The purpose of this study was to investigate secondary muscle spindle afferents from the triceps-plantaris (GS) and posterior biceps and semitendinosus (PBSt) muscles with respect to their fusimotor reflex control from different types of peripheral nerves and receptors. The activity of single secondary muscle spindle afferents was recorded from dissected and cut dorsal root filaments in alpha-chloralose anaesthetized cats. Both single spindle afferents and sets of simultaneously recorded units (2-3) were investigated. The modulation and mean rate of firing of the afferent response to sinusoidal stretching of the GS and PBSts muscle were determined. Control measurements were performed in the absence of any reflex stimulation, while test measurements were made during reflex stimulation. The reflex stimuli consisted of manually performed movements of the contralateral hind limb, muscle stretches, ligament tractions and electrical stimulations of cutaneous afferents. Altogether 21 secondary spindle afferents were investigated and 20 different reflex stimuli were employed. The general responsiveness (i.e. number of significant reflex effects/number of control-test series) was 52.4%, but a considerable variation between different stimuli was found, with the highest (89.9%) for contralateral whole limb extension and the lowest (25.0%) for stretch of the contralateral GS muscle. The size of the response to a given stimulus varied considerably between different afferents, and, in the same afferent, different reflex stimuli produced effects of varying size. Most responses were characterized by an increase in mean rate of discharge combined with a decrease in modulation, indicative of static fusimotor drive (Cussons et al., 1977). Since the secondary muscle spindle afferents are part of a positive feedback loop, projecting back to both static and dynamic fusimotor neurones (Appelberg Et al., 1892 a, 1983 b; Appelberg et al., 1986), it is suggested that the activity in the loop may work like an amplified which, during some circumstances, enhance the effect of other reflex inputs to the system (Johansson et al., 1991 b).