On 25 September 1998 a moderate-size mbLg 5.2 earthquake occurred in northwestern Pennsylvania, near the Ohio border. Early source analyses suggested an unusual, non-double-couple component to the faulting mechanism, but subsequent checks of the near-real-time solutions suggested the non-double-couple component may have been an artifact. We investigate the size of the non-double-couple faulting component, improve mechanism and depth estimates for the Pymatuning event, and explore the reason(s) why early estimates contained large non-double-couple source components. We conclude that the non-double-couple component, while possible, is unnecessary and can be interpreted as a consequence of small amplitude features in near-nodal Rayleigh waves. We show that the Pymatuning earthquake can be explained with a pure double-couple faulting mechanism, corresponding to a near-vertical, mostly strike-slip fault with planes striking 110° and 13°, with dips of 70° and 71° and rakes of 20° and 159°. The estimated moment is 5.6 × 1022 dyne-cm (Mw = 4.5). Regional waveforms constrain the depth of the event to be shallower than 7.5 km, and a short period teleseismic P waveform from northwest Russia suggests a shallower 2-4 km source. The roughly east-west- or north-south-striking vertical strike-slip mechanism agrees well with existing estimates of the stress field and is similar to the 1986 eastern Ohio earthquake mechanism.