[en] The US Bureau of Mines borehole deformation gauge (BMG) was designed in the early 1960's to allow rock stress measurements by the overcoring method. Since that time it has become a de facto standard against which the performance of other borehole deformation gauges is often judged. However, during recent in situ stress studies in the Climax Stock at the Nevada Test Site a strange ''negative hysteresis'' in the order of 300 to 500 microstrains was observed in standard calibration data. Here, the relaxation curve lies below the indentation (compression) curves as if the system were to somehow respond with an energy release. Therefore, a precision micro-indentation apparatus has been designed and used to perform a series of tests allowing a better understanding of the BMG button to cantilever interaction. Results indicate that the hysteresis effect is caused by differential motion between the button base and the cantilever resulting from the geometric motion inherent in the cantilever. The very large apparent hysteresis is mainly caused by cycling opposing cantilevers through the instrument's entire dynamic range, and the fundamental imprecision inherent in use of the standard micrometers to calibrate the BMG. Laboratory mean hysteresis magnitudes for a polished cantilever typically range from 3 to 25 microstrain for 100 and 1000 microstrain relaxations on 1000 microstrain deflection loops intended to simulate typical field data. The error percentage is thought to remain fairly constant with deformation loop size, and is sufficiently small such that it can be safely ignored. The hysteresis effect can probably be reduced, and instrument stability improved by machining a small 90 degree cone in the cantilever in which a slightly larger mating cone on the base of the indentation button would reside. 5 refs. 26 figs., 1 tab