Morphogenesis
Reaction–diffusion you paint into; one rule, and the spots decide to happen.
Paint a smear of chemical into a still dish and walk away. It does not fade. It splits, spots, branches, and settles into a pattern that looks designed, from one rule with no designer in it. This is how a leopard decides where its spots go.
The concept
Alan Turing's last great idea: two substances, one that activates and one that inhibits, diffusing at different speeds, can break a uniform state into a pattern all by themselves. The activator feeds itself but also feeds its own inhibitor; the inhibitor spreads faster and fences the activator into islands. Local self-amplification plus long-range suppression is the whole recipe for a spot.
The math
The Gray-Scott model is two coupled reaction-diffusion equations for concentrations $u$ and $v$:
The $\nabla^2$ terms are diffusion; the $u v^2$ term is the reaction where $v$ consumes $u$ to make more of itself. $F$ is the feed rate, $k$ the kill rate, and the ratio $D_u / D_v$ sets the inhibitor's reach. Tiny moves in $(F, k)$ flip the dish between spots, stripes, mazes, and self-replicating blobs.
Why it stays strange
Nothing in the equations says "spot" or "stripe." The pattern is not stored anywhere; it is what the system relaxes into, and the same two numbers that grow a leopard's rosettes grow a maze a hair's width away in parameter space. Form here is not a blueprint being executed. It is an instability being expressed.
Further reading
- Turing (1952), "The Chemical Basis of Morphogenesis."
- Pearson (1993), "Complex Patterns in a Simple System" - the Gray-Scott zoo.
@article{turing1952,
author = {Turing, Alan M.},
title = {The Chemical Basis of Morphogenesis},
journal = {Philosophical Transactions of the Royal Society B},
volume = {237},
number = {641},
pages = {37--72},
year = {1952}
}
@article{pearson1993,
author = {Pearson, John E.},
title = {Complex Patterns in a Simple System},
journal = {Science},
volume = {261},
number = {5118},
pages = {189--192},
year = {1993}
}