The possible role of precipitation losses in eroding stormtime ring current is subject to debate. To explore this controversy, the recovery phase of the February 6--10, 1986, great magnetic storm is examined, when intense ion precipitation was observed at midlatitudes by NOAA-6 and DMSP satellites. This storm period is particularly interesting because the ring current exhibits distinctive two-phase decay as seen in the Dst index, the early rapid timescale decay corresponding to the intense ion precipitation period described above. Hamilton et al. [1988> concluded, from close agreement between the observed timescale for ring current decay and the theoretical timescale for O+ charge exchange loss, that rapid early recovery phase of this storm resulted from the charge exchange loss of high energy O+; the second and longer decay phase was equated with H+ charge exchange loss. A model of the ring current evolution during this great magnetic storm [Fok et al., 1995> failed to reproduce the observed ring current decay rates, a puzzling result because charge exchange losses were well represented in the ring current model and initial and boundary conditions were taken from the same data set used in the Hamilton et al. [1988> study. A simple energy balance calculation for the global ring current is carried out using (1) either an energy input predicted from upstream solar wind parameters or one calculated from the drift-loss model output, (2) collisional loss timescales extracted from the drift-loss model, and (3) precipitation losses estimated from NOAA-6 and DMSP observations. The energy balance model replicates the evolution of the ring current energy content derived from Active Magnetospheric Particle Tracer Explorers/Charge Composition Explorer (AMPTE/CCE) observations when ion precipitation losses are included and model energy input function is reduced to agree with predictions based upon upstream solar wind parameters. The O+ charge exchange losses and observed global precipitation losses were of equal magnitude in early recovery of the ring current during this great magnetic storm. Later longer decay timescales in the model resulted from a combination of O+ and H+ charge exchange losses; O+ charge exchange losses remained important throughout the model time interval. The present model produces agreement with the AMPTE/CCE estimates of ring current kinetic energy content versus time. Disagreement between the Dst* inferred from the AMPTE/CCE particle measurements and the observed Dst* is an interesting issue needing further explanation. ¿ 1998 American Geophysical Union