More Options

Fire-Breathing Dinosaurs?


Philip J. Senter

Skeptical Inquirer Volume 41.4, July/August 2017

Mainstream geologists and biologists accept the abundant physical evidence that the Earth is billions of years old; that all organisms are evolutionary descendants of a common ancestor; and that non-avian dinosaurs became extinct sixty-five million years ago (e.g., Gradstein et al. 2004; Prothero 2007). In contrast, young-Earth creationist (YEC) authors have long maintained that the Genesis account of creation and the biblical timeline are literally correct, placing the creation of the Earth and all types of organisms at approximately 6,000 years ago. A corollary of this position is that dinosaurs and humans were created on the same day and must therefore have encountered each other. The claim that dragon legends are based on such encounters has long been a 
mainstay of YEC literature, and in 1977, biochemist and YEC author Duane Gish took this concept up a notch in his children’s book Dinosaurs, Those Terrible Lizards, by positing that dinosaurs breathed fire. Other YEC authors followed suit (see references below), and dinosaurs now breathe fire in seventh-grade biology textbooks from BJU Press (Batdorf and Porch 2013; Lacy 2013).

 In support of the idea that a real animal can produce fire, Gish (1977) cited the defense mechanism of bombardier beetles (Brachinus spp.), which spray a mixture of hydrogen peroxide and hydroquinone into the faces of would-be predators. Chemical catalysts cause the mixture to reach a scalding 100º C (Aneshansley et al. 1969). Subsequent YEC authors followed Gish’s lead and added imaginary details such as sparks or explosions or flame (Phillips 1994; Hamp 2000; Isaacs 2010; Paul 2010). In reality, the beetles merely spray hot liquid—which scalds but does not produce flame—and therefore provide no biological precedent for organic fire production.

 Some YEC authors have cited bioluminescent animals and electric eels as biological precedent for fire production (Morris 1984; Petersen 1986; Morris 1988; Niermann 1994; Morris 1999; DeYoung 2000; Petersen 2002). However, the processes that produce bioluminescence (Haddock et al. 2010) and bioelectrogenesis (Pough et al. 2013) are chemically unrelated to combustion and generate little or no thermal energy. These processes are therefore irrelevant to fire production and provide no biological precedent for it.

 Proponents of the fire-breathing dinosaur hypothesis have suggested a number of potential mechanisms of fire production, reviewed below, each of which is a separate hypothesis in its own right. Below, I use scientific data to evaluate each hypothesis. Each such hypothesis implicitly predicts that the mechanism is not only physically possible but that it will not cause serious injury to the animal. Each hypothesis is falsified if either of those predictions is not met.

Exhalation of Pyrophoric Gas

 Henry Morris (1984) and James Gilmer (2011) suggested that a fire-producing reptile could breathe out gases that would ignite upon contact with oxygen. A substance that ignites upon contact with air is said to be pyrophoric. When released into air from a container, a pyrophoric gas explodes after traveling only a few centimeters or millimeters from the opening. The wider the opening, the nearer to it the explosion occurs. For example, the pyrophoric gas silane (SiH4) generates an explosion 5–80 mm from the mouth of a tube 4.32 mm in diameter and 5–30 mm from the mouth of a tube 3.5 mm in diameter (Tsai et al. 2010). Given this, a pyrophoric gas released from the nostril or throat of a large dinosaur—an opening greater than a few millimeters in diameter—would have exploded immediately (Figure 1), burning the animal’s face or throat. The serious harm that this would have caused the animal falsifies this fire-production hypothesis.

 The hypothesis could be tenable if the nostrils were protected by a fire-resistant tissue, but animals produce no such tissue. Many animals produced heavily keratinized epidermal derivatives (hair, feathers, horn sheaths, etc.) that protect from abrasion, but these burn when exposed to fire, as do cuticles of collagen or chitin. If dinosaur snouts produced a protective, nonflammable, mineral covering (e.g., calcium carbonate or calcium phosphate), this covering would have fossilized, as is usual with hard animal parts. No such covering is present on the snout of any dinosaur fossil.

Ignition of Belched Methane

 Herbivorous mammals emit large eructations (belches) of methane daily, averaging twenty-six per hour in cattle and forty-two per hour in sheep (Colvin et al. 1958; Ulyatt et al. 1999; Koch et al. 2009). John Morris (1999) suggested that herbivorous dinosaurs did likewise and secreted a pyrophoric material from a gland to ignite the belched methane.

