The clinical value of latency measurement of tendon reflexes in neurological patients has been reported by several authors. However, normal values are not readily comparable. In the present study, latencies and amplitudes of patellar (PTR) and ankle tendon reflexes (ATR) were measured at rest and after facilitation in 102 normal controls. A manually operated reflex hammer, tipped with electrically conductive rubber, ensured an immediate start of the sweep of the oscilloscope. Latencies showed a significant correlation with height (r = 0.70 for PTR and r = 0.72 for ATR, P < 0.0001) and to a lesser degree with age (r = 0.16 and r = 0.30, P < 0.0001). While amplitudes were highly variable, rendering them less useful for diagnostic purposes, latencies showed minimal intra-individual variability (CV 1.5 and 0.8%, respectively). Correlation of ATR-latency with the H-reflex latency of the soleus muscle was very high (r = 0.97, P < 0.0001). Comparison with three other hammer types yielded corresponding results with a hammer supplied with a piezo-electric element; however, significantly shorter latencies were found with a hammer with a microswitch, and with another hammer with a spring-contact, due to a delay from the tap on the tendon until the start of the sweep of the monitor.