Flores’ Pleistocene Fauna & the Dwarf Stegodon (van den Bergh et al.)

📖 18 min read~3272 words

When paleontologists excavated the limestone cave of Liang Bua on the island of Flores in the early 2000s, they expected to find hominin bones. What they also uncovered was a complete ecological portrait — a Pleistocene faunal community dominated by dwarfed elephants, giant rats, Komodo dragons, and eventually two distinct species of human. The fossil sequence, documented in meticulous detail by van den Bergh and colleagues, has since become one of the most important datasets in island biogeography and directly informs ongoing debates about why Varanus komodoensis grew as large as it did. This is an original editorial summary by the Komodo Guide science team, pointing readers to the primary source for full data and methodology.

Paper Overview

Van den Bergh and a ten-author international team published this paper in late 2009 (online December 2008) as part of a landmark special issue of the Journal of Human Evolution dedicated to the Liang Bua excavations. The full citation is: van den Bergh, G.D., Meijer, H.J.M., Due Awe, R., Morwood, M.J., Szabó, K., van den Hoek Ostende, L.W., Sutikna, T., Saptomo, E.W., Piper, P.J., & Dobney, K.M. (2009). "The Liang Bua faunal remains: a 95 k.yr. sequence from Flores, East Indonesia." Journal of Human Evolution, 57(5), 527–537. DOI: 10.1016/j.jhevol.2008.08.015.

The paper reports on systematically excavated vertebrate, bird, and invertebrate remains recovered from a sequence of sediment layers — each independently dated by luminescence and U-series methods — spanning approximately 95,000 years. Unlike many paleontological studies that present isolated finds, the Liang Bua dataset constitutes a near-continuous ecological record. It lets researchers ask not just what animals lived on Flores, but when they arrived, how their body sizes shifted over time, and what caused the dramatic extinction event that terminated the Pleistocene community around 17,000–12,000 years ago.

For Komodo dragon research specifically, the paper is invaluable: it documents Varanus komodoensis bones throughout the entire 95,000-year sequence, confirming the monitor lizard as a persistent apex predator in the same ecosystem as multiple generations of dwarfed Stegodon. This temporal overlap is the empirical backbone of the prey-driven gigantism hypothesis — the idea that Komodo dragons grew large precisely because large prey was reliably available over geological timescales.

The Dwarf Stegodon: Island Evolution in Action

The proboscidean fossil record from Flores is among the most compelling illustrations of insular dwarfism known from the Quaternary. The island was colonized at least twice by distinct stegodon lineages, each subsequently evolving toward reduced body size. The Liang Bua sequence captures the terminal phase of this process: Stegodon florensis insularis, the dwarf subspecies that co-existed with Homo floresiensis and V. komodoensis during the Late Pleistocene.

Earlier Flores sites — particularly those in the So'a Basin studied by van den Bergh and colleagues in complementary publications — document the full trajectory. Ancestors reaching the island may have weighed several tonnes; their insular descendants were reduced to the size of a large water buffalo, or approximately 300–500 kg. This is not merely smaller than their continental relatives: it represents a reduction of 70–80% of ancestral body mass within the Pleistocene timeframe, one of the most dramatic documented size-reduction events in Proboscidea.

Van den Bergh's team attributes this dwarfing to the classic island-rule mechanisms: limited food resources on a small landmass, reduced predation pressure on juvenile and adult body sizes (at least from large felids and canids, which never colonized Flores), and potentially accelerated generation times. The result was a population of pygmy proboscideans that, paradoxically, became the most abundant large-bodied prey item available to the island's apex predator — V. komodoensis.

Stegodon lineages documented from Flores (simplified from van den Bergh et al. and related publications)

Taxon	Age (approx.)	Estimated body mass	Key site
Stegodon sondaari	>900 kya	~200–300 kg (pygmy)	Tangi Talo, So'a Basin
Stegodon florensis	~840–700 kya	~1,000–2,000 kg	Mata Menge, So'a Basin
Stegodon florensis insularis	95–17 kya	~300–500 kg	Liang Bua

The Pleistocene Community at Liang Bua

The faunal assemblage recovered from Liang Bua is strikingly impoverished by mainland standards — a hallmark of oceanic island communities that lack continuous land bridges to continental source populations. Van den Bergh and colleagues identified the following major components of the Pleistocene Flores community documented in the 95,000-year sequence:

Varanus komodoensis (Komodo dragon): Present throughout the entire sequence, from the oldest layers to approximately 12,000 years ago. The authors note that Komodo dragon remains occur alongside Stegodon bones that show cut marks, implying that H. floresiensis was processing carcasses that both hominins and monitors competed over. This co-occurrence of large monitor lizard, dwarf elephant, and a small-bodied hominin in the same taphonomic context is unique in the Pleistocene fossil record.

