The spontaneously explosive decomposition of gaseous chlorine dioxide has been investigated over the temperature range 42–86°C. The emphasis is on direct measurements of the temperature-time histories of the reaction using a very fine thermocouple probe with a fast response. Pressure-time histories and induction periods preceding explosion have also been measured. The spontaneous ignition of chlorine dioxide is characterized by long induction periods. The pressures of chlorine dioxide used range 1–35 Torr, but no critical pressure is observed; explosion inevitably occurs if one waits long enough, although the wait may be several hours. All the temperature-time histories show negligible self-heating prior to explosion, suggesting directly that the spontaneous explosion of chlorine dioxide is branched-chain rather than thermal in character. This conclusion is supported indirectly by the pressure-time histories and by the induction periods observed for the explosion of mixtures of chlorine dioxide with helium and hydrogen as diluents. Hydrogen acts essentially as an inert diluent in the pre-explosive decomposition of chlorine dioxide, although it reacts during the explosion. Pressure-time histories and induction periods preceding explosion exhibit two distinct types of behavior. When the ambient temperature is above about 50°C, there is negligible pressure change prior to explosion and induction periods decrease monotonically as the pressure of chlorine dioxide is increased. When the temperature is below 50°C, the pressure of the reacting gas rises practically continuously throughout the course of the pre-explosive reaction and only a small fraction of the initial chlorine dioxide remains at the onset of the explosion; moreover, the induction periods preceding explosion increase with increasing initial chlorine dioxide pressure when the latter exceeds about 2 Torr. Possible mechanisms for the decomposition of chlorine dioxide are discussed in the light or these now findings.