Leopard Frequency

The skull on Naomi Osei’s desk was a replica, but the puncture marks were accurate to the tenth of a millimeter. Two holes in the parietal bone, 36.2 millimeters apart — the distance between the upper canines of a leopard, recorded in the cranium of a Paranthropus robustus who died in what is now South Africa, three and a half million years ago. Naomi had measured the replica herself with digital calipers on her first day at the lab, and the number had stayed with her the way certain numbers do: not as data but as an address. 36.2 millimeters. The space where a predator’s mouth met a primate’s brain. The place where fear was invented, or discovered, or first made necessary.

She kept the skull between her monitor and a photograph of her mother, who had died of a stroke at fifty-nine after a decade of sleeping no more than four hours without waking to check the locks. Her mother’s doctors had called it generalized anxiety disorder, which Naomi understood now was like calling a forest fire “generalized combustion.” The diagnosis named the symptom and missed the architecture entirely.

The architecture was what Naomi had spent eleven years mapping. Not the amygdala itself — everyone knew the amygdala — but the specific frequency at which it fired when processing predator threat versus social threat versus abstract threat. Her work showed that these were not the same signal processed differently, as the modular model suggested, but genuinely different signals stacked on a common carrier. The oldest signal — predator threat, the leopard — ran at the base frequency. Social threat modulated on top of it. Abstract threat, the modern kind, rode on top of that. Three layers. Three evolutionary epochs. One circuit.

The breakthrough was not the mapping. The breakthrough was the separation.

The therapy was called amygdalar frequency isolation, or AFI, and it worked by introducing a synthetic peptide that selectively dampened the carrier frequency — the predator layer — without disturbing the modulations above it. Naomi had published the mechanism in Nature Neuroscience in January, and by March she had approval for a Phase I clinical trial with forty-two participants, all diagnosed with anxiety disorders that had resisted conventional treatment.

She had a replica of SK-54 made for the intake room. Patients sat across from it during their initial consultation, and Naomi would pick it up, turn it in her hands, point to the puncture marks. “This is what we’re treating,” she’d say. “Not your anxiety. Your leopard. The circuit that makes you scan for threats that aren’t there — the reason you can’t sleep, the reason your chest tightens on the subway, the reason you check your phone forty times a day for bad news. That circuit is three and a half million years old, and it is responding to a predator that no longer exists. We’re going to turn it down.”

She believed every word. The trial data confirmed her belief so completely that it should have made her suspicious.

The first patient to complete the twelve-week protocol was a woman named Dara Josephs, thirty-one, a paralegal from Southeast D.C. who had not slept more than five hours in three years. Dara described her anxiety as a humming — not a metaphor, she insisted, an actual auditory sensation, like a fluorescent light left on in a room you can’t find. She had tried SSRIs, benzodiazepines, cognitive behavioral therapy, EMDR, meditation, weighted blankets, magnesium supplements, and a sound machine that played recordings of rain forests. Nothing touched the hum.

After the fourth AFI session, the hum stopped.

Naomi watched the change in Dara’s bloodwork, her cortisol curves, her sleep architecture. The numbers were not merely improved; they were transformed. Dara’s nighttime cortisol dropped from 18.2 to 4.1 mcg/dL. Her REM latency normalized. Her galvanic skin response to standardized threat stimuli — images of predators, heights, oncoming traffic — flattened to near-baseline. She was, by every measurable index, a person who had been cured.

“I didn’t know,” Dara said at her eight-week follow-up. She was sitting in the consultation chair, and the skull was on the table between them, and Dara was holding it. She held it the way you might hold a seashell — an interesting object, a curiosity, nothing more. “I didn’t know there was this much quiet in the world. I thought the noise was the world. I thought everyone heard it.”

“Most people do,” Naomi said. “They just have different volumes.”

“I keep trying to explain it to my sister, and she looks at me like I’m describing a color she can’t see.” Dara set the skull down gently, centered it on the table like a vase. “She asked if I was afraid it would come back. I said no. She asked if I was afraid of anything. I had to think about it. I said, ‘I’m afraid of being unkind.’ She didn’t know what to do with that.”

Naomi noted the exchange in Dara’s file under qualitative response. The warm feeling in her chest was not, she told herself, personal satisfaction. It was the recognition of a hypothesis confirmed.

