How a seed decides where to land
Most of the slow plant cycle had been built one careful link at a time: a plant is born, grows through its stages, dies, comes apart, and feeds the ground it grew on. But for a long while the chain only ran one way, downhill toward an ending. The last link — the one that turns an arc into a true circle — is the smallest and quietest moment in the whole thing: a mature plant casting a seed, and that seed deciding where to come to rest.
That a plant’s death feeds the next thing to root, and what it means for a meadow to renew itself across a decade, is its own story. This one is narrower. It’s about a single seed leaving a single plant, and the short journey it takes before it becomes ground the world will accept as a place to grow.
Only the grown, and only now and then
A plant doesn’t scatter seed for most of its life. Only a mature one does — not a seedling, not a juvenile, not one already gone to seed and dying back — and even then, only every so often. Seeding rides a slow recurring beat of the world’s ecology clock, not every tick. A meadow grass plant, when its turn comes around, lets go of about a single seed and then falls quiet again until the next pass.
That restraint matters more than it looks. An early version was far more eager — it cast several seeds on a fast beat, and a fresh patch of grass filled the available ground almost as soon as you looked away. We tuned the rate down hard, to roughly a single seed on a long interval, so that a meadow spreads at a pace you could sit and watch rather than one that overruns the land while your back is turned. A grass plant lives mature just long enough to do this once, or not at all. The whole future of a meadow turns on these occasional, sparing throws.
Two ways to fall
When a seed does leave, it needs two things: how far, and which way. We build that as a small handful of vectors — little pushes, each with a length and a heading — added together into a single offset from the parent. For meadow grass, the first plant we proved all of this on, two of those pushes are real and doing work.
The first is plain gravity: the seed simply drops and tumbles a short way from the plant that made it, with no preference for any direction. This is the right fit for low ground cover — a tuft of grass seeding mostly into the soil around its own feet. The second is wind: a push that biases the throw in the prevailing direction and carries it farther. Between them, a seed comes to rest somewhere between roughly a meter and a half and six meters from its parent — close enough to thicken the patch it came from, far enough to step out onto fresh ground beside it.
There’s a subtlety in the wind that took a moment to get right. If you let wind nudge a seed anywhere within a full circle around the prevailing heading, the pushes cancel and you’re left with no real direction at all — just a rounder version of the same scatter. So the wind throw is held to a wedge opening downwind, a span of roughly ninety degrees around the heading. That bounded arc is the whole reason wind reads as wind: seeds drift down the breeze and not back into it, and over many plants over many seasons a patch leans, almost imperceptibly, the way the air has been moving.
Two more ways for a seed to travel — carried on water, or carried by an animal — are written into the design as named places to fill in later. They aren’t loose in the world yet. For now, a seed either falls near home or rides the wind, and that’s enough to give a meadow its shape.
The ground has the final say
Picking an offset only proposes a spot. Before anything grows there, that spot has to pass through the same honest question every living thing in this world must answer: is this ground you can live on? A seed that lands outside the world, or on terrain this species can’t take — the wrong kind of country, open water, bare unsuitable ground — is simply dropped. No seedling appears. The throw happened; it just didn’t take.
And a seed that does land somewhere it can grow is settled onto the real surface there — pinned to the actual height of the ground at that point, not floating at the height it was thrown from. Only then does it become a seedling, the youngest stage of a plant, ready to begin the same slow climb its parent made. Most of a plant’s reach into the world is hopeful and lost this way, which feels right: scatter widely, and let the land keep only what fits.
Random to look at, fixed underneath
Here is the part we care about most. All of this scattering looks random — seeds going this way and that, a patch creeping outward in a way you’d never quite predict — but it’s determined, not random: grow the same world twice and its meadows scatter their seeds to exactly the same places, every time. (How a living, surprising ecology stays perfectly reproducible is its own story.) New seedlings are applied at clean boundaries between the world’s steps so the spread stays reproducible no matter how much is being born at once — the same deferred-to-a-clean-frame discipline the rest of the world runs on.
A meadow’s spread only looks like random scatter. Grow the same world again and every seed lands in exactly the same place — the churn is real, but it’s reproducible.
The smallest link, closing the circle
With seeds finally landing, the plant cycle stopped being an arc that only ran down: a throw the ground accepts is the one moment that hands the next generation forward. What that turnover means for a meadow — how a death feeds the next thing to root, and how the whole population renews itself across a decade — is the loop’s own story. (A far-off ceiling on the order of several thousand plants sits underneath purely so the spread can’t run away; where a meadow actually comes to rest is the patch’s question, not the seed’s.)
We proved all of this on meadow grass first, the simplest ground cover, then carried the same machinery to other plants — and a slow, long-lived tree disperses very differently from a quick carpet of grass, but how the world grew into a forest is its own story.
A seed deciding where to land is about the smallest thing the simulation does. It’s one plant, one throw, a meter or a few, kept or dropped depending on what the ground says back. But it’s also the hinge the whole slow cycle turns on — the difference between a world that quietly winds down and one that renews itself, year after patient year. You don’t win this world, and you don’t force it to grow. You tend it, and let the seeds find their own way down.
