The feasibility of using the nucleation properties of combustion- generated particulates for the identification of their sources was investigated using the cloud condensation nuclei (ccn) spectrometer. The investigated materials included coal, wood, belt materials, electrical insulation, brattice cloth, and diesel exhaust. The materials were heated in an oven using filtered airflow, and their time, temperature, and the critical supersaturation spectrum of condensation nuclei (also referred to as the ccn spectrum) produced by the combusting material were recorded. Some data were obtained on a typical metal mine background containing blasting products. The data were processed and analyzed in an attempt to specify parameters and limits that would allow the identification of the combustion sources. It appears that under controlled test conditions it is possible to identify the source on the basis of data derived from the ccn spectrum. More general use of the idea for fire detection under uncontrolled practical circumstances in an actual mining environment does have potential but would require further research.