The importance of neural activity and motoneuronal cell death in the formation of specific synaptic connections between muscle afferents and motoneurons was studied in chick embryos. Patterned neural activity was blocked by applying d-tubocurarine (dtc) chronically to embryos during the period when sensory-motor connections are formed [stages (St) 28– 42]. Dtc blocks neurogenic muscle contractions, thereby abolishing any temporal correlation between neural activity in motoneurons and stretch- sensitive afferents. The normal pattern of motoneuronal bursting is also blocked (Landmesser and Szente, 1986), as is motoneuronal cell death (Pittman and Oppenheim, 1979). Dtc applications were started more than 1 d before muscle sensory afferent collaterals make anatomical contact with motoneuronal dendrites and continued until St 38–42, when the pattern of synaptic connectivity was examined by recording synaptic potentials intracellularly from identified lumbosacral motoneurons upon stimulation of identified populations of muscle afferents. In both normal and dtc-treated animals, large monosynaptic excitatory potentials were evoked in homonymous motoneurons (those that supply the same muscle as the sensory afferents) and were often observed in motoneurons that supplied synergistic muscles. Monosynaptic potentials were uncommon in motoneurons supplying antagonistic muscles. The overt patterns of sensory-motor connections in normal and dtc-treated embryos were essentially identical. However, the amplitudes of the composite EPSPs recorded in dtc-treated animals were consistently about twice as large as normal. These observations suggest that neither normal patterns of neuronal activity nor motoneuronal cell death play a large role in determining the specificity of connections between the sensory and motor neurons involved in the stretch reflex.