Many peripherally administered substances which modulate memory do not freely enter the brain. Such compounds may act through peripheral receptors that send messages centrally through vagal afferents. To explore this hypothesis, rats were chronically implanted with cuff electrodes on the left cervical vagus nerve. Each animal was trained 48 h after surgery on a one-trial inhibitory-avoidance task with a 0.75-mA, 1.0-s footshock. Immediately following training, each animal received either no stimulation or vagal stimulation (0.5-ms biphasic pulses; 20 Hz, 30 s) at one of three intensities (0.2, 0.4, 0.8 mA; eight animals per group). Retention was tested 24 h later. Neither the 0.2-in or 0.8-mA groups (22.1-s; 53.7-s median latency) showed altered retention performances, whereas the 0.4-mA group showed significantly improved retention (881.0 s) compared to unstimulated controls (21.1 s; U = 6, p < .01). This inverted-U shaped function indicates that vagal activation during memory consolidation modulates retention for memory tasks.