Twenty-two patients with polymorphous light eruption were prophylactically treated with ultraviolet A (UVA) with and without trimethylpsoralen in the first randomized double-blind study in this subject. Twelve of the patients were treated during 2 consecutive springs with placebo during one spring and psoralens during the other. Eighteen of the patients improved after the therapy, but there was no clear-cut difference between the 2 regimens. As many as 12 patients got light eruptions during the treatment, but all but one continued with the therapy. This study indicates that UVA alone is as good prophylactic therapy for polymorphous light eruption as PUVA with trimethylpsoralen. However, because of the high incidence of provoked eruptions during therapy, the treatment may be difficult to handle for the patients themselves, at least during the initial treatment.