📖 17 min read~3090 words
This is an original editorial summary prepared by the Komodo Guide team; readers seeking the primary data should consult the published article directly. In 2013, a research group led by clinical microbiologist Ellie J. C. Goldstein subjected the oral cavities of captive Varanus komodoensis to the most thorough bacterial culture protocol yet applied to this species — culturing for both aerobic and anaerobic organisms, not just the easily grown aerobes that earlier surveys had examined. The result was methodologically decisive: the dragons' mouths harboured ordinary carnivore flora, not a uniquely deadly bacterial arsenal, and no virulent pathogen capable of causing rapid lethal infection was recovered from any animal. That single finding, reported as a direct challenge to the "bacteria as venom" model, aligned squarely with what Fry et al. had argued from a biochemical direction in 2009 — that the real killing mechanism is venom, not microbes.
Paper Overview
The full citation for the article reviewed here is: Goldstein, E. J. C., Tyrrell, K. L., Citron, D. M., Cox, C. R., Recchio, I. M., Okimoto, B., Bryja, J., & Fry, B. G. (2013). "Anaerobic and aerobic bacteriology of the saliva and gingiva from 16 captive Komodo dragons (Varanus komodoensis): new implications for the 'bacteria as venom' model." Journal of Zoo and Wildlife Medicine, 44(2), 262–272. DOI: 10.1638/2012-0022R.1. PMID: 23805543.
The research team was led from the R. M. Alden Research Laboratory in Culver City, California, and notably included Bryan G. Fry — the co-author of the landmark 2009 venom paper — as a senior contributor. That combination of clinical microbiology expertise and herpetological toxinology gave the study unusual breadth. The team's explicit goal was to test whether the oral microbial community of captive V. komodoensis was sufficiently pathogenic to support the decades-old claim that bacteria-laden saliva functions as a lethal injection for prey.
| Item | Detail |
|---|---|
| Journal | Journal of Zoo and Wildlife Medicine |
| Year & Volume | 2013; Vol. 44, Issue 2, pp. 262–272 |
| Lead author affiliation | R. M. Alden Research Laboratory, Culver City, CA, USA |
| Animals sampled | 16 captive dragons — 10 adults (aged 2–17 yr), 6 neonates (aged 7–10 days) — across three U.S. zoos |
| Culture types | Both aerobic and anaerobic (anaerobic transport media, reference-laboratory processing) |
| Total isolates (adults) | 128 isolates representing 39 aerobic and 21 anaerobic species |
| Virulent species found | None |
| DOI | 10.1638/2012-0022R.1 |
Study Design & Methods
Sixteen captive animals were enrolled from three zoological institutions in Los Angeles, Honolulu, and Houston. Zoo veterinary staff collected swabs from both the saliva and the gingival margin of each animal and placed them into anaerobic transport media — a critical step that preserved oxygen-sensitive organisms during courier shipment to the reference laboratory. Earlier culture surveys of dragon oral flora had relied exclusively on aerobic methods, meaning obligately anaerobic bacteria thriving in gingival pockets would have been entirely invisible. The Goldstein team designed the protocol explicitly to close that gap.
At the laboratory, isolates were identified by standard microbiological methods and 16S ribosomal RNA gene sequencing for strains not assignable by biochemical profiling alone. Results were tabulated separately for adults and neonates, allowing the team to test whether age and captive environment shaped the microbial community.
Why Anaerobes Matter
Anaerobic bacteria are of particular interest in wound-infection biology because they dominate deep tissue infections, bite-wound abscesses, and gangrenous processes in mammals. If any microorganism in a dragon's mouth were to function as a genuine lethal agent, anaerobes would be the prime candidates. The inclusion of anaerobic culture in Goldstein's protocol was therefore not a minor methodological refinement — it was the study's defining contribution over all prior surveys.
What the Cultures Found
In the adult animals, 128 isolates were recovered and assigned to 60 distinct taxa: 39 aerobic and 21 anaerobic species. The dominant aerobic organisms were Staphylococcus sciuri and Enterococcus faecalis, both recovered from the majority of adults (8–9 of 10 animals, depending on the species). Gram-negative Enterobacteriaceae — a bacterial family broadly distributed in the gut contents and tissues of vertebrate prey animals — made up much of the remaining aerobic diversity. The anaerobic fraction included clostridia and other obligate anaerobes typical of gastrointestinal and environmental origin.
