This study was designed to determine intraobserver and interobserver reliability of maximal muscle strength for upper and lower extremity muscle groups. Four healthy subjects were tested with a portable muscle dynamometer on two separate occasions by three separate examiners to determine maximal isometric strength of lateral pinch between thumb and index finger, elbow flexion, elbow extension, shoulder flexion, hip flexion, hip extension, hip abduction, knee flexion, and ankle dorsiflexion. A break test was also used to evaluate the strength of cervical flexor muscles and extensor hallucis longus. Pearson correlation coefficients for inter- and intraexaminer testing was good for upper extremity test values from 0.85 to 0.99. The variation coefficient of the methodology error (CV) in all upper extremity muscle tests was between 5.1% and 8.3%. These results seemed reliable for clinical muscle strength testing. However, correlation coefficients for lower extremity testing were poor, with values ranging from -0.20 to 0.96. The CV in these tests was much greater than in the upper extremities and ranged from 11.3% to 17.8%. Both break tests also had high CV, with each being greater than 17%. Our results demonstrate that although the dynamometer is reliable for testing upper extremity muscle groups, it is unreliable for testing lower extremity muscle groups. Further work is needed to evaluate muscle strength quantitatively in the clinical setting in an accurate, valid, and reliable manner.