Sunflower seeds, pinecone scales, pineapple skin tiles, artichoke leaves. Each new seed is placed at angle α from the previous one, at radius √n. For nearly every α, this produces ugly clumping. Only at α = 1/φ² (the golden angle, 137.5°) do the seeds pack with no gaps and no aligned spirals. The seed count along visible spirals: always consecutive Fibonacci numbers.
angle α (degrees) 137.50°
seeds N 1200
size 1.00
why the golden angle is the only one that works
Place seeds at successive angles k·α for k = 1, 2, 3, …, at radius √k (so each ring of seeds has the same area). For most α, two seeds far apart in k end up at nearly the same angle if α happens to be near a rational p/q — they collide along q rays, leaving wasted area between.
The angle that most resists rational approximation is the one whose continued-fraction expansion is all 1s — i.e., 1/φ = [0; 1, 1, 1, …]. Since the golden ratio φ is the "most irrational number" (its convergents F_n/F_{n+1} are exactly the Fibonacci ratios — the slowest possible CF convergence), any multiple of 2π/φ² is maximally far from every rational angle. 137.508° is the angle 2π(1 − 1/φ) ≈ 137.50776°.
The Fibonacci-spiral count: when you stare at a sunflower head, you see two families of spirals — one going clockwise, one counterclockwise. Count them: always two consecutive Fibonacci numbers like (34, 55) or (55, 89). This isn't sunflower mysticism, it's geometry: the visible spirals correspond to the denominators of the best rational approximations of 1/φ², which are exactly Fibonacci numbers.
Move the α slider by 0.5° away from 137.5° — the packing breaks instantly. Move it to 120° — three rays jump out. The golden angle isn't aesthetic preference; it's the unique optimal solution to "pack points on a disk without leaving rays."