstrates how simple the molar row of Hulitherium is compared to bamboo-feeding taxa and how specialised their teeth need to be in order to process the tough fibrous nature of bamboo.

Image: White et al 2021.
© White et al 2021.

Using a combination of dental complexity and dental microwear analysis, we were able to reconstruct the diet of this ancient and misunderstood giant wombat. Our research shows that Hulitherium most likely did not eat bamboo like giant pandas, and instead likely fed on softer plants like grasses and leaves. This study also shows the importance of applying quantitative methods to test for assumed convergence between distantly related taxonomic groups, such as marsupials and placentals.

Joshua White, PhD candidate, Australian National University and Australian Museum Research Institute.

White, J.M., DeSantis, L.R., Evans, A.R., Wilson, L.A. and McCurry, M.R., 2021. A panda-like diprotodontid? Assessing the diet of Hulitherium tomasettii using dental complexity (Orientation Patch Count Rotated) and dental microwear texture analysis. Palaeogeography, Palaeoclimatology, Palaeoecology, p.110675.


  1. Flannery, T.F. and Plane, M., 1986. A new Late Pleistocene diprotodontid (Marsupialia) from Pureni, Southern Highlands Province, Papua New Guinea. BMR Journal of Australian Geology & Geophysics, 10(1), pp.65-76.
  2. Evans, A.R., Wilson, G.P., Fortelius, M. and Jernvall, J., 2007. High-level similarity of dentitions in carnivorans and rodents. Nature, 445(7123), pp.78-81. https://www.nature.com/articles/nature05433
  3. Eronen, J.T., Zohdy, S., Evans, A.R., Tecot, S.R., Wright, P.C. and Jernvall, J., 2017. Feeding ecology and morphology make a bamboo specialist vulnerable to climate change. Current Biology, 27(21), pp.3384-3389. https://doi.org/10.1016/j.cub.2017.09.050
  4. DeSantis, L.R., Field, J.H., Wroe, S. and Dodson, J.R., 2017. Dietary responses of Sahul (Pleistocene Australia–New Guinea) megafauna to climate and environmental change. Paleobiology, 43(2), pp.181-195. https://doi.org/10.1017/pab.2016.50