1. Using a rabbit cerebellar slice preparation, we stimulated a classical conditioning procedure by stimulating parallel fiber inputs to Purkinje cells with the use of a brief, high-frequency train of eight constant-current pulses 80 ms before climbing fiber inputs to the same Purkinje cell were stimulated with the use of a brief, lower frequency train of three constant-current pulses. In all experiments, we assessed the effects of stimulation by measuring the peak amplitude of Purkinje cell excitatory postsynaptic potentials (EPSPs) to single parallel fiber test pulses. 2. Intradendritically recorded Purkinje cell EPSPs underwent a long-term (> 20 min) reduction in peak amplitude (30%) after paired stimulation of the parallel and climbing fibers but not after unpaired or parallel fiber alone stimulation. We call this phenomenon pairing-specific long-term depression (PSD). 3. Facilitation of the peak amplitude of a second EPSP elicited by a parallel fiber train occurred both before and after paired stimulation suggesting that the locus of depression was not presynaptic. Depression of the peak amplitude of a depolarizing response to focal application of glutamate following pairings of parallel and climbing fiber stimulation added support to a suggested postsynaptic locus of the PSD effect. 4. The application of aniracetam potentiated EPSP peak amplitude by 40%, but these values returned to baseline as a result of pairings. With the removal of aniracetam from the bath 20 min after pairings, normal levels of pairing-specific EPSP depression were observed, indicating that the effect did not result from direct desensitization of alpha-amino-3-hydroxy-5-methyl-4-isoxazole-proprionic acid (AMPA) receptors. 5. Incubation of slices in the protein kinase inhibitor H-7 potentiated EPSP peak amplitudes slightly (9%), but peak amplitudes returned to baseline levels after pairings. The net reduction in EPSP peak amplitude of < 10% after pairings suggested that H-7 partially blocked PSD and that, in turn, PSD involved protein kinases. 6. The means of induction and the specificity of those means suggest that the phenomenology of PSD is fundamentally different from that of long-term depression. PSD only occurs with pairings of trains of parallel fiber and climbing fiber stimulation; it occurs without the need for bicuculline; and it can overcome the blocking effects of aniracetam. 7. Nevertheless, the involvement of protein kinases and the potential role of calcium suggest that the mechanisms involved in the induction of PSD and long-term depression have a number of features in common. 8. Because of the pairing-specific nature of the long-term synaptic depression observed in these experiments, PSD provides a mechanism that may contribute to the role of the cerebellar cortex in classical conditioning.