Stegodon florensis insularis: The most numerous large-mammal taxon in the sequence. Dental, cranial, and postcranial remains confirm substantial dwarfing relative to earlier Flores stegodonts. Bite marks on Stegodon long bones are consistent with both hominin butchery and carnivore ravaging — the latter plausibly attributable to V. komodoensis.

Giant murids: Rodents of unusual size, including species in the Hooijeromys lineage, exemplify the reverse pattern to stegodon — small-bodied colonizers that grew larger on the island, a phenomenon known as island gigantism. These rats likely formed a secondary prey resource for both V. komodoensis and avian predators.

Large avifauna: The assemblage includes remains of storks (Leptoptilos), marabou-like birds capable of consuming carrion alongside monitors, and small corvid-sized species. Their presence indicates a diverse scavenging guild competing with V. komodoensis at carcasses.

Freshwater turtles and bats: Completing the community, chelonian and chiropteran remains confirm a diverse vertebrate fauna well-adapted to the humid cave environment and surrounding forest-savanna mosaic inferred from palynological data.

The Terminal Pleistocene Extinction Event

One of the most striking findings reported by van den Bergh et al. is the sharp faunal discontinuity visible in the Liang Bua sedimentary sequence at approximately 17,000–12,000 years ago. Below this horizon, the assemblage is dominated by Stegodon florensis insularis, V. komodoensis, and H. floresiensis. Above it, modern humans appear, H. floresiensis disappears, Stegodon disappears, and the large Komodo dragon also vanishes from the Liang Bua record — though the species survives on Flores to the present day in reduced range, and currently persists on Komodo, Rinca, Gili Motang, and Padar.

The cause of the extinction event remains debated. Van den Bergh and colleagues note that a major volcanic eruption, evidenced by a tephra layer in the sequence, coincides roughly with the disappearance of the Pleistocene fauna. This suggests that volcanic activity may have devastated the local habitat, eliminating the small Stegodon populations on which the entire predator community depended. The subsequent arrival of modern humans — with their more sophisticated tool technologies and hunting strategies — may have administered the final blow to already-stressed remnant populations.

For Komodo dragon conservation, this extinction episode is directly relevant. It demonstrates that V. komodoensis populations can collapse regionally in response to prey base collapse, whether driven by volcanism, hunting pressure, or habitat loss. The species' current restriction to a handful of small islands in the Lesser Sundas is itself consistent with a long-term contraction from a formerly wider range that once extended across Java and possibly Timor.

What the Fossil Record Means for Prey-Driven Gigantism

The Liang Bua faunal sequence bears directly on one of the most contested questions in Komodo dragon biology: did giant body size in V. komodoensis evolve because large prey was available, or despite the absence of large prey? Two competing frameworks have dominated this debate:

The prey-driven gigantism hypothesis holds that selection for larger body size in ancestral varanid populations was driven by access to large-bodied prey — specifically, megafaunal species such as dwarf elephants, pygmy hippos, and giant extinct bovids that colonized island Southeast Asia during the Pleistocene. Larger monitors could subdue and consume larger prey, giving them a competitive advantage that drove directional selection toward increased body mass over many generations.

The competitive-release hypothesis, associated with work by Hocknull and colleagues (see the companion review at /research/hocknull-megalania-origins-2009/), proposes instead that large body size evolved on the Australian continent in a very different ecological context, and that the species subsequently colonized island Southeast Asia, where it encountered dwarf megafauna as an already-large predator.

The Liang Bua data do not definitively resolve this debate, but they provide crucial constraints. The 95,000-year co-occurrence of V. komodoensis and Stegodon florensis insularis demonstrates that a stable predator-prey relationship between giant monitor and dwarf elephant was ecologically viable on Flores for at least the duration of the Late Pleistocene. Whether that relationship generated the gigantism or merely maintained it requires examination of older fossil sites — including the So'a Basin localities where Stegodon florensis (the larger, earlier taxon) and V. komodoensis overlap in deeper time. Van den Bergh's broader research program, of which the Liang Bua paper is one component, continues to build that deeper record.

Note that the related hypothesis about Jared Diamond's pygmy-elephant prey model is discussed separately at /research/diamond-pygmy-elephant-1987/, and the island gigantism framework for body size evolution is treated at /komodo-dragon/island-gigantism/. This page focuses specifically on the Flores fossil evidence.

