Sunscreens are employed to mitigate the adverse effects of sunlight on skin but are primarily designed to prevent ultraviolet-B-associated burning and damage. The increasingly recognized role of ultraviolet A in aging, and possibly melanoma, highlights the need to include ultraviolet A screens; however, validation remains difficult. We have used a novel method to establish the efficacy of sunscreens, by measuring ultraviolet-A-induced free-radical production (thought to contribute towards ultraviolet-A-related aging and malignant change). Electron spin resonance spectroscopy was used to detect free radicals directly in human Caucasian skin during irradiation with levels of ultraviolet comparable to solar intensities. Using this system the protection afforded by three high factor sunscreens (sun protection factor 20+) that claim ultraviolet A protection was examined. Each sunscreen behaved similarly: at recommended application levels (> or = 2 mg per cm2) the ultraviolet-induced free radicals were reduced by only about 55%, and by about 45% at 0.5-1.5 mg per cm (0.5 mg per cm2 reported for common usage). A "free-radical protection factor" calculated on the basis of these results was only 2 at the recommended application level, which contrasts strongly with the erythema-based sun protection factors (mainly indicative of ultraviolet B protection) quoted by the manufacturers (20+). The disparity between these protection factors suggests that prolonged sunbathing (encouraged by use of these creams) would disproportionately increase exposure to ultraviolet A and consequently the risk of ultraviolet-A-related skin damage.