The temporal signature of thermal emission from a volcano is a valuable clue to the processes taking place both at and beneath the surface. The Galileo Near Infrared Mapping Spectrometer (NIMS) observed the volcano Prometheus, on the jovian moon Io, on multiple occasions between 1996 and 2002. The 5 micron (μm) brightness of this volcano shows considerable variation from orbit to orbit. Prometheus exhibits increases in thermal emission that indicate episodic (though non-periodic) effusive activity in a manner akin to the current Pu'u 'O'o-Kupaianaha (afterwards referred to as the Pu'u 'O'o) eruption of Kilauea, Hawai'i. The volume of material erupted during one Prometheus eruption episode (defined as the interval from minimum thermal emission to peak and back to minimum) from 6 November 1996 to 7 May 1997 is estimated to be 0.8 km3, with a peak instantaneous volumetric flux (effusion rate) of 140 m3 s−1, and an averaged volumetric flux (eruption rate) of 49 m3 s−1. These quantities are used to model subsurface structure, magma storage and magma supply mechanisms, and likely magma chamber depth. Prometheus appears to be supplied by magma from a relatively shallow magma chamber, with a roof at a minimum depth of 2–3 km and a maximum depth of 14 km. This is a much shallower depth range than sources of supply proposed for explosive, possibly ultramafic, eruptions at Pillan and Tvashtar. As Prometheus-type effusive activity is widespread on Io, shallow magma chambers containing magma of basaltic or near-basaltic composition and density may be common. This analysis strengthens the analogy between Prometheus and Pu'u 'O'o, at least in terms of eruption style. Even though the style of eruption appears to be similar (effusive emplacement of thin, insulated, compound pahoehoe flows) the scale of activity at Prometheus greatly exceeds current activity at Pu'u 'O'o in terms of volume erupted, area covered, and magma flux. Whereas the estimated magma chamber at Prometheus dwarfs the Pu'u 'O'o magma chamber, it fits within expectations if the Pu'u 'O'o chamber were scaled for the greater volumetric flux and lower gravity of Io. Recent volumetric eruption rates derived from Galileo data for Prometheus were considerably smaller than the rate that produced the extensive flows formed in the 17 years between the Voyager and Galileo missions. These smaller eruption rates, coupled with the fact that flows are not expanding laterally, may mean that the immediate heat source that generates the Prometheus plume is simultaneously running out of available volatiles and the thermal energy that drives mobilization of volatiles. This raises the question of whether the current Prometheus eruption is in its last throes.