1. Ca2+ imaging and simultaneous intracellular recording were performed on CA3 pyramidal neurons in hippocampal slice cultures and standard acute slices. Both fura-2 and a dextran conjugate of fura-2 (MW = 10,000) were used in the Ca2+ measurements to control for compartmentalization artifacts. Experiments were performed under conditions giving minimal ligand- and voltage-gated Ca2+ influx, with the use of competitive and noncompetitive antagonists of ionotropic glutamate receptors and steady- state depolarization, respectively. 2. Tetanic stimulation of stratum lucidum evoked dendritic Ca2+ transients with rapid onset that were blocked by the noncompetitive N-methyl-D-aspartate (NMDA) receptor antagonist. MK-801 (25 μM), but not by the competitive α-amino-3-hydroxy-5-methyl-4- isoxazolepropionic acid (AMPA) receptor antagonist, 6-cyano-7- nitroquinoxaline-2,3-dione (CNQX) (10-50 μM). Zn2+-containing mossy fiber terminals (assessed by Timm's staining) and postsynaptic structures (thorny excrescences) are preserved in s. lucidum of hippocampal slice cultures. 3. A Ca2+ store loading protocol, consisting of brief repolarizations followed by steady depolarization, primed most of the neurons so that a subsequent tetanus gave a Ca2+ increase in the presence of MK-801 that was reported by both fura-2 and the dextran conjugate. The onset of the Ca2+ increase was significantly delayed (by 2-3 s) with respect to the MK-801-sensitive increase, and often had a different spatial pattern within the neuron. Response characteristics were similar in slice cultures and acute slices. 4. The delayed Ca2+ increase showed a steep rundown with subsequent stimuli, but was restored by further priming by the Ca2+ store loading paradigm. Postsynaptic currents evoked by the tetani under these conditions were not correlated with the magnitude of the delayed Ca2+ transients. 5. Delayed Ca2+ increases were observed in 44% of the neurons dialyzed with normal intracellular solution at room temperature. The success rate of observing delayed Ca2+ transients was increased to 86% in neurons maintained at 30°C, and dialyzed with an inhibitor of the inositol-triphosphate-3-kinase. 6. The delayed Ca2+ transients could not be initiated after inhibition of endosomal Ca2+-ATPase-mediated uptake by thapsigargin. 7. Both fura-2 and the dextran conjugate reported increases in resting Ca2+ levels after the loading protocols, that were absent after priming in thapsigargin, and decreases in resting Ca2+ levels after successive tetani in MK-801, suggesting that the Ca2+ changes were largely cytosolic. 8. The present results support the hypothesis that these synaptically mediated, delayed Ca2+ transients represent release from intracellular Ca2+ stores that can be loaded and depleted repeatedly, and are evoked by presynaptic release of endogenous neurotransmitter.