The lack of a suitable system for hot gas cleaning is the greatest obstacle to the development of gasification combined-cycle power generation. In this work, a shallow 0.15 m diameter fluidised bed of 1 mm sand particles was used as a filter for 0.5-10 mum gasifier char particles redispersed in nitrogen at 700-900°C and 1 bar a. Earlier work has suggested that, for a correctly-designed low free-area distributor, initial collection efficiencies greater than 99% at 2 could be realised but that poor retention limits the overall filtration efficiency. As an aid to retention, heavy fuel oil was sprayed continuously into the bed using a concentric-tubs atomiser. With this arrangement, theoretically-predicted filtration efficiencies were approached for particles below; 7 mum in size, but the overall performance was seriously marred by secondary aerosol originating from the addition of the oil itself. The results suggested that two modes of filter operation are possible; at lower gas and retention-aid flowrates the challenging aerosol is retained on the collector; at higher gas and retention-aid flov/rates the challenging aerosol is agglomerated and re-emitted from the bed in a distribution with a larger mean size. The operation of various instruments for particle size measurement below 10 mum has been reviewed, and their capabilities have been compared by experiment. The method selected (collection in a liquid impinger, followed by off-line analysis by "Coulter Counter") is described in detail. The addition of a liquid retention-aid to a fluidised bed can cause modification of its fluidisation behaviour, leading eventually to catastrophic defluidisation. Theoretical and experimental aspect of these effects are discussed, leading to the development of an experimental method for direct measurement of interparticle forces, and an expression for the transition condition between Geldart's (1973) groups A and B.