To compare vibrotactile thresholds between five international test centres, to determine causes of variability, and to provide normative data for use by health professionals. Vibrotactile thresholds were measured on various fingers in 1,008 subjects at 31.5 Hz (2,531 measurements) and at 125 Hz (2,807 measurements). Three centres used the up-and-down method of limits, rate of change of stimulus 3dB/s, to obtain vibrotactile thresholds. The push force on the surround was 2 N, the probe contact force was 1 N and the probe diameter was 6 mm concentric to a 10-mm diameter hole in the surround. One centre used an up-and-down method of limits with a 5-dB step and a 1.26-mm diameter probe with no surround. Another centre used a stepping algorithm, no surround, and the force on the 6-mm-diameter probe was unspecified. The 31.5-Hz thresholds and 125-Hz thresholds, expressed in dB, were similar between centres using similar methods. There were small (<5 dB) differences in vibrotactile thresholds between male and female subjects, between white-collar and blue-collar workers and between fingers. Both 31.5-Hz and 125-Hz thresholds increased with age; the mean effect of age on 31.5-Hz thresholds was approximately 2 dB over 25 years, while the effect of age on 125 Hz-thresholds was as high as 10 dB over 25 years. Vibrotactile thresholds are influenced by measurement method but can be similar at different centres if similar methods are used. Effects of measurement location and gender are often negligible. Thresholds depend on age, but over the range of ages considered, only the effect on 125-Hz thresholds is sufficient for correction; methods of correcting for age are given. These normative data may be used to assist the diagnosis of peripheral neuropathy involving vibrotactile sensation.