When a fine-powder gas-solid fluidised bed is employed as a catalytic chemical reactor, mass transfer from the bubble to the dense phase could be limiting. This transfer process is often rapid in the grid region and, due to bubble growth, very much slower in the upper reaches of the bed. A device is put forward to overcome the disadvantages of this natural gas-solids contact distribution by horizontally splitting the bed into two zones and allowing half of the total gas flow to flow through each. The insert, referred to as a bubble collector, has the basic form of an inverted funnel and is shown to operate in a stable and robust manner in a two-dimensional bed. The incentive for developing this kind of insert is examined on the basis of published bubble size/velocity data and reacting tracer results. It is concluded that throughput increases (at constant conversion) of around 70% are possible under certain conditions or, conversely, that increases in conversion (at constant throughput) of around 15 - 17 percentage points are achievable. The incentive for considering the installation of such a device appears to be strongest when the reaction itself is fast. The incentive is considerably weaker for systems which operate with a limiting maximum stable bubble size, though for this kind of process throughput is unlikely to be a problem in the first place.