With the continued rise in skin cancers worldwide there is a need for effective skin protection against sunlight damage. It was shown previously that sunscreens, which claimed UVA protection (SPF 20 ), provided limited protection against UV-induced ascorbate radicals in human skin. Here the results of an electron spin resonance (ESR) investigation to irradiate ex vivo human skin with solar-simulated light are reported. The ascorbate radical signal in the majority of skin samples was directly proportional to the irradiance over relevant sunlight intensities (0.9–2.9 mW cm⁻²). Radical production (substratum-corneum) by UV (wavelengths <400 nm) and visible components (>400 nm) was ∼67% and 33% respectively. Ascorbate radicals were in steady state concentration at low irradiance (∼1 mW cm⁻² equivalent to UK sunlight), but at higher irradiance (∼3 mW cm⁻²) decreased with time, suggesting ascorbate depletion. Radical protection by a four star–rated sunscreen (with UVA protection) was optimal when applied as a thin film (40–60% at 2 mg cm⁻²) but less so when rubbed into the skin (37% at 4 mg cm⁻² and no significant protection at 2 mg cm⁻²), possibly due to cream filling crevices, which reduced film thickness. This study validates ESR determinations of the ascorbate radical for quantitative protection measurements. Visible light contribution to radical production, and loss of protection when sunscreen is rubbed into skin, has implications for sunscreen design and use for the prevention of free-radical damage.