Estimates of neuron numbers have been useful in studies of neurodegenerative disorders, and in their animal models, and in the computational modeling of hippocampal function. Although the retrohippocampal region (presubiculum, parasubiculum, and entorhinal area) is an integral part of the hippocampal circuitry and is affected selectively in a number of disorders, estimates of neuron numbers in the rat retrohippocampal region have yet to be published. Such data are necessary ingredients for computational models of the function of this region and will also facilitate a comparison of this region in rats and primates, which will help to determine how well we may expect rat models to predict function and dysfunction in primate brains. In the present study, we used the optical fractionator to estimate the number of neurons in the rat retrohippocampal region. The following estimates were obtained: 3.3 x 10(5) in presubicular layers II and III, 1.5 x 10(5) in parasubicular layers II and III, 2.2 x 10(5) in the combined pre- and parasubicular layers V and VI, 6.6 x 10(4) in medial entorhinal area (MEA) layer II, 1.3 x 10(5) in MEA layer III, 1.9 x 10(5) in MEA layers V and VI, 4.6 x 10(4) in lateral entorhinal area (LEA) layer II, 1.2 x 10(5) in LEA layer III, and 1.4 x 10(5) in LEA layers V and VI. A surprising finding was the large numbers of neurons in the pre- and parasubiculum, which indicate an important role of these areas in the control of the entorhino-hippocampal projection. A comparison of the numbers of neurons in the hippocampus and entorhinal areas in rats with similar estimates in humans revealed that gross input-output relations are largely conserved. Differences between rats and humans may be accounted for by more prominent entorhino-neocortical projections in primates and consequent increases in the number of neurons in the hippocampus and retrohippocampal region, which are dedicated to these projections.