He discovered that he could regenerate lost tissue on a morning when he was six weeks into the intensive practice phase, tired in the specific way of someone whose secondary magical pool had been drawn down daily for weeks and was replenishing through genuine rest but not quite fast enough to feel fully recovered.
The subject was a Bark Lizard — the species was becoming his most frequent test case, for the practical reasons he had originally selected it and for a new one: he had been working with the same territorial group long enough that the individual lizards had, he suspected, some memory of previous interactions. They did not like Shadow. They also did not flee Shadow immediately the way lizards in unfamiliar territory did. There was a category of wariness that was specific and informed rather than generic, which he found interesting as an observation about the cognitive capacity of creatures that the world tended to categorize as minor monsters.
This particular lizard had lost most of its tail in a territorial dispute — not a partial wound this time but a clean removal at the midpoint, the kind of injury that Arthur's initial experiments had dealt with in the form of partial wounds but that he had not yet attempted in this severe a form.
He healed the wound first. Closed the end of the remaining stump cleanly, addressed the tissue damage around the removal site, stabilized the structure. This part was within his current competency and took approximately seven minutes — dramatically faster than his first attempts, with considerably more precision and control.
Then, on an impulse that was not strictly experimental — more the kind of impulse that came from pushing against a boundary to see if it moved — he kept going.
Past wound closure. Past stabilization. He directed the healing sense toward the stump end and pushed it further, not into the damaged tissue but outward, beyond the physical boundary of what was currently there, toward where the rest of the tail should have been.
The healing magic, to his considerable surprise, followed.
It reached past the body's current edge and began building something new from the template of what should be there — not creating from nothing, but reading the body's own biological information about what that missing tissue was supposed to look like and using that information as a blueprint. Slowly. At enormous energy cost. Requiring more of his secondary pool than anything he had yet attempted and leaving him noticeably depleted afterward. But the tissue grew.
He watched it through Shadow's perception, the slow emergence of new Bark Lizard tail from the stump, segment by segment, scale by scale, taking forty minutes to produce approximately an inch of new growth before he ran out of energy to sustain the process and had to stop.
He released the lizard — which bolted immediately and comprehensively, which was fair — and spent the rest of the morning sitting with what he had discovered.
He could regenerate.
With enough energy and time, he could grow back things that had been lost.
He thought about Lyra's lungs, about the years of accumulated scarring and impaired tissue that the chronic infection had been producing slowly and steadily since before Arthur had arrived in this world. He thought about what the healing sense had shown him when he finally performed the diagnostic scan — the specific texture of tissue that had been damaged and had healed improperly and had been damaged again, the layering of repair-upon-repair that left the structure less functional with each cycle.
He thought: I can address that. Not just the infection. The damage the infection has caused.
He thought: I am going to need to be better at this than I am currently, and I am going to need a great deal more practice, and Lyra is going to need a great deal more patience from me before I trust my technique with the tissue of someone I love.
He thought all of this and did not rush. Rushing was the thing he was most careful not to do, because he had understood from the beginning that being careful with Lyra required being more careful than he was capable of being when he had just discovered something exciting.
He went back to practicing on forest creatures.
◆ ◆ ◆
The central problem crystallized itself with uncomfortable clarity approximately two months into the practice phase.
He had become, by this point, a competent healer of things he could identify. Surface wounds, internal injuries with visible or palpable structure, specific breakages and tears with discrete sites. His technique had improved to the point where the Bark Lizard's tail regeneration, on the second attempt, had taken twenty-two minutes and produced growth of two and a half inches before the pool gave out — and he had estimated that with another month of practice he would be able to regenerate a complete tail in a single session.
All of this was entirely irrelevant to Lyra's condition.
Because Lyra's condition was not a wound. It was not an injury with a site. It was a dynamic state — a population of infecting organisms in a complex interaction with a host immune system that had been fighting them for years and had reached a balance that was slowly, invisibly, progressively damaging. You could not point at the infection the way you could point at a wound. The infection was everywhere it had established itself, in varying concentrations, at varying stages of activity, distributed through the architecture of Lyra's lungs in a pattern that you could not address without first mapping.
