The rule is almost insultingly simple. Take the last two numbers and add them to get the next one. That's the entire instruction — no primes, no calculus, nothing hidden. Start with two ones and it hands you 2, then 3, 5, 8, 13 — a sequence that needs no memory beyond its last two entries, and yet never repeats and never stops.
Divide each new number by the one before it, and something else appears, uninvited: a ratio that overshoots, undershoots, and slowly stops moving. It closes in on 1.6180339887…, a number that cannot be written as a fraction of two whole numbers no matter how far the division is carried.
The rule is arithmetic. Its limit is not.
01 — The Construction
Each square is drawn once, at a constant size. Everything already built steps back to make room for it. The arc that runs through them is one continuous curve, compass-drawn, corner to corner.
F(n) = F(n−1) + F(n−2)
An animated diagram builds a Fibonacci square spiral one square at a time — 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89 — while a running ratio of the two most recent numbers converges toward 1.6180339887. The full sequence, with each number's two parents and the ratio at each step, is listed in the table below the diagram.
STAGE 00 / 10
1
—
F(n) ⁄ F(n−1)
The first mark. Nothing to divide yet.
The sequence, plainly
| N | Value | Sum | Ratio |
|---|---|---|---|
| 00 | 1 | seed | — |
| 01 | 1 | seed | 1.000 |
| 02 | 2 | 1 + 1 | 2.000 |
| 03 | 3 | 2 + 1 | 1.500 |
| 04 | 5 | 3 + 2 | 1.667 |
| 05 | 8 | 5 + 3 | 1.600 |
| 06 | 13 | 8 + 5 | 1.625 |
| 07 | 21 | 13 + 8 | 1.6154 |
| 08 | 34 | 21 + 13 | 1.6190 |
| 09 | 55 | 34 + 21 | 1.6176 |
| 10 | 89 | 55 + 34 | 1.61818 |
02 — The geometry of the rule
x² = x + 1
Here is why the ratio settles where it does. If each new square's side is the sum of the previous two, then in the limit the ratio of consecutive terms — call it x — must satisfy x = 1 + 1/x, the same relationship the sequence is built from, rearranged. Multiply through and it becomes x² = x + 1: an equation with an irrational root, φ = (1 + √5) / 2 = 1.6180339887…
Turn that equation into a shape and it becomes a rectangle: cut a square from one end of it, and what's left is a smaller rectangle with exactly the same proportions as the one you started with. That self-similarity — not any mystical property of the number — is the entire reason the spiral keeps its shape as it grows. Every fold looks like the last fold for the same reason a sheet of A-series paper does: the ratio reproduces the whole when a piece is cut away.
03 — In the field
Where it actually shows up
Look at the head of a sunflower and count the spirals — the ones curving one way, the ones curving the other. They come out as consecutive Fibonacci numbers: 34 one way and 55 the other, sometimes 89 in a large seed head. Pinecones do the same in miniature, running 8 spirals one way and 13 the other.
The cause isn't decorative. Each new seed or scale is laid down at a fixed angle from the last — roughly 137.5°, the golden angle, the angle that divides a full turn in the golden ratio. Because φ is irrational, that angle never divides evenly into the circle, so the pattern never closes into straight rows or repeating gaps. It's the packing arrangement that wastes the least space, arrived at by growth, not by design — a genuine optimisation, not a coincidence dressed up as one.
04 — A correction
Where it doesn't
It's worth saying plainly: the golden ratio has been found in a great deal that never contained it. The nautilus shell is a logarithmic spiral, which is a real and lovely thing, but its growth ratio is nowhere near φ — that claim doesn't survive a tape measure. The Parthenon and the Mona Lisa acquired their golden rectangles later, from admirers holding a ruler loosely enough to find whatever they were looking for. Petal counts lean Fibonacci more often than chance, but it's a tendency, not a law — plenty of flowers simply don't comply.
The real cases don't need the borrowed ones. A seed head and a pinecone earn their place in this story because there's an actual mechanism — the golden angle, the packing, the growth — not because the number is owed reverence. It's more interesting this way, not less.
05 — Both directions
It runs backwards, too
Extend the sequence before the first 1 and the rule still holds: …, 5, −3, 2, −1, 1, 0, 1, 1, 2, 3, 5, … The signs alternate, the magnitudes are the same familiar numbers, and the ratio of consecutive terms still converges on the same φ. The sequence doesn't begin where we started counting it.