How Do We Know What Dinosaurs Looked Like?
By the BrainSnail editorial team. How these articles are written and checked, and how to tell us when one is wrong.
How do we know what dinosaurs looked like when no human has ever seen one? The answer is a chain of reasoning that starts with bones and works outwards, using living animals as a guide at every stage. Some of it is very solid, some of it is careful inference, and scientists are usually clear about which is which.
What bones alone can tell us
Bones give the skeleton, and the skeleton gives more than shape. Joint surfaces show how limbs could move and how far. The size and position of each bone constrains posture, so a reconstruction that drags a tail on the ground can be ruled out if the tail vertebrae could not bend that way.
Bone surfaces also carry scars where muscles and tendons attached. A large rough patch means a large muscle, and its position shows which way it pulled. This is how researchers rebuild a body outline that is not simply a guess.
Using living animals as a check
Palaeontologists compare dinosaur anatomy with birds and crocodiles, the two living groups most closely related to them. If a feature appears in both, it is reasonable to expect it in the animals between them on the family tree.
This is why modern reconstructions give many dinosaurs thicker necks, fuller faces and lips covering the teeth. Dissecting living animals shows how much soft tissue sits on top of bone, and older reconstructions consistently underestimated it.
When skin and colour are preserved
Occasionally fossils preserve more than bone. Skin impressions show scale patterns. Exceptional deposits preserve feathers, stomach contents and even the outline of the body.
A few of these fossils retain melanosomes, the tiny structures that carry pigment, and their shape can be compared with those in living birds to estimate colour. This has produced supported reconstructions of a striped ginger tail on one small dinosaur and a grey and white pattern on another.
What we still guess at
It is worth being honest about the limits. Several things are hard or impossible to recover:
- •Colour for most species, since pigment is rarely preserved
- •Sounds, since vocal soft tissue almost never fossilises
- •Behaviour, which has to be inferred from trackways and bone beds
- •Fine details like lips, wattles or crests made only of soft tissue
Why the pictures keep changing
Reconstructions have shifted a lot. Dinosaurs went from sluggish tail dragging lizards to active, upright, often feathered animals, because the evidence changed and the comparisons improved.
That is a feature of the science rather than a flaw. Each new fossil bed tightens what is known, and the honest position is that a dinosaur reconstruction is a best current estimate with clearly different levels of confidence in different parts of it.
The shrink wrapping problem
Palaeoartists have criticised a long standing habit of drawing dinosaurs with skin pulled tight over every bone, showing the shape of the skull and vertebrae clearly through the body.
The critique is straightforward. Applying that approach to living animals produces monsters. A shrink wrapped elephant, swan or camel looks nothing like the real thing, because living animals carry substantial fat, muscle and soft tissue that hides the skeleton. Modern reconstructions therefore add much more bulk, which is one reason recent dinosaur art looks so different from the versions many people grew up with.
The takeaway
We know what dinosaurs looked like by reading skeletons for posture and muscle attachment, comparing them with birds and crocodiles, and using rare fossils that preserve skin, feathers and even pigment, while being clear about which details remain uncertain.