Combinatorial remarks on a classical theorem of Deligne