Stimpunks Foundation × More Realms · L★S · Field Guide No. 18

A Field Guide to
Unfinished Animals

Sixteen animals we reconstructed wrong, sorted by the kind of mistake — and every card says whether the argument is over. There is no final draft. One was drawn upside down for fourteen years and then back-to-front for twenty-four more. One was three animals in three different groups. One had its skull mounted on the end of its tail. One was two arms and nothing else for forty-four years. Three of these are still being argued about in print. The animals were never wrong. The drawings were, and the drawings are signed and dated.

L★S
How to use this guide Each entry gives an animal, what we said it was, what the evidence turned out to say, who published the correction and when — and a status: whether that correction is finished. Click any entry to open its notes.
Where this came from The companion to Thirteen Changes (No. 61), which follows one category boundary — Brontosaurus — through a hundred and forty years of being drawn, erased and drawn again. That zine asks what an honest boundary owes you. This one asks the prior question: how often does the record actually move? The answer is constantly, and in public.
Why this one breaks the house formula Eleven of these guides arrive at there is no standard star, nest, egg, shark, way to rest. This one cannot, because what varies here is not the animals — it is us. A sauropod is not unfinished. Our drawing of it is. So the line is there is no final draft, and it is a claim about the record rather than about kin.
Why two cards are dashed Because the apparatus stops there. Sixteen entries get a reframe table because a verdict was reached about an animal and the verdict was wrong. Mary Anning is not a specimen, and the last card has nothing to correct yet — that is the point of it.

Five kinds of mistake, and one place the apparatus stops

The sort is ours and it came out of the entries rather than ahead of them. We tried sorting by date of correction, which put a 1991 inversion next to a 1991 synonymy that had nothing in common with it. We tried sorting by taxon, and it separated two animals that were misread in exactly the same way. What varies cleanly is the shape of the error: the pieces assembled in the wrong order, the animal known from almost nothing, the boundary that moved, the name that outlived its own correction, and the story that got tidier than the evidence. None of these ranks above another, and none of them is a failure of science — every single correction on this page was published, dated and signed, which is the only reason we can list it at all.

One · Assembled wrong
Every piece real, the arrangement invented
Two · Known from almost nothing
Two arms, a whorl, a specimen that burned
Three · The line moved
Merged, split, resurrected — sometimes twice
Four · The name outlived it
Science corrected, label permanent
Five · Tidier than the evidence
The version that travelled better
Where the apparatus stops
Two cards with nothing to reframe

And the status on every card

This is the one thing this guide adds to the house form, and it is borrowed from How We Got Here, where every joint in a chain is marked documented, contested or leap. A catalogue of corrections is dishonest if it will not say which corrections are finished. Without the notation this page would read as a list of settled triumphs, which is exactly the smoothing it is complaining about.

Revised
The correction is published and broadly accepted. Ten cards.
Contested
The argument is live in the literature right now. Four cards — and one of them is the animal the companion zine is named for.
Name stuck
The science was corrected and the label is permanent anyway, because nomenclatural priority does not care. Two cards.
One · Assembled wrong

Every piece was real. The arrangement was ours.

Four animals where nobody misread a bone — they misread how the bones went together. Hallucigenia was published upside down and then, once righted, still back-to-front. Anomalocaris spent ninety-three years as three animals filed in three different groups. Cope put a skull on the end of a tail. Mantell put a thumb on a nose. Three of the four errors were made from real, well-described fossils by people doing careful work — which is the point: an assembly error is not a failure of observation, it is a failure of the arrangement you brought with you.

Two · Known from almost nothing

When the evidence is a fragment, the gap gets filled anyway

Three animals reconstructed from far too little, and each shows a different way that goes. Deinocheirus was two enormous arms for forty-four years and the honest answer stayed we do not know. Helicoprion's spiral of teeth was placed on the snout, on the back, on the tail and in the jaw, because a whorl with no skull around it will support almost any guess. And the Spinosaurus holotype no longer exists — it burned in 1944 after its describer's requests to move the collection were refused, so every reconstruction since argues from photographs and drawings.

