Background: Many patients who give a history of a dermatitis reaction to jewelry or metal contact with skin are negative to metals on standard patch testing. Some may be showing false-negative reactions. Objective: Our purpose was to determine whether patients with a history of jewelry reactions but whose standard patch tests were negative have a false-negative reaction or are allergic to metals other than nickel, cobalt, or chromium. Methods: Four hundred forty-nine patients were studied who gave a history of reacting to jewelry or metal. Of these, 210 were tested to the metals in the European standard series (ie, nickel, cobalt, and chromate), and 239 were tested to the metals in the standard series and to an extended metal series of palladium, gold, platinum, a second nickel salt, and a nickel/cobalt mixture. These were compared with 752 patients who did not give a history of jewelry or metal reactions, of which 50, besides the standard series, were also treated with the additional metal series. Results: A higher proportion of jewelry-reactive patients tested with the extended series reacted to nickel (and to other metals) than those who were tested only with the European standard series: (61% vs 38%; P < .0001 ). The use of the extended series showed that palladium allergy was common, present in 34% of nickel-allergic patients, but it always occurred with nickel sensitivity. Gold allergy coexisted with nickel sensitivity in 10% of cases. Testing simultaneously with separate patches containing 5% nickel sulfate and 5% nickel chloride showed a concordance of 71% in identified nickel-sensitive patients. Nickel sulfate was more likely than nickel chloride to detect nickel sensitivity. The use of a combined preparation of 2.5% nickel sulfate and 0.5% cobalt chloride in petrolatum revealed only 3 jewelry-reactive patients who were negative to other metals. There was a slightly higher proportion of atopic patients in the patch test–negative jewelry reactors group than in the positive group; however, the difference was not significant and it was not sufficient to account for the negative findings. Conclusion: Some jewelry reactors who had negative patch tests are likely to be subclinically allergic to nickel. We suggest that the higher number of antigens, or perhaps the larger nickel load, in the extended metal series, resulted in a larger proportion of patients reacting. To better demonstrate nickel allergy in jewelry reactors, patients should be patch tested to a metal series that contains palladium and gold salts and perhaps a second nickel patch because these may reveal the presence of nickel sensitivity when standard patch tests would otherwise have been negative. (J Am Acad Dermatol 2000;43:31-6.)