Short answer
Air molecules are far smaller than the wavelength of light and scatter it in proportion to roughly the inverse fourth power of wavelength — Rayleigh scattering — so short wavelengths are redirected across the sky much more than long ones. The sky is blue rather than violet because sunlight contains less violet, some is absorbed high up, and the eye is far less sensitive to it; by day ozone contributes only a few per cent.
Sunlight crossing the atmosphere meets nitrogen and oxygen molecules a thousand or so times smaller than a wavelength of visible light. Particles that small scatter light with a strength that rises steeply as wavelength falls — in proportion to about one over the wavelength to the fourth power. By that law alone, 400 nm violet is scattered roughly nine times as strongly as 700 nm red, and 450 nm blue about six times. Look at any part of the sky away from the sun and you are seeing sunlight that has been knocked sideways by molecules, weighted heavily towards the short end of the spectrum. With no atmosphere, as on the Moon, the sky is black even in full sun.
If scattering alone decided, the sky would look violet. Three things intervene. The sun emits less violet than blue to begin with, and some of the shortest wavelengths are absorbed high in the atmosphere. The eye's sensitivity falls steeply below about 450 nm, so violet contributes little to the colour we perceive. And skylight is not a single wavelength but a broad mixture weighted towards the blue end; the visual system reads that mixture as sky blue, slightly desaturated by the green and even red it still contains. The result is not a spectral colour at all, which is why no single paint or screen colour quite reproduces a clear sky.
A 2023 radiative-transfer study revisited a claim first made by Hulburt in 1953: that ozone, not only scattering, colours the sky. Ozone absorbs weakly across the orange and green in its Chappuis bands. With the sun high, the study found this adds only a few per cent to the zenith's blueness; Rayleigh scattering dominates. With the sun on the horizon, the balance flips and about two thirds of the zenith's blue comes from ozone absorption. That is the reason the deep blue of twilight is covered on the site's golden hour and blue hour page rather than here.
Overhead, you look through the shortest path of air, and most light reaching you has been scattered once. Towards the horizon the path is many times longer; light is scattered repeatedly, blue light is scattered out of the line of sight as well as into it, and the mixture that arrives is closer to white. Aerosols — dust, sea salt, pollution, water haze — add wavelength-independent scattering that whitens it further. A deep, saturated blue overhead with a pale band near the horizon is therefore a sign of clean, dry air; a milky sky all the way up means plenty of particles. The same aerosols matter to the colour of sunsets.
| Wavelength | Scattered relative to 700 nm |
|---|---|
| 400 nm | 9.4 |
| 450 nm | 5.9 |
| 500 nm | 3.8 |
| 550 nm | 2.6 |
| 600 nm | 1.9 |
| 650 nm | 1.3 |
| 700 nm | 1.0 |
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
FactStrong evidence
Rayleigh scattering by air molecules is more effective at short wavelengths, so light scattered down from the sky away from the sun is weighted towards the blue end of the spectrum.
Source: HyperPhysics (optics, atmospheric optics and vision pages); The Color of Clouds (JetStream)
FactStrong evidence
Radiative-transfer modelling attributes about 3–4% of the zenith sky's blue colour to ozone absorption when the sun is high (solar zenith angles of 10–50°), rising to about 66% with the sun at the horizon.
Method: Radiative-transfer simulations analysed in CIE colour space
Caveat: Values depend on the ozone column and aerosol load assumed.
Colourwise analysisStrong evidence
Under an inverse-fourth-power law, 400 nm light is scattered about nine times as strongly as 700 nm light, and 450 nm about six times.
Based on: Computed by Colourwise as (700/λ)⁴; see the table.
Caveat: Scattering efficiency only; it ignores the sun's spectrum, absorption and the eye's sensitivity, which together set the colour we see.
Source: HyperPhysics (optics, atmospheric optics and vision pages)
Reviewed 29 September 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.