Investigators involved in a clinical study on dental implants being conducted by the Dental Implant Clinical Research Group participated in a two-part in vitro evaluation of the Periotest system. The first part involved each of 35 investigators taking three readings for each of 16 models designed to simulate an implant in place in the oral cavity and to cover most of the effective measuring range of the instrument. The second part, with which this article is concerned, was conducted by three study investigators to determine the level of agreement in readings for six different Periotest instruments. Readings were taken according to the manufacturer's instructions until two coincident readings were obtained. The readings were compared to evaluate inter- and intrainstrument reliability. Instrument repeatability was evaluated by examining the number of attempts required to obtain the two coincident readings. The Periotest system generally demonstrated a high degree of repeatability and reliability, with higher variability associated with specific model samples.