The authors point out errors in an earlier paper by Di Gesu` and Sacco [1] in which a statistic to test for degree of departure from a uniform distribution of points in a finite region of a multidimensional space was proposed. The primary flaw in [1] is the assumption that lengths of edges in a minimum spanning tree will be independent under randomness; this assumption is systematically violated and thus the proposed statistic is invalid. The Di Gesu` and Sacco paper also contains other errors and misleading statements. Researchers interested in tests for uniformity vs. clustering would do well to consult recent statistical work in the classification literature, such as that of Hartigan [2].