Why is the sky blue, not violet?

Violet wins the scattering race

Sunlight contains every colour⁶. Tiny nitrogen and oxygen molecules in the air scatter short waves much more than long ones^{6, 9}. Air molecules send this scattered light off in all directions, so it reaches us from all over the sky¹⁶. This is Rayleigh scattering, and its strength scales as one over the wavelength to the fourth power⁸. Lord Rayleigh, then called John William Strutt, derived this law in 1871³. Only in 1899 did he show that air molecules alone are enough to colour the sky³.

Do the maths and violet at 400 nm scatters about 1.6 times more than blue at 450 nm, and roughly 9 to 10 times more than red at 700 nm⁸. So why isn't the sky violet? Physicists point to several factors working together⁹.

Less violet arrives, and ozone isn't the culprit

The Sun does not shine equally in all colours^{8, 9}. Measurements from a NASA research aircraft flying at about 12 km, corrected to the top of the atmosphere, give about 1.5 watts per square metre per nanometre at 400 nm, but about 2.1 at 450 nm⁵. That is roughly a third more blue than violet⁵. Scattering more than makes up for it. A rough estimate of our own, multiplying these sunlight values by the scattering law, gives about 20% more scattered light at 400 nm than at 450 nm, before losses on the way through the air, which shrink that lead^{5, 8}. A 2026 preprint, using the same kind of model, likewise states that skylight contains more violet than blue¹³.

A popular claim says ozone eats the violet. The evidence points elsewhere. The ozone layer absorbs nearly all of the most harmful ultraviolet light, but it is mostly transparent to visible light¹². Its visible absorption band, the Chappuis band, is centred around 600 nm, in the orange³. In 1953 E. O. Hulburt found that ozone has little effect on the daytime sky's colour⁴. But at sunset and twilight it helps keep the overhead sky blue⁴. A 2023 computer study confirmed that, with the Sun on the horizon, ozone causes about two-thirds of the overhead blue, and around half when the air is hazier³.

Your eyes do the final mix

For engineer Glenn Smith, this is the half of the story that usually gets left out^{1, 2}. Our eyes have three kinds of colour-sensing cones, with peaks near 442, 543 and 570 nm². By the standard 1924 brightness curve, the eye responds about 100 times more weakly at 400 nm than at 450 nm; newer, corrected curves put the gap closer to 15 to 20 times^{10, 15}. The eye's lens also blocks most ultraviolet, and it absorbs more short-wave light as it yellows with age¹¹.

In 2005 Smith showed that skylight is a 'metameric' match to unsaturated blue¹. Put simply, your eye can't tell the sky's complicated spectrum from a mix of pure blue and white². Calculations place that matching blue around 470 to 480 nm¹³. So the sky looks neither violet nor deep blue, but pale blue^{1, 8}.

Near the horizon, light crosses more air, so the blue looks paler^{6, 7}. At sunset, the long path scatters blue away, leaving reds and oranges^{6, 7}. On Mars, fine dust flips the script: sunsets there glow blue around the Sun, while the rest of the sky looks yellow to orange¹⁴.