Does thermal explosion really get delayed by natural convection?

A thermal explosion is said to occur in a region if heat generated steadily therein cannot be conducted to its boundaries. Our interest is in the effect of natural convection on the conditions for thermal explosion. It is surprising, first, that adding a second mechanism for heat loss lowers the steady rate of heat generation and, second, that natural convection, therefore appears to do little to change the point of explosion. This is due to the fact that the flow, by lowering the heat generation, is self-weakening.