1. Intracellular recordings from ninety-nine motoneurones have been made in an in vitro hemisected spinal cord preparation. Their mean resting membrane potential in normal artificial cerebrospinal fluid (CSF) was -71 +/- 0.5 mV (+/- S.E.M.). The mean amplitude of the action potential was 84.0 +/- 1.4 mV (n = 50), and the mean input conductance was 101 +/- 7 nS (n = 49). 2. Both membrane potential and input conductance were sensitive to changes in [K+]o, [Na+]o, [Cl-]o and [Ca2+]o. 3. Replacement of extracellular Ca2+ by Mn2+ resulted in less than 1 mV hyperpolarization and a decrease in input conductance from 102 +/- 7 to 93 +/- 6 nS (n = 15). 4. At high [K+]o (greater than 10 mM) the membrane potential followed the potential predicted by the Nernst equation for K+ ions with a slope of 58 mV per 10-fold change in [K+]o. At low [K+]o (less than 10 mM) there was significant deviation from K+ equilibrium potential (EK). 5. [K+]i was found to be 106 mM when estimated from the reversal potential of the after-hyperpolarization of the antidromic action potential. 6. The reversal potential of the recurrent inhibitory postsynaptic potential (IPSP) in normal CSF was used to calculate [Cl-]i. This was 6.6 mM, which is less than would be expected if Cl- was passively distributed, indicating the presence of an outwardly directed Cl- pump. 7. Decreasing [Cl-]o from control (134 mM) to 4 mM resulted in a depolarization of 6.9 +/- 0.9 mV and a decrease in input conductance from 102 +/- 5 to 90 +/- 5 nS (n = 14) in 3 mM [K+]o. 8. Decreasing [Na+]o from 156 to 26 mM by substitution with choline resulted in a 6.2 +/- 0.5 mV hyperpolarization and a decrease in input conductance from from 102 +/- 4 to 76 +/- 4 nS (n = 5) in 3 mM [K+]o. 9. The input conductances for Na+, Cl- and K+ at the resting potential were calculated. After allowing for a microelectrode leak conductance, the relative input conductances were gNa/gK = 0.13 and gCl/gK = 0.25.