Fireflies flash in unison, cardiac cells beat together, and audiences clap in sync.
Set a roomful of metronomes on a shared platform and within minutes they lock into step. The coupling through the platform is enough.
Yoshiki Kuramoto's 1975 model captures this with $N$ oscillators, each with its own natural frequency, each pulled toward the group's mean phase. One equation, one coupling parameter, a sharp phase transition.
Each oscillator has a phase and a natural frequency. At $K = 0$ they spread uniformly around the circle. Past the critical $K$, oscillators near the mean frequency lock together first, forming a cluster that grows with $K$. The order parameter $r$ measures coherence — 0 incoherent, 1 fully synced.
Figure 1. 50 oscillators on a unit circle, colored by natural frequency. Drag the coupling slider to watch them synchronize. The order parameter $r$ measures how tightly they cluster.
$r$ stays near zero until $K_c$, then lifts sharply — a phase transition like ferromagnetic magnetization or superconductivity onset. For a Lorentzian frequency distribution with half-width $\gamma$, $K_c = 2\gamma$ exactly. Drag the vertical line to pick $K$; the oscillator circle above updates to match.
Drag the vertical line to set $K$
Figure 2. The order parameter $r$ as a function of coupling strength $K$. The dashed line marks the theoretical critical coupling $K_c$. The transition is sharp: a small increase in $K$ near the threshold produces a large jump in coherence.
A narrow frequency distribution syncs easily; a wide one resists. A bimodal one produces partial synchronization: two clusters rotate at different speeds and never merge.
Figure 3. Different frequency distributions produce different synchronization behavior. Bimodal distributions can sustain two competing clusters that never fully merge.
Southeast Asian fireflies blink in unison through pulse coupling: each phase rises toward a threshold, and a flash gives neighbors a small phase advance.
Raise the coupling and random twinkling organizes into traveling waves. Click to perturb a region and watch the field heal.
Click on the grid to perturb a region
Figure 4. Pulse-coupled firefly oscillators on a grid. Each dot flashes when its phase crosses the threshold, nudging neighbors forward. Raise coupling to see random twinkling organize into synchronized waves.
Section 1 assumed mean-field coupling — every oscillator feeling the average. Most real coupling is local. A ring-coupled Kuramoto model (each drummer hearing only its two neighbors) is how a drum circle locks rhythm, how CPG neurons entrain, how a starling adjusts to its flanking birds. Local coupling is slower to sync and can sustain traveling phase waves, polyrhythmic clusters, or full unison.
Click a drummer to knock its phase out of place. Recovery time shrinks as coupling rises. Turn sound on to hear each flash as a click, pitched by ring position.
Click an oscillator to knock it out of phase
Figure 5. Ring-coupled oscillators with nearest-neighbor influence. The mean-phase arrow shows global coherence. Click to perturb; the recovery timer reports when order returns above 0.8.
The Millennium Bridge in London wobbled on its 2000 opening day because pedestrians unconsciously synced their steps to the bridge's lateral sway — a Kuramoto-like feedback loop. The fix was dampers that broke the coupling.