Glenoid reaming is a challenging bone machining operation performed in shoulder joint replacement surgery and represents a suitable target for further developments in haptic surgery simulation. Tool vibration is an important mode of feedback for surgeons to perform glenoid reaming, and thus it is quintessential to be included in virtual replications of this operation. However, there is currently little known about the characteristics of vibrations in glenoid reaming and how it can be reproduced in haptic simulations. To address this, the current study presents an empirical investigation of glenoid reaming vibrations using robot-driven experiments on human cadaveric glenoids. Time- and frequency-domain characteristics of vibration signals obtained from the experiments are analyzed. The power of the vibration signals are predicted as a function of feed force and bone density and the dominant frequencies are identified. The obtained results serve as an important step towards the development of haptic simulators for upper limb orthopedic procedures.