Under the assumpation that the constant biphasic appearance of H-wave is one of the useful parameters in H-reflex study, this study was designed to compare the 3 different recording techniques of H-reflex with surface electrode. The H-reflexes were recorded in 20 normal healthy male subjects from the posterior side of soleus; midpoint between medial malleolus and popliteal crease (methold 1), the medial head of gastrocnemius muscle (method 2), and the soleus, just medial to tibial crest, at the midpoint of tibia (method 3). Also, by usin g the above 3 methods, intrapersonal variations of H-reflexes were compared with one another. The conclusions were as follows.

1) H-latency at the medial gastrocnemius (method 2) was shortest as 28.01 msec and H-latency of posterior soleus was similar with that of medial soleus.

2) H-amplitude and H-area at the medial soleus (method 3) were largest as 5.98 mV and 21.76 ՌVsec, respectively.

3) The shape of H-reflex was biphasic with initial negative deflection in method 3, but triphasic with initial positive deflection in method 1 and method 2.

4) There was no significant difference in intrapersonal variabilities among the above e methods.

So, we think it is better to record H-reflex at the medial side of soleus (method 3) than posterior side. And H-reflex recording at the medial gastrocnemius can be used as useful as that of soleus.