Nassiri met me in the doorway.
"Assessment," he said.
"The gates can't be suppressed by closing them," I said. "They're part of a larger structure. Closing the surface apertures is the equivalent of blocking one nostril — the system compensates and opens others."
He looked at me with the expression of someone who had been hoping for better news and had enough operational experience not to be surprised.
"What's the right approach," he said.
"I don't know yet." I held up a hand before his expression could fully form. "I know what the wrong approach is, which is more useful than it sounds. The whole city has been trying to close gates. Militaries globally have been trying to close gates. Every approach has either produced no effect or produced temporary suppression followed by new aperture formation."
"The Paris approach. The London approach."
"Both produced gate migration, not gate elimination. They were treating the symptom." I pulled out a notebook — I'd started carrying one, another thing that had changed about me that I was slowly cataloguing — and showed him the map I'd built of the root system. "The organism's anchor is deep and currently unreachable by conventional means. Until someone figures out how to address the anchor directly, the surface apertures are going to keep appearing wherever the network finds useful access points."
He studied the map. His expression was the controlled version of what less disciplined people would let become frustration.
"Which means."
"Which means the correct strategy is containment and pressure management, not elimination. Keep the existing apertures mapped, monitor the Hound populations near them, maintain perimeter discipline to prevent expansion into new districts." I tapped the harbor area on the map. "And don't try to close the port cluster. Those three apertures have been stable for twenty-seven days. Stability is worth something. If you force migration now, the new apertures will be unpredictable."
"And in the long term."
"In the long term, I'm working on it."
He looked at me steadily. "That's not a complete answer."
"No. But it's an honest one." I met his eyes. "I'll have more when I have more. What I can give you right now is a reason to stop doing the thing that's been making the problem worse."
A pause. Then: "The global coordination call is at Hour 21. I can include your analysis in the summary packet."
"Do that."
He nodded. "Get some sleep. You look like you haven't in a while."
"I'm logging it as within operational parameters," I said.
He gave me the flat look of someone who had heard that particular form of self-deception before, from younger people, in earlier wars.
"Get some sleep," he said again, and went back to his maps.
---
**Earth: Day 31, Hour 3**
The coordination post at three in the morning was as quiet as it ever got, which was not very quiet — someone was always on communication watch, and the building itself held the ambient sounds of eight hundred people's ongoing survival infrastructure at various removes.
I sat against the far wall of the bunk room with the notebook on my knee and ran the Vassal-Link diagnostic.
The residual was still active. Fifteen percent of the encoding, running in the quarantine partition, broadcasting a degraded signal that was essentially static with a recognizable carrier wave underneath. From the Tower's monitoring perspective, it would look like a relay node operating at severely degraded efficiency. From a signal analysis perspective, it would look like damage — a Hero who had undergone some form of transit corruption and was broadcasting at reduced capacity.
This was not the same as looking like deliberate sabotage.
But it was not invisible.
The Tower's Earth-side operation had been running since Day 2. I didn't know its exact structure — Vasir had mentioned it but hadn't known the details, and Elara's intelligence had focused on the Lens architecture rather than the colonization logistics. What I knew: at least one Tower-affiliated agent was embedded in Earth's surviving institutional structure, with access to mana detection equipment, reporting back to someone in the Council's colonial division.
If that agent had access to a Vassal-Link monitoring receiver, they had known I was on the ground since six hours after I transited.
It had now been twenty-one hours.
No contact. No surveillance I could detect. No anomalous mana signatures in the surrounding blocks that would indicate a Tier-capable mage operating covertly.
Either the agent didn't have monitoring receiver access, or they had identified my signal and were waiting, or they were moving with more patience than I'd have expected given the stakes.
I added it to the daily assessment and updated the threat model: *Tower Earth-side contact: probable within 48-72 hours. Approach: currently unknown. Recommended action: maintain operational visibility to avoid appearing evasive, identify all incoming contact attempts before committing to response posture.*
The other item in the daily assessment was worse.
The Architect's final message was locked behind a milestone I hadn't hit. The second reference I'd found in the Maw Gate search had said: *do not engage the surface apertures without understanding the root structure.* I'd met that condition — I understood the root structure, or at least enough of it. But the third reference had said: *the anchor is not a machine. Protocol for engagement: see Primary Disciple instructions, final message.*
The final message had not unlocked.
Which meant understanding the root system wasn't the milestone. Something else was.
I went through the Library, running the Stone's analysis function against the milestone trigger logic I'd reverse-engineered from previous message activations. Each previous message had triggered on:
Core integration milestones. Contact with specific elemental phenomenon. Structural architecture completion markers.
The final message had none of those triggers. It had something different — a composite trigger, not a single event but a combination of conditions. I could see the structure of the trigger in the Stone's architecture but couldn't read the content, the way you can see that a lock is there without being able to see what key it requires.
One of the conditions was present. I was on Earth. That was one condition.
Another appeared to require contact with the Maw Gate anchor structure directly — not surface analysis, but physical contact with the convergence node forty meters below the harbor.
A third condition was partially met: I had seven elements at functional capacity and two at incomplete status. The trigger appeared to require something different about that state — not completion necessarily, but something about how the incomplete elements were held. The Time seed and the Space Tier 2 were both running in a specific mode that the trigger was monitoring.
I didn't have enough information to fully map it. The Architect, true to form, had designed a system that would not give me its last instructions until I was standing exactly where I needed to be to use them.
This was very much in keeping with what I'd come to know of the Architect's approach to disciple management. Posthumously controlling, with excellent reasons that would only become clear in retrospect.
I wrote a note to myself in the Library: *the final message requires: Earth contact (met), anchor proximity (unmet), third condition (partially met, nature unclear). The anchor is forty meters below the harbor, accessible through the root system, in territory controlled by the Maw Gate network. Approach is the problem.*
Then I put the notebook down, lay back on the bunk, and gave sleep my best attempt.