Of the tidal stations in the world, as is present in Fig. 1 and Table II, were picked out 36 stations where they have made annual observations continuously for more than 40 years, in principle. With the data from those chosen tidal ststions the secular variation and long-periodic variation of the sea level have been investigated. As for the method of the investigation, to get the secular variation of the sea level, the changes in the annual mean sea level were applied in a linear equation, and then, to get the periodicity of the long-periodic variation, it was calculated that the serial correlation' coefficient of the annual mean sea level subtracted the linear variation on sea level from the observed values. The annual rate of the rise in the sea level and the serial correlation coefficient for each tidal station are given in Table III and Fig. 3 respectively. (1) At least in the northern hemisphere, the sea level has been in a rising tendency for these scores of years, and the annual quantity is 1 milliliter. (2) No common period of variation for all the stations could not be found, and the pre-eminence of the 19-yearly periodicity by the shifting of the moon's ascending node could not be explained neither. (3) The amplitude of the 19-yearly