Is there any fast sound in water?

We measured the dynamic structure factor S(Q,omega) of liquid and undercooled water down to 253 K in the Q approximately 0.02-0.1 nm;{-1} momentum transfer region. We observe the neat departure of the apparent speed of sound from the adiabatic regime as a function of decreasing temperature. Our evaluation of the infinite-frequency limit of sound velocity, c_{infinity}, matches with the results obtained in the high momentum transfer limit by inelastic neutron and x-ray scattering. These results strongly support the viscoelastic interpretation of the dynamics of water. Hence, we propose to call c_{infinity} the high-frequency speed of sound and to abandon the term fast sound, which recalls a propagation mechanism through lighter atoms, like in gas mixtures.