When rehabilitating patients with back dysfunction, extension exercises that are presumably "passive" for the erector spinae muscles are frequently used. The purpose of this study was to record electromyographic (EMG) activity from back extensor muscles to determine if these muscles are truly inactive during these maneuvers. Surface EMG was recorded bilaterally from lower lumbar muscles in 62 pain-free subjects. The root-mean-square EMG activity, recorded in microV, for the four positions tested was (mean +/- SD): lying prone (1.268 +/- 0.902), extension in lying prone (6.713 +/- 6.976), standing neutral (4.760 +/- 3.282), and extension in standing (3.558 +/- 2.273). One-way ANOVA for repeated measures and Tukey's post hoc test were used to analyze the data. The results indicated that EMG activity was greatest for extension in lying prone (p < .05), equivalent between the two standing positions (p > .05), and least when lying prone (p < .05). The results of this study demonstrated that "passive" extension exercises were not truly passive for lumbar back extensor muscles. From a clinical perspective, if the performance of passive back extension is important, extension in lying prone may not be the exercise of choice and having patients lying prone may be the most beneficial.