Ulnar motor and sensory nerve conduction velocities (NCV) were studied bilaterally in 20 able-bodied subjects for below elbow (BE) and across elbow (AE) segments to assess the effect of 4 different elbow positions on NCV (0 degrees, 45 degrees, 90 degrees, and 135 degrees). Although constant skin stimulation marker points were used, the AE segment length became progressively longer with increased elbow flexion. At 0 degrees flexion the AE segment motor NCV was found to be slower, and at 45 degrees it was found faster than the BE NCV. At each subsequent elbow flexion position (90 degrees and 135 degrees) there was an erroneous increase in motor and sensory NCV for the AE segments (p less than 0.01). This increase in AE NCV with elbow flexion was mostly due to stretching of skin over the flexed elbow. The nerve itself was observed in 4 cadaver specimens to slide distally with respect to the above elbow skin marker. Since 45 degrees elbow flexion was the position of least variation in motor NCV for AE and BE segments, this degree of elbow flexion appears to be optimum. From these measurements and from literature review neither short AE segment length (less than 10 cm) nor long AE segment length (greater than 15 cm) is optimum for measurement of AE NCV in the assessment of compressive neuropathy at the elbow. Short segments are subject to increased NCV variation while long segments may not detect pathological slowing of NCV only occurring over a short portion of the nerve.