Brendan said: > I don't see how that works. Suppose you have a complete graph with > a perfect matching of edges with multiplicity 2 You are right.. it doesn't work; now I know why I settled on using the subdivision technique !! Gordon