I think that the clone properties to look for in proportional representation methods are:
Increasing the clones in a clone set should not decrease the number of winners from the clone set and should not make a looser that is outside the clone set into a winner. and Decreasing the clones in a clone set should not increase the number of winners from the clone set and should not make a winner that is outside the clone set into a looser. I don't know if the STV method I proposed or other STV methods that derive from clone-proof single winner methods satisfies these or not. ---- Election-Methods mailing list - see http://electorama.com/em for list info
