The Biological (Im)Plausibility of Dragons

So I made an off-handed remark about the biological plausibility of dragons in one of my Game of Thrones posts, and got several comments, posted and via e-mail, essentially amounting to “Wait! I want to read that post!”. Ask and you shall receive: I called in some help from my friendly neighborhood developmental biologist blogger, and she gave me some background information to run with.

I’m going to get to the fire-breathing thing eventually, but I’m going to start with something that is easier to explain away both mythologically and possibly genetically: the fact that many western pictures of dragons seem to be six-limbed vertebrates: two arms, two legs, two wings.

By Gian Luca Ruggero, aka Actam (Own work) GFDL ( or CC-BY-SA-3.0-2.5-2.0-1.0 (, via Wikimedia Commons
By Gian Luca Ruggero, aka Actam (Own work). This Welsh dragon has six limbs, which is rare for a vertebrate.

Now, the obvious counterpoint to the above is that not all dragons, even in western depictions, have arms as well as wings. There are plenty of four-limbed dragons. Not only Chinese depictions, which don’t have wings, but also wyverns, which don’t have arms. The dragons in the HBO Game of Thrones television series fall into this latter category. But the fact remains that the traditional western dragon body plan is… a bit off for a vertebrate.

This Wikipedia and Wikimedia Commons image is from the user Chris 73 and is freely available at // under the creative commons cc-by-sa 3.0 license.
This Japanese dragon has only four limbs, like other vertebrates. How it would fly, of course, is an exercise left to the reader. This Wikipedia and Wikimedia Commons image is from the user Chris 73 and is freely available at // under the creative commons cc-by-sa 3.0 license.

To explain why that is, I want to talk about how vertebrate body plans are organized, and how extra limbs can be formed. And for that I need to get into some definitions. We have three axes: dorsal/ventral, which goes from your back to your stomach, anterior/posterior, which goes from your head to your foot, and right/left, which (unsurprisingly) goes from your right to your left (while your outsides might be roughly symmetrical, your insides definitely aren’t). The anterior/posterior axis is controlled by a family of genes called Hox genes. Hox genes are conserved all the way to flies and worms, so they’re pretty universal (and obviously pretty flexible in terms of things like number of limbs). How this works is that vertebrates, and arthropods (like flies), are segmented: very early in development we split the embryo into something like stripes, and then you get some repeated and some different structures in each segment. So, for example, you get repeated structures like ribs and non-repeated ones like arms. But it’s the same family of genes whether you’re a person or a fly or a worm. So the master regulators of head-to-tail development don’t really care whether you have zero limbs, four, or eight. But other gene pathways, farther down the line, do.

By Meyers Konversionlexikon (Meyers Konversionlexikon 1888) [Public domain], via Wikimedia Commons
Image from 1888 Meyers Konversionlexikon. In this diagram of a whale skeleton, you can see vestigial limb bones labelled in “c”. Even when they don’t use them, vertebrates tend to have four limbs.

And now we get to the cool genetics thought experiments. How would you get another set of limbs? One of the most obvious ways would be to duplicate a couple Hox genes. Hox genes are kind of cool in that they’re laid out on the chromosome in the same order that they get transcribed, from anterior to posterior. So maybe you get a gene duplication or two and end up with two shoulder/upper body segments. Of course, then you’d end up duplicating a lot more than just limbs: all the organs and such that go between the limbs would also be duplicated. Which would be a little bit awkward. So we can get more specific. Is there a master-regulator that says “put a limb here!” instead of one that says “this is about where your shoulders should be”? Yes. It’s called FGF-8, and in chicken eggs if you add in FGF-8 expression where it shouldn’t be, you get additional wings or legs. (Whether you get a wing or a leg depends on where the extra expression is — right back to hox genes) So maybe you get some kind of regulatory co-option event that allows FGF-8 expression in two splotches near the shoulder blades: one for wings and one for arms. Then you can start maybe grabbing other genes from other pathways to make the two limbs different from each other. Or something. There’s some talk that this could have been what happened to give a pterosaur with six limbs.

For more on what that would take, check out the cool experiments in chickens documented here.

