The effects of axotomy on the size and number of rat dorsal root ganglion cells was studied using stereological methods. Twenty adult Wistar rats were axotomized by transection of the right fifth lumbar spinal nerve approximately 7 mm distal to the fifth lumbar dorsal root ganglion (DRG-L5). The corresponding ganglia from the nonaxotomized side served as controls. The DRG-L5 were removed for study 4, 8, 15, and 45 days after axotomy. The number of neurons in each DRG-L5 was determined from estimates of the numerical density, NV, made with disectors and estimates of the volume of the ganglion using the Cavalieri principle. The mean cell body volume was determined with the vertical planar rotator method. There was a progressive loss of nerve cells during the postoperative period. There was a loss of 6% (not significant) after 4 days, 19% (not significant) after 8 days, 22% (2P < 0.05) after 15 days, and 35% (2P < 0.005) after 45 days. The relative reduction in cell number 45 days after axotomy was larger for B-cells (43%) than for A-cells (15%). The mean nerve cell body volume for the entire DRG-L5 cell population was reduced by 33% (2P < 0.005) 4 days after axotomy and remained so throughout the experimental period. The distribution of the individual cell volumes in the ganglia appeared to be uniformly shifted to lower values. It is concluded that permanent axotomy of the fifth lumbar spinal nerve results in a substantial loss of dorsal root ganglion cells and is well-suited as a model for studying the potential protective effects of neurotrophic factors using modern stereological techniques.