She came back to herself in the east room.

This had not happened before. She had always thawed in the study — the desk, the ceiling with its named water stains, the particular light of the season she had come back to. She had records of every thaw location across five cycles, documented in the transition journals as a minor data point she tracked for completeness rather than significance. The east room was a departure.

She registered it without alarm. The east room was one of the rooms she spent most time in; it was not surprising that her body had found its way there in the moments before consciousness reconsolidated. She noted it in the transition journal — *thaw location: east room secondary desk, atypical, cycle transition note* — and then she looked up.

The room was different.

Not subtly different. Differently different from the way the house was always different when she returned to it — the ordinary accumulation of a decade's modifications, the furniture positions adjusted, the surfaces changed by use and time. This was a different quality of difference: present everywhere simultaneously, as if the room had not been modified in specific places but had been reconfigured around a new set of principles. She could feel it before she could see it, the way she felt the quality of light before she identified the source.

She stood up from the secondary desk.

The secondary desk had a new lamp on it. She had noted this lamp in a previous cycle — it had appeared in Cycle Three, positioned at the angle she used for close work, selected by someone who had been observing how she worked. The lamp itself was the same. What was different was the way the room organized itself around it: the new placement of the secondary shelving unit, moved to reduce visual obstruction from the reading position; the removal of the rug she had placed in Year Four of the previous cycle, which she had mentioned, once, in a letter written while discussing something else entirely, as slightly wrong for the proportions of the room. She had mentioned it once. He had removed it.

She stood in the east room and looked at the removed rug's ghost on the floor — the slight difference in the wood's surface color where the rug had been — and felt something she could not classify as data.

She went to the library.

The library held three new books on the shelf she used for current work. She pulled the first one: a volume on the relationship between spatial continuity and temporal cellular maintenance, published during her suspension decade, by a researcher she had cited in a footnote as pursuing adjacent work. She had been working toward the same problem from a different angle. She had not known this researcher had arrived there first. She pulled the second: the rebuttal to her footnote argument that she had found in the previous active decade, which she had read and which had changed the direction of her current work. It was on the shelf. She had not placed it there — she had her own copy, worn and annotated, in the research cabinet. This was a new copy. A second copy. He had put a second copy on the shelf where she would encounter it in the current research arc, at the point in the sequence where she would be likely to need it again.

She put the book back.

She went to the kitchen.

The kitchen light was different. The north-facing window had new glazing — she recognized the quality of it immediately, the specific clarity that came from a glass formulation that addressed diffusion rather than simply filtering it. She had a note in her Cycle Three journals about the north window's afternoon light problem. She had mentioned it once, briefly, while describing the house's current light conditions as context for a comment about her research productivity. She had not asked for it to be fixed. She had mentioned it once and he had fixed it with a glazing solution that was more sophisticated than the standard approach, which required either significant research into the available options or the kind of professional familiarity with glass specification that came from working in architecture.

She stood in the kitchen and looked at the corrected light.

Then she checked her wrist.

The fracture had grown.

She measured it with the small measuring instrument she kept in the research cabinet, the one she had been using for the monitoring protocol since Year One. Fourteen centimeters had been the baseline: stable across the previous active decade, the evidence on which she had based her cautious assessment that the protocol was working, or at least not failing. It was no longer fourteen centimeters. She measured it twice, adjusting the angle of the instrument on the second measurement to account for the curvature of her wrist. Both readings were consistent.

Seventeen centimeters.

Three centimeters of progression in a single suspension period.

She sat down on the kitchen floor. Not dramatically — she was not the kind of person who sat on floors dramatically. She sat because the kitchen floor was where she was when she needed to sit, and because the chairs were in the other rooms and getting to them required moving through the house and she was not ready to move through the house yet.

She sat on the kitchen floor and looked at her wrist and thought about the progression rate.

Three centimeters across one suspension decade was a rate she had not seen in any of her case files. The fastest progression she had documented in a living subject was one and a half centimeters per decade, in Case Seven — the researcher whose counterpart had stopped writing, whose cells had continued reaching toward a stimulus that had ceased. She had documented Case Seven's progression with the careful attention of a researcher observing a particularly clear example of the mechanism she was studying. She had not, at the time, been able to imagine that the example would be surpassed in her own cells.

She had not been able to imagine it because she had been running a protocol.

She opened the research notebook — new volume, picked up from the cabinet where she kept her supplies — and wrote, in the transition section: *Day 1, Cycle 5, Year 21. Thaw complete. Location: east room. Fracture measurement: 17cm. Previous measurement: 14cm. Progression: 3cm across one suspension decade. Note: progression rate significantly exceeds previous case file maximum.*

She read this.

She added: *Protocol assessment: ineffective. Fracture progression during suspension period indicates that modification of active-decade behavior is insufficient to address the mechanism. The trigger is not active-decade engagement with stimulus materials. Revised hypothesis: trigger operates during the suspension period itself, through a mechanism not previously documented in the literature.*

She sat with this.

A trigger that operated during suspension would be a different category of problem from what she had been treating. The protocol had been designed on the assumption that the trigger was active-decade exposure — that the cellular response was generated during periods when she was awake and engaging with his letters and his objects and his house. If the cellular response was also occurring during suspension, then the trigger was not exposure. The trigger was the condition of attachment itself, operating independently of behavioral expression, continuing through the suspension period the way the house continued through it — present, waiting, accumulating.

She stood up.

