Between the 1783 February 5 and 1783 March 28, five earthquakes struck the southern part of Calabria. The main shock (February 5) and the first aftershock (February 6) devastated the region ENE of the Messina Strait. The greatest damage occurred along the foot of the Aspromonte Mountains south of San Giorgio Morgeto, and along the Tyrrhenian coast south of Palmi. A surface break about 18km long, with several feet of downthrow to the west, formed along the Cittanova (Santa Cristina) Fault as a result of the main shock. On February 7, a third large shock ruined villages at the foot of the Serre Mountains north of San Giorgio Morgeto. Morphological and structural evidence, combined with a reassessment of observations made at the time of the earthquakes, suggest that these three shocks were shallow (<=20km) and related to slip on the west-dipping, NE-striking Cittanova-Sant'Eufemia, Palmi-Scilla and Serre normal faults, respectively, which juxtapose the basement of the Aspromonte and Serre mountains with the Pleistocene deposits of the Gioia Tauro and Mesima basins, and border the Palmi coastal high. The three faults belong to an active rift that stretches from northern Calabria to offshore the Ionian coast of Sicily. The spatial coupling between the 1783 events is investigated by resolving changes of Coulomb failure stress. The main shock (1783 February 5, M ~7), on the Cittanova and Sant'Eufemia faults, increased that stress by several bars on the Scilla Fault, triggering the 1783 February 6 earthquake (M ~6.5). The cumulative effect of these two shocks was to raise the Coulomb stress by more than 1bar on the SW part of the Serre Fault, which was subsequently the site of the 1783 February 7 shock (M ~6.5). In turn, the first three events increased the stress by about 1 bar on the NE part of this latter fault, leading to the 1783 March 1 shock (M ~5.7). The gap between the 1783 February 7 and 1783 March 1 events may be related to the previous occurrence of an earthquake 124yr before (1659 November 5, M ~6), which had already released stress locally. The occurrence of the last 1783 event (28 March) is not as simply accounted for by Coulomb modelling, in part because it remains unclear which fault slipped and how deep this event was. Overall, the 1783 sequence increased the Coulomb failure stress by several bars south of the Messina Strait and north of the epicentral region of the 1693 SE Sicily (Catania-Noto) earthquakes. 125yr later, this same region was the site of the 1908 Messina earthquake, also a normal faulting event. Our study thus provides one convincing example in which Coulomb stress modelling brings insight into the spatial dynamics of seismic sequences.