Moon echoes at a frequency of about 20 Mc have been studied, mainly to obtain information on low-angle transmission through the ionosphere. Using a broadcast transmitter in its free time, with fixed aerial systems, thirty experiments were carried out in just over a year, echoes being received in twenty-four cases. Experimental results have been compared with those to be expected from an orthodox ray-theory treatment, and the following major discrepancies were found: (1) Observed echo intensities were well below theoretical values. (2) Minimum altitudes, at which echoes were first detected, were unexpectedly high. (3) Ray deviations of several degrees in a vertical plane apparently occurred. The anomalous values of echo intensities and minimum altitudes for detection correlate closely with the critical frequency of the F2region, suggesting that the anomalies arose in that portion of the ionosphere. The cause may lie in departures from horizontal stratification in the F region, or in failure of the ray-theory treatment at very oblique incidence. The received echoes showed two types of fading. One type was consistent with an ionospheric origin. The second has been shown to be due to the moon's libration, effective reflection taking place over a large proportion of the moon's surface. This, together with the elongation of short pulses on reflection, demonstrates that the moon is a "rough" reflector at this frequency.