The corona ignition voltage of an electrical discharge in air of atmospheric pressure depends on the presence of (moisture) particles, which may increase the corona losses. A relation between the corona ignition voltage and the particle size when tested shows unexpected results. With the corona ignition voltage in air as observed by Rose and Wood our calculations do not give particle sizes of 50 μm (as used by Rose and Wood), but sizes of about 1 A corresponding to the diameters of the molecules of the component gases in air. Our conclusion is that these molecules align in a conductive channel from the axial wires to the particle considered. In this way the charge transfer from the axial wire to the particles may be explained.