She moved through the house.

Every room stopped her.

This was not a figure of speech. She moved through the study and found the new arrangement of her working materials — reorganized, she determined after several minutes of standing with her hand on the cabinet, to match the progression of her current research arc rather than the archival system she had established in Cycle Two. He had reorganized her research materials into the order she currently needed them in, which required understanding the current state of her research, which required reading not just her publications but the private working logic of the arc she was pursuing. She could not determine how he had known the progression of the private arc. She suspected it was her footnotes again. She suspected that the footnotes of her papers were, to him, a more reliable index of where she was going than the papers themselves.

The library held the three books he had selected and also, on the second shelf from the bottom, a new gap where two volumes had been removed. She checked the gap against her catalogue and identified the missing volumes: two references she had been intending to return to a colleague and had kept for three cycles beyond the expected return date. He had returned them, apparently, at some point during his active decade. He had returned her borrowed books.

He had returned them. In his own time. Without being asked.

She stood at the gap in the shelf.

She thought about the fracture: seventeen centimeters, three centimeters of progression, a rate that exceeded her maximum documented case. She thought about the protocol she had been running for five years, the cold letters and the empty shelves and the stripped correspondence and the careful removal of herself from the material fabric of the house. She thought about the result of the protocol: a fracture that had not decreased, had not stabilized, had grown at the fastest rate she had on record across a period when she had been suspended and had not been engaging with trigger materials at all.

The protocol had been wrong.

Not wrong in its implementation — she had implemented it correctly. Wrong in its premise. She had been treating the cause as behavioral, and the cause was not behavioral. The cause was constitutional: a long-term attachment of the absence-pathway variety so established in her cellular structure that it continued operating during suspension, without behavioral input, without conscious engagement. She was not causing the fracture by reading his letters. The fracture was caused by the fact of the attachment, which continued regardless of whether she engaged with its expression.

She went to the corridor adjacent to the storage room.

The corridor wall was different. She placed her hand on it immediately, because the texture was different under her palm — a different plaster compound, denser, with the specific acoustic quality of a formulation she knew from her research literature. She knew this formulation. She had written about it in the paper he had read, the one on temporal cross-sections and the material properties of stable interfaces. She had written that a plaster formulation with these specific thermal and acoustic properties was associated with more stable cross-section presentations.

He had used her research to modify the wall.

He had used her research, published in a paper he had read, to make the anomaly more stable. He had told her about this in the letter she had received at the end of his previous decade — she had read it, had filed it under the protocol documentation as *external modification, noted, further assessment required.* She had assessed it now. The modification worked: she could feel the difference in the corridor immediately, the specific quality of the air at the interface, the way the thermal gradient was more gradual and more stable than it had been.

She opened the storage room door.

The Zone was present and clear — the strongest, most stable presentation she had observed. The light from Cycle One came through with the quality of the original, not filtered or attenuated the way the Zone had sometimes presented in the previous cycles. The temperature differential was precise and bounded. She stood at the threshold and looked into the past and thought about the corridor wall he had modified with her own research specifications, and thought about the seventeen centimeters on her wrist, and thought about the protocol that had been running for five years and had not been working because it had been designed to address the wrong thing.

She thought: I have been trying to reduce my exposure to him, and I have been suspended for a decade in a house he built for me, and my cells have been reaching toward him in my sleep.

She thought: the protocol was never going to work. The protocol was treating the symptoms and the cause is in the walls.

She thought about the case files. She thought about Case Two — the researcher whose counterpart had stopped writing, whose cells had continued reaching. She thought about her own hypothesis about Case Two: that the fatality had come not from the love itself but from the sustained effort of resisting something large while pretending it was small. She had written this in the unsent letter. She had written it as a hypothesis. She was beginning to think it was a finding.

She stood at the threshold of the Zone and looked into the light of Cycle One.

The fracture was seventeen centimeters. The protocol had not worked. The house around her was the most complete record of another person's sustained attention to her existence that she had ever encountered, and her cells were reaching toward the person who had built it with the same consistency they reached while she was awake, because the reaching was not a choice she was making. It was what she was.

She was, she understood with the full weight of someone who had spent fifteen years studying this mechanism in other people, Case Ten.

She stepped back from the Zone.

She closed the door.

She stood in the corridor with the modified plaster under her hand and the seventeen-centimeter fracture on her wrist and the understanding that had been approaching across five years of protocol documentation and had finally arrived, complete and irrefutable and organized around a single question she had been avoiding for forty years: not *can I survive loving him* but *what kind of survival am I trying to have?*

She went back to the study.

She opened the research notebook and wrote, in the section she had been using for protocol documentation, which she was now, she decided, going to stop calling protocol documentation: *Day 1, Year 21. Assessment complete. Previous hypothesis incorrect. Correct hypothesis: the cost of this attachment is a fixed cost, not a behavioral one. It does not decrease with reduced engagement. It may decrease with resolved engagement. This is not a variable I have tested.*

She looked at this.

She added: *The protocol was designed to avoid a cost. The cost cannot be avoided. The question of whether it can be changed is different from the question of whether it can be avoided, and I have been treating them as the same question for five years.*

She sat with the notebook in the study of the house that was his attention made architectural, surrounded by the evidence of being known across forty years, with the fracture that was the cost of it on her wrist, and she looked at the question she had written, and she thought: *resolved engagement. I have not tested that variable. I have been running the wrong experiment.*