Six Japanese cases of griseofulvin (GF)‐induced photodermatitis were reported. All of the patients had received GF from 4 days to 4 months as treatment for tinea unguium. Clinical features were basically eczematous, but differed somewhat case by case. Patient 2 showed pigmentation and depigmentation and patient 5 showed pellagra‐like eruptions. From the clinical view, we could not differentiate precisely between phototoxic and photoallergic reactions. The difference may also depend on response severity, duration of skin eruption, or skin color. Blood and urine porphyrins were within normal ranges. An action spectrum study revealed that UVA was responsible for eliciting a photosensitivity reaction. Patch and photopatch tests were performed on all 6 patients. The results of the photopatch tests were positive for 3 patients and negative for the other 3. The degrees of reaction to photopatch tests differed so widely that it was difficult to determine suitable conditions for UVA intensity or GF concentrations for photopatch testing. Moreover, half of the patients showed negative photopatch tests. Although there is merit in the use of photopatch tests for studying the causes of systemically administered drug‐induced photodermatitis, there may also be limits to their usage. When performed carefully, drug readministration is a more reliable method. From the results of positive photopatch tests, photocross reactions between GF and penicillin, persistent light reactions and distant flare up phenomena, we concluded that in some of our cases, reactions were photoallergic. Histological findings showed non‐specific inflammatory changes. Direct immunofluorescence studies showed that immunoglobulin and complement were deposited at papillary perivascular areas and the dermo‐epidermal junction.