The ankle clonus test, a method for evaluating the integrity of the spinal cord during operations for scoliosis, is predicated on the finding that patients recovering from general anesthesia normally have temporary ankle clonus bilaterally. An absence of transient ankle clonus has been shown to indicate neurological compromise. The test was performed for 1006 patients who were being managed with spinal arthrodesis and instrumentation and 115 control patients who had an operation under general anesthesia because of a condition that was unrelated to the spine. The six patients in whom a neurological deficit developed all had had a so-called positive result on the ankle clonus test (that is, an absence of transient ankle clonus). There were no false-negative results and three false-positive results; the test therefore had a sensitivity of 100 per cent and a specificity of 99.7 per cent. The ankle clonus test was found to be more accurate than the wake-up test and monitoring of somatosensory evoked potentials for predicting neurological compromise.