Combustion toxicology of fires is one of the most difficult subjects currently facing researchers, manufacturers, fire brigades, architects and regulatory/control bodies, and is a subject where there is a great deal of misunderstanding. It is important because many of the fatalities in fire are now attributed to the effects of the atmosphere (rather than burns and other injuries) and attention has been focussed on the problems by a number of recent and very serious fires throughout the world. Combustion toxicology involves the whole spectrum of physiological interactions of fire gases and smoke, from sublethal effects of low concentrations to very high concentrations where unconsciousness and death may result within minutes. It is complex because smoke and gases may contain many hundreds of different chemical species even from the burning of a simple polymer. Because of the great complexity, the understanding of the problem needs to be advanced on several fronts. The paper will outline briefly the approaches being adopted internationally, including pathological studies, biological work, analytical observations and model/fullscale fire experiments to probe the mechanisms of incapacitation, injury and death. Some of these topics (pathological and biological) will be presented by later speakers. The paper will therefore concentrate on the information being gained from laboratory decomposition and fire experiments (particularly with analytical monitoring) to understand the nature of the fire atmosphere. As part of this, the paper will describe a series of fullscale fire experiments (involving both natural and synthetic polymers) with measurements of oxides of carbon, oxygen, smoke, and ‘organic’ products using GC-MS ‘finger-printing’ techniques. Where possible, the results will be presented in a form suitable for estimating the potential contamination of areas remote to the fire from volume flow data of smoke and gases. This work provides not only an insight into the chemical nature of the fire atmosphere but is an essential part of the understanding of the overall hazards presented by fires and can show the wider ways in which the problems of combustion products can be solved. This is important because there is a great deal of interest internationally in the development of simple tests for combustion toxicity (using biological procedures) and methods proposed in other countries will be mentioned briefly. The possible role of these tests will be discussed within a framework of recent research which is attempting to understand the relationship of the tests with practical fire conditions.