The aim of this study is to describe the reliability of measuring maximal strength of eight muscle groups of the lower limb by a hand-held dynamometer, according to a standard assessment protocol. The study population consisted of 26 patients (14 males and 12 females; age range 60-90 years) admitted to a geriatric hospital. Multiple assessments of muscle strength by two different examiners were compared to estimate test-retest and inter-rater reliability. The range of strength evaluated across the eight muscle groups was 2.1-29.8 Kg/force. Overall, short-term (same day) and long-term (one week apart) test-retest and inter-rater reliability were very high, with 60% of the intraclass correlation coefficient values above 0.8, and the majority above 0.7. No significant differences in strength were found comparing the left and the right side of each muscle group. Differences between values collected in the same subject by two different examiners, and by the same examiner at different points in time were similar, not influenced by the average strength of the muscle group, and significantly larger for long-term than for short-term comparisons. By using a standardized measurement protocol, reliable measurements of muscle strength can be obtained by a hand-held dynamometer in frail older persons.