By week twelve, all forty-two patients had completed the protocol. Thirty-nine showed significant improvement. Three showed no change, which was within expected parameters for peptide delivery variance. The thirty-nine were sleeping. They were working. They were driving without gripping the steering wheel until their knuckles whitened. Several of them reported happiness for the first time in memory, although happiness was not, Naomi reminded them, a clinical endpoint.

She submitted the Phase I results and began designing Phase II.

One detail troubled her at the margins, small enough to dismiss. A patient named Reva Okonkwo, thirty-seven, who had entered the trial because she could not ride the Metro without panic attacks, mentioned during her twelve-week appointment that she had stopped flinching when her husband raised his voice. She said this cheerfully. She said it the way you describe the resolution of a symptom — the rash is gone, the cough has stopped, I no longer flinch when my husband shouts.

Naomi asked, carefully, whether the husband shouted often.

“Oh, he has a temper,” Reva said. “But it doesn’t bother me now. Before, my whole body would tense. My hands would go cold. I’d lose twenty minutes afterward just trying to breathe. Now I hear him yelling and I think, he’s upset, the way you’d think it’s raining. A fact about the weather.”

“Has his behavior changed?”

“No. But my response to it has. Isn’t that what we wanted?”

Naomi wrote a note in the file: Patient reports diminished startle/freeze response to domestic verbal aggression. Social cognition intact. Behavioral response: observation without protective withdrawal. She stared at the note. Reva’s social circuits were reading the room correctly. She knew her husband was angry. She simply couldn’t produce the body’s ancient instruction to make herself small, to go quiet, to locate the nearest exit. The leopard had carried that instruction.

She added the note to the follow-up database and did not flag it.

The first anomaly she did recognize surfaced in the six-month follow-up data, hiding inside all the data that looked like health.

Patients reported increased trust. They described strangers as “less threatening” and social situations as “lighter.” Three patients in therapy for codependency reported their therapists were concerned — they were being “too open” with new acquaintances, sharing personal information too readily, failing to maintain what the therapists called “appropriate boundaries.” One patient, a man named Solomon Garrett, had lent eight thousand dollars to a coworker he’d known for three weeks. He did not seem distressed about this. He described the coworker as “someone who needed help,” and when Naomi asked whether he’d considered the possibility of being exploited, he gave her an expression she couldn’t categorize.

“Why would I consider that?” he asked. Not defensively. Genuinely. As if the concept had been presented to him in a language he’d once spoken and had since forgotten.

Naomi administered the standard social-threat battery — images of angry faces, audio clips of raised voices, scenarios describing betrayal and exploitation. Solomon’s responses were within normal range. His social fear circuits were functioning. He could recognize anger, identify deception in hypothetical scenarios, predict the behavior of hostile actors in narrative vignettes.

But he no longer felt any of it.

The distinction was critical, and Naomi spent two weeks trying to understand it before the answer arrived, not from the data but from a conversation with her postdoc, Jamie Whitelaw, who said: “It’s like he can see the color red but it doesn’t mean stop anymore.”

“That’s not how color works.”

“It’s not how fear works either. Or it isn’t supposed to be. But you separated the signal from the carrier, and the carrier was providing something the signal can’t provide on its own. Salience. Urgency. The this matters part of the message. He can still detect social threats — his neocortex is doing that — but without the predator frequency underneath, the detection doesn’t translate to avoidance. He knows intellectually that the coworker might not pay him back. He just doesn’t care. Not in a reckless way. In a genuinely unbothered way, like it doesn’t have enough weight to change his behavior.”

Naomi looked at the skull on her desk. The puncture marks. 36.2 millimeters.

“You’re saying the predator frequency was providing emotional weight to all threat processing,” she said. “Not just predator threats.”

“I’m saying it was providing the flinch. The part that comes before conscious evaluation. You see something wrong and your body moves before your mind decides. That’s the predator layer. It’s not processing the social information — it’s providing the substrate on which all threat information becomes actionable. You removed the substrate.”

“The social circuits are intact.”

“The social circuits are running on empty. They’re still computing, still producing outputs, but the outputs don’t connect to the motor system, to the gut, to the part of behavior that makes someone actually change what they’re doing. You silenced the carrier wave. The audio is still playing in their prefrontal cortices. It’s just not reaching their bodies.”

Naomi did not sleep well that night. Her own predator frequency was entirely intact.

The disappearances did not begin dramatically. They began with addresses.

