Thirteen Changes

It is one of the small standard humiliations of a childhood spent loving something. You say Brontosaurus. Somebody older says actually, it's Apatosaurus. And the lesson attached is never really about taxonomy. It is that your enthusiasm was the amateur thing, and being corrected is what growing up feels like.

Here is the paper trail. Othniel Charles Marsh named Apatosaurus ajax in 1877. Two years later, in 1879, he named Brontosaurus excelsus. In 1903 Elmer Riggs looked at both and folded the second into the first, and the reason he gave is worth reading slowly, because everything else here hangs off it. The differences between the two type specimens were, in the words the revision of 2015 uses to summarise him, usually considered insufficient to justify generic distinction.

Insufficient. It is a quantitative word doing a job with no quantity behind it.

Riggs was not being careless — he was doing what everyone did, which was to look hard at bones and form a judgement. But read the sentence again and ask the obvious question, the one a nine-year-old would ask and an adult has usually been trained out of asking: insufficient by what measure?

For eighty-odd years nobody could answer that, because the measure did not exist.

In April 2015, Emanuel Tschopp, Octávio Mateus and Roger B. J. Benson published A specimen-level phylogenetic analysis and taxonomic revision of Diplodocidae in PeerJ. It ran to 298 pages. It scored 81 operational taxonomic units against 477 morphological characters — by its own abstract's description, one of the most extensive phylogenetic analyses of sauropod dinosaurs.

It was submitted on 30 January 2014 and published on 7 April 2015. It is open access under CC-BY, which means you can go and read every page of it right now, for nothing, including the parts where the authors argue with themselves.

And rather than look at bones and form a judgement, they did something else first. They stopped and asked what a genus is, and they wrote the answer down in a methods section instead of leaving it in their heads:

Read what that admits. Not our field is rigorous and yours is not. The opposite: the thing we have been doing is arbitrary, the subjectivity is especially great, and here is what we are going to do about it. Printed, in the paper, above the result.

This is the spread the whole zine is for. They needed a number: how many differences make two animals a different genus rather than a different species? Any number they picked would be a choice. So they picked one, and then they published where it came from.

They took the pairs that everybody already agreed about — well-preserved, long-accepted species sitting inside the same genus — and counted the differences the analysis found between them.

Twelve, and eleven. Both of those pairs are agreed to be one genus. So the bar for being two genera was set at thirteen — one step past the largest gap anybody was willing to call the same.

The species bar was built the same way, from specimens everyone agrees are the same species: one difference, one, and five. So six became the species threshold, deliberately generous, and they said why: a wider margin is preferred herein in order to be more cautious.

Nobody has to like thirteen. That is exactly the point. You can argue with a number whose derivation is printed. You cannot argue with insufficient.

Having built a threshold and used it to reorganise a famous family of animals, the authors did the rarest thing in this whole story. They put a fence around their own instrument.

Not thirteen is how many differences make a genus. Thirteen is how many differences make a genus in this matrix, with these characters, in this software. Change the inputs and the number is void — and they say so, in advance, in writing, in the paper that depends on it.

Then the second criterion, the same discipline again. Pairwise dissimilarity: within this subfamily, specimens in one genus scored below 0.181 and different genera 0.222 and up. Two numbers, a gap between them, and the gap named as the thing the judgement rests on.

A threshold that declares its own range of validity is a threshold behaving honestly. Hold that thought for four spreads. It is the one property the labels we get handed have never once had.

Not the same argument as No. 40. Five Sigma is about how much evidence you need before you claim a discovery, and why that bar should scale with the stakes. This is a different question — not how sure are you but where does one kind end and another begin. The two zines share a suspicion of round numbers and nothing else.

Run the ruler over the specimens and Brontosaurus excelsus comes out far enough from Apatosaurus ajax to hold its own genus. After a hundred and twelve years of children being corrected, the name was not wrong after all.

And look at how the authors phrased their own headline result, in the abstract, where a lesser paper would have put a flag up:

Not is valid. Considered valid by our quantitative approach — the claim carries its own instrument in the sentence. Four words that keep the door open.

Which is just as well, because the argument is genuinely live. The reinstatement has published critics — Donald Prothero's Is “Brontosaurus” Back? Not So Fast! came out four months later, under his own name. Nobody in this dispute is hiding.

And the ruler was not a rescue machine. This matters, or the whole thing is just fandom with a bibliography. The same method, in the same paper, took things away: “Dinheirosaurus” from Portugal was folded into Supersaurus, and “Diplodocus” hayi was moved out into a brand-new genus, Galeamopus. The instrument cut in every direction, including against its makers' own earlier work.

Here is the part almost nobody tells, and it is better than the comeback.

The same revision found a second problem. The type specimen of Diplodocus longus — the specimen the entire genus Diplodocus is legally anchored to — is a partial tail. The paper concludes it cannot be diagnosed to species level at all, and calls it a nomen dubium, noting drily that this leaves the otherwise well-known genus Diplodocus typified by a dubious species.

So they did the constitutional thing. They petitioned the International Commission on Zoological Nomenclature to swap the anchor over to D. carnegii, which is known from a nearly complete, articulated skeleton. That is Case 3700. There was a comment period. Other paleontologists filed in support.

On the last day of 2018, the Commission ruled. They said no. Opinion 2425: Diplodocus longus maintained as the type species.

Sit with the shape of that. The people who wrote 298 pages, who built the ruler, who won Brontosaurus back — asked for one more thing and were refused, in print, with a number on it.

Not overruled in a corridor. Not quietly ignored. Refused by a standing body, in a numbered Opinion, in a journal, which you can look up and cite and be annoyed about for the rest of your life.

