The bound derived in the above paper(1) is shown via two new stability criteria not to be the tightest possible for ail unstructured perturbations. A minor error of their theorem is also corrected.