The purpose of this study was to determine whether surface electromyography (EMG) from the erector spinae muscles could correctly identify individuals with low back pain without a population of elite athletes. A similar technique had previously been successful in identifying low back pain patients within a non-athletic population. A Back Analysis System was used to compute the median frequency of the EMG power density spectrum to monitor metabolic changes in back muscles associated with muscle fatigue. Twenty-three members of a men's collegiate varsity crew team consisting of port (N = 13) and starboard (N = 10) rowers were tested in a laboratory during a fatigue-inducing isometric contraction sustained at a relatively high, constant force. Six of the rowers tested were further classified as having low back pain. A brief test contraction was repeated at a fixed interval following the fatiguing contraction to monitor recovery. A two-group discriminant analysis procedure correctly classified 100% of the rowers with low back pain and 93% of the rowers without back pain on the basis of the median frequency data. The median frequency parameters related to recovery were the best discriminators of back pain. A similar analysis correctly classified 100% of the port rowers and 100% of the starboard rowers on the basis of their spectral parameters. The best discriminating variables in this instance were the median frequency parameters relating to both fatigability and recovery. Results from this study demonstrate that low back pain and asymmetrical muscle function in rowers can be assessed on the basis of EMG spectral analysis.