Objective - To clarify the optimum trace count for jitter calculation. Materials and methods - Single-fiber electromyography was performed during voluntary contraction (VC-SFEMG) of extensor digitorum muscle in six healthy and five myasthenic subjects. Twenty single-fiber action potential (SFAP) pairs were recorded from each subject. For each pair, 100 consecutive traces were recorded and stored for off-line analysis. Ten different jitters were calculated in 10 consecutive steps for each SFAP pair by increasing the number of traces included to analysis ( from 10 to 100). The jitter calculated in each step was compared with that calculated from 100 traces. Results - Sixty was the optimum number of trace count for both myasthenics and normal controls. Using less than 60 traces for jitter calculation revealed lower values. Conclusion - Jitter calculation should be performed from at least 60 traces in VC-SFEMG. Less numbers yield lower jitter and may cause diagnostic error.