Background: Nonammoniated latex, ammoniated latex, and rubber glove extracts are the only sources of natural rubber (Hevea brasiliensis) latex that have potential for use as skin testing reagents in the diagnosis of latex allergy. Their diagnostic sensitivity and specificity as skin test reagents are unknown. Objective: We conducted a phase 1/2 clinical study to examine the safety and diagnostic accuracy (sensitivity and specificity) of nonammoniated latex, ammoniated latex, and rubber glove extracts as skin test extracts to identify the most efficacious source material for future skin test reagent development. Methods: Twenty-four adults not allergic to latex, 19 adults with hand dermatitis or pruritus, and 59 adults with a latex allergy were identified by clinical history. All provided blood and then received puncture skin tests and intradermal skin tests with nonammoniated latex, ammoniated latex, and rubber glove extracts from Malaysian H. brasiliensis latex by use of sequential titration. A glove provocation test and IgE anti-latex RAST were used to clarify positive history–negative skin test response and negative history–positive skin test response mismatches. Results: All three extracts were biologically safe and sterile. After normalization to 1 mg/ml of total protein, all three extracts produced equivalent diagnostic sensitivity and specificity in puncture skin tests and intradermal skin tests at various extract concentrations. Optimal diagnostic accuracy was safely achieved at 100 μg/ml for puncture skin tests and 1 μg/ml for intradermal skin tests (e.g., nonammoniated latex: puncture skin test sensitivity 96%, specificity 100%; intradermal skin test sensitivity 93%, specificity 96%). The presence of IgE antibody in skin was highly correlated with IgE anti-latex in serum (nonammoniated latex: r = 0.98, p < 0.001; ammoniated latex: r = 0.94, p < 0.001; rubber glove extract: r = 0.96, p < 0.001). All five available subjects with a positive history, negative skin test response, and absence of IgE antibody in serum had a negative glove provocation test response, indicating no clinical evidence of latex allergy. No systemic or large local allergic reactions were observed with puncture skin tests or intradermal skin tests. Conclusions: Equivalent diagnostic sensitivity and specificity were observed with the nonammoniated latex, ammoniated latex, and rubber glove extract skin test reagents after normalization for total protein; nonammoniated latex may be considered the reagent of choice on the basis of practical quality control and reproducibility considerations. (J A LLERGY C LIN I MMUNOL 1996;98:872-83.)