The paper addresses the establishment of a steady inventory of bed inert material in a fluidized bed combustor when the bed consists of fuel ash only. The analysis is based on the concept of the primary ash particle size distribution (PAPSD), defined as the PSD of primary ash particles liberated by fuel particles upon complete burn-off. An experimental procedure for the quantitative assessment of the rate of primary ash generation and of the size distribution of primary ash particles (PAPSID) is presented. The application of the procedure is demonstrated with reference to a set of solid fuels, including coals and biogenous fuels. The reliability and accuracy of the technique was satisfactorily tested, with the exception of two biomass fuels which gave rise to bed agglomeration making the application of the test procedure unsuccessful. A simplified theoretical analysis is presented, based on material balances on lumped classes of primary ash particles around the CFBC. The analysis aims at clarifying the role of the PAPSD and of the cyclone efficiency to the establishment of bed inventory, particle size distribution and solids circulation rate. Lower and upper bounds to the solids circulation rate that can be established in the steady operation of a CFBC are related to the PAPSD and to the cyclone performance.