ABSTRACT. The T-wave of the patellar tendon reflex (PTR) was recorded in 24 neurologically normal dogs. The surface electromyogram (EMG) was recorded as the T-wave from the vastus lateralis muscle (VL) in response to percussion of the patellar tendon. The di stance of the reflex arc (DRA) was measured along the straight line between the spinous process of L5 and the greater trochanter (GT), and between GT and the patellar ligament (PL). There was a significant correlation (P<0.001) of the latency with the DRA on each side, but no difference in the slopes of the relationships between right and left VL was shown. The regression line between the DRA a nd the latency of all data was Y = 0.0216X + 1.693, where Y = latency in ms, X = DRA in mm. The mixed sensory-motor conduction velocit y was estimated as 84.6 ± 5.5 m/s. In contrast, there was no significant correlation between the DRA and the amplitude of the T-waves. The mean (mean-CV) and standard deviation (SD-CV) of all CV (coefficient of variation) in each dog were 9.14 ± 3.65% in latency and 3.54 ± 1.14% in amplitude, indicating that the use of a simple hand-held reflex hammer is sufficient to record the reproducible T-wave of the PTR even in unanesthetized dogs. This method was applied to a case with minimal paraparesis, and the latency of the T-wave of the PTR in the right hind limb with slight proprioceptive deficit was outside of the upper limit of the 95% confidence interv al between latency and the DRA. In conclusion, this method may be used in neurological diagnosis to quantify more precisely the PTR in dogs. KEY WORDS: canine, paraparesis, patellar tendon reflex, T-wave, 95% confidence interval.