In the preceding report (Alstermark and Sasaki 1986) it was shown that a stimulus of 500 microA applied in the lateral reticular nucleus (LRN) evokes a maximal or near monosynaptic EPSP (LRN EPSP) in forelimb motoneurones. This EPSP which is assumed to be selectively mediated by C3-C4 propriospinal neurones (PNs), was used to estimate the strength of the excitatory projection from C3-C4 PNs. A systematic comparison was made of the size and time course of the maximal LRN EPSP in various species of forelimb alpha-motoneurones innervating shoulder, elbow, wrist and digit muscles. The LRN EPSP was evoked in all investigated species of forelimb motoneurones. When either the peak amplitude or the underlying area of the LRN EPSP was compared, a three-fold range was found with some tendency for the size to vary in the order of wrist greater than shoulder approximately equal to elbow greater than digit greater than intrinsic paw motor nuclei. Generally, a positive correlation was found in each motor nucleus between the peak amplitude of the LRN EPSP versus the monosynaptic homonymous group Ia EPSP, input resistance and afterhyperpolarization duration respectively (cf. Alstermark and Sasaki 1986). It is therefore postulated, that the LRN EPSP peak amplitude is correlated with motor unit type. Comparison of the time course of the LRN EPSPs was made by measuring the time-to-peak (T-t-p) and half-width (H-w). The finding in the preceding report that the T-t-p and H-w is longer in slow than in fast motoneurones was confirmed and extended to all the investigated motor nuclei. The hypothesis that both fast slow motoneurones receive projection from a group of fast C3-C4 PNs, while slow motoneurones receive an additional projection from a group with lower conduction velocity, can therefore be applied to all forelimb motor nuclei. In addition, it is proposed that some slow shoulder, wrist and digit motoneurones receive projection from a special subpopulation of C3-C4 PNs with very slow conduction velocity.