Dara Josephs moved in the ninth month, leaving her apartment in Southeast D.C. for a collective house in Columbia Heights that she’d found through a friend of a friend. The collective was composed of six people who shared cooking, rent, and what Dara described as “a willingness to be present with each other.” Three of the six were AFI patients. The other three were not, but they had been drawn into the household by the deep calm that the treated patients projected — a gravitational pull that Naomi, reading Dara’s update, recognized with growing unease as the social consequence of absent wariness.

Untreated people were attracted to the treated ones. Of course they were. A person without the ambient tension of threat-readiness felt safe to be around. Their body language was open. Their voice was unhurried. They listened without the micro-flinches of a nervous system scanning for danger in every social encounter. The calm was contagious — mirror-neuron entrainment, Naomi’s discipline called it. She understood the mechanism and was not comforted by it.

In the eleventh month, two patients stopped responding to follow-up calls. Not all at once — the first, a retired teacher named Pauline Morrow, missed one appointment, then another. When Jamie drove to her listed address, a neighbor said Pauline had gone to stay with “some people up in Baltimore” and that she’d seemed fine, happy even, and had given the neighbor her television before leaving.

The second was Solomon. His phone had been disconnected. His apartment had been vacated, the lease broken without notice. He had left no forwarding address, but a colleague at his former workplace said he’d mentioned “a place where people live without walls,” and when pressed had only smiled.

Naomi called every patient in the trial. Twenty-nine answered. Of the remaining ten, six had moved without leaving contact information, three had disconnected their phones, and one — a twenty-six-year-old graduate student named Kit Delacroix — had transferred everything in their bank account to a GoFundMe for a stranger’s medical bills and then, according to their roommate, walked out of their apartment with nothing but a coat and a water bottle and had not come back.

“Where did they go?” Naomi asked the roommate, and the roommate said, “They said they were going to find out what was on the other side of being afraid,” and Naomi sat in her car in the parking lot of Kit’s apartment complex for twenty minutes with the engine running, feeling in her chest the precise sensation that her patients could no longer feel: the old flinch, the leopard’s signature, the body’s ancient and wordless insistence that something was wrong.

She found the house in Baltimore by following Pauline Morrow’s neighbor’s directions to a street in Waverly where three rowhouses had been connected by removing interior walls. The front door was open. Not unlocked. Open. Propped with a brick, in March, letting the damp air in.

Inside, fourteen people were living in a space designed for three families. Some were AFI patients; she recognized Solomon sitting cross-legged in what had been a dining room, reading a paperback with his shoes off. Others she didn’t know. They ranged in age from early twenties to late sixties, and they moved through the connected rooms with the ease of people who had lived together for years, though Naomi doubted any of them had been there for more than a few months.

Nobody locked anything. There were no individual bedrooms — just sleeping areas with mattresses arranged in loose clusters. Food was communal, shelved in an open kitchen. Wallets, phones, and keys sat in a bowl by the front door. The bowl was overflowing.

A woman Naomi didn’t recognize offered her tea.

“I’m looking for Pauline Morrow,” Naomi said.

“Pauline left last week. She went south.”

“Where south?”

The woman considered this with the untroubled concentration of someone who had genuinely never needed to track another person’s movements. “She said she wanted to see the ocean. I think she took the bus.”

“She just left? Alone? Without telling anyone where she was going?”

“She told us she was going. She told us when she left.”

“But not where.”

“Why would she need to tell us where?” She was not being evasive. She was asking a real question. The concept of needing to know — the monitoring impulse, the background hum of what if something happens — had departed from her entirely. She was standing in an unlocked house with fourteen strangers, in a neighborhood where the adjacent block had three boarded windows and a stripped car, and she was offering tea with the unfortified hospitality of someone who had never been given a reason to expect harm.

Naomi drank the tea. It was good — loose-leaf, steeped correctly, served in a clean mug. The house was clean. The people were healthy. Nobody seemed coerced or confused or impaired. They touched each other’s arms when they passed, made eye contact that lasted a beat longer than social convention permitted. The neurologist in Naomi catalogued this: without predator-response dampening social approach behavior, the mammalian drive toward affiliation would express itself more freely. Grooming behavior. Pair bonding. Proximity maintenance. All elevated.

But the door was open and the neighborhood was not controlled, and Naomi counted the people in the bowl-by-the-door’s worth of wallets and came up with more wallets than people in the house and could not make herself ask whose wallets had no owners anymore.

