ADA specification No. 32 for determining the range (elastic limit) of orthodontic wires uses the bending of a wire section treated as a cantilever beam. An alternative method for defining the range of orthodontic wires proposed by Waters (1981) is to wrap wire sections around mandrels of varying diameters and measure the deformation imparted after unwrapping. Four brass mandrels with a total of 46 test diameters ranging from 3.5 to 60.0 mm were used in this study. Wire sections 9 cm in length were rolled on the mandrel with a hand lathe. The mandrel cross section required to produce a predetermined amount of deformation (2 mm arc height for a 5 cm chord) was defined as the yield diameter for that particular wire. No individual wire was tested twice so as to avoid introduction of strain history. Test samples of 488 different orthodontic wires supplied by nine commercial distributors were evaluated (a total of 4,747 samples). Stainless steel wires of identical dimensions had a large variation in range, depending on the state of strain hardening and heat treatment. For example, 0.020 inch round wire had yield diameters ranging from 22.8 mm for Australian special plus orange (TP Laboratories) to 42.9 mm for Nubryte gold (G.A.C. International). Chromium cobalt wires had less range than stainless steel before heat treatment, but increased greatly in range after heat treatment. Nitinol (Unitek) had the greatest range of all wires tested (yield diameter of 8.7 mm for 0.016 inch Nitinol). Multistranded stainless steel wires had yield diameters between 9.0 and 14.0 mm.