The first direct evidence that patterns of basaltic eruptive activity may be very similar on Io and Earth was seen on November 26, 1999, when the solid-state imaging system on board the Galileo spacecraft obtained high-resolution images of a fissure eruption taking place in one of the calderas of the Tvashtar Catena complex. We analyzed the dynamics of the fissure eruption from the morphology of the lava fountains and pyroclastic deposits and derived an estimate of the eruption rate. Our best estimate of the volume eruption rate per unit length of fissure lies in the range 0.7 to 7 m3 s−1 m−1, and multiplying this rate by the fissure length of ∼25 km yields a total volume eruption rate of magma of ∼2 × 104 to 2 × 105 m3 s−1. The upper end of the inferred eruption rate is very close to the upper end of the range of eruption rates observed or inferred for fissure eruptions on Earth. Theoretical models of the possible dike systems feeding the fissure show that for eruptions of positively buoyant magma, the elastic properties of rocks are such that a magma flux of 6 to 7 m3 s−1 m−1 is the largest value that can ever be realized in such eruptions, this limiting value being independent of planetary gravity and thus the same on all silicate bodies. The eruption rate inferred for Tvashtar can also be achieved by a negatively buoyant magma, but in that case the source depth cannot be much greater than a few tens of kilometers. We find that some eruptive events are consistent with the effusion of magmas which are buoyant relative to the crust, whereas others appear to involve magmas which are not. This implies that the distribution of crustal volatiles is laterally very inhomogeneous. Our analyses suggest that the depths from which magmas make their final ascent to the surface of Io are generally of order tens, rather than hundreds, of kilometers.