Solomon saw her looking and came over. He was thinner than she remembered. Not gaunt — lean, the way people get when they stop eating from anxiety and start eating only when hungry.

“Dr. Osei.”

“Solomon. How are you?”

He sat down across from her at a table that had been pushed against the wall to make room for floor cushions. “I know why you’re here. You’re worried the treatment broke something.”

“I’m concerned about the follow-up data.”

“The data is fine. We’re fine. The people who left — they’re doing what people do when they’re not afraid. They go places. They meet people. They don’t report back to their doctors because they’ve forgotten that they’re patients.” He leaned forward. “You know what the strangest part is? I can still feel anxious. When I’m about to give a presentation, my stomach tightens. When someone’s angry with me, I feel it. The social stuff is all there. I just don’t feel hunted anymore. And I didn’t know I felt hunted until it stopped.”

“Solomon, Pauline left without telling anyone where she was going. Kit Delacroix gave away their entire savings and walked out of their apartment. That’s not — ”

“That’s not what? Not normal? Normal is sitting in an apartment with three locks on the door watching the news until you’re too exhausted to be afraid anymore and calling that sleep. Normal is walking past a person sitting on the sidewalk and feeling a flash of danger instead of a flash of recognition. I lived that way for thirty-two years and it was eating me alive.”

“Normal is also checking that the person sitting on the sidewalk isn’t going to follow you home.”

He studied her. She could see his social cognition working — he was reading her face, her posture, the tension in her jaw. The social circuits were intact, just as the data had shown. He knew she was afraid. He simply couldn’t produce the same feeling in himself.

“You think we’re being naive,” he said. “You think the world is going to hurt us because we can’t see it coming. And maybe you’re right. Pauline might get on the wrong bus. Kit might trust the wrong person. But Dr. Osei — ” He paused, and his face did something that was not quite a smile and not quite grief. “I used to cross the street when I saw a young man in a hoodie. Every time. I’m a Black man crossing the street to avoid a Black teenager because the leopard couldn’t tell the difference. You took the leopard away and I stopped crossing. Tell me that’s a disease.”

Naomi had no answer for that. It was not, she realized, a rhetorical question. He was asking her to diagnose something, and the thing he was asking her to diagnose might actually be health.

She returned to the lab and pulled the twelve-month data and found what she had known she would find, laid out in the clean columns of a spreadsheet like a body on a table.

Of thirty-nine patients who had responded to treatment: twenty-three were still contactable and functioning well by every clinical measure. Their anxiety scores were near zero. Their sleep was normal. Their social functioning was, by standard metrics, improved. They held jobs. They maintained relationships. They reported satisfaction with their lives at rates significantly above the general population.

The other sixteen had entered what Naomi reluctantly labeled drift. They had moved — usually toward other treated patients, forming the kind of household she’d seen in Baltimore. They had reduced their possessions. They had become, in the language of her discipline, increasingly affiliative and decreasingly avoidant. None of them showed signs of psychosis or mania. None of them were using substances. They were lucid, coherent, and capable of explaining their choices in terms that were, within their own framework, perfectly rational.

They were also becoming harder to find.

The demographics of the drift population told a story that Naomi could have predicted if she’d been paying attention to anything other than her peptide curves. Of the sixteen who had drifted: eleven were Black or Latino. Fourteen lived in neighborhoods where the median household income was below $40,000. Twelve had experienced at least one violent crime in their lifetime.

The seven treated patients who lived in Northwest D.C. or the Virginia suburbs had not drifted. They had better sleep. They exercised more. They had, as one of them told Naomi, “finally stopped worrying about everything.” They were thriving in the way that people thrive when the world around them is already organized for their safety — when the locks are decorative, the streets are lit, the neighbors know your name, and the police respond to your calls.

The patients for whom the locks were not decorative — for whom the streets were not reliably lit, the neighbors not always known, the police not always helpful — those patients had lost the only defense system that ran on internal power. Their wariness had been environmental, calibrated by experience, earned through the body’s education in how the world actually worked. The treatment hadn’t distinguished between irrational fear and rational caution, because the frequency was the same. The leopard didn’t care whether the threat was real. The leopard just flinched.

Naomi wrote this up in a memo to her IRB, and then she deleted it and wrote it again, and deleted that version too, because every version sounded like an argument for putting the fear back. And she could not, sitting in her office with the skull and the photograph and the data, bring herself to argue that the correct clinical response to a treatment that eliminated suffering was to reverse the treatment because the world outside the clinic door was insufficiently safe.