One quick note before we get to fire: this is assuming that once you got wings and two arms and two legs, the wing area to body weight ratio and general bone density and such are enough to support flight. Which isn’t necessarily likely for something as large as an adult dragons The biggest flying bird is the wandering albatross, and it only weighs 46 pounds. One scientist in Japan calculated that the heaviest an animal could possibly be and still fly – as in, long soaring flights, not quick take-offs like a turkey – is 88 pounds. The size of a large dog. Certainly not the size of a pterosaur or a dragon. That’s because a heavier bird needs longer wings to support it, which need more muscle to flap quickly and navigate between air currents, which therefore adds more weight. Muscle strength grows with cross-sectional area, but weight grows with volume, so after a certain point the strength to flap fast enough to keep yourself aloft is outpaced by the weight-gains necessary to flap that fast. There’s a bit of a controversy there, and if you want to read more about it this is a good place to start. But regardless of the details of the calculation, a full sized dragon would take quite a bit of lift, and whether or not we can get there with muscle and sinew is yet another thing to consider.

By Heinrich Harder (1858-1935) (The Wonderful Paleo Art of Heinrich Harder) [Public domain], via Wikimedia Commons
By Heinrich Harder, an image of a Pteranodon. Whether these guys would actually be seen soaring out of sight of land is up for debate.

Now, for fire. This is the biological mechanism my dev bio buddy suggested: let’s start from something like a spitting cobra. This seems nicer because it’s a vertebrate. We’ll have it spray something very flammable instead of the usual mix of neurotoxins, hemotoxins, and other nasty stuff used by cobras. Ideally, something vaguely like gasoline or lamp oil: hydrocarbon, liquid, highly flammable. And you’ll need to protect the dragon itself, probably with heat-resistant and fire-retardant scales around the mouth and face.

By Peter Engelen (Wikipedia:Contact us/Photo submission) [CC-BY-SA-3.0 (], via Wikimedia Commons
By Peter Engelen. This beauty spits a potent mix of neurotoxins, cardiotoxins, and cytotoxins. That’s kind of like fire, right?

Believe it or not, that’s the easy part. We also need a spark. The best we could come up with was some kind of source of elemental sodium or magnesium that could be uncovered at the same time as the hydrocarbon spray. It would certainly spark, since most life is mostly water and sodium and magnesium both react strongly to water. And there’s maybe almost kind of some evidence for something similar evolving: certain wasps have zinc or manganese-coated ovipostors. But on the other hand, zinc and manganese are far less dangerous to water-filled life than sodium or magnesium. The dragon would need to keep its sodium reservoir in a totally oxygen-free, water-free pocket, very close to the source of the fuel. Possibly a lipid pocket: possibly in a bubble of something vaguely like gasoline or lamp oil.

Mix in a little water, for example from saliva in the mouth, and it all goes boom.

By Stonagal (Own work) [CC-BY-SA-3.0 ( or GFDL (], via Wikimedia Commons
By Stonagal. Sodium, drool, and lamp oil might work like the torch used in this animatronic dragon.

Of course, we’ve gone increasingly far afield: it would also mean that a relatively small injury to the dragon’s fire-breathing organ would cause the entire thing to ignite. Which might not be evolutionarily favorable to the dragon. But might be advantageous to any people who could get close enough.

Elizabeth Finn

Elizabeth is a geneticist working for a shady government agency and therefore obliged to inform you that all of the views presented in her posts are her own, and not official statements in any capacity. In her free time, she is an aerialist, a dancer, a clothing designer, and an author. You can find her on tumblr at, on twitter at @lysine_rich, and also on facebook or google+.

Related Articles


  1. Actually, ignition is probably simpler than that. Hydrogen peroxide + hydrazine is pretty much pyrophoric at room temperature. Both are produced by known biological processes, and could be stored in separate vesicles and combined during exhalation.

  2. There is one example of a four-legged vertebrate with wings in the fossil record, but its a bit of a cheat. Coelurosauravus had what look like a set of modified ribs (or possibly dermal rods) with a membrane stretched over them. It wasn’t capable of powered flight and it was only about a foot long, though. So, more like a Permian flying squirrel than a dragon. Still, it’s an interesting example of pseudo-limbs developing from existing structures.

