The adsorption of H2O on a Si(111) surface has been studied in situ by electron-energy-loss and photoemission spectroscopies. Two different phases of water are prepared on this surface as determined by photoemission, which have been previously interpreted as molecular and dissociatively adsorbed water. Unlike photoemission spectra, which reveal strong differences between these two states, the electron-energy-loss spectra are very similar. We analyze the differences and find that the existence of molecular water is reasonable and discuss why molecular water may not be detectable in the electron-energy-loss data. © 1986 The American Physical Society.