That was not a medical argument. That was an indictment.

She drove back to Baltimore three weeks later and the house in Waverly had seven people in it instead of fourteen. The woman who’d given her tea was gone. Solomon was gone. A new person she didn’t recognize was sitting on the porch, looking at the street with the wide-open gaze of someone watching weather.

“Where did they go?” Naomi asked the remaining residents, and the answers were all variants of the same shape: south, west, to a friend, to a place someone had told them about, away. Not fleeing. Dispersing. The way seeds disperse — carried by whatever wind is available, landing wherever the soil will accept them.

She found Kit Delacroix through an accident of social media. Someone had posted a photograph of a community garden in Richmond, and in the background, kneeling in the dirt with their hands in the soil, was Kit. Naomi drove to Richmond. The garden was in Church Hill, in a lot between two abandoned houses. Kit was not there, but a man tending the squash said Kit had stayed for about a month, had been “the calmest person I’ve ever met,” and had left heading west with someone they’d met at the garden.

“Were they okay?”

“They were great. Happiest person here. We all wanted to be around them.” He pushed his cap back and squinted at her. “Kit had this thing — they’d talk to anyone. Didn’t matter who. There was a guy hanging around the garden for a few days, and the rest of us could tell something was off. You know the feeling. Hairs on the back of your neck. Kit walked right up and offered him water, sat with him, talked for an hour. Nothing happened. The guy left. But I kept thinking — what if the next guy is different? Kit wouldn’t know. Or they’d know but it wouldn’t register the way it’s supposed to.”

“Did you say anything to them about it?”

“I tried. They said, ‘I can hear you’re telling me to be careful. I just can’t feel what you’re feeling when you say it.’”

“Did you worry about them?”

The man straightened up. “Yeah. I worried. They didn’t seem worried, though. Not about anything.”

She drove home through the spring landscape of Virginia, and the dogwoods were blooming, and she thought about the Taung child — the skull of a three-year-old Australopithecus africanus, found in 1924, with puncture marks from the talons of a large raptor. An eagle had picked up that child and carried it away and eaten it, and the bones had survived for three million years, and the child’s mother had presumably survived too, and had lived the rest of her life watching the sky.

How many hours of that mother’s remaining life had been spent watching the sky instead of looking at the food, the ground, the faces of her living children? The eagle had taken one child. The fear of the eagle had stolen what remained of the others — not their lives, but their mother’s attention, her sleep, her capacity for anything other than vigilance.

That was the case for the treatment. Not the clinical data. The mother, watching the sky forever. An existence organized around the anticipation of predation.

Except the mother was alive. The flinch kept her moving, kept her children in her sight, kept her sleeping with one ear toward the dark. The eagle had taken one child. The fear of the eagle had kept the rest.

Naomi drove and could not hold both versions at once. You were watching the sky or you were looking at the dogwoods. You were locking the door or you were leaving it open.

At fourteen months, Pauline Morrow was found dead.

Not murdered. Not assaulted. She had died of exposure in a state park in the mountains of West Virginia, where she had been hiking alone, in late October, without adequate gear. The ranger who found her said she had walked off the marked trail and continued walking, apparently without distress, for six miles into terrain that became progressively steeper and more remote. She had not built a fire. She had not attempted to retrace her steps. She had sat down under a hemlock tree and died in the night when the temperature dropped to twenty-eight degrees.

She was sixty-three. She had a heart condition that made hypothermia more dangerous. She had known about the heart condition. She had known about the cold. What she had not felt — what she could no longer feel — was the body’s screaming insistence that cold plus darkness plus solitude equaled threat. The sequence of evaluations that a functioning amygdala would have produced — this is dangerous, turn back, find shelter, the dark is coming — had been silenced. Not the knowledge. The urgency. She knew, in the way that one knows a fact, that she should not be alone in the mountains at night with a bad heart. But knowing and flinching are different operations, and the flinch was gone.

Naomi sat in the office of the Chief of Institutional Review and listened to herself explain the mechanism, the carrier wave, the separation of signal from substrate, and heard the words come out of her mouth in the precise, ordered sequence of a scientist describing a system she understood, and felt the gap between her understanding and Pauline’s body like a seam in the world where the air was a different temperature on each side.

