Perforated and non-perforated synapses in the molecular layer of rat parietal cortex have been assessed morphologically and quantitatively using three-dimensional reconstructions of the postsynaptic terminal. Perforated synapses were analyzed at nine ages, ranging from 0.5 to 22 months of age, and non-perforated synapses at three ages--0.5, 12, and 22 months. Examination of the reconstructions shows that perforated synapses increase in size and complexity with increasing age. This increasing complexity is reflected in a break-up of the postsynaptic density, which is punctuated by larger, branched perforations. In the most extreme cases the result is the appearance of isolated islands of postsynaptic density separated by, and also surrounded by, a synaptic contact zone. Spinules are especially prominent at around 12 months of age in perforated synapses, and the overall negative curvature of the young junctions is replaced by positively curved junctions from 4 months onwards. The non-perforated synapses are relatively small and show few changes with increasing age. Using the measurement option in the reconstruction program, the following trends emerged. All parameters of perforated synapses increased in size with increasing age, whereas the corresponding parameters of non-perforated synapses remained relatively unchanged over this age range. In addition, the percentage of the synaptic contact zone surface area occupied by the postsynaptic density decreased with increasing age in perforated synapses, but increased in non-perforated synapses. The total postsynaptic density surface area of non-perforated synapses per unit volume of molecular layer was double that of perforated synapses at 0.5 months, but the situation was reversed at 12 months. This parameter was similar in the 2 populations at 22 months. This suggests that perforated synapses contribute more to the total surface area of the postsynaptic density in mid- to late-adulthood than do non-perforated synapses, despite non-perforated synapses outnumbering perforated by 2-3:1 at these ages. These data provide more specific evidence that perforated and non-perforated synapses constitute separate synaptic populations from early in development, and that perforated synapses are responsible for the maintenance of neuronal postsynaptic density surface area from mid-adulthood onwards.