Activations and disruptions of dark Ha filaments are very common phenomena on the Sun. They precede the most powerful two-ribbon solar flares, but they also appear far from any active region without any chromospheric flaring. Therefore, until very recently, filament disruptions were considered as interesting, but physically insignificant, flare precursors. Only Skylab observations have shown that the filament disruptions actually represent one of the basic and most important mechanisms of solar activity. These observations have revealed (1) that many coronal transients originate in eruptive filaments without chromospheric flares, (2) that Bruzek’s slow-mode waves originate in disrupted filaments and not in flares themselves, and (3) that many coronal X-ray enhancements outside active regions are also tops of newly formed loops, similar to the post-flare loops observed after filament disruptions in active regions. An interpretation of these data stems from Kopp & Pneuman’s theory of postflare loops: the process that disrupts a filament opens the magnetic field and causes a greatly enhanced mass-flow along the field lines. The open field lines subsequently reconnect, starting from the bottom of the corona and proceeding upwards. This process can last for many hours. Hot loops are first seen in X-rays, later in extreme ultraviolet (e.u.v.) lines, and, after an appropriate cooling time, in Hx as the loop prominence systems. The visibility of loops depends on plasma density. Several observed properties of solar flares indicate that the primary acceleration occurs as the field lines reconnect. Thus the process of particle acceleration in two ribbon flares can last for hours. Because reconnection is accomplished after essentially all filament disruptions, ‘disparitions brusques’ outside active regions should also accelerate particles.