A submarine eruption in Cumberland Bay, Robinson Crusoe Island, was reported by Thomas Sutcliffe, the former British Governor, shortly after the earthquake that struck the coast of Chile on 20 February 1835. This episode was described by Charles Darwin in his Voyage of the Beagle and extensive mention has been made since then, especially stimulated by a renowned painting by J.M. Rugendas. Because of the apparent causal relation, this event has also been widely cited as an example of remote tectonically triggered eruption. However, there are inconsistencies that pose doubts about the actual occurrence of an eruption. Here we present evidence against the hypothetical eruption based on both the absence of any geological evidence and a reinterpretation of the historical accounts. We first observe that no bathymetric anomaly is present immediately below the place of the depicted ‘eruptive column’. We also note the absence of any deposit or recent volcano morphology and then unravel some incompatibility between the expected volcanological parameters and the featured column. In addition, we analyse the historical records and conclude that they are compatible with a tsunami entering the bay. By means of numerical simulations we further demonstrate that the accounts well match with the expected behaviour of a distant earthquake-triggered tsunami. We infer that some tsunami-related processes (sound waves, rockfalls, lightning) may have been misunderstood at that time. The latter corresponds to the current knowledge of natural processes but also could have been deliberatively amplified in Sutcliffe’s report. Our multidisciplinary approach provides full consistent geographical evidence of a fact that did not happen. This finding is relevant from the hazard’s perspective, but also for the science of earthquakes and eruptions, or the knowledge of processes that control the late secondary volcanism at oceanic islands and seamounts.