Abstract
The record of eighteenth and nineteenth century explorers' references to Aboriginal fire in Queensland was stratified according to fourteen vegetation typcs and season of fire. It was demonstrated that references to ‘current’ fire (i.e. flames or smoke) may not represent traditional Aboriginal activity and that many fires were lit to frighten or harm, to protect themselves from, or to signal to kinfolk the presence of the European intruders. Because of this interpretational difficulty the records to ‘current’ fire were treated separately from ‘past’ fire (i.e. burnt ground). The data were analysed as the number of observations per 100 km spent in each vegetation type for any one season to compensate for bias created by differing amounts of travel. The record suggests highest frequency of burning in grassland around the Gulf of Carpentaria, relatively high fire frequency of most coastal and subcoastal vegetation types and relatively infrequent burning of inland Queensland. The analysis indicates a propensity for winter and autumn fue relative to spring and summer fire in all vegetation types combined and in most individual vegetation types.