 However, once liberated from a container (such as a dinosaur’s oral or nasal cavity), a gas immediately spreads in all directions; the methane would quickly surround the dinosaur’s head. The pyrophoric gas would not only produce a nasal or oral burn as it emerged but would also ignite the cloud of methane-infused air around the animal’s head, burning the surface of the head in the resulting fireball (Figure 1). The serious harm that this would have caused falsifies this fire-production hypothesis.

 Human cases confirm that ignition of combustible belches causes facial burns. Eructation in humans usually involves the emission of swallowed air, which is methane-free and is not combustible. However, in adult cases of pyloric stenosis, combustible gases can accumulate in the stomach due to the fermentation of food in the stomach when passage to the duodenum is obstructed. When a patient belches those gases while smoking, the lit cigarette ignites the gases, and the resulting fireball causes facial burns (Galley 1954; MacDonald 1994).

Ignition of Methane by an Electric Organ

 Several fish species have electric organs made of modified muscle cells called electrocytes, which are arranged in series so that their voltages are summed (Gallant et al. 2014; Sillar et al. 2016). Most electric fish species generate weak, harmless pulses that are used to navigate, detect other fishes, and convey social signals (Lissmann 1958; Sillar et al. 2016). In contrast, electric eels (Electrophorus electricus) and torpedo rays (Torpedo spp.) produce pulses strong enough to stun prey (Sillar et al. 2016) or to cause serious harm or death to humans (Copenhaver 1991; Carlson 2015).

 DeYoung (2000) and Gilmer (2011) proposed that dinosaurs possessed an electric organ, as in the electric eel, and used it to generate a spark to ignite metabolically produced methane. However, as we have seen, ignition of belched methane would envelop the animal’s head in a fireball. Also, electric organs do not produce sparks.

 A spark is an electrical discharge into air. The permittivity (a measure of how easily electric current flows through a material) of air and of methane are both low, approximately 1.0 (Wohlfarth 2013), whereas that of water is over 60 (Harvey 2013). For this reason, current generated by an electric fish travels easily through water or through biological tissue—which is mostly water—but not through air or methane. Because electrical current follows the path of least resistance, current flowing from one part of an animal to another will flow through the animal’s tissues and will not jump an air-filled or methane-filled gap between body parts as a spark. For example, if a dinosaur had sufficient voltage between its upper and lower jaws for an electrical current to flow from one jaw to the other and the dinosaur opened its mouth and belched some methane, no spark would leap into the methane cloud; instead, the current would flow from one jaw to the other through the dinosaur’s jaw muscles. There would be no spark, flame, or any other visible evidence that the electrical event had occurred (Figure 1). This fire-production hypothesis is therefore falsified.

Figure 1. The lambeosaurine dinosaur Corythosaurus casuarius, showing hypothesized mechanisms of fire production in dinosaurs and the results that would occur in a real animal. Note that some proposed mechanisms would cause harm to the animal, and the others would not actually produce fire.

Ignition of Fuel by a Spark Produced by Friction

 Gilmer (2011) hypothesized that a dinosaur could use friction to produce a spark to ignite a combustible gas such as methane “in the mouth, throat, [or] internal organs.” However, no animal produces a material that creates sparks in response to friction. The tough materials that animal bodies make—calcium phosphate, calcium carbonate, chitin, and keratinized integumentary derivatives—do not spark when rubbed together. Flint, which produces sparks when rubbed together, is a form of silica (SiO2), a chemical that some microbes can precipitate (Erlich and Newman 2009). However, even if flint-producing microbes inhabited dinosaur mouths or throats, any methane ignited there would explode there, causing serious internal injury.

 Internal organs are anoxic environments and therefore do not allow flame production. The one exception to this rule is the respiratory tract, but flame generated here causes severe injuries or death (Wöllmer et al. 2010). Even the digestive tract has too little O2 gas in it to support flame (Levy 1954; Cunha et al. 2011) (Figure 1). This hypothesis is therefore falsified.