Three · The line moved

Merged, split, and in two cases put back

Four boundaries between kinds, drawn and redrawn on the evidence. Brontosaurus was folded into Apatosaurus in 1903 and separated again in 2015. Torosaurus was proposed as merely a grown-up Triceratops in 2010, rejected in 2012, and the question is still open. Nanotyrannus was dismissed for forty years as a juvenile Tyrannosaurus and reinstated in 2025 — with a second species nobody had noticed. Tanystropheus was one species until a CT scan made it two. Two of these four went the way the sceptics did not expect, and all four were argued in public under names.

Four · The name outlived it

Corrected on the science, stuck with the label

Two animals carrying names that the evidence disproved and the rules will not let anyone change. Oviraptor means egg seizer, and it was sitting on its own clutch. Basilosaurus means king lizard, and Richard Owen established within five years that it was a mammal. In both cases the correction is published, accepted and old — and in both cases the animal will carry the accusation in every textbook forever, because nomenclatural priority protects stability rather than accuracy. This is the band the companion zine cares about most: a label can outlive its own refutation by rule, and it can outlive it by repetition, and only one of those two is written down anywhere.

Five · Tidier than the evidence

The version that travelled better than the paper

Three cases where the popular account is smoother than anything anybody published. The fish-crawls-out-of-water story has a date problem: footprints eighteen million years older than the earliest tetrapod body fossils — though those footprints are themselves disputed, which the tidy retelling of that also leaves out. The film raptor is a different animal from the fossil one. And the sluggish tail-dragging dinosaur of the old picture books was not corrected by one discovery but by a whole field changing its mind, starting with a single 1969 monograph.

Where the apparatus stops

Two cards with nothing to reframe

Every card above carries three columns: a trait, what it sounded like, and what turned out to be the case. That table exists because a verdict was passed on an animal and the verdict was wrong. These two have no verdict to overturn. One because the correction owed is to the human record rather than to a fossil, and running sounds like / actually over a woman whose finds were published under other men's names would file her with the misreconstructed animals — which is precisely the injury. One because its correction has not been published yet: on this page's own argument, some of the entries above are currently wrong, and saying actually before anybody knows which is bluff rather than rigour.

What this guide refuses

Not a list of scientists being stupid. Almost every error here was made carefully, from real fossils, by someone working at the edge of what the evidence could support. Cope had jaw fragments at one end and reasoned from them. Conway Morris had a flattened smear in shale. Being wrong from good evidence is what the method looks like from inside. Not an argument that science is unreliable. The opposite, and it is the whole point of the page: every correction listed here is published, dated and attributable, which is the only reason a catalogue like this can exist at all. Try assembling the equivalent list for the categories people get sorted into. Not a claim that the animals were unfinished. They were complete. Our record of them is the unfinished thing, which is why the arrival line here is about the record and not about kin. Not a set of gotchas about names. Oviraptor and Basilosaurus are not scandals; nomenclatural stability is a real good bought at a real price, and the band about them says so. Not settled where it is not settled. Four cards are marked contested and stay that way, including the one the companion zine is named for.

Held honestly

The status notation is ours, and it is a judgement about the state of a literature rather than a fact in it — where the reading is close, the card says so in its notes rather than in the chip. Two cards deliberately carry the awkward version of a famous story. The Elasmosaurus anecdote is usually told with Marsh catching Cope; it was Joseph Leidy, on 8 March 1870, with Marsh in the room, and Marsh's claim survives only in a newspaper piece almost twenty years later. And the Devonian trackways that displace Tiktaalik in the popular retelling are themselves disputed — Lucas (2015) reads some of the Zachełmie structures as fish traces, and a 2020 study agrees for part of the surface — so the card carries the dispute instead of the punchline. The Velociraptor quill-knob counts that circulate in press coverage are not printed here, because No. 61 could not reach that paper either. Where a claim rests on a secondary account, the notes say which.

Companion & sources

Companion to Thirteen Changes (No. 61), which reads the 298-page Diplodocidae revision in detail and takes the argument about published provenance, declared range and appealable rulings. The Brontosaurus card here is the short version of that zine; the Oviraptor card is the short version of its spread eight. The status notation is borrowed from How We Got Here, where every joint in a chain is marked. Every figure on this page is sourced in the field notes of the card that carries it, and where a figure would not stand up it is described rather than printed. There is no final draft.