The discussion of the modern observations of the Sun, Mercury and Venus by Spencer Jones, when taken in conjunction with Jeffreys's theory of tidal friction, is shown to lead to a total dissipation of energy in the oceanic tides which is three times that calculated by Jeffreys in his discussion of tidal data for shallow seas. The true (negative) acceleration of the Moon corresponding to this value of the dissipation has been included in the expression for the Moon's tabular mean longitude in the Improved Lunar Ephemeris . An investigation of the value of this acceleration in ancient times using Hipparchus' eclipse and equinox observations leads to a value twice as large as that given by the modern observations. This result is very satisfactorily confirmed by the observed magnitudes of partial lunar eclipses recorded in the Almagest . It is pointed out that if the acceleration does change, then ephemeris time derived from lunar observations will not be uniform in the Newtonian sense.