Emission of a Hypergolic Pair of Chemicals

 Isaacs (2010), Batdorf and Porch (2013), and Lacy (2013) hypothesized that dinosaurs could produce fire by emitting a pair of chemicals that would ignite in the air upon contact with each other, after being sprayed from the mouth or nose. A pair of chemicals that spontaneously ignite when combined, without a separate ignition source, is termed hypergolic. A hypergolic pair of chemicals includes a fuel chemical and an oxidizing chemical. Numerous hypergolic pairs of chemicals are known, but most such chemicals are man-made and either do not occur in nature or—as in the case of liquid oxygen—must be chilled to temperatures that animal bodies cannot withstand. The two exceptions are hydrogen peroxide (an oxidizer) and ammonia (a fuel), both of which are produced by organisms. However, the fuels with which hydrogen peroxide is hypergolic—kerosene, pentaborane, or mixtures including hydrazine plus methanol (Sutton 2006)—are extremely toxic to organisms. Likewise, the oxidizers with which ammonia is hypergolic—liquid oxygen and liquid fluorine (Sutton 2006)—have boiling points too low (-183 ºC and -188 ºC, respectively [Compressed Gas Association 1990; Hammond 2013]) for organisms to withstand their presence as liquids. There is therefore no known hypergolic pair of chemicals the production of which organisms can withstand.

 Furthermore, emission of a pair of hypergolic chemicals would harm the animal if one or both chemicals were gaseous. Because a gas diffuses in all directions immediately upon release, a pair of gaseous hypergols would surround the dinosaur’s head and envelop it in a fireball as they reacted with each other. If one hypergol were gaseous and the other liquid, then upon release the gaseous hypergol would immediately spread in all directions and reach the liquid hypergol’s point of exit from the body, burning the animal in that location. These versions of the hypergol hypothesis are therefore falsified.

 If both chemicals were liquid, they would have to be sprayed at such angles that the two streams would converge sufficiently far from the animal not to burn it. However, this version of the hypothesis is nonetheless falsified by the lack of any pair of hypergolic chemicals the presence of both of which animal bodies can withstand.

Lambeosaurine Dinosaur Crests

 The duckbilled dinosaurs of the subfamily Lambeosaurinae had hollow crests, and in several YEC publications—including seventh-grade biology textbooks published by BJU Press (Batdorf and Porch 2013; Lacy 2013)—these crests are interpreted as storage or mixing chambers for combustible gases (Gish 1977; Petersen 1986; Morris 1988; Niermann 1994; DeYoung 2000; Petersen 2002). However, the crests’ hollow passages are part of the respiratory tract (Evans et al. 2009). Anything stored or mixed there would have obstructed airflow, causing suffocation. In addition, the left and right passages are separate, which precludes mixing, except in a rear compartment that housed the olfactory epithelium (Evans et al. 2009). Anything mixed and combusted there would have destroyed the animal’s sense of smell in addition to causing internal burns (Figure 2).

Figure 2. Corythosaurus casuarius, showing hypothesized combustion-related functions for the crest. Note that all proposed functions would cause harm to the animal.

 Connected to each of the left and right nasal passages in the lambeosaurine skull is a lateral diverticulum, a blind sac extending upward from the nasal cavity (Evans et al. 2009) (Figure 2). The diverticulum is lateral to the airway, so that a gland or storage facility housed there would not obstruct the airway. However, in today’s archosaurs (crocodilians and birds), diverticula from the nasal passages house only air, not glands or storage structures (Witmer 1997; Witmer and Ridgely 2008). Furthermore, fire-production methods involving lambeosaurine lateral diverticula would have injured the animal. If the lateral diverticula housed glands that released pyrophoric chemicals into the air in the respiratory passages, this would have generated combustion inside the crest (Figure 2), causing internal burns. If a duct conducted pyrophoric chemicals to the nostril, ensuring that combustion occurred outside the crest, the tissue around the nostril would suffer burns (Figure 2). If each lateral diverticulum housed a gland that produced one of a pair of hypergolic gases, then upon exiting the nostrils the gases would diffuse in all directions and combust, enveloping the animal’s snout (or possibly the entire head) in a fireball.

 If each lateral diverticulum housed a gland that produced one of a pair of hypergolic liquids rather than gases, and if ducts conducted the liquids to the nostrils, and if the liquids were projected by muscular squeezing, and if the two streams of liquid converged far enough beyond the snout to avoid burning it when the two streams contacted each other and ignited, then the animal might have avoided injury. However, there is no bony indication of the presence of the requisite glands, ducts, or muscles in the lambeosaurine skull (Evans et al. 2009). Moreover, there is no pair of hypergolic chemicals for which animal bodies can withstand the production of both. This hypothesis is therefore falsified.

Cartilaginous Blasting Caps?

 Isaacs (2010) hypothesized that dinosaurs possessed a “chunk of cartilage” that extended beyond the bony snout, as in mammals, and “may have housed a mixing region for chemicals and oxygen used for combustion.” However, nasal cartilages do not host fire-production mechanisms in known animals. More important, if a cartilaginous “chunk” hosted combustion, the structure would have suffered burns, which falsifies this hypothesis.

