A series of control, IgE myeloma containing and atopic sera were tested by the radioallergosorbent test [RAST] using identical lots of commercially available reagents following 3 different RAST protocols: Phadebas, Fadal-Nalebuff and a new method (Hoffman protocol). A new RAST tracer was also evaluated in the Phadebas and Fadal-Nalebuff methods. The Phadebas technique, using the standard 0.35 PRU [Phadebas RAST unit] cut-off, was the least sensitive at 68% but its sensitivity was raised to 84% using an 0.14 PRU cut-off. The specificity was 100% at 0.35 and 95% at 0.14 PRU. The Hoffman laboratory technique gave a sensitivity of 84% and a specificity of 100%. The Fadal-Nalebuff method had the highest sensitivity at 92% but the lowest specificity at 92%. The Fadal-Nalebuff technique leads to a significant number of false positive scores, particularly in sera with IgE > 500 U/ml or with RAST discs that have significant non-specific binding. The Fadal-Nalebuff method can be used reliably if modified to include suitable negative controls for each antigen tested.