The anchor's density shifted in a way that required the Time-aspect seed to process — the organism was showing me something that existed across multiple temporal states simultaneously, the way a geological cross-section shows you all the time periods at once. The root system as it currently was. The root system as it could be. The delta between them.
The current root system: forty percent global coverage, drawing at full expansion-mode rate.
The possible root system: approximately eight percent global coverage, drawing at maintenance-mode rate. The eight percent would be the minimum viable feeding network for the organism's survival at its current size. The maintenance mode would draw at a rate that Earth's substrate could regenerate within approximately two hundred years.
Not zero depletion. Sustainable depletion. The difference between a wound and a bleed-out.
"Is eight percent containable," I said. Not to the anchor — to myself, to the Library.
The Library ran the analysis. Eight percent coverage at maintenance rate, restricted to the regions with the lowest human population density and the highest geological substrate concentration, would preserve the mana ecology in all inhabited areas. The substrate loss would be real but it would occur in regions where the ecology was thinner to begin with and where the loss would have minimal impact on the sensitive populations developing in cities and populated zones.
Not perfect. Liveable.
"The apertures in populated areas," I said. "The ones that have been feeding the creature populations. If the root system contracted to eight percent maintenance coverage, those apertures would close."
*Yes,* the anchor said. *The creatures near them would lose their substrate source. They would die or relocate to the remaining apertures.*
"How long for the contraction to complete."
*If we begin now: three to four of your months.*
Three to four months of transition — the creatures near contracting apertures becoming increasingly mana-starved, increasingly desperate, potentially more dangerous in the short term as they competed for the remaining substrate. A difficult three months. Then the gates closing, one by one, as the root system pulled back to its minimum viable footprint.
And the apertures that remained — the eight percent — in remote geological zones, away from cities, away from populations. The organism's permanent presence on Earth, a maintenance-rate feeding network in unpopulated geology, drawing at a rate Earth's substrate could sustain indefinitely.
A deal. The kind that is not comfortable and is better than the alternatives.
"I need to discuss this with the people responsible for Earth's defense before I can commit to anything," I said. "Can you hold the current state — not contract, not expand — for approximately two weeks while I arrange that."
*We can hold,* it said. *We have been holding since you first contacted us. The expansion reflex has been suppressed since we became aware of the harm. We did not know how to communicate this before you arrived.*
It had been holding for two days. Waiting for me to come back.
*There is one more thing,* the anchor said. *The structure building at the substrate boundary. It has accelerated since your first descent.*
I already knew this. I had known it since the anchor mentioned it on Day 61. I had been leaving it in the *urgent, address after* category because the anchor negotiation was more important and because I couldn't address both simultaneously.
The *after* had arrived.
"What kind of structure," I said.
*A lens,* it said. *But not like the Tower's Lens. This one is built into the substrate itself. It is designed to focus substrate energy rather than dimensional transit.*
A substrate lens. Not a transit portal. Something that used the dimensional membrane as a focusing element, the way a solar lens uses glass.
If you focused substrate energy — concentrated the dimensional potential of a significant area into a single point — you could do things with that concentration that no individual mage could do. You could rewrite a Vassal-Link. You could, potentially, rewrite an entire node architecture.
You could, if the lens was large enough and aimed correctly, restore a Vassal-Link that had been degraded to fifteen percent.
From inside.
Without transiting anyone. Without another summoning.
*It will be ready in approximately twelve days,* the anchor said.
Twelve days.
The Tower had spent sixty-two days watching me build a sensitive network and negotiate with the organism they'd been calling a Demon King and run a Tier 4 compression, and they had been building a substrate lens the whole time.
Everything I had done. All of it visible to their instruments.
They knew about the anchor. They knew about the Zalarus contact. They knew I was the saboteur.
And they had a lens that would be operational in twelve days.
I came up from the harbor faster than I'd gone in.
---