Early historic documents can provide valuable records on eruptive events. The information, however, is often ambiguous, vague and sometimes erroneous. Early records for São Miguel, Azores, refer to eruptions at Sete Cidades at the west end of the island in the 15th century. Geological studies have so far been unable to substantiate these events. These early observations, probably from passing ships, are vague accounts and often mar have been written down many decades after the event. It is also true that during this early phase of colonisation maritime countries were secretive about their discoveries. In a study of the eruptive history of Furnas Volcano, the Azores, we made a detailed analysis of historic records of the 1630 AD eruption (Queiroz et al. 1995; Cole et al. 1995). From this study we were able to establish the extent and time period of the seismic activity prior to the onset of the eruption and from the contemporary accounts descriptions of probable pyroclastic surges. This information proved critical in our assessment of the hazard posed by Furnas and the recognition of likely precursory phenomena for a future eruption. Careful study of the records also revealed descriptions of an earlier eruption of Furnas not previously recognised in accounts of historic activity. The first Portuguese to settle on São Miguel are thought to have established themselves c.1444 AD in Povoação, the next valley to the east of Furnas. Their observations clearly refer to eruptive phenomena in the Furnas Caldera. Integration of these accounts with field data suggests that this was the dome building phase of the Furnas I eruption which followed shortly after the Furnas H event. Radiocarbon dates place the Furnas H eruption in the early 15th century. It may be that these are the eruptions on São Miguel that were first recorded by mariners in the 15th century.