The neonate animals presented a strikingly different picture: only aerobic bacteria were recovered from the six hatchlings, with no anaerobes detected at all. This developmental disparity is consistent with what microbiologists observe in other reptiles and mammals: the anaerobic component of oral flora is acquired progressively with age and dietary history, not inherited or established at hatching. In the context of the bacteria-as-venom hypothesis, the absence of anaerobes in neonates is particularly telling — if bacterial lethality were a biological adaptation shaped by natural selection, one would expect it to appear at some developmental stage regardless of the animal's feeding history.
The researchers were emphatic on one point: across all 16 animals and both sampling sites, no virulent species were isolated. The organisms present were attributable to the skin and gut flora of the animals' recent meals and to the captive environment generally. Where pathogens of any note appeared in earlier literature — species such as Pasteurella multocida, which had sometimes been cited as evidence for dragon dangerousness — the Goldstein team found these could be explained as transient contaminants reflecting what the dragons had recently eaten, not a stable, cultivated oral microbiome.
Implications for the Bacteria-as-Venom Model
The "bacteria as venom" framework rested on a chain of reasoning that Goldstein et al. systematically undermined at each link. The chain was: (1) dragons carry uniquely dangerous bacteria; (2) these bacteria are stable residents of the oral cavity, not transient passengers; (3) transmission to bitten prey initiates a fatal infection within the observed time frame of prey collapse. The 2013 study provided evidence against all three premises.
On point one, no uniquely dangerous species were found. The flora recovered — staphylococci, enterococci, enterobacteria, and environmental anaerobes — represents the kind of mixed community that any omnivorous or carnivorous animal accumulates on its mucosal surfaces after handling food. On point two, the variation across the three zoos and between adults and neonates suggests a community shaped by local environment and feeding history rather than by stable selective pressure toward pathogenicity. On point three, the authors judged the recovered community "unlikely to cause rapid fatal infection," a conclusion consistent with what veterinary wound-infection specialists know about the time course of serious bacterial sepsis in large mammals.
Goldstein and colleagues framed their conclusion explicitly as a challenge to the bacteria-as-venom model and as support for the venom hypothesis advanced by Fry et al. (2009). That earlier paper had demonstrated, through MRI anatomy and proteomic analysis, that V. komodoensis possesses mandibular venom glands secreting anticoagulant and hypotensive compounds. Goldstein's bacteriological evidence did not directly prove that venom kills prey, but it removed the competing explanation and shifted the evidential balance decisively toward the biochemical mechanism.
Originality Note
This page focuses specifically on Goldstein et al.'s 2013 contribution. For the earlier bacteriological catalogue of dragon oral flora (Montgomery et al., 2002), see our Montgomery saliva bacteria page. For the ecological modelling of the "deadly drool" effect in wild populations (Bull et al., 2010), see our Bull deadly drool page. For the venom anatomy and biochemistry that Goldstein's study complements, see the Fry 2009 venom paper page.
Comparison with Other Carnivores
A central weakness of the bacteria-as-venom hypothesis was that it never specified in what sense dragon oral bacteria were special relative to those of other carnivores. The Goldstein team found that bacterial load and species diversity were not meaningfully elevated compared with reports for other captive large carnivores, including lions, and were in some measures lower. If a lion's mouth presents equivalent microbial complexity, invoking bacteria to explain Komodo predation success is no more justified than invoking it for lion predation. Captive dragons were also observed to spend extended time after feeding in lip-licking and head-rubbing, behaviours consistent with active oral clearing — the opposite of maintaining a cultivated bacterial weapon.
| Animal | Aerobic + Anaerobic Survey? | Broadly Comparable Flora? |
|---|---|---|
| Komodo dragon (Goldstein 2013) | Yes — first full survey | Yes — carnivore-typical flora, no virulent species |
| Captive lions & large felids | Partially | Yes — similar mixed aerobe/anaerobe community |
| Wild Komodo dragons (Montgomery 2002) | Aerobic only | Partially assessed; diversity attributed to prey & habitat |
| Captive Komodo neonates (Goldstein 2013) | Yes | Aerobes only; no anaerobes — flora not yet established |
Where This Study Sits in the Research Timeline
The bacteria-as-venom model was articulated from field observations in the 1970s and formalised in Auffenberg's influential 1981 monograph. By the early 2000s it had achieved the status of received wisdom — appearing in documentaries and zoo signage — without ever having been rigorously tested. Montgomery et al. (2002), reviewed on our Montgomery saliva bacteria page, first showed that bacterial diversity in wild dragons correlated with diet and habitat rather than being a stable intrinsic feature of the oral cavity. Bull et al. (2010), reviewed on our Bull deadly drool page, challenged the model further through ecological modelling. Fry et al. (2009) then provided the positive alternative by demonstrating functional mandibular venom glands. Goldstein et al. (2013) completed the bacteriological counter-case by addressing the one remaining objection — that aerobic-only surveys had missed the dangerous anaerobic fraction — and finding it wanting. Together these four papers form a cumulative argument that has effectively retired the septic-bite hypothesis from serious scientific consideration.
