The rates of oxidation of tiny particles (initial diam. similar to 115 mum) of a nonporous graphite have been measured in beds of silica sand fluidized by different mixtures of O-2 + N-2 at 1 atm. Temperatures of 973-1173 K were used. Before burning, the graphite was fully characterized by B.E.T. analysis. Such a fluidized bed has several advantages for measuring the rate of burning of solid carbon: it proved straightforward to ensure that combustion was kinetically controlled, and that such small particles were both isothermal and at the same temperature as the fluidized bed. Also, the oxidation of CO to CO2 is inhibited by the large area of sand, which removes radicals, such as HO2 and OH. It is concluded that in this situation oxidation occurs in C-s + 1/2 O-2 --> CO (I) whose rate was measured to be half-order in O-2 over the temperature range used and for a variety of concentrations of O-2. Writing the rate of consumption of carbon in reaction (I) as k(1) [O-2](1/2) per unit area of carbon, it was found that k(1) = 1.1 X 10(4) exp (-179 kJ mol(-1)/RT) mol(1/2) s(-1) m(-1/2) correct to 50%. The implications of this are that initially O-2 chemisorbs dissociatively on carbon to yield tightly bound oxygen atoms. These are not desorbed; instead, they appear to be in equilibrium with less tightly bound atoms of oxygen, which desorb to give mainly gaseous O-2 or, less probably, CO. In such a fluidized bed, CO subsequently burns in the gas phase (usually in bubbles) far away from the original carbon particle. The graphite particles used here were initially not spherical; their shape was found to remain fairly constant during burning, as expected for combustion being kinetically controlled. In combustors other than fluidized beds, the radical HO2 or even OH well above 1300 K, might be alternative reactants instead of O-2 in (I). (C) 2001 by The Combustion Institute.