The Ely Test (or Duncan-Ely test) has been accepted as a clinical tool to assess rectus femoris spasticity by passively flexing the knee rapidly while the patient lies prone in a relaxed state. In this retrospective review, patients' dynamic knee range of motion (ROM) during gait and an electromyogram (EMG) were compared with the results of the Ely test. Data for 70 patients (44 males, 26 females; 104 limbs) were included. Mean age of patients was 13 years, SD 9 years, range 4 years 5 months to 54 years. All patients were diagnosed with cerebral palsy (spastic diplegia, n=42; spastic quadriplegia, n=15, and hemiplegia, n=13). All patients were ambulatory (50 independent, 20 with assistive devices). A standard matrix was used to calculate sensitivity and specificity of the Ely test as well as its positive and negative predictive value. For the gait variables examined (decreased dynamic knee ROM, timing of peak knee flexion, and abnormal EMG in swing) the sensitivity of the Ely test ranged from 56 to 59% and the specificity ranged from 64 to 85%. For the same variables the positive predictive value ranged from 91 to 98% and the negative predictive value ranged from 4 to 19%. The Ely test was shown to have a good positive predictive value (i.e. the certainty about the presence of rectus spasticity in patients with a positive Ely test result) for rectus femoris dysfunction during gait.