On Edwards' theory of powders