The excitotoxic hypothesis of Huntington's disease pathogenesis suggests that selective striatal neuronal loss results from excessive activation of striatal excitatory amino acid receptors. Using a microdialysis probe mated to an Alzet 2002 mini-osmotic pump three different concentrations of quinolinic acid or vehicle were administered to the striata of rats over a 3-week period. Animals that received a total of 3.3 μmol of quinolinic acid had significant striatal atrophy that could be attributed to two distinct areas of neuronal loss. First, an area of necrosis surrounding the probe was marked by inflammatory infiltrate and a lack of neurons. In the second region, surrounding the necrotic area, there was a significant reduction in nissl-stained cells, with relative sparing of NADPH-diaphorase-staining neurons. In addition, there was a reduction in cytochrome oxidase staining throughout both of the areas of cell loss. Beyond the area of cell loss, the striatum appeared normal in all respects. The striata of animals that received 880 nmol quinolinic acid appeared identical to those that received vehicle. The striata of animals that received 8.8 μmol quinolinic acid showed severe nonselective atrophy of the striatum and some surrounding structures. We conclude that dialytic delivery of 3.3 μmol quinolinic acid produces an area of neuronal destruction that resembles the selective neuronal loss seen in Huntington's disease. This selective neurodegeneration produced by chronic exposure to quinolinic acid simulates more closely the course of Huntington's disease than previously described methods.