№ 165 · physics

A rainbow is one reflection inside every raindrop, piled up at 42°

A rainbow is not a coloured object hanging in the sky. It is sunlight that has gone into raindrops, bounced once off the back of each drop, and come out bent by close to the same angle from every drop. Your eye picks out the drops that sit at that angle, and they form a circle around the point directly opposite the sun.

What one drop does

Follow one ray of sunlight into a round drop of water. It bends as it enters, crosses the drop, reflects once off the inside of the far surface, crosses back, and bends again as it leaves. That is two refractions and one reflection. Jagla and Rojo give this as the ray-optics account of the main bow in a 2025 paper. In his Opticks, Isaac Newton credits the idea to Archbishop Antonius de Dominis, in a book published in 1611, and says Descartes pursued it and improved the account of the second bow.

A drop has no single exit direction. A ray that hits the drop dead centre comes straight back toward the sun. A ray that grazes the edge comes out at a different angle again. Every ray between those two leaves at its own angle, so one drop sprays light over a wide range of directions.

Predict first

You look at the red band of a rainbow. What colours of light are the raindrops in that band sending out?

Check your answer

The last one. Every drop splits all of sunlight, but each colour leaves the drop at its own angle. From the drops at about 42° from the point opposite the sun, red comes to your eye most strongly. Their violet leaves at a slightly smaller angle and misses you. Newton gives this account of the inner bow in his Opticks.

The angle that light piles up at

Start with the central ray and move the entry point outward. The angle between the ray coming out and the line pointing straight back along the sunlight grows, then reaches a largest value and shrinks again. Near a turning point like that, a whole band of entry points sends its light out at almost the same angle. So the light from a drop is not spread evenly. It piles up at that largest angle. Jagla and Rojo work through this with Snell's law and get about 42° for water. Past it there is no once-reflected light at all, so the rainbow is the sharp edge between a bright region and a dark one. Physicists call an edge like that, where rays bunch together, a caustic.

Colour enters because water bends each colour by a slightly different amount. Newton took the bending of red and of violet from his own prism measurements and worked the angle out for each. Red's largest angle was 42° 2′ and violet's was 40° 17′. So red always forms the outer edge of the main bow, with violet on the inside. Each colour also lights the whole disc inside its own edge, a little more faintly. That is why the sky inside a bright rainbow looks lighter than the sky just outside it.

Interactive One raindrop, with sunlight arriving from the left. Slide where the ray enters, from the centre of the drop to its edge, and watch the angle it leaves at. Switch between red and violet light and between one and two reflections. The bars on the right count where all the rays end up. Every control works with Tab, Enter and the arrow keys.

where the ray enters0.60 of the way out
sunlight → dashed: straight back toward the sun
Move the slider to send a ray into the drop at a different place.

A cone with your eye at the tip

Now step back from the drop to the sky. Draw a line from the sun, through the back of your head, and on out in front of you. It points at the spot directly opposite the sun, which is where the shadow of your head falls. Every drop that sits 42° away from that line, in any direction, sends its red light to you most strongly. Those drops lie on a cone with your eye at its tip, and the cone looks like a circle. This is Newton's construction in Opticks.

Two things follow from the geometry. First, the person beside you sees a rainbow made by different drops, because their cone has its tip at their eye. Second, the centre of the circle is always below the horizon when the sun is up. When the sun is more than about 42° above the horizon, the whole main bow is below it, and from flat ground you see no bow at all.

The seven colours are a count, not seven bands

The spectrum inside a rainbow does not come in seven stripes. Newton himself describes the colours as passing gradually into one another, as they do from a prism. When he needed to divide them up, he cut the circle of colours into seven parts sized to match the seven intervals of a musical scale: red, orange, yellow, green, blue, indigo and violet. The seven are names laid over a smooth run of colour, not seven separate bands in the sky.

Real bows smear the colours further. Jagla and Rojo describe each wavelength as making its own disc of light, brightest at its edge and fading inward, and these discs overlap. The smaller the drops, the wider each disc spreads, so the colours mix. Large drops, such as those after a summer thunderstorm, give sharp, vivid bows. Very small drops give fuzzy, whitish ones.

The second bow, and the dark band between them

Some light reflects twice inside the drop before it leaves. That light has a smallest angle instead of a largest one: about 51° for red and about 54° for violet, by Newton's figures. So the second bow sits outside the first, with its colours in the opposite order, red on the inside. It is fainter because, as Newton puts it, light becomes fainter at every reflection.

Between about 42° and about 51°, neither path sends any light. Once-reflected light cannot reach past 42°, and twice-reflected light cannot come in closer than 51°. That strip of sky is darker than the sky on either side, and it is called Alexander's dark band. Selmke and Selmke show how a classroom version of this, a round flask of water lit through a hole in a screen, can hide the dark band. Up close, the two bows still overlap. The band opens up only at a large enough distance, and the flask's glass wall, even an ordinary thin one, narrows it further.

Where rays stop being enough

Look just inside the main bow on a good day and you may see faint extra arcs, separated by darker gaps. The ray picture cannot explain them. Jagla and Rojo explain that for every angle inside the bow, two different rays leave a drop heading the same way. Light is a wave, so these two can add or cancel. Thomas Young and then George Airy used this to explain the extra arcs, which are called supernumerary bows. Their spacing depends on drop size, which is why, Jagla and Rojo argue, they show up best near the top of the bow, where the falling drops are still smallest.

In short

Each raindrop takes in sunlight, reflects it once off its back, and sends it out over a range of angles that piles up at about 42° from the point opposite the sun. Red piles up a little farther out than violet, so the colours separate into a ring. The ring is a cone of drops with your eye at the tip, so everyone sees their own. Two reflections give a fainter, reversed bow near 51°, with a dark band between. The seven colours are Newton's way of dividing a smooth spectrum.

Where this comes from

  1. Common principles behind rainbows and boat wakes reuse permitted with attribution
    Eduardo A. Jagla and Alberto G. Rojo · arXiv:2508.05644 · 2025
    arxiv.org/abs/2508.05644
  2. Revisiting the round bottom flask rainbow experiment linked only, not reproduced
    Markus Selmke and Sarah Selmke · arXiv:1612.09563 · 2016
    arxiv.org/abs/1612.09563
  3. Opticks, 4th edition, Book I Part II, Props. VI and IX linked only, not reproduced
    Isaac Newton · 1730
    www.gutenberg.org/ebooks/33504