Wavelets everywhere, all the time
Christiaan Huygens proposed in 1678 a strikingly simple picture of how a wave advances: every point on a wavefront is itself the source of a small secondary wavelet, spreading outward at the wave's own speed, and the new wavefront a moment later is just the common envelope — the surface tangent to all of those wavelets at once. It sounds almost too simple to be useful, yet building the envelope over and over again reproduces straight-line propagation, reflection, refraction and diffraction from one repeated geometric operation, with no separate rule needed for each phenomenon.
Deriving Snell's law from wavelet geometry
Huygens' most powerful trick is deriving the law of refraction without ever invoking rays. When a wavefront strikes a boundary between two media at an angle, part of the front reaches the second medium — where the wave travels slower or faster — before the rest of it does. The part already across the boundary radiates wavelets at the new (slower, for a denser medium) speed while the rest keeps radiating at the old speed, so the envelope tilts: the whole wavefront pivots toward the boundary's normal on the slow side. Working through the resulting triangle geometry gives exactly Snell's law:
n₁ sinθ₁ = n₂ sinθ₂ n = c / v (refractive index = vacuum speed / speed in the medium) Larger n → slower wave speed → wavelets bunch up → wavefront bends closer to the normal.
The same reasoning, applied to a boundary where the wave simply bounces back without changing speed, gives the ordinary law of reflection: angle of incidence equals angle of reflection, again purely from the geometry of matching up wavelet envelopes on either side of the mirror.
Diffraction: the construction meets an edge
Point a wave at a narrow slit and the wavelets that would have come from the blocked parts of the wavefront simply are not there, so the envelope on the far side of the opening can no longer close into a clean flat front. Wavelets from the unblocked strip spread the wave sideways into what would otherwise be geometric shadow — this bending around an edge is exactly what diffraction means, and it is completely absent from a picture of light as straight, unbendable rays. The narrower the slit is relative to the wavelength, the closer the transmitted wave gets to a single point source radiating in every forward direction.
From wavelets to an interference pattern
Downstream of a slit, wavelets launched from different points across the opening have travelled slightly different distances to reach any given observation angle, so they arrive out of step with each other by an amount that depends on that angle. Summing all of those phase-shifted contributions gives strong reinforcement straight ahead and a series of dark fringes where the contributions cancel:
a sinθ = mλ single-slit diffraction minima, m = ±1, ±2, ... d sinθ = mλ double-slit interference maxima, m = 0, ±1, ±2, ...
Put two slits side by side and each one is its own Huygens source of a spreading diffracted wave; the two overlapping waves interfere with each other on top of the single-slit spreading each one already has on its own, producing the familiar fine interference fringes riding inside a broader diffraction envelope — everything traceable back to the same rule of summing wavelet contributions.
The gap Fresnel had to close
Huygens' original construction has a loose end: nothing in it explains why wavelets shouldn't also radiate backward, toward the source, which would predict a spurious backward-travelling wave that is never observed. Augustin-Jean Fresnel closed the gap by weighting each wavelet's contribution with a directional (obliquity) factor that suppresses the backward direction, turning a qualitative geometric sketch into the quantitatively correct Huygens-Fresnel principle used in modern diffraction calculations, later placed on rigorous mathematical footing by Gustav Kirchhoff working directly from the wave equation.
Frequently asked questions
How does Huygens' principle explain Snell's law of refraction?
Picture a wavefront crossing a boundary at an angle: the part still in the fast medium keeps racing ahead while the part already in the slow medium lags behind, so the wavelets on the slow side bunch up and the whole envelope pivots toward the normal. Working through the geometry of that pivot gives exactly n1 sin(theta1) = n2 sin(theta2), derived purely from the fact that wave speed differs between media, with no assumption about light being made of particles or rays.
Why do you need Fresnel's correction to the original construction?
Huygens' raw wavelet construction has no built-in reason wavelets shouldn't also radiate backward toward the source, which would predict a nonsensical reflected wave travelling back the way the original wave came. Fresnel fixed this by adding a directional weighting (an obliquity factor) that suppresses backward radiation and correctly weights each wavelet's forward contribution, and this combined Huygens-Fresnel principle is what actually gets diffraction intensities right, not just diffraction's rough shape.
Does the wavelet construction also explain simple mirror reflection?
Yes, by the same logic as refraction but without a speed change. Each point on the wavefront that reaches the mirror becomes a new wavelet source reflecting back into the original medium at the same speed it arrived, and building the envelope of those reflected wavelets reproduces the ordinary law of reflection, angle of incidence equals angle of reflection, directly from the geometry of the construction.
Try it live
Everything above runs in your browser — open Huygens' Principle and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Huygens' Principle simulation