So-called anti-irritants (AI) are widely used in cosmetic formulations, with the aim of reducing irritation from substances in the formulation. It may also be claimed that they are 'soothing' and 'healing' ingredients. However, the proof for these claims is circumstantial. The dose-response effect of 4 alleged AI (nifedipine, (-)-alpha-bisabolol, canola oil and glycerol) was studied on experimentally induced acute irritation in healthy volunteers, and only glycerol showed dose-related response and effects potentially better than no treatment. The acute irritation model only allowed a small window of opportunity in which to demonstrate efficacy. Therefore, the effect of AI was studied in a cumulative irritation model by inducing irritant dermatitis with 10 min daily exposures for 5+4 days (no irritation on weekend) to 1% sodium lauryl sulfate on the right and 20% nonanoic acid on the left volar forearm. AI ointments were applied twice daily. Clinical scoring was performed daily, evaporimetry (Trans Epidermal Water Loss), hydration and colourimetry were measured at baseline (D0), in the middle and at the end of treatment. The glycerol ointment was the only treatment statistically better than both 'no treatment' and vehicle.