Comment on "Metric Fluctuations in Brane Worlds"

Recently, Ivanov and Volovich (hep-th/9912242) claimed that the perturbation of $h_{\mu\nu}$ with nonvanishing transverse components $h_{5\mu}$ is not localized on the brane because $h_{\mu\nu}$ depends on the fifth coordinate $z$ linearly. Consequently, it may indicate that the effective theory is unstable. However, we point out that such linear dependence on $z$ can be {\it gauged away}. Hence the solution does not belong to the physical one. Therefore, even if one includes $h_{5\mu}$, Randall and Sundrum's argument for the localized gravity on the brane remains correct.