Myths vs Facts
| Common Claim | What Goldstein et al. (2013) Found |
|---|---|
| "Komodo dragon mouths are packed with uniquely deadly bacteria." | No virulent species were isolated across 16 animals at three zoos. The flora reflects diet and environment, not a cultivated weapon. |
| "Earlier surveys only found aerobes because anaerobes hadn't been looked for — the real danger is anaerobic." | This was the study's explicit purpose. When anaerobes were properly cultured, they included clostridia and other environmental organisms — not pathogens capable of causing rapid lethal sepsis in large prey. |
| "Dragon saliva is more bacterially dangerous than that of other carnivores." | Bacterial load and diversity were comparable to — and in some measures lower than — those of other captive large carnivores such as lions. |
| "Dragon mouths are filthy and self-cleaning does not occur." | Captive animals were observed to spend extended time lip-licking and head-rubbing after feeding, behaviours consistent with active oral cleaning. |
| "Hatchlings are born with dangerous oral flora." | Neonates grew only aerobes; no anaerobes were detected in any of the six hatchlings, showing that oral microbial community develops with age and diet rather than being a fixed biological trait. |
| "Bacteria are the primary killing mechanism." | The oral flora was judged unlikely to cause rapid fatal infection in prey. The authors explicitly cite the venom hypothesis as the more plausible explanation for prey incapacitation. |
Key Takeaways
- The study provided the first complete aerobic + anaerobic bacteriological profile of Komodo dragon oral cavities, closing a methodological gap that had allowed the bacteria-as-venom claim to persist in the face of incomplete evidence.
- No virulent pathogens were found. Across 16 animals at three zoos, every species recovered could be explained as fauna derived from meals, environment, or normal commensal colonisation — not from a selectively maintained arsenal of bacterial weapons.
- Neonates carried only aerobes, consistent with an environmentally acquired rather than genetically programmed microbial community, and inconsistent with any adaptive model of bacterial predation.
- Dragon oral flora is qualitatively similar to that of other large captive carnivores. The absence of a uniquely dangerous microbial community removes the premise on which the septic-bite hypothesis depended.
- The study explicitly supports the venom hypothesis. By eliminating the competing bacteriological explanation with a thorough and appropriate methodology, Goldstein et al. provided negative but important evidence that the mechanism described by Fry et al. (2009) is the correct one.
- Captive versus wild caveats apply. The dragons studied lived in controlled zoo environments where diet and microbial exposure differ from those of wild animals on Komodo Island. The authors acknowledged this limitation, though they also noted that the captive setting, if anything, might be expected to introduce more environmental bacteria than a wild habitat — making the clean results more, not less, impressive.
Frequently Asked Questions
What exactly did Goldstein et al. culture, and why does it matter?
The team cultured both saliva and gingival samples under aerobic and anaerobic conditions from 16 captive dragons at three zoos. The anaerobic component was the critical addition: all prior surveys of dragon oral bacteria had used aerobic culture only, meaning any oxygen-intolerant organisms — which include many of the most dangerous wound-infection bacteria in mammals — had simply been invisible. Goldstein's study was designed specifically to fill that gap, so when no dangerous anaerobes were found, it represented a genuinely informative negative result rather than a failed positive one.
Could captive dragons have different bacteria from wild dragons?
Yes, and the authors were explicit about this caveat. Captive animals eat food of known origin in controlled enclosures, whereas wild V. komodoensis consume large prey including water buffalo and deer, and drink from watering holes that can harbour environmental bacteria. Earlier work (Montgomery et al., 2002, reviewed on our Montgomery saliva bacteria page) found that bacterial diversity in wild dragon saliva correlated with diet and habitat rather than being an intrinsic oral feature — suggesting that if anything, wild dragons in bacterially richer environments might show more diversity, but still not a uniquely deadly suite of species.