The IRB suspended the trial. Naomi did not argue. She drove home and sat in her kitchen and called the twenty-three patients who were still answering their phones and told them the trial was suspended and that she would like to schedule follow-up appointments to discuss options.

Eighteen agreed. Five said they did not need follow-up. They were fine. They had never been better. One of them, a financial analyst from Arlington, said: “Dr. Osei, I understand that something terrible happened to that woman, and I’m sorry. But I’m sleeping through the night for the first time since I was eleven years old. I’m not giving that back. You’d have to hold me down.”

“I’m not asking you to give anything back,” Naomi said. “I’m asking to monitor you.”

“Monitor what? My doorlocks? Whether I check for cars before I cross the street? I still check. I still lock. I’m not in the group that doesn’t.”

No. He wasn’t. He lived in Arlington, in a three-bedroom colonial with a lawn service and a mortgage he could afford. His fear had been irrational — the leopard firing at tax deadlines and climate headlines and nothing that could actually hurt him in the space between his car and his front door. Removing the leopard had, for him, been an uncomplicated good.

Pauline’s fear had not been irrational. Pauline’s fear had been the only thing standing between her body and the world’s actual indifference to whether she survived. The treatment couldn’t tell the difference because the frequency was the same. Three and a half million years of fidelity. The leopard in the dark of the Pliocene and the cold in the mountains of West Virginia singing the same note, and Pauline unable to hear either one.

Jamie Whitelaw brought her a coffee the morning after the IRB meeting and said, “Have you considered that the treatment works perfectly and the problem is the world?”

Naomi looked at him over the rim of the mug.

“I’m serious. The problem isn’t the treatment. The problem is that we still live in a world that contains leopards. They just look like hypothermia, and bad neighborhoods, and men who can spot someone without defenses from across a room.”

“Pauline is dead.”

“The sixteen patients in Arlington and Silver Spring are sleeping for the first time in decades. Those are outcomes too.”

“And the sixteen who drifted?”

Jamie didn’t answer that. He took his coffee and went back to the bench. Naomi sat with the spreadsheet. Twenty-three functional, sixteen drifting, one dead. The distribution falling along the same fault lines as every other distribution — income, geography, race, the gradient of safety that determined whether openness was a luxury or a liability.

She had not invented that gradient. She had not built the world in which Pauline’s neighborhood required locks and the analyst’s neighborhood didn’t. But she had introduced a treatment that interacted with the gradient as if it didn’t exist, and the gradient had responded the way gradients always do — sorting outcomes along its slope, concentrating harm at the bottom, benefit at the top.

That was not a side effect. That was the main effect, operating in a context she had excluded from her model.

Six months after the trial suspension, Naomi received a photograph by mail. No return address, postmarked from Tennessee. A small house with a garden, and in the garden three people she recognized — Solomon, Kit, and a former patient named Aisha Kendrick whom she had last spoken to at month nine. They were planting something. Beans, maybe. They were not looking at the camera.

On the back of the photograph, in Solomon’s handwriting: We are not lost. We just don’t need to be found.

She pinned the photograph to her corkboard next to the data printouts and the IRB suspension notice. Three people in a garden, unafraid, in a place where they’d arrived by drifting. They looked healthy. They also looked like people who had left their front door open in a country where that choice carried different costs depending on the color of the hand that opened it.

Naomi began designing a revised protocol. Not a reversal — she would not put the leopard back. A calibration. A way to restore the urgency without the suffering, to let Dara sleep through the night and still lock her door. She worked on it for four months and made no progress, because the frequency was the frequency. You could not have the flinch without the fear, and you could not have the fear without the insomnia, the chest-tightening, the vigilance that stole hours and years from the people it kept alive.

On a Tuesday in November she found herself at her desk at eleven p.m., running the same failed models, and she realized she had not eaten since noon and that her hands were shaking and that the fluorescent light above her bench was humming at a frequency she had never noticed before, or had always noticed and had only now identified. She turned it off. The hum stopped. She sat in the dark and thought about Dara, who had described her anxiety as a hum in a room you can’t find.

She turned the light back on.

The data said she had cured thirty-nine people and that seventeen of them had walked into the world without armor. The data did not say what happened to Solomon in Tennessee, whether Kit had found somewhere to land, whether the garden in the photograph was still there or had been sold out from under them. The data ended where the follow-up ended, which was where the patients stopped answering their phones.

Naomi pulled up the revised protocol and changed a variable and ran the model again. It failed again. She changed another variable.