We are not going to pretend the paperwork is kind. It is not, and the clearest case in paleontology is a small feathered animal that has been slandered for a century.

In 1924 Henry Fairfield Osborn named Oviraptor philoceratops. Oviraptor means egg seizer. The specimen had been found on top of a clutch of eggs, and the eggs were assumed to belong to something else, so the animal was named for theft — and the species name, philoceratops, doubles down: lover of ceratopsians, meaning fond of eating them.

Then the evidence came in. Norell and colleagues showed in Science in 1994 that eggs of that type from those beds contain oviraptorid embryos. In Nature in 1995 they described an oviraptorid preserved over a clutch in a posture like a brooding bird. The animals were not raiding those nests. They were sitting on their own.

The science was corrected. The name was not, and cannot be. Nomenclatural priority makes it permanent. There is no mechanism, and no appetite, for renaming an animal because the name libels it. It will be egg seizer in every textbook forever.

So there are two ways a label survives its own refutation. Priority — the rule says the first name stands. And repetition — it has simply been said so many times that saying it again requires no evidence at all. Only one of those two is written down anywhere.

Everything above is a category boundary drawn by people, in public, on purpose. So is a functioning label. So is a support tier, a percentile, a reading group, a cut-off score, a line between disorder and difference. The difference is not that one is science and the other isn't. The difference is which properties the boundary is willing to show you.

A cut-off score arrives already applied. It does not say who chose it, or from what, or where it stops being meaningful — and then it travels anywhere: across ages, countries, languages, and purposes it was never built for. The ruler that would not travel one clade sideways gets less latitude than the one we hand to a five-year-old for life.

And there is no Case 3700. Nowhere to file. No comment period, no numbered Opinion, nobody obliged to write down we considered this and here is our answer. Deficit ideology moves on repetition, not evidence, which is why it never needs a hearing. The pathology paradigm does not publish corrections.

The other great revision of a childhood favourite went the other way — not a name restored, but an animal made stranger.

Velociraptor mongoliensis was feathered, and we know it from bone. Turner, Makovicky and Norell reported quill knobs on the forearm in Science in 2007 — the small ridges where, in living birds, large feathers anchor to the ulna by ligament. Feathers themselves rarely survive. The fixings for them do.

So: turkey-sized rather than man-sized. Sickle-clawed, sharp-toothed, and plumed. Much less like a monster and much more like something bad-tempered in a farmyard — which is not a metaphor, because birds are dinosaurs. Not descended from. Are.

Nobody in this conversation grew out of dinosaurs. There are dinosaurs at the feeder right now.

Notice what every one of these revisions did. Brontosaurus came back. Velociraptor got feathers. Oviraptor turned out to be a parent on a nest. Not one correction made the animal less interesting. Every single one made it stranger, more specific, harder to hold in one hand.

That is what revision does when the enthusiasm is still in the room. The animals were never wrong about themselves. We were wrong about them, and being wrong out loud is what let them get better.

Ask who does this work. Two hundred and ninety-eight pages on the neck vertebrae of long-dead animals; a lifetime learning to tell one sauropod from another; a formal petition, a comment period, and a loss absorbed in public in 2018.

That is not what growing out of it looks like. That is what it looks like when nobody grows out of it — when the interest was allowed to keep going until it became load-bearing enough to reorganise a family of animals and correct a hundred and twelve years of textbooks.

The demand that we get serious almost never means go deeper. Monotropic attention already goes deeper into a subject than most curricula ever ask of anyone. What serious asks is that the interest become convertible — point the same attention at something billable and it is expertise; point it at sauropods and it is a hobby you should have outgrown. The demand is not that we stop being absorbed. It is that we stop being absorbed for free.

Read the second half again, because it inverts the complaint entirely. Sustained childlike curiosity is not developmental lag. It is a distributed adaptive capacity — the thing that stops a group's collective intelligence from setting into whatever the highest-ranking person already believes. We carry more of it. That is a contribution, not a delay.

The facts moved. The names moved. Twice, the paperwork refused to move. The enthusiasm was the one part that was right the whole time — and it is also, on the evidence of 298 pages, the instrument that did the correcting.

One last thing about the animal at the centre of all this, and it is the reason it belongs in this collection rather than a museum guide.

Apatosaurus louisae has been estimated at around 41 tonnes; its relative Supersaurus at thirty-five to forty metres long. That skeleton was calcium phosphate — and calcium and phosphorus are made in stars. Forged in the burning shells of massive stars and scattered when they died, exactly like the calcium in the bone your hand is resting on.

So the sauropod was star stuff, at forty-one tonnes of it, walking around the Late Jurassic on a skeleton assembled from dead stars. Nothing in that femur was invented for it. Every atom was second-hand, and every atom is still in circulation now — which is the only immortality any of us gets, and it is not a small one.

Bone is also piezoelectric: squeeze it and it produces charge, the finding Iwao Yasuda published in the 1950s and this whole project is built on. That was true of a sauropod's bone too, because it was bone — collagen and mineral, the same trick. Which means those animals were writing their own skeletons by walking on them, load turning into signal turning into more bone exactly where the pressing was.

And we still cannot fully explain how they got that big. Sander and colleagues call it an evolutionary cascade: no single adaptation, a whole interlocking set, none of which was for gigantism. Their review states flatly that there is currently no direct evidence on sauropod food and food processing. The biggest animals ever to walk, and we do not know how they ate. The mystery is a hundred and fifty million years old and still open, and being unexplained never made them less real.

Which is where the phrase this project signs off with reaches further than usual. Down to is not because of — the star stuff is how far the loving goes, never why it is owed. And it goes back a long way. Love you down to your star stuff — yours, ours, and a forty-one-tonne earthling's, made of the same dead stars, still in the yard, still at the feeder.