Sera took the number the way she took everything. With open eyes and a straight back.
"Four-point-eight percent," Voss said. They were in the Corps facility's briefing room, the field team assembled, Ryn's individual cost readouts displayed on the wall screen. No averages. No smoothing. Each person's name beside their number. "The arch's load distribution is frequency-dependent. Participants with lower Thread Sight frequency bear more of the restoration cost than participants with higher frequency. You and Tomek, as the most recent developers, carried the heaviest share."
Tomek Rais looked at the numbers. His cumulative cost across the three operations was 8.6%. His bridge-builder's mind was doing the math — at that rate, twenty-five operations would deplete him completely. He sat with the calculation for ten seconds, then nodded once, the way a man nods when the stress report on a structure comes back worse than expected but within the range of engineering responses.
"Then I need to get stronger," Sera said. "So the arch stops dumping the load on me."
She said it without anger. Without self-pity. With the particular directness of a person who had identified the variable in the equation and was already thinking about how to change it. She had been a teacher for eight years. She knew what remediation looked like. She knew that being behind was not the same as being unable.
"The frequency gap between you and the veterans is the source of the imbalance," Mira said through the speaker. She had been running the analysis since the transport landed. "At your current Thread Sight frequency, the arch routes approximately thirty-five percent more load to your position than to Lyle's. If your frequency increases by ten to fifteen percent, the routing differential drops below ten percent. The arch would treat you as a structural peer, and the cost distribution would equalize."
"How do I get a ten percent frequency increase?" Sera asked.
"Training. Proximity to active nodes. Time."
"How much time?"
"At the current training rate, two to three weeks."
Sera looked at Tomek. Tomek looked at Sera. The exchange lasted one second and communicated everything that needed communicating between two people who had walked into the same building on the same day and signed up for the same job and were now discovering that the job was harder than they had been told.
"We'll do the training," Tomek said. "But we want to keep operating while we train. Lower frequency or not."
Voss considered this. The medic's answer was to pull them from operations until the frequency gap closed. The operational answer was that he needed every weave participant he had, because the Gradients were not going to wait for his civilians to finish their homework.
"Modified deployment," he said. "You train near active nodes four hours per day. You deploy in weaves where the restoration target is low-density — endpoint nodes, not junctions. The residue at endpoints is lighter. The cost per operation is lower. You build frequency and field experience simultaneously."
Sera nodded. Tomek nodded. Neither of them asked for easier assignments. They asked for the training they needed to handle harder ones.
---
Mira designed the accelerated frequency development program in two days.
The principle was environmental immersion. Thread Sight frequency correlated with the strength of the ambient substrate signal — the Loom's radiation in the local environment. Stronger signal meant the resonance was louder, which meant the neural architecture had to work harder to process it, which meant the processing pathways strengthened. The same mechanism that made muscles grow under load.
Active doorway nodes produced the strongest local substrate signal. Mira identified four nodes within the metropolitan area that were running at above-average capacity — junction nodes with multiple active connections, where the thread-energy throughput created a concentrated radiation field within a hundred-meter radius. She designated these as training sites.
Civilian weave candidates trained at the sites for four hours daily, running Pell's breath ladder and frequency hold exercises inside the amplified field. The results were measurable within days. Sera's Thread Sight frequency, which had plateaued at the weave target level during standard training, began climbing again. By the end of the first week, she had gained seven percent. By the end of the second week, twelve percent. Her cost differential in the test weaves dropped from thirty-five percent above Lyle's to fourteen percent. Still not equal. Close enough that the arch treated her position with more respect.
Tomek gained nine percent over the same period. His structural engineering metaphors grew more specific as his Thread Sight strengthened — he started describing the weave's resonance patterns in terms of load paths and compression vectors, translating the experience into the vocabulary of his profession. Voss found the translations useful. A bridge-builder's language was not far from a surgeon's when both were describing how structures held together under stress.
The broader Corps expansion did not wait for the training to finish.
---
Screening stations opened in Port Maren, the eastern coastal hub where Thane Orr's deployment force had established a Corps presence. In the southern agricultural capital of Greenhollow, where Lena dispatched a logistics team to convert a community center into a testing facility. In the northern mining city of Stonewall, where the pre-existing population of former Rift Defense Corps Carvers provided a nucleus of experienced candidates who already had dormant Thread Sight potential.
Each station ran the same protocol. Mira's resonance screening. The breath ladder training. Frequency development near the nearest active node. Qualification testing for the weave target frequency.
The numbers came in daily. Port Maren processed six hundred candidates in its first week. Sixty-three passed screening. Of those, twenty-one reached the weave frequency within ten days. Greenhollow was slower — smaller population, fewer active nodes nearby, the candidates developing at the rural substrate level rather than the urban one. Stonewall was faster — the ex-military Carvers had dormant neural pathways that reactivated under the resonance training like old roads cleared of undergrowth.
By the end of the second week after the southern corridor operation, the Carver Corps had grown from eight Thread Sight carriers to forty-seven. Not all of them were weave-qualified. Thirty-one had reached the target frequency. The rest were in development.
The first civilian weave team deployed without Voss.
Lyle anchored it. She was the strongest resonance carrier in the original Corps — her natural frequency was the closest to Voss's, though the gap between them was the gap between a candle and a bonfire. She led an eight-person team to restore two nodes drained by the eastern Gradient fragment. Standard formation. Breath ladder. Link. Restoration.
