Continuous recordings of the instantaneous frequency of the strongest received signal component of WWV-20 (4000 km distant) were made for a period of 30 days in the spring of 1960 at Palo Alto, California. On normal days regular morning and evening frequency shifts, amounting to roughly 1 part in 107, were observed. The received frequency was higher than normal in the morning and lower than normal at night. On magnetically disturbed days, the maximum observed frequency shifts probably amounted to 3 parts in 107, a figure frequently cited by others [Essen, 1935; Booth and Gregory, 1948; Jackson, 1952; Lincoln, 1960]. On days when only the one-hop mode could propagate over the path, the observed frequency shifts were considerably smaller than when the two-hop mode was present. This circumstance may be of practical interest. On many days short-enduring upward frequency shifts amounting to as much as 3 cps, 1.5 parts in 107, were observed. The upward portion of these shifts typically took place in 1 to several minutes; they were followed by small downward shifts lasting several times as long. It is tentatively suggested that these shifts are caused by entry into the ionosphere of the downward-moving clouds of ionization first identified on ionograms by Wells, Watts, and George [1946] and subsequently studied by Bibl [1952]. During magnetically disturbed days, occasional periods were found during which changes in the instantaneous frequency of WWV-20 seemed to correlate reasonably well with changes in the earth's total magnetic field as measured at Palo Alto.