Excitation of microseisms is generally considered to be due to pressure change at ocean bottom, for which Longuet-Higgins derived his celebrated formula in 1950. Use of this formula is an approximation, however. Comparison with a more rigorous normal-mode formula shows that this conventional approach is acceptable for ocean depths less than 1 km but fails in deep oceans. On the other hand, there seems to be a multitude of evidence that source region for double-frequency microseim is near the coast and thus is generally in shallow water. An evidence from buoy data for nonlinearity in ocean waves is presented to support this view. If a source region is in shallow water, use of the Longuet-Higgins pressure formula at ocean bottom for the excitation of microseisms is justified, although one should pay attention to ocean depths very carefully.