Although the following proof is implicit in Garsia and Milne's paper [I], it is scj elegant that we felt that it should be presented by itself for the benefit of the general mathematical public. The idea behind the proof was further exploited by Remmel [2] and Wilf [3]. Consider a set A of elements each of which possess a (possibly emptyj subset of the properties (1,. .. , n}. The inclusion-exc...