Microdialysis probes were utilized to follow the appearance and disappearance of dialyzable aluminum (Al) in rat and rabbit brain and liver extracellular fluid compared to blood after iv Al lactate or Al citrate injection. Dialyzable Al was assumed to be the fraction not protein bound or self-associated into complexes greater than the molecular weight cutoff of the dialysis membrane. Aluminum concentrations peaked in brain frontal cortex and ventral hippocampus and in the liver in the first 20-min dialysis sample, indicating rapid Al penetration into the extracellular space of these organs. In vitro recovery experiments conducted with microdialysis probes at room temperature revealed an average dialysis efficiency of about 10% for both Al lactate and citrate. At 37 degrees C Al recovery increased for both Al lactate and citrate. In vivo Al recovery from rabbit blood averaged 5.15% for Al lactate and 3.25% for Al citrate. These observations are consistent with results from recovery studies of other substances showing an increased recovery with increased temperature but an overestimate of recovery by in vitro methods. Tissue/blood Al ratios (TBR; representing dialyzable extracellular tissue Al divided by dialyzable blood plasma Al) for liver were approximately 1, suggesting unhindered diffusion of Al between blood and liver. In contrast, brain TBR were less than 1, demonstrating a partial blood-brain barrier to Al. The brain TBR for Al lactate was greater than TBR for Al citrate, suggesting that Al citrate did not preferentially penetrate the blood-brain barrier. Higher TBR were seen in the rabbit than the rat, perhaps contributing to the greater susceptibility of the rabbit to Al-induced neurobehavioral toxicity. Metals can be repetitively sampled in the extracellular space using microdialysis, enabling metal toxicokinetic determinations in these compartments.