"It can," I said. "And the feeding process is ending that relationship. At current rates, in approximately seventy years, the substrate beneath this world will be below the threshold required to sustain the ecology that makes that relationship possible."
A pause. Long, by the anchor's temporal standards — perhaps fifteen seconds of real time, which given how the organism experienced duration was significant.
*We did not know,* it said. *We would not have chosen this.*
Three words that were not three words — the density shift that communicated them had the quality of something genuinely troubled. Not guilt, exactly — guilt requires a prior moral framework, and the organism was only now constructing the relevant moral framework in real time. But something adjacent: the awareness that an action has caused harm that the actor would not have chosen to cause.
"I know," I said.
*Can it be stopped.*
"That's what I'm trying to work out," I said. "I need to understand your structure. The root system, the apertures, the feeding channels. Whether there's a way to redirect the process without ending it entirely."
*We will show you,* the anchor said. *But we should tell you: the Tower is also aware of this location now. We can feel their instruments at the substrate level. They have been watching since before you arrived.*
I held very still in forty meters of harbor water.
"How long have they been watching this location."
*Since we became aware of the blank slate's presence on this plane,* it said. *Approximately thirty days.*
Day 30. The day I landed.
They'd known where the anchor was for a month. They'd let me work — let me build the sensitive network, let me talk to Mardus, let me run my assessments — while watching.
Which meant they had not been idle. They had been preparing.
*There is something building at the substrate boundary,* the anchor said. *We do not know what it is. But it is large, and it is directed at this location.*
I filed this under: *urgent, address immediately.*
Then I filed it under: *after I understand the root system, because going up now without that understanding means coming back later under worse conditions.*
"Show me," I said.
---
**Earth: Day 62, Hour 4**
I came up from the harbor at Hour 23 on Day 61, eleven hours after I'd gone in.
Nassiri was on the seawall. He had a thermal flask and the expression of someone who had decided to wait without having been asked to wait, which was a decision that said something about where we were in the working relationship.
He handed me the flask without preamble. I drank. It was tea, which was not what my body wanted but was what was available, and the warmth was real.
"Eleven hours," he said.
"I know."
"Sera came by at Hour 14. Didn't say what she wanted. Said it could wait."
"I'll find her in the morning."
He looked at the harbor. The apertures were doing their nighttime pattern — low output, clustering, the Hounds conserving. "The gates look different to you now," he said. It wasn't a question.
"Yes."
"Different how."
I thought about how much to say. Nassiri had been operating on partial information since Day 30, making good decisions with it. Giving him more information was always the right call. The question was sequence — what he needed now versus what he could usefully act on later.
"They're not portals," I said. "They're not mechanical at all. The organism creating them doesn't know it's causing harm." I looked at the harbor. "I made contact tonight. It's willing to cooperate. What we do with that cooperation is the problem I need to solve next."
He absorbed this with the controlled patience I'd come to recognize as his version of significant surprise.
"What do you need from me," he said.
"Twenty-four hours," I said. "I need to finish something before I can be useful to anyone else."
He nodded once and walked back toward the coordination post without asking what the something was. This was, I thought, the most useful quality a commanding officer could have: the judgment to know when more information would help and when it would only create obstacles.
---
The Tier 4 compression had been running at sixty percent intensity since Day 59. Three days of gradual pressure, the architecture collapsing inward toward the new density in the structured way the Architect's truss design required.
What eleven hours in the root system had done to it: the composite perception — eight elements running as one — had imposed a new geometry on the compression. Not faster. More precise. The synthesis framework that the Architect had designed for full-element simultaneous operation was, it turned out, also a more efficient compression architecture than running the eight channels in managed parallel.
When all eight elements understood themselves as aspects of one observational system, the compression force was distributed not across eight separate pathways but through the unified framework as a single load. The truss held differently. Better.
I ran the diagnostic at Hour 4 on Day 62, sitting in the coordination post's bunk room with the composite perception showing me my own architecture from the inside.
*Mana capacity: 4,102 units.*
The transition had completed while I was in the harbor.
Not with announcement. Not with the surge of something major shifting — or rather, with exactly that, but I had been too occupied with the anchor contact to register it as a distinct event. The Tier 4 threshold had arrived and the architecture had crossed it the way you cross a watershed: you're going uphill, then you're going downhill, and the moment of transition is only visible in retrospect.
The world was different at Tier 4.