  3. I’ve always thought that a much more feasible – and probably also really cool – body plan for a “dragon” would be insect-like: six limbs and pair of wings or two. You’d still need to take quite a few liberties, of course.

  4. Andreas: We talked about Hydrogen Peroxide reactions; particularly ala the bombadier beetle (these guys are crazy). There were a couple reasons we went a different route: first of all, we wanted to start with a vertebrate rather than a beetle, so the spitting cobra presented some advantage there. And more importantly, one disadvantage of most of the peroxide reactions is that in addition to creating lots of heat, they also generate water as a product. I double checked the hydrazine reaction — its products are water and nitrogen. Which might extinguish the flame that the heat just created. But yes — oxidation of many things using hydrogen peroxide is a highly exothermic reaction using biomolecules and might be another way to get a spark.

    Steve: The rib modifications! I read about those guys too and I loved it! There are also flying snakes that basically flatten themselves midair to turn their entire bodies into an airfoil. Which is pretty awesome.

    Kiki: At first glance, you could maybe get centaurs by some strange Hox duplication. BUT you’d have to go through a really creepy six-legged-horse intermediate.

    Zhankfor: Oh, totally. Then you could go the bombadier beetle route and use hydrogen peroxide to start a fire! Of course, it wouldn’t look anything like most pictures of dragons. And giant insects have their own biological constraints (due to either problems with the hemocoel or the exoskeleton… I don’t recall which).

  5. An explanation as to why there are no dragon fossils is due to the biological immunities of the dragon’s “fire organ” failing after they pass away. The dragon carcass simply dissolves leaving noting behind.

    **Did a paper in high school about how dragons could exist if they were real** 😛

  6. Of course, if Targaryen humans can have genetic fire immunity, dragons should be able to also, right? Could it be an actual related gene, via some kind of cross-species virus at the dawn of the Targaryen line?

  7. Good gosh, the gene stuff was A-MA-ZING! I’m making up some plausible dragon anatomy and physiology for my Fantasy RPG Forum and I’ve found this article really useful. Thank you!
    But, as for the fire: hydrazine is a really, really bad idea, seeing how it’s highly toxic, caustic, deadly, and basically it does bad stuff to you. Dragon cells still have to be built as any other eucaryote cells and there’s no possibility for the dragon to store it in a gland without the gland dissolving. It couldn’t even have different cells procuding a protective mucus (that’s what the stomach does), because the mucus would quite possibly alter hydrazine’s chemical properties.
    I’m also not so sure about hydrogen peroxide, seeing as it would also be very difficult and dangerous to store inside of a body, and a reaction being exothermic is only half of the success–the products of the reaction also have to react with the oxygen and burn, ’cause otherwise you’d just get a very, very hot breath, not fire.
    Another huge problem is that they cannot burn inside the dragon’s mouth, but outside of it, so the glands should spurt their secretions with a lot of force, they should not be able to work if the mouth is closed, and the gushes of secretions should cross just before the mouth and go up in flames in contact with the air. BUT in a state of a gland inflammation (and like any other very specialized structure in a body, they would be affected), the glands would probably leak their secretions into the mouth, allowing them to mix there, so I think there should be some sort of neutralizing glands that would produce something to deactivate the power mixture that would probably work all the time there, and just intensify their production when the fire glands would work. Because otherwise the whole thing would be able to burn inside the mouth and if the dragon inhaled their own flames, their lungs wouldn’t be well able to work anymore.
    Jerodast, I’m not exactly sure it is even possible for the Targaryens to be immune to flames. The only way would be for the skin to produce some sort of a fluid that would be secreted alongside sweat and protect their whole bodies with a fire-resistant cover. But it’s certainly not possible to protect every cell by itself, so if a Targaryen were to inhale the flame, they would be burnt the same way. But if there was a gene to give such a protection to the skin, a cross-species virus is certainly the way it would go!
    Kiki, the least of the problems with Centaurs is their skeleton, because you just make a really long spine with a really long ribcage, and get going from there. The real problem is the internal organs, ’cause they can’t possibly have two stomachs, and that’s pretty much where I get lost.

Leave a Reply

Check Also
Back to top button