A note on finite sets of terms closed under subterms and unification