Quantitative studies of the vestibular system with serially sectioned human temporal bones have been limited because of difficulty in distinguishing hair cells from supporting cells and type I from type II hair cells. In addition, there is only a limited amount of normative data available regarding vestibular hair cell counts in humans. In this study, archival temporal bone sections were examined by Nomarski (differential interference contrast) microscopy, which permitted visualization of the cuticular plate and stereociliary bundle so as to allow unambiguous identification of hair cells. The density of type I, type II, and total numbers of vestibular hair cells in each of the 5 sense organs was determined in a set of 67 normal temporal bones that ranged from birth through 100 years of age. The mean total densities at birth were 76 to 79 cells per 0.01 mm2 in the cristae, 68 cells per 0.01 mm2 in the utricle, and 61 cells per 0.01 mm2 in the saccule. The ratio of type I to type II hair cells at birth was 2.4:1 in the cristae and 1.3:1 in the maculae. There was a highly significant age-related decline in all sense organs for total, type I, and type II hair cell densities that was best fit by a linear regression model. The cristae lost type I cells with advancing age at a significantly greater rate than the maculae, whereas age-related losses for type II cells occurred at the same rate for all 5 sense organs. Hair cell densities in the cristae were significantly higher at the periphery than at the center. There were no significant sex or interaural differences for any of the counts. Mathematical models were developed to calculate the mean and 95% prediction intervals for the total, type I, and type II hair cell densities in each sense organ on the basis of age. There was overall good agreement between the hair cell densities determined in this study and those reported by others using surface preparation techniques. Our data and related models will serve as a normative database that will be useful for comparison to counts made from subjects with known vestibular disorders.