How dinosaur sizes are reconstructed, and why they keep moving
Every extinct animal in this game carries a figure that is a reconstruction rather than a measurement. How firm that reconstruction is varies enormously from one to the next, and the difference is worth knowing before you quote any of them.
Three grades of confidence
Extinct animals in this catalogue carry figures of very different reliability, and the difference comes down to how much of the animal has been found.
- Several substantially complete skeletons. T. rex, Triceratops, Stegosaurus. The figure is an average across known individuals and the uncertainty is about variation between animals rather than about anatomy.
- Partial skeletons with a missing or disputed section. Diplodocus, Quetzalcoatlus. The anatomy is understood and one part of it has to be inferred.
- Teeth, or very little else. Megalodon. The body length is calculated from a relationship measured in living relatives.
T. rex, the well-documented case
The figure used here is 12 m nose to tail, and it is an average adult. Several near-complete specimens are known, and they differ by metres: the largest individual is considerably longer than the average, which is why published figures for the species range up toward 13 m.
That variation is real biology rather than measurement error. It is also the reason a single number for a species is always a choice. T. rex against an African elephant at 3.2 m shoulder height gives 3.75, mixing a length and a height, and against a saltwater crocodile at 4.5 m total length gives 2.67, which is the cleaner comparison because both figures are lengths taken the same way.
The best like-for-like dinosaur comparison in the set: Stegosaurus at 7 m against Velociraptor at 1.9 m, ratio 3.68. Both nose-to-tail, both from reasonably complete material, both averages. Nothing in that comparison is doing any inferring.
Diplodocus, and the tail nobody has all of
Diplodocus is quoted at 25 m, and close to half of that is tail. That is the problem: the tail tapers to vertebrae the size of a fingernail, and the last of them are almost never preserved. Total length therefore depends on how many caudal vertebrae the reconstruction assumes, and specimens differ.
The result is a figure that has moved over the decades and will move again. Against a blue whale at 27 m the ratio is 1.08, which is close enough that a revision of the dinosaur figure could flip which one is longer. Against a tennis court at 23.77 m, ratio 1.05, the court is the fixed half and the dinosaur is the moving one.
Megalodon, reconstructed from teeth
This is the softest figure in the catalogue and it deserves to be flagged every time it appears. Megalodon is known almost entirely from teeth. Shark skeletons are cartilage and rarely fossilise, so there is no complete animal to measure.
Body length is estimated from tooth size using a relationship established in living sharks, principally the great white. That relationship is reasonable and it is an extrapolation well beyond the range it was fitted on, which is exactly the situation where statistical relationships become unreliable. Published estimates run from about 15 m to over 20 m depending on the method and the tooth.
The figure used here is 16 m, a mid-range choice. Against a blue whale at 27 m that gives 1.69. Take the upper estimate instead and the ratio drops to about 1.35, which is a materially different picture of the animal.
Quetzalcoatlus, where the pose is the inference
The stated 10.5 m is a wingspan, and a wingspan for an extinct animal is a reconstruction of a pose as much as of a skeleton. The bones are known; how far the wing membrane extended and how fully the animal could spread it are inferred.
Estimates for this species have moved substantially over the last few decades and have generally come down from the largest early figures. Against a wandering albatross at 3.11 m the ratio is 3.38, comparing a reconstructed span to a measured one.
Films are not evidence
One species in the set has a size most people know from a film rather than from a fossil. Velociraptor is quoted here at 1.9 m nose to tail, which makes the animal roughly the size of a large turkey. The creatures that carry the name on screen are a different genus entirely, scaled up considerably for effect.
That is worth stating plainly because it is the most widespread size misconception about any extinct animal. A great white shark at 4.75 m against a Velociraptor at 1.9 m gives 2.5, and the shark is two and a half times the dinosaur, which surprises almost everyone who learned their raptor from a cinema.
How to read an extinct animal's figure
Three questions, and they mirror the three grades.
- How many individuals are known? One skeleton means the figure describes that animal, not the species.
- Which part is inferred? A tail, a wing membrane, or the entire body.
- Is the number an average or a maximum? For well-known species both circulate, and they can differ by a fifth.
None of this makes the figures unusable. It makes them provisional, which is a different thing, and it is worth writing down alongside the number rather than leaving the reader to assume a precision that does not exist.