I had once solved this problem at length. The problem also arises when you consider an amount *m* of air in a closed container. Since the air molecules are moving all around and not settled at the bottom of the container how is it that the weight of the container still includes the weight *mg* of air?
For the falling grain of sand if you consider the momentum transfered by the grain when it impacts the bottom of the container it can acount for the weight mg of the grain of sand. You will have to take an average over a long period of time if the collision is not perfectly inelastic. The result is however true (as it should be) whether the collision is elastic, inelastic, or perfectly ineleastic.