Evidence of large-scale episodic venting of hydrothermal fluids was initially discovered in August 1986 in the form of a 130-km 3 radially symmetric ''megaplume'' over the southern Juan de Fuca Ridge. We report here on the discovery in September 1987 of a second, smaller megaplume about 45 km north of the location of the first megaplume. The 3He/heat, 3He/dissolved Mn, and 3He/dissolved silica ratios in both megaplumes were typical of high-temperature vent fluids. Evidence from long-term records of current flow over the southern Juan de Fuca Ridge, and from the mineralogy and Mn chemistry of megaplume particles, makes it unlikely that the second megaplume was a reencounter of the first. A plume model that relates the heat flux to the observed plume rise height of ~1000 m finds that the total heat content of the fluids that formed the megaplumes was 1016--1017 J, or equivalently a fluid volume of 3--8¿107 m3 at 350¿C. The geometry and suspended particle population of the first megaplume imply that such features are formed within a few days time. The extraordinary heat and volume fluxes associated with megaplumes (102--103 greater than ordinary vent fields), as well as their typical hydrothermal chemistry, suggest that they resulted from tectonic or hydraulic fracturing that suddenly increases the permeability of the hydrothermal fluid reservoir in the axial crust. The flux of hydrothermal heat from continuous venting and episodic megaplumes on the southern Juan de Fuca Ridge is presently 4-10¿109W, a factor of 5--10 greater than various geophysical model calculations for this ridge segment. This imbalance may be symptomatic of a recent surge in the local cycle of magmatic activity. ¿ American Geophysical Union 1989