11. Why These Molecules

Run the morning again. The alarm at 6:12, and the histamine that answers it is riding a receptor scaffold older than the bilaterian line, though the wake-up job itself has no date outside mammals. The want of coffee arrives on receptors quadrupled 450 million years ago, doing a computation nobody has caught below the teleost fishes. The nick on your chin closes with serotonin moved by a transporter whose protein fold worked in bacteria. The fish sauce at breakfast presses on the taste side of a molecule that was moving nitrogen through metabolism four billion years back; the truck that runs the light is answered by noradrenaline through a system wired at the base of the vertebrates, with adrenaline — a hundred million years younger — arriving second; the sneeze at the gate is the control group, working a binding site invented at least three separate times; the text from someone you love lands on the derived copy of a vasopressin-like original; and the toe that stopped hurting stopped, in whatever part, through a system that shipped whole in the great duplication. Same morning. Now every moment wears a date and a grade. One question is left standing, and it is the deepest one: why these? Why is the inventory made of these eight families — and why, across six hundred million years in which everything else about animals was rebuilt again and again, was the inventory never replaced?

Chapter one left three answers on the table. The chemical-constraint answer: there was never a choice — the cast is what amino-acid chemistry offered cheaply, and no other store was open. The frozen-accident answer, Crick's phrase borrowed from Vol. I's account of the genetic code: the assignments were arbitrary, and then a thousand dependencies made them permanent. The selection answer: the cast is still here because it kept winning. The temptation is to crown one. The evidence, laid out ledger by ledger, does something more interesting: it hands each hypothesis a different layer of the stack.

Constraint owns the bottom. It is no accident that the brain's principal words are glutamate, its own decarboxylation product GABA, and a handful of amines that are each one or two cheap enzymatic steps from an amino acid the cell already stockpiles. The "go" signal is the busiest metabolite in nitrogen chemistry; the "stop" signal is what you get when you snip one carboxyl group off "go." Serotonin is a modified tryptophan, the catecholamines are a production line run on tyrosine, histamine is histidine minus the same group GABA lost. Even the bonding peptides obey a structural budget: nine amino acids in a ring, a shape so constrained that one substitution separates the water-balance hormone from the one in the labor-ward drip. Chemistry proposed a short menu, and everything on the menu got used. Where the constraint story runs out is exactly where the ledgers' grades are strongest: it explains the cast, and explains nothing about the casting.

The casting — which molecule got which job — is where the accident froze. Nothing about an indoleamine makes it a thermostat rather than an alarm; nothing about a catecholamine makes it wanting rather than rank. The proof is the book's cleanest natural experiment, met in chapter six: noradrenaline and octopamine, chemical near-twins built by one enzyme lineage, running the same alarm brief in two halves of the animal kingdom for six hundred million years — the protostomes locked onto one, the deuterostomes onto the other, and neither side ever switched. Two frozen answers to one question, standing side by side since the bilaterian ancestor. That is not constraint, because chemistry offered both to everyone. It is not ongoing selection choosing a winner, because in six hundred million years no winner was chosen. It is what an accident looks like after it freezes: arbitrary, then load-bearing, then permanent.

What the control group proves

Selection's share is subtler, and histamine is what makes it visible. The control group proves that the freeze is not a law of nature. Binding sites for histamine were re-invented at least three times; bacterial synthesizing enzymes are still being traded laterally today; the system is a patchwork, and it works. Reinvention is chemically possible and evolutionarily survivable — which means the seven frozen families are not frozen because change is impossible. They persist either because every mutant that tampered lost (selection standing guard over the accident), or because the dependency web priced tampering out entirely (the frozen accident's own logic). The ledgers cannot fully separate those two, and this book will not pretend they can. What the ledgers do show is selection working the edges of the frozen core: duplicated receptor inventories retained and specialised after 2R, four opioid systems kept where one might have done, an analgesia circuit whose fight-through-the-injury logic is as close to an adaptive story as anything in the book. The core froze; the copies were groomed.

So the three answers are not rivals for one throne. They answer three different questions hiding under one word:

A book that dates its claims owes you the ways it could die, and the ledgers have been accumulating them all along. Collect the movers from every chapter and they read as one standing bet against the book's own frame. A confirmed opioid precursor gene in a snail or a worm. A functional dopamine receptor in a ctenophore. A PNMT ortholog in a lamprey, erasing adrenaline's hundred-million-year lag. Serotonin-specific signalling — own synthesis, own selective receptor, blockable behaviour — in any non-bilaterian. A cloned basal-bilaterian channel that answers to GABA itself. A shared ancestral histamine-binding motif, promoting the control group into the frozen cast. An oxytocin-family gene in a jawless fish. Any one of these lands, and a chapter of this book gets rewritten. That is not a weakness of the method. It is the method — a ledger that cannot lose is not a ledger, and the last column exists precisely so that someone, someday, can fill it in against us.

The code is still running. It ran while you read this: glutamate carrying the sentences, GABA shaping them, dopamine deciding whether you would finish the chapter, and a histamine tide already scheduling tonight's sleep. Four billion years of commits, and not one of the authors knew what the codebase would someday compute. Nature never documents her code. She leaves that to the latest processes running on it — the first ones in the whole history of the repository able to read the log, grade it, and sign their names to the places where they might be wrong.

Overturned if: the bets start landing together — if receptors and precursors keep turning up below every floor this book has drawn, then the cast was never frozen, only thinly sampled, and the right title for this volume would have been not The First Commits but The Commits We Had Found So Far.

Sources

• 2023 · Goulty, Botton-Amiot, Rosato, Sprecher & Feuda, "The monoaminergic system is a bilaterian innovation" — Nature Communications
• 2020 · Huang et al., genome-wide search finding no confirmed protostome opioid genes (via ch. 09, where the contested claims are graded)
• 2009 · Gwee, Tay, Brenner & Venkatesh, "Characterization of the neurohypophysial hormone gene loci in elephant shark and the Japanese lamprey" — BMC Evolutionary Biology (via ch. 08 — the vasopressin-first result)
• 2024 · Edens, Stundl, Urrutia & Bronner, "Neural crest origin of sympathetic neurons at the dawn of vertebrates" — Nature (via ch. 06)
• The adjudication table and the standing bets compress the chapter ledgers, which carry the full sourcing; no claim appears here that a family chapter does not source.