The strength of lower limb musculature, as determined by clinical muscle testing, influences assigned lesion levels, prognosis, and the subsequent orthopedic and rehabilitative management of children with meningomyelocele. The purpose of this study was to determine the accuracy of muscle testing at various ages based on clinical impressions of therapists; and to determine how the stability of measurements over time varies with age and muscle tested. A total of 3,084 examinations were performed on 825 children with meningomyelocele. Therapists coded the degree of accuracy of the last 1,282 examinations. Of these patients 177 with six or more serial examinations were studied by generating Spearman rank order correlation coefficients for comparison of strength measurements at given ages with median strengths during five-year follow-up periods. The percentages of patients who had examinations at a given age equal to, and within one grade of individual best estimates of ultimate muscle strength were also calculated. It was concluded that: therapists are markedly more confident about assessments in children five years and older; stability of muscle testing measurements over time generally increases from birth to age five or six when peak stability is approached; the probability that a single muscle test precisely predicts future muscle strength varies with age and muscle tested; and the probability that a muscle test will predict future strength to within one grade is not strongly influenced by age or muscle tested.