The oil industry has inherited a mixed history of success and failure of application of air injection as an enhanced-oil-recovery method. Close scrutiny of these projects shows that in order to conduct a successful in-situ-combustion-oil-recovery project, sustained propagation of the combustion front within the reservoirs is necessary. Sufficient air must be supplied to maintain the propagating combustion front in the desired bond-scission (carbon-oxide-forming) mode, otherwise unfavourable oxygen addition [i.e., low-temperature-oxidation (LTO) reactions] will consume oxygen and immobilize oil. When this happens, the combustion process is deemed to be exhausted. Quantification of the minimum air flux required for sustaining combustion-zone propagation is needed to properly match the capacity of the air-injection facility to the volume of the reservoir that is to be swept by the thermal zone. Undersizing the air-injection capacity causes the in-situ-combustion process to become inefficient at a point when only a small portion of the reservoir has been "burned." One-dimensional combustion tubes (CTs) are conventionally used to obtain important combustion parameters required for designing an air-injection project. Because of the high heat capacity of laboratory equipment designed for elevated-temperature and -pressure operation, oxen addition or LTO reactions are promoted by the heat transfer through the core-holder walls when the laboratory tests are performed at low air-injection rates. Therefore, when operated at elevated pressures, the CTs are unable to operate at the low air fluxes required to establish the minimum possible air-injection flux while maintaining the combustion reactions in an effective mode. To address this issue, a state-of-the-art combustion cell was conceived and used as a way of addressing the previously mentioned constraints associated with high-pressure ID CTs. A conical combustion-cell design was built because it enables continuous air-flux reductions without having to adjust the air-injection rate. The heater control strategy was also modified in order to address the lag-lead operation often used for 1D CTs. To date, the unit has operated at air fluxes down to 3 std m(3)/m(2.)h. The experimental work described in this paper provides insight into the limitations in laboratory investigations of in-situ combustion and the expected behaviour of field applications of the in-situ-combustion process.