1. The synaptic connections of 44 single identified muscle spindle Ia afferents and of 21 Golgi tendon organ (Ib) afferents from medial gastrocnemius (MG) were studied in 46 cats by the spike-triggered averaging of synaptic noise in 803 motoneurons of various types. 2. The well-known monosynaptic Ia excitatory connections were confirmed and their characteristics examined in 113 cells. The method was used at greater sensitivity than before and revealed that, in addition to the larger EPSPs of the order of 300 muV, there were many below the previously reported lower 17-muV limit. 3. By studying the Ia disynaptic inhibitory pathway with quick stretch-evoked Ia volleys and by spike-triggered averaging (STA), it was shown that the latter method can reveal disynaptic and possibly trisynaptic excitatory connections. This is believed to depend on having continuous activity in the relevant interneurons. 4. Latencies of individual connections showed broad distributions and arguments are advanced for setting working limits to mono- and disynaptic paths for Ia excitation and inhibition. Monosynaptic EPSP latency from cord entry was 0.4-1.1 ms and disynaptic inhibition was 1.2-2.4 ms. It was recognized that the boundaries are not rigid and monosynaptic Ia EPSPs may have latencies up to 1.5 ms. 5. Rise times of disynaptic PSPs were, on average, significantly longer than monosynaptic, but individual disynaptic responses could have values within the monosynaptic range. 6. A small diphasic wave shortly preceding the monosynaptic EPSPs was interpreted as a presynaptic spike. Its timing was consistent with this and, as such, permitted estimates to be made of central conduction time. 7. An early negative wave (latency less than or equal 1.1 ms) of small emplitude was sometimes detected in antagonist motoneurons when triggering from Ia afferents. It was found tha extracellular fields could be detected due to single Ia afferent excitations and efforts were made to see if the early negative wave could be explained by this. In a few cases there was evidence that a very short-latency IPSP might be occuring. This evidence and its implications are discussed with attention to the new factors which have to be considered in using the spike-triggered averaging method at very high sensitivity. 8. Ib effect were di- or trisynaptic. They were excitatory to 18% of synergists and to 28% of antagonists. They were inhibitory to 41% of synergists and to 19% of antagonists. The Ib IPSPs were larger than the EPSPs.