Heteromeric α1β channels are the main isoform of glycine receptors (GlyRs) in adult mammalian neurones. We have recently formulated a detailed activation mechanism that accounts for the single-channel properties of recombinant α1β GlyRs (1). The aim of the present experiments is to determine whether the same kinetic scheme can provide a satisfactory description of the behaviour of native GlyRs. The lumbar spinal cord was dissected rapidly post mortem from juvenile rats (P12-15) and slices (350 µm thick) cut using a vibratome. Single channel recordings were performed from motoneurones (visually identified by their morphology and location in the ventral horn) in the cell-attached or in the outside-out configuration. Surprisingly, only a small minority of cell-attached patches contained glycine-activated channels (14/~1200 patches). On the contrary, GlyRs were observed in the majority of outside-out patches, possibly because outside-out patches have a much larger surface area. Channel properties were found to be substantially different in the two configurations. In cell-attached mode, only one current amplitude was observed in the presence of 30, 100 or 1000 µM glycine. Native GlyR amplitudes varied from patch to patch (range 1.8-3 pA) reflecting differences in the intracellular chloride concentration. Nevertheless, the single channel conductance (estimated applying a 1s voltage ramp) was similar to that measured in cell-attached experiments from recombinant heteromeric GlyRs (2) (41±1 pS (n=14) vs 39±1 pS (n=5); all data expressed as mean±S.E.M.). On the other hand, in excised patches, glycine channels opened to several different amplitudes (typically 2 or 3). Average chord conductance values were similar for GlyRs from spinal motoneurones (28, 38 and 46 pS, n=7, 1 mM glycine) and for recombinant α1β channels from HEK293 cells (27, 34 and 45 pS, n=9). In our investigation of the kinetic properties of native glycine channels, we focused on the cell-attached configuration, as this is closer to physiological conditions. Thus we selected two cell-attached experiments in which enough transitions were recorded and the signal to noise ratio allowed detection of transitions as short as 30 µs. At 1 mM glycine, the open period distribution was described by a single exponential (τ=7.0±0.3 ms) and 98±1% of the shut times distribution is described by an exponential with time constant of 11 µs, in good agreement with values from recombinant α1β GlyRs. The gating rates for the fully liganded native receptor, calculated from maximum likelihood fitting (HJCFIT program, http://www.ucl.ac.uk/Pharmacology/dc.html), were found to be similar to those estimated for recombinant receptors (1). This suggests that the gating behaviour of fully liganded native channels closely resembles that of recombinant α1β glycine receptors.