The Round Mountain gold-silver deposit is one of the world's largest volcanic-hosted precious metal deposits; reserves plus production total at least 16 million oz (500,000 kg) Au. The deposit occurs in ash-flow tuff that is ponded within its source caldera, along and overlying the caldera ring fracture. Host rocks have undergone propylitic, potassic, and high-level silicic and intermediate argillic alteration. These are interpreted either as a temporal sequence from early propylitic to late silicic and argillic or as a spatial progression from a potassic core outward to a propylitic halo. 40 Ar/ 39 Ar ages determined by single crystal fusion of sanidine from host tuff and incremental heating of adularia indicate a most likely duration of hydrothermal activity of approximately 0.1 m.y., possibly no more than 0.05 m.y., and a maximum of 0.5 m.y. Eruption of host ash-flow tuff and caldera collapse occurred at 26.5 Ma; a late or postalteration tuff that rests upon mineralized rock is 26.0 Ma. Hydrothermal circulation is unlikely to have been established immediately after caldera collapse, so duration was probably considerably less than 0.5 m.y.Eight adularia samples give ages between 25.94 + or - 0.09 and 26.09 + or - 0.05 Ma, and one is 26.20 + or - 0.05 Ma; the mean of all nine is 26.03 + or - 0.08 Ma. The maximum and minimum values and their uncertainties allows a duration of as much as 0.4 m.y., but for the following reasons, the longevity is probably much less. The absolute ages of seven samples that can be placed in the possible temporal alteration sequence do not correlate with that sequence. This does not disprove the temporal sequence but does indicate that alteration occurred in a time span which is not resolvable even with precise 40 Ar/ 39 Ar dating. Furthermore, all samples show flat age spectra, which themselves argue against a long, continuously active system. Kinetic studies of two adularia reveal diffusion domains with Ar closure temperatures between approximately 230 degrees and approximately 390 degrees C. Adularia that crystallized and remained at temperatures of 275 degrees C, the maximum temperature inferred from fluid inclusion data, would yield age spectra with age gradients spanning the lifetime of the system. No spectra are consistent with a thermal history that incorporates sustained (>0.1 m.y.) temperatures at or above 275 degrees C. Episodic activity to account for the one older adularia date cannot be precluded but is inconsistent with most data. Therefore, the hydrothermal system probably was active for less than the uncertainty in the ages of individual samples, possibly as little as 50,000 years.We conclude that hydrothermal activity at Round Mountain was a brief event at approximately 26.0 Ma and was closely linked to caldera structure and magmatism. This activity followed initial ash-flow eruption by 0.5 m.y. and may have been driven by a ring fracture intrusion.