Other Mesozoic Reptiles

 Booker (2005) hypothesized that the enlarged cavity at the tip of the snout of the Cretaceous crocodyliform Sarcosuchus housed a fire source. Wieland (2005) and Paul (2010) suggested that the cavity was used to mix combustible gases. In seventh-grade biology textbooks, Batdorf and Porch (2013) and Lacy (2013) also implied that the cavity was used in fire production.

 The appearance of an enlarged cavity at the tip of the snout in adult Sarcosuchus is due to developmental widening of the snout. In Sarcosuchus, juveniles are narrow-snouted, and the entire length of the snout widens during development, so that the snout tip is widened not by itself but along with the rest of the snout (Sereno et al. 2001) (Figure 3). The wide adult snout tip is therefore developmentally related to the width of the snout as a whole, not to any special function of the snout tip. Narrow-snouted crocodilians eat fishes nearly exclusively, whereas medium- and wide-snouted species are generalist predators with diets that include large prey (McHenry et al. 2006). The change in snout proportions in Sarcosuchus is therefore functionally related to a dietary switch as the animal grew.

Figure 3. Snouts of juvenile (left) and adult (right) Sarcosuchus imperator in dorsal view, showing that the great width of the nostril in the adult is an artifact of ontogenetic widening of the entire snout. Modified from Sereno et al. (2001).

 The widening of the snout, in combination with the lack of a bony internarial bar between the nostrils, yields the illusion of an enlarged cavity at the snout tip in Sarcosuchus. Most extant crocodilian species also lack a bony internarial bar and instead possess a soft-tissue septum that separates the nostrils (Iordansky 1973). There is no reason to suspect otherwise in Sarcosuchus, so a single, enlarged cavity is very unlikely in the fleshed-out animal. Even if it had existed, production of fire within it would have burned the animal, which falsifies this hypothesis.

 Gilmer (2011) hypothesized that the crests of “some pterosaurs, such as the Pteranodon” “could store flammable gas . . . for use on demand.” This hypothesis is anatomically unrealistic; in some crested pterosaurs it is possible that the crest contained air-filled diverticula of the nasal cavity or the middle ear cavity (Bennett 2001), but the crests are extremely thin in cross-section, with insufficient room for storage, and in extant archosaurs the nasal diverticula contain only air (Witmer 1997; Witmer and Ridgely 2008). Furthermore, in Pteranodon itself, there is no indication that the frontal bone, which makes up the entire crest, is invaded by nasal diverticula in the first place (Bennett 2001). This hypothesis is therefore falsified.

The Real Origin of Dragon Legends

 The biological reality behind the origin of dragon legends has long been known. The word dragon is derived from the ancient Greek δράκων (drakōn) and the Latin draco, both of which meant “serpent.” Many ancient Greek and Roman artifacts depict drakōn/draco myths, in all of which the depicted animal is a snake (Ogden 2013; Senter 2013). Greek and Roman works of natural history described the drakōn/draco as a snake. Some such works restricted the term to large, nonvenomous constrictors, especially the Aesculapian snake (Zamenis longissima) or African and Indian pythons (Senter 2013; Senter et al. 2016).

 European dragon lore evolved in the Middle Ages. Rumors that dragons could fly and could produce fire had begun by the fifth century (Senter et al. 2016). By the tenth century, dragons were routinely depicted with feathered wings and a pair of limbs (Temple 1976; Mittman 2006). Depiction of dragons as quadrupeds with bat-like wings began in the thirteenth century and became common during the Renaissance (Allen and Griffiths 1979; Benton 1992; Absalon and Canard 2006; Morrison 2007). By the time nineteenth-century naturalists gave dinosaurs their first scientific descriptions, European artists had been regularly portraying dragons with an uncannily dinosaurian or pterosaurian appearance for about four centuries. Speculation that human encounters with these animals had inspired dragon legends naturally followed, and YEC literature now routinely portrays dinosaurs as fire-breathers. But as shown here, that is unrealistic, and its continuance in science textbooks is downright irresponsible.


Philip J. Senter

Philip J. Senter's photo

Philip J. Senter is a vertebrate paleontologist with a specialty in dinosaur paleobiology. He earned a PhD in Biological Sciences from Northern Illinois University in 2003 and teaches biology courses at Fayetteville State University in North Carolina. He has published more than sixty articles on dinosaur paleontology, reptile biology, and the creation-evolution debate. He can be contacted at