Postnatal neuron addition, if it occurred, would have profound implications both for the conceptualization of developmental processes and for efforts directed at replacing neurons that were lost to injury or disease. Although dorsal root ganglia (DRGs) offer the advantages of clear boundaries and functional homogeneity, studies comparing neuron number in the DRGs of animals of different ages or sizes have yielded conflicting results. In the present study, neuron number in DRGs L3-L6 was compared in neonatal (approximately 11 days old, mean weight of 24.5 g, mean volume of 25 cm3) and adult (approximately 80 days old, mean weight of 373.5 g, mean volume of 346 cm3) male Sprague-Dawley rats. Estimates of neuron number were derived by using both stereological (physical disector) and profile-counting (one or more nucleoli within a nucleus) methods. The reliability and validity of the two methods were evaluated by comparing estimates of neuron number with those derived from three-dimensional reconstruction of a subset of neurons. The recommended protocol for using the physical disector was found to give accurate estimates of neuron number, but the heterogeneous distribution of neurons in the ganglion led to sampling errors of up to 50%. Reliability was improved by increasing the number of disector pairs examined. Counts of nuclear/nucleolar profiles were more reliable, but introduced a bias that worked against the experimental hypothesis in that estimates of neuron number in neonates exceeded actual values. Nonetheless, both methods indicated that adult rats had more DRG neurons than did neonates. Profile counts were 19% higher in adults (P < .01, two-tailed t-test); and data obtained by using the physical disector showed that adult rats had 28% more neurons than did neonates (P < .05). The difference in neuron number between adults and neonates could be due either to neuron proliferation or to late differentiation of neurons that do not assume a typical appearance until adulthood.