Evidence for gradual external reconnection before explosive eruption of a solar filament

We observe a slowly evolving quiet-region solar eruption of 1999 April 18, using extreme-ultraviolet (EUV) images from the EUV Imaging Telescope (EIT) on the Solar and Heliospheric Observatory (SOHO) and soft X-ray images from the Soft X-ray Telescope (SXT) on Yohkoh. Using difference images, in which an early image is subtracted from later images, we examine dimmings and brightenings in the region for evidence of the eruption mechanism. A filament rose slowly at about 1 km s(-1) for 6 hours before being rapidly ejected at about 16 km s(-1), leaving flare brightenings and postflare loops in its wake. Magnetograms from the Michelson Doppler Imager (MDI) on SOHO show that the eruption occurred in a large quadrupolar magnetic region with the filament located on the neutral line of the quadrupole's central inner lobe between the inner two of the four polarity domains. In step with the slow rise, subtle EIT dimmings commence and gradually increase over the two polarity domains on one side of the filament, i.e., in some of the loops of one of the two sidelobes of the quadrupole. Concurrently, soft X-ray brightenings gradually increase in both sidelobes. Both of these effects suggest heating in the sidelobe magnetic arcades, which gradually increase over several hours before the fast eruption. Also, during the slow preeruption phase, SXT dimmings gradually increase in the feet and legs of the central lobe, indicating expansion of the central-lobe magnetic arcade enveloping the filament. During the rapid ejection, these dimmings rapidly grow in darkness and in area, especially in the ends of the sigmoid field that erupts with the filament, and flare brightenings begin underneath the fast-moving but still low-altitude filament. We consider two models for explaining the eruption: breakout,'' which says that reconnection occurs high above the filament prior to eruption, and tether cutting,'' which says that the eruption is unleashed by reconnection beneath the filament. The pre-eruption evolution is consistent with gradual breakout that led to ( and perhaps caused) the fast eruption. Tether-cutting reconnection below the filament begins early in the rapid ejection, but our data are not complete enough to determine whether this reconnection began early enough to be the cause of the fast-phase onset. Thus, our observations are consistent with gradual breakout reconnection causing the long slow rise of the filament, but allow the cause of the sudden onset of the explosive fast phase to be either a jump in the breakout reconnection rate or the onset of runaway tether-cutting reconnection, or both.