I have a few fossil molluscs from the Devonian — they’re fairly common orthocerids, these cone-shaped shells that once housed mighty ancient cephalopods. Mine are small, but some of these shells get to be 5 or more meters long. We have to imagine big eyes and swarms of arms writhing out of the broad end of the cone, because those squishy bits don’t fossilize well. Well, not just imagine, because we do have data that lets us reasonably infer what the animal looked like. Here’s an excellent post that describes how this kind of reconstruction of Endoceras was done.
That’s not guesswork. Using trace fossils and phylogenetic bracketing and assembling bits of evidence from multiple specimens, you can make an informed estimation of the main features of the animal.
And it is awesome. Bring ’em back.