Some remarks on uniqueness in the calculus of variations