Mira's monitoring showed the results: average per-participant cost of 2.8%. Higher than the 1.8% of a Voss-anchored weave. Lower than the 3.1% of the hardest field operations. The restoration was clean. The nodes came back to eighty-seven and eighty-nine percent capacity.
The weave worked without Voss.
The cost was higher. The efficiency was lower. But it functioned. Lyle could anchor. With more trained participants and stronger civilian frequencies, the numbers would improve. The system did not depend on one man in the center of the arch. It depended on the arch itself — the principle, the formation, the shared resonance of people who chose to contribute.
Three more civilian-anchored weaves deployed in the following week. Two in the metropolitan area. One in Port Maren. All successful. All at the 2.5-to-3.0% cost range. The learning curve was visible in the data — each team performed better than the last, as the participants gained experience and the training curriculum incorporated field lessons.
Voss read the deployment reports from his office. He was running the logistics now — scheduling weave teams, coordinating with Mira's monitoring on Gradient fragment tracking, managing the pipeline from screening to training to qualification to deployment. An administrator. Not a field operative. The work he had always resisted was the work the situation demanded.
Dex, who had taken over as field coordinator for the metropolitan deployment teams, reported daily. His assessments were concise, his observations sharp, his judgment reliable. The man who had once been nothing but a weapon had become something the Corps needed more than weapons: a manager who understood both the human cost and the operational necessity.
Ryn ran the medical oversight for all deployed teams. Her protocols were strict, her monitoring was continuous, and her intervention threshold was low. Any participant whose biometric readings exceeded safety parameters was pulled from the weave immediately. No exceptions. No appeals. The policy generated friction with some candidates who wanted to push through. Ryn did not negotiate with people who wanted to damage themselves.
---
Three weeks into the expansion.
The Carver Corps had sixty-two Thread Sight carriers across four cities. Forty-four were weave-qualified. Six weave teams were operational, rotating through restoration assignments as Gradient fragments continued to propagate through the network. The propagation rate had slowed — not because the Gradient was weakening but because the sacrifice-and-isolate operations at the major junction nodes had reduced the number of available high-energy paths. The fragments were being funneled into corridors where weave teams could reach them.
The network's total node loss stood at twenty-three percent. Of that, the weave teams had restored nine percent. Net loss: fourteen percent. Not good. Not terminal. The race between restoration and decay was still running, and the decay was still ahead. But the gap was narrowing. Every node restored was a piece of the substrate healed, the dimensional fabric strengthened, the Gradient's food supply reduced.
Mira called at 0347.
Voss was not asleep. He was in his office, reviewing the next day's deployment schedules, his coffee cold on the desk beside the data tablet he had been using left-handed because his right hand was tired. The hours were unsustainable. Ryn had told him so. He had filed the observation and continued working.
"I'm looking at node 7-31," Mira said. Her voice had the quality it took on when data was behaving in a way her models did not predict. Not the bad-data quality — she had a separate register for that. This was the other one. The quality that meant something new was happening.
Node 7-31 was in the first Gradient's isolation zone. The northern sector. Part of the thirty-two-node sacrifice operation from three weeks ago — the nodes that had been deliberately cut off and left to starve after the bait drew the Gradient. The isolation zone was supposed to be dead. The Gradient had fed on the bait, consumed the isolated nodes, and was now dissipating in the depleted zone. Nira Sol's estimate was one to three years before the zone could be rebuilt.
"What about it?" Voss asked.
"It's showing an energy signature."
He put down the coffee. "Define."
"The monitoring system pinged node 7-31 at 0300 as part of the isolation zone's routine check. The routine check verifies that isolated nodes remain at zero output — confirmation that the Gradient is still dissipating and the zone is safe. Node 7-31 should be reading zero. It's reading zero-point-four terajoules."
Zero-point-four. Not zero. Not even close to the 6.8 terajoules that a standard node produced at full capacity. But not nothing. A faint pulse of energy in a space that should be completely dark.
"Equipment malfunction?" he asked.
"I checked the sensor calibration. It's accurate. I ran the reading three times. Zero-point-four terajoules, stable, consistent with a substrate radiation source." She paused. "Voss, there is no substrate radiation source in the isolation zone. The nodes are disconnected from the network. The Gradient consumed all available energy. There should be nothing there generating thread-energy."
"But something is."
"Something is." She paused again. Longer this time. "The energy signature is not a node pattern. It's not Weaver architecture. It's not ambient substrate recovery — that would be orders of magnitude weaker and diffuse, not concentrated at a specific location."
"What does the signature look like?"
"Biological."
He stared at the wall of his office. The deployment schedules forgotten. The coffee cooling further.
"Biological," he repeated.
"The frequency profile matches the thread-energy output of a living biological neural architecture. Not human — the pattern is different. But biological. Something alive, generating organized thread-energy, in a zone where the Gradient consumed everything and should be dissipating."
Something was growing in the dead zone. Something alive. Something that had not been there when the isolation perimeter sealed.
Something that was generating the one thing the Gradient was supposed to have consumed completely.
"I'm going to the site," he said.
"It's inside the isolation zone. The Gradient residue—"
"I'm going. Send me the coordinates."
He was already reaching for his field jacket when Mira's coordinates arrived on his tablet. Node 7-31. Forty kilometers north. Inside a perimeter that was supposed to contain nothing but dead architecture and the fading stain of something that ate dimensions.
Whatever was growing there, it was alive. And it was producing thread-energy in a place where thread-energy had been erased.
The night was dark outside his office window. The city's nodes glowed in the distance. Somewhere north, in the dead zone, something new was waking up.
