During the 1989 eruption of Mt. Etna, one of the most important in the last 20 years in terms of effusion rate, a pair of fractures formed on the slopes of the Southeast Crater cone and propagated in a few days trending ca. northeast and' southeast, with large quantities of magma being drained from the northern fracture. In contrast the southern fracture, after a short initial eruptive phase, grew by 5 km in five days without eruptive or seismic activity. When the fracture system crossed the southern scarp of the Valle del Bove wall, some 6 km SE of the active crater, an intense seismic swarm started in the same area. This lasted about four days and consisted of several hundred events clustering in a small focal volume. Although such a sequence of failure events contains elements of compatibility with a shallow dyke intrusion, it did not lead to the expected flank eruption which would have threatened settlements just a few kilometres from the distal end of the fracture. The question arises, therefore, whether or not the fracture system was dyke-related, and if so to what extent. We attempt here to answer such a question by integration and interpretation of the available field observation and geophysical data. Indeed, field observation constraints on the southern fracture system and the typical depth (> 2 km) of the swarm do not contain evidence for shallow dyke emplacement beneath the lowermost part of the SSE fracture. Conversely, the inferred direction and shape of the tremor source, the location and extent of a gravity anomaly observed a few months before the eruption, the near-horizontal migration of the seismic focii during the swarm, and the change with time of related focal mechanisms, lead us to hypothesize that the extreme development of the fracture system at its south was due to a deeper intrusion beneath the southern slope of the volcano.
