In her last summer, the keepers put a rope around Martha's cage. Not to keep her in. She was old, and alone, and there was nowhere for a passenger pigeon to go anyway. The rope was to keep the visitors back. People came to the Cincinnati Zoological Gardens to see the last one, and when she sat still on her perch, some of them threw sand at her to make her move.
She was the last known passenger pigeon on earth. Her wings drooped. She trembled. A male named George had shared the cage until he died in 1910, and after that she had no one of her kind. A thousand-dollar reward for anyone who could bring her a mate went unclaimed, because there was no mate. She died in the early afternoon of September 1, 1914.
Dead, she was worth more than she had been alive. Her body was packed in a three-hundred-pound block of ice and sent by train from Cincinnati to Washington. The man who dissected her wrote that she arrived in fine condition, a few tail feathers gone. He decided not to cut open her heart. He wanted to keep it whole: the heart, he wrote, of the last of her kind the world would ever see alive. Her skin went onto an artificial body, probably wire, shredded wood, and tightly wound string. Her organs went into jars of alcohol on a shelf.
A hundred years before that shelf, this was the most abundant bird in North America, perhaps forty percent of every bird on the continent. A naturalist riding toward Louisville in 1813 tried to count the passing flocks, marked a hundred and sixty-three in twenty-one minutes, and gave up. He wrote that the noon light went out as in an eclipse, that the droppings fell like melting snow, that the buzz of the wings nearly put him to sleep. In 1871, one nesting in Wisconsin held an estimated one hundred and thirty-six million birds across eight hundred and fifty square miles of forest. The roosts were heavy enough to break the limbs off the trees.
Then men with guns and railcars and telegraph wires learned where the flocks were, and shipped the birds to the cities by the barrel, and cut the nesting forests down. It took about fifty years. The sky emptied. What remained of the most numerous bird on the continent was one trembling animal in a cage, and a crowd throwing sand to watch her walk.
A lab can now plausibly promise something close enough to call her back.
Not her, exactly. But there is a standing proposal to rebuild the passenger pigeon by reading its genome out of preserved museum specimens, editing the differences into its nearest living cousin, the band-tailed pigeon, and breeding toward the lost bird. Millions of dollars, at least a decade. And it belongs to a larger moment. A company that this spring added a vanished blue antelope to its roster, that edits elephant cells toward a woolly mammoth, whose chief executive says a living dodo is five to seven years off, has raised more than half a billion dollars and reports a valuation above ten billion. The work is real. It is not a movie.
And that is the problem. The danger in de-extinction is not that it will fail. The danger is that it will partly work.
Because partial return is a worse teacher than clean loss. It lets you mourn without ever finishing the sentence. It is gone. If some version of her can be made, then she is not gone, only away, and away is a thing you wait out instead of grieve.
Start with what the machine can actually do. Nobody has a living cell from Martha. What survives is DNA, and DNA in a dead thing does not keep. It breaks into fragments a few hundred letters long, and the letters themselves corrode. So the work is not copying. It is reconstruction. You read the torn text, lay it against the genome of a living cousin, guess at the missing words, and edit some of the differences into the living bird's cells. What you get carries some of the old traits. It does not carry the old animal.
How wide is the gap? Scientists took a recently extinct rat, one with a close living relative and well-preserved DNA, and sequenced it deeply, sixty times over. Even then, about five percent of the genome could not be recovered at all, and twenty-six genes were simply missing. The gaps were not random. The genes for smell and for immune defense were hit hardest, which are among the genes an animal most needs to survive in a real place. One geneticist said it plainly: there will always be some kind of hybrid. Another said the tools bring back extinct traits, not extinct species.
There is one case where it went past traits. A wild goat called the bucardo went extinct in 2000, when the last female was crushed by a falling tree; they found her by the beep of her radio collar. But a year earlier someone had frozen a scrap of her skin. So they tried. They built two hundred and eighty-five embryos, then more, and put them into surrogate goats. Seven pregnancies. One went to term. A cloned bucardo was delivered by cesarean, a normal-looking female, genetically the last one over again. She had a lung defect, a solid extra lobe. The scientist held her while she fought to breathe, her tongue pushed out of her mouth, and she died about ten minutes after she was born.
