The measurement of dust in chimny stacks and of airborne particles is an urgent matter with respect to the problem of air pollution. The cascade impactor, originally developed by K. R. May, is a very effective device for measuring the amount of aerosol particles. Until now, many investigators have drawn their attention mainly to the collection efficiency of dust particles. However, the phenomena arising from the recoil action of particles at the target plate remains unclear and, in connection with the above recoil motion, particles stick to both the surrounding walls and the lower part of the nozzle block. In some cases, this causes serious errors in the measurement of collection efficiency, especially when dust has a wide range of particle sizes. In this report, we are interested in clarifying these problems and in collecting useful information which might lead to its resolution.