Why were no anaerobes found in the neonate dragons?
The six neonates, aged 7–10 days at the time of sampling, had not yet accumulated the feeding history and environmental exposures that build an anaerobic oral community. This is a normal developmental trajectory seen across vertebrate groups. The absence of anaerobes in hatchlings is inconsistent with a biologically programmed weapon system and suggests the oral flora is shaped by external exposure throughout life rather than being present from birth as an adaptive defence or hunting tool.
Did the study find any pathogens at all?
No virulent species were isolated, according to the paper's own summary. Species such as Staphylococcus sciuri and Enterococcus faecalis — the most frequently recovered organisms — are ordinary commensals found across a wide range of animals and are not typically associated with rapidly fatal systemic infections in healthy large mammals. Where pathogens had appeared in earlier surveys, the 2013 team attributed them to transient contamination from prey tissues rather than stable oral residency.
How does this paper relate to Fry et al. (2009)?
The Fry et al. (2009) paper — reviewed in detail on our venom predator page — argued from anatomical and biochemical evidence that V. komodoensis possesses functional mandibular venom glands and that venom, not bacteria, is the physiological mechanism responsible for prey incapacitation. Goldstein et al. (2013) approached the same question from the opposite direction: if venom is the answer, then bacteria should prove insufficient, and the 2013 study confirmed that insufficiency with thorough culture data. Bryan G. Fry appears as a co-author on both papers, giving them continuity of authorial intent.
Does this mean Komodo dragon bites are not dangerous?
Not at all — dragon bites carry serious risks, but the mechanism is mechanical trauma, venom-mediated anticoagulation and hypotension (as described by Fry et al., 2009), and the normal wound-infection risk posed by any large tooth wound, not a specially cultivated bacterial payload. Any person bitten by a Komodo dragon requires immediate emergency medical care, including wound management and monitoring for systemic effects.
Was there any variation between the three zoos?
The paper noted some variation in species composition across the three institutions, consistent with the hypothesis that environmental and dietary factors — which differ between zoo facilities — shape the oral microbial community. This cross-zoo variability itself argues against a stable, selectively maintained oral bacterial weapon: a biological adaptation would be expected to be present consistently regardless of housing conditions.
Has any subsequent study challenged these findings?
No peer-reviewed study has produced bacteriological evidence sufficient to revive the bacteria-as-venom hypothesis in its original form. The cumulative weight of evidence from Montgomery (2002), Bull (2010), Fry (2009), and Goldstein (2013) has led to broad acceptance within herpetology and wildlife medicine that oral bacteria are not the primary killing mechanism in V. komodoensis predation.
Sources & Further Reading
- Goldstein, E. J. C., Tyrrell, K. L., Citron, D. M., Cox, C. R., Recchio, I. M., Okimoto, B., Bryja, J., & Fry, B. G. (2013). "Anaerobic and aerobic bacteriology of the saliva and gingiva from 16 captive Komodo dragons (Varanus komodoensis): new implications for the 'bacteria as venom' model." Journal of Zoo and Wildlife Medicine, 44(2), 262–272. https://doi.org/10.1638/2012-0022R.1 — Primary source for this review.
- Fry, B. G., et al. (2009). "A central role for venom in predation by Varanus komodoensis (Komodo dragon) and the extinct giant Varanus (Megalania) priscus." Proceedings of the National Academy of Sciences, 106(22), 8969–8974. https://doi.org/10.1073/pnas.0810883106 — Reviewed on our venom predator page.
- Montgomery, J. M., et al. (2002). "Aerobic salivary bacteria in wild and captive Komodo dragons." Journal of Wildlife Diseases, 38(3), 545–551. PMID 12238371 — Reviewed on our Montgomery saliva bacteria page.
- Bull, J. J., et al. (2010). Ecological and population-level analyses of the "deadly drool" model. Reviewed on our Bull deadly drool page.
- Auffenberg, W. (1981). The Behavioral Ecology of the Komodo Monitor. University Presses of Florida. — Foundational field study that gave the bacteria hypothesis much of its early authority.
- Fry, B. G., et al. (2006). "Early evolution of the venom system in lizards and snakes." Nature, 439, 584–588. https://doi.org/10.1038/nature04328