The animal is real enough to hold, and real enough to lose.
So a kind of animal came back and went extinct again in ten minutes. That is the honest picture of partial return.
The apparatus that makes an ending feel like an option has grown large and strange. There is a freezer in San Diego, running since 1975, holding living cells and gametes at three hundred and twenty degrees below zero Fahrenheit, some from animals no longer alive: a genetic line kept on as an option, outlasting the body it came from. There is a lab in Dallas, fifty-five thousand square feet, whose address the company will not let a reporter print and where no photographs are allowed, its nuclear-transfer team running two to three hundred transfers a day to train machines to run more, so that making a proxy becomes a process rather than a miracle. There is a silicone egg with a clear window, built so you can watch a bird take shape without a shell. Each object moves a death one step further from final: stored, scaled, and put behind glass to be watched.
The strongest case for these tools is not stupid, and on the facts it is partly right. The same tools that stage a proxy can rescue the living. The clearest case is a small North American weasel, the black-footed ferret. Every one alive, until recently, descended from just seven animals, and the inbreeding was a slow death sentence. Then they cloned a female from cells frozen in 1988, cells that carried three times more unique variation than the living population's average. In 2024 one of those clones bred naturally and had two healthy kits. That is not spectacle. That is a real gene pool getting deeper, a living species pulled a step back from the edge. The people doing it are careful to say the cloning does not replace habitat, or reintroduction, or the plain grinding work of keeping animals alive in a place.
It is help for the living. It is not the reversal of death. The ferret's kind was still here. Willa, whose cells they thawed, had been dead for decades; her species had not. The cells went back into a population that still breathed and bred and dug burrows. No one raised a species from the dead, because the species had never died. That is a different promise from the one being sold when a company puts a lost animal on a list and a headline says it is coming home.
So the tools cut both ways, and the line between the cuts is money and attention. They help when they are additive, when they bring in dollars or notice that would never have reached a living animal otherwise. They cost something when they do the reverse, when the spectacle of the dead pulls money and attention off the species still here. And there is a lot still here to pull it from. Reversing the current wave of extinction would take something like six to eight hundred billion dollars a year, and the money is not there. Nearly fifty thousand species sit on the threatened list right now. Emperor penguins were just moved to endangered. A population of Antarctic fur seals fell by more than half in about twenty-five years, from over two million to under a million. One study of the tradeoff found that public money spent on resurrected species would leave fewer living species saved, not more. One company selling the return of the dead reports a valuation above ten billion dollars. The fur seals, halved, got no valuation, no countdown, no story about coming home. The dead are photogenic. The dying are not.
Grief has a job. Not the private kind, the sort you feel for a person, but the public kind: the shared story a culture tells itself about whether a vanished animal is gone for good or merely out of stock for a while. That story is how a public accepts a loss as final, sets it down, and carries the weight of what it did. It only works if the loss stays final. A loss you might undo is not something you grieve. It is something you keep a tab open for. And what de-extinction offers, precisely because it partly works, is a permanent open tab. Not the bird, but the option on the bird. Not her, but the idea that some version of her is recoverable, filed under later, next to the mammoth and the dodo and the blue antelope no living person ever saw.
A scientist at the museum that keeps Martha's body said the thing most worth keeping. A passenger pigeon in a glass, he said, even if it were possible, would still be just a passenger pigeon in a glass. The bird was never only its genes. It was a hundred and thirty-six million birds in one forest, a sound like a gale in a ship's rigging, a weight that snapped branches, a way of being a pigeon that a chick learned from a whole sky of other pigeons. None of that is in a cell. None of it comes out of a surrogate hen. You could hatch the traits and still not have the thing that put out the noon light, because the thing that put out the noon light was the number, and the number is what we killed.
So keep the rope in mind. People came to see the last passenger pigeon and threw sand to make her move. We are close to being able to make something enough like her move again. The trouble is not that it will not work. The trouble is that it might work just well enough that we never have to feel the sky go quiet, and never have to ask what we did to empty it.