The Empirical Implications of a Utility Tree: A Further Comment