The two Hyuganada, Japan, earthquakes of 1961 and 1968 (M=7.0 and 7.5, respectively) each broke 80-km segments of the thrust zone abutting to the southern end of the Nankaido 1946 rupture. Mean annual sea levels (period 1942--1973) at eight tide gage stations along a 400-km segment of the Japanese coast centered in the epicentral area were compared to each other. It was found that mean annual sea level differences between the northernmost and southernmost stations (Kochi and Kagoshima) remained constant over a 16-year period, with a standard deviation from the 16-year mean of 1.2 cm. Compared with these northern and southern reference stations, four central tide gages located within 50 km of the two Hyuganada ruptures showed the following changes: (1) At the station between the two earthquakes (Hosojima), sea level dropped by about 5 cm in 1957, and it rose by about 5 cm after the earthquake in 1968. (2) At two stations north of the ruptures (Tosashimizu and Uwajima), sea level rose by about 4 cm in 1957. (3) At a station close to the southern end of the ruptures (Aburatsu), sea level remained constant with a standard deviation of about 1.1 cm between 1950 and 1966. These sea level changes are interpreted to reflect local crustal uplift at Hosojima and subsidence in the Tosashimizu-Uwajima area in 1957. If one assumes that these vertical movements were percursors to the two Hyuganada earthquakes, the data support the dilatancy-diffusion hypothesis: Near the center of the ruptures, dilatancy may have caused uplift (Hosojima), and beyond the northern end of the aftershock area, subsidence may have occurred due to reduction in crustal pore pressure. Precursor times of these magnitude 7.0 and 7.5 earthquakes may have been approximately 3.5 and 5 years, respectively.