He did not know how to map it.
He had the healing sense. He had extended it into Lyra's breathing while she slept, twice, carefully and briefly, feeling the general state of her lung tissue with the attention of someone who understood the difference between healthy and damaged. He knew things were wrong. He could not tell specifically where, or how wrong in each location, or what the distribution of the infection looked like in the detail he needed to address it precisely without the risk of missing the worst areas or overworking the less affected ones.
He needed to see inside her the way the healer had seen inside her during the examination. He needed the diagnostic technique he had watched from his cradle and stored as a behavioral observation without any understanding of its mechanism.
He had been stuck on this for four weeks when Shadow killed the Thornback Boar.
◆ ◆ ◆
The Thornback Boar was large — not the largest creature Arthur had hunted, but meaningfully larger than anything he had attempted since the Stone Boars of the deep forest, and the decision to take it had required the kind of careful situational assessment that he applied to targets that pushed his current capability. Shadow's power had grown with Arthur's leveling; what would have been impossible at Level 7 was manageable at Level 12, though not comfortable. The kill had required a more complex dark-magic construct than his usual spear — a multi-point application that addressed the Thornback Boar's defensive biology, which involved a layer of magically reinforced hide around the torso that a single-point impact would not reliably penetrate.
He had been working on multi-point constructs for three weeks in anticipation of this hunt. The investment paid off.
The absorption, when it came, was the richest single event since he had begun systematically hunting the deep forest. The Thornback Boar had been old — old in the way of large magical creatures that survived to accumulate age, which meant years of encounters, years of territory, years of a life that had been long and varied and had intersected with human presence more than most forest creatures' lives did. The deep forest near Thornwick was not wilderness in the absolute sense; humans had been moving through it for generations, and the creatures that lived in it had memory of that movement.
Arthur processed the absorption over four days. This was longer than his usual processing time, reflecting both the volume of the material and the specific decision to go through it more carefully than he typically did — to attend to the fragments rather than extracting the summary and releasing the rest. He had learned that the most valuable information often lived in the fragments.
On the third day he found it.
It was not prominently stored — it did not have the vividness of a recent or emotionally significant memory. It was an old encounter, filed under the cognitive category that large territorial animals used for 'unusual human behavior observed near my range,' the specific alertness that such animals maintained because unusual human behavior was a data point about whether it was safe to continue using a particular territory.
A clearing, possibly twenty years past in the boar's life. Two humans — one injured, lying still; one kneeling over the first with hands that were glowing. The boar had watched from the tree line, assessing whether the humans were a threat or merely interesting. It had decided interesting. It had watched for approximately three minutes before concluding that nothing alarming was going to happen and moving on.
Three minutes.
In those three minutes: the kneeling human had placed both palms flat against the injured human's chest. The light at his hands had changed quality — from the warm directional glow of healing to something broader, flatter, less focused. A pulse of that flatter light had extended from the healer's hands into the injured body, and the injured body had briefly illuminated from within, not uniformly but in a pattern that corresponded — Arthur was nearly certain, from the shape of it — to the body's interior structure rather than its surface.
The healer had kept his eyes closed throughout.
His expression was not the expression of someone projecting power outward. It was the expression of someone receiving information — the unfocused, slightly tilted quality of a person listening to something they needed very much to hear clearly.
A diagnostic pulse was not a projection. It was a question asked of the body in a language the body could answer. You sent it in and it came back having touched everything it passed through, carrying the signature of what it had found, and you read the return the way you read the return on sonar — not by seeing but by interpreting pattern.
He understood this immediately. He understood it in the specific way that you understand something that your brain has been building toward without your direct knowledge — the sudden completion of a structure that had been assembling itself in the background, recognizing itself as finished when the last piece arrived.
He spent the fourth day building the mechanism from the behavioral evidence.
◆ ◆ ◆