The objective of this study was to define adequate patch test materials to evaluate mercury allergic contact dermatitis. We applied 0.1% and 0.05% mercuric chloride, and 0.5% and 0.2% mercury in petrolatum to systemic eczematous contact-type dermatitis (baboon syndrome), and gold-dermatitis patients. All baboon-syndrome patients reacted not only to mercuric chloride but also to metallic mercury. In gold-dermatitis patients, significantly more patients reacted to mercuric chloride than to metallic mercury (21 of 35, 60%, versus 2 of 19, 10.5%, p < 0.0005). We speculated that sensitization to mercury may be of 2 types: one a reaction to ionized mercury only, the other to both ionized mercury and non-ionized mercury. The possibility that the phenomenon is caused by differences in bioavailability or percutaneous penetration between ionized and non-ionized mercury cannot be ruled out, but could be explored by penetration measurement. For the evaluation of mercury hypersensitivity, it may be more reliable to apply both ionized and non-ionized mercury, rather than only mercuric chloride or ammoniated mercury.