Myths vs Facts

Common Misconception	What the Fossil Evidence Shows
Komodo dragons always lived only on Komodo island.	Fossil evidence confirms V. komodoensis occupied Flores, Java, and possibly Timor during the Pleistocene; the current range is a post-extinction remnant.
Stegodon on Flores were full-sized elephants equivalent to modern African elephants.	Late Pleistocene Flores stegodonts (S. florensis insularis) had undergone 70–80% body mass reduction from mainland ancestors, comparable in size to large cattle.
The Pleistocene extinction on Flores was caused solely by human hunting.	A volcanic tephra layer at the extinction horizon implicates volcanic disruption; modern human arrival and hunting pressure may have compounded the collapse of already-stressed populations.
Komodo dragons and Homo floresiensis never interacted.	Both species are recovered from the same Liang Bua sediment layers. Cut-marked Stegodon bones and dragon skeletal elements at the same taphonomic level indicate they competed for the same prey and carcasses.
Island dwarfism only reduces body size gradually over millions of years.	The Flores stegodon record shows substantial dwarfing occurring within Pleistocene time spans, with multiple colonization events each producing rapid size reduction.
The Flores fauna was rich and diverse like a mainland tropical community.	Oceanic island isolation produced a highly impoverished fauna with very few mammal species, dominated by murids, a single proboscidean lineage, and one large reptilian predator.

Key Takeaways

• The Liang Bua faunal sequence is the most complete Pleistocene ecological record from Flores. It spans 95,000 years and documents the co-existence of V. komodoensis, Stegodon florensis insularis, and Homo floresiensis in a single, well-dated stratigraphic context.
• Stegodon dwarfing on Flores was extreme and rapid. Multiple colonization events each produced proboscideans reduced to a fraction of mainland body mass, driven by the classic island-rule mechanisms of resource limitation and altered predation regimes.
• Komodo dragons were persistent apex predators in this impoverished island ecosystem. Their presence across the full 95,000-year sequence, alongside dwarf elephants as the only large mammalian prey, is the strongest fossil evidence for a long-term predator-prey relationship relevant to gigantism hypotheses.
• The terminal Pleistocene extinction was abrupt and probably multi-causal. Volcanic disruption and the arrival of modern humans together eliminated both the dwarf elephant prey base and the hominin competitor, causing V. komodoensis to retreat to its current restricted range.
• This paper does not resolve the gigantism debate, but narrows it. It shows the predator-prey relationship was ecologically stable for tens of thousands of years; whether it was causative of gigantism requires deeper fossil evidence from earlier Flores and Java sites.

Frequently Asked Questions

What is Liang Bua and why is it important?

Liang Bua is a large limestone cave near the town of Ruteng in western Flores, Indonesia. It became world-famous in 2004 when excavations led by Mike Morwood and Raden Awe Due yielded skeletal remains of Homo floresiensis — the so-called "Hobbit" — a diminutive hominin species that survived on Flores until at least 60,000 years ago according to revised chronologies. The van den Bergh et al. (2009) faunal paper is the companion study that documents all the other animals living alongside H. floresiensis, providing the ecological context for understanding the Hobbit's world.

Were Komodo dragons responsible for accumulating the Stegodon bones at Liang Bua?

The taphonomic evidence is mixed. Many Stegodon bones in the Liang Bua deposit show anthropogenic cut marks consistent with stone-tool butchery by H. floresiensis. Van den Bergh and colleagues conclude that hominins were the primary bone accumulators in many layers. However, dragon gnawing marks are also present on some specimens, consistent with V. komodoensis scavenging carcasses already processed by hominins — a pattern still observable in Komodo National Park today.

How does Stegodon florensis insularis compare in size to living elephants?

Adult Stegodon florensis insularis individuals are estimated to have weighed roughly 300–500 kg, comparable to a large water buffalo or a small horse in terms of biomass available to a predator. This contrasts with the African bush elephant (Loxodonta africana) at 4,000–6,000 kg, and even the smaller Asian elephant (Elephas maximus) at 2,700–5,000 kg. The Flores dwarf stegodon was, in essence, a medium-large ungulate-equivalent in terms of prey value for a large carnivore.

Did Komodo dragons ever hunt Homo floresiensis?

There is no direct fossil evidence of V. komodoensis predation on H. floresiensis. Given the small body mass of the hobbits — estimated at roughly 25–30 kg for adults — they would have been within the size range of prey that modern Komodo dragons occasionally attack. However, the archaeological evidence from Liang Bua suggests hominins were capable tool-users who processed large prey, implying they were not simply passive prey items. The relationship between the two species remains an open question.

What happened to the Stegodon on Flores, and why don't they survive today?

The entire Flores stegodon lineage became extinct at roughly 12,000–17,000 years ago, coinciding with a volcanic tephra layer in the Liang Bua sequence and the arrival of modern humans. Proboscideans are particularly vulnerable to extinction because of their low reproductive rates, long gestation periods, and dependence on large tracts of intact habitat. On a small island already experiencing volcanic disruption, even modest hunting pressure from modern humans — with their projectile weapons and coordinated hunting strategies — could have driven remnant populations below viable thresholds within a few generations.

Why don't Komodo dragons live on Flores today in significant numbers?

Modern V. komodoensis populations on Flores are small and restricted primarily to Wae Wuul Nature Reserve on the western tip. The primary driver of their decline on Flores is human competition, habitat conversion to agriculture, and loss of prey species such as deer and pigs to hunting pressure. The fossil record suggests that after the extinction of Stegodon, the Komodo dragon was deprived of its dominant prey species and subsequently contracted to areas where alternative large prey (principally introduced ungulates and the Rusa deer Cervus timorensis) remained available — most successfully on the smaller, less-human-impacted islands of Komodo and Rinca.

How does van den Bergh's Liang Bua work relate to his So'a Basin research?

Liang Bua documents the final 95,000 years of the Flores faunal sequence. The So'a Basin sites — Mata Menge, Tangi Talo, Boa Leza, and others — extend the record back to approximately 840,000–1,000,000 years ago, documenting the early and middle Pleistocene stegodon communities alongside evidence of hominin activity (stone tools). Together, van den Bergh's research across both regions provides a near-complete picture of Flores faunal evolution from the initial Pleistocene colonization through the terminal extinction, spanning roughly one million years.

Is the prey-driven gigantism hypothesis the scientific consensus?

Not definitively. It remains a well-supported hypothesis, not a settled conclusion. The Flores fossil evidence is consistent with prey-driven gigantism — large monitors and dwarf proboscideans co-existed for geologically long periods — but the Australian fossil evidence discussed by Hocknull and colleagues (see /research/hocknull-megalania-origins-2009/) suggests that the ancestor of V. komodoensis was already a large-bodied predator before colonizing the island chain. Current evidence is best interpreted as consistent with gigantism being maintained and possibly reinforced by island prey dynamics, with the initial evolution of large body size possibly predating arrival in island Southeast Asia.

Sources & Further Reading

1. van den Bergh, G.D., Meijer, H.J.M., Due Awe, R., Morwood, M.J., Szabó, K., van den Hoek Ostende, L.W., Sutikna, T., Saptomo, E.W., Piper, P.J., & Dobney, K.M. (2009). "The Liang Bua faunal remains: a 95 k.yr. sequence from Flores, East Indonesia." Journal of Human Evolution, 57(5), 527–537. https://doi.org/10.1016/j.jhevol.2008.08.015
2. Brown, P., Sutikna, T., Morwood, M.J., et al. (2004). "A new small-bodied hominin from the Late Pleistocene of Flores, Indonesia." Nature, 431, 1055–1061. https://doi.org/10.1038/nature02999 — Original description of Homo floresiensis from the same excavation.
3. van den Bergh, G.D., Kurniawan, I., et al. (2009). Palaeontology and Archaeology of the So'a Basin, Central Flores, Indonesia. Edited by F. Aziz, M.J. Morwood, & G.D. van den Bergh. Bandung: Geological Survey Institute. — Companion volume documenting the older, Early–Middle Pleistocene Flores fauna.
4. Hocknull, S.A., Piper, P.J., van den Bergh, G.D., et al. (2009). "Dragon's paradise lost: palaeobiogeography, evolution and extinction of the largest-ever terrestrial lizards (Varanidae)." PLoS ONE, 4(9), e7241. https://doi.org/10.1371/journal.pone.0007241 — Australian fossil context for Komodo dragon origins; reviewed at /research/hocknull-megalania-origins-2009/.
5. Sutikna, T., Tocheri, M.W., Morwood, M.J., et al. (2016). "Revised stratigraphy and chronology for Homo floresiensis at Liang Bua in Indonesia." Nature, 532, 366–369. https://doi.org/10.1038/nature17179 — Revised dating placing H. floresiensis extinction at ~50,000 BP.
6. van den Bergh, G.D., et al. (2016). "Homo floresiensis-like fossils from the early Middle Pleistocene of Flores." Nature, 534, 245–248. https://doi.org/10.1038/nature17999 — Extends the hominin record from Flores to 700,000 years ago.
7. Diamond, J. (1987). "Did Komodo dragons evolve to eat pygmy elephants?" Nature, 326, 832. — The original prey-gigantism hypothesis; reviewed at /research/diamond-pygmy-elephant-1987/.
8. Turvey, S.T., et al. (2017). "Quaternary vertebrate faunas from Sulawesi, Indonesia: documenting Wallace's Line in the fossil record." Proceedings of the Royal Society B, 284, 20170267. https://doi.org/10.1098/rspb.2017.0267 — Comparative context for island faunal evolution across Wallacea.