The loss of neurons is viewed as one of several causes of the deterioration of neural function during ageing. However, the existing experimental evidence for an age-related decrease in the neuronal number may be misinterpreted due to the way the cells are counted and to the interference of unsuspected degenerative pathology of the animals studied. To reinvestigate this question we have quantified as easily identifiable population of neurons, the cerebellar Purkinje cells, in very old but healthy rats. The number of Purkinje cells in the cerebellum was assessed in two populations of rats: control (10 months) and old (42 months) rats from the Wistar/Louvain strain. In both groups, paraffin-embedded brains were cut serially in the sagittal plane. Purkinje cells were counted every 15 or 22 sections under the light microscope at a magnification of 1250 x. The raw value of cell counts were corrected according to the method of Hendry (21) in order to avoid the overestimation due to splitting of the nucleus during sectioning. The latero-lateral extent of the cerebellar cortex, obtained by multiplying the thickness of the section by the number of sections in which Purkinje cells were counted, was not statistically different (mean +/- standard deviation): 12.8 +/- 1.16 mm (n = 6) for the control rats and 12.0 +/- 1.02 mm for the old animals (n = 8) (Student's t-test, p = 0.18). The corrected number of the Purkinje cells (mean +/- standard deviation) was 330,350 +/- 35,448 cells (n = 6) for the control animals and 299,019 +/- 50,223 (n = 8) cells for the old rats. The 9.5% reduction observed in aged animals was not statistically significant (p = 0.22). In conclusion, our studies indicate that, in the cerebellum of the Wistar/Louvain rat, if anything, the effect of ageing upon the Purkinje cell population is very mild.