Been thinking about Jeff W's suggestion to add a unique index variable to the tree record. Everything seems to point to this being a good idea, but then I got to thinking about models that do pseudo-stochastic tripling (or other record duplicating methods). Now the correspondance between the index value and what that tree represents is ambiguous. We could still have an index variable and add new unique values for each "created" record, but does this violate the intent of the index in the first place?
Food for thought.