Recently, when investigating the effect of graded stimulus strengths on the size of the conducted monophasic spike in muscle afferent fibres, Brock, Eccles & Rall (1951) found that with biceps-semitendinosus nerve the range of stimulus strengths from threshold to that giving the maximum group I spike was much larger than with other muscle nerves. There was a three-to fourfold increase of strength in contrast to an increment of 1-6-2-3 for other nerves. It has also been an almost invariable finding that, when recorded triphasically (a contact electrode against an indifferent electrode), an afferent volley from the biceps-semitendinosus or quadriceps nerves showed a double negative spike (cf. Brock, Coombs & Eccles, 1952, fig. 9), which was never observed for afferent volleys from the leg muscles-gastrocnemius, flexor digitorum longus, extensor digitorum longus, etc. Since the fast afferent fibres of muscle (the group I) are known to have two major components of quite different reflex function, which respectively may be designated group I a from the annulo-spiral endings and group I b from the Golgi tendon organs (Matthews, 1933;Barker, 1948;Granit, 1950; Granit & Strom, 1951; Hunt & Kuffler, 1951 a, b;Hunt, 1951 Hunt, , 1952Laporte & Lloyd, 1951, 1952Hagbarth & Naess, 1950), it was of interest to determine whether the double negative spike was attributable to this dual composition. If this was so, then the discrimination between the reflex functions of groups Ia and Ib would become simpler because the relative group I a and I b composition of an afferent volley would be directly determinable.
METHODCats, anaesthetized with pentobarbitone sodium (Nembutal), have been used in all experiments. The method, and the dissection of nerves and their mounting on electrodes, have resembled in general those procedures described by Brock et al. (1951). Special precautions have been taken to ensure that the stimulus did not 'escape' to nerves other than those on the electrodes. For example the sartorius and saphenous components of the femoral nerve were dissected up and cut about 2 cm proximal to the cathodal electrode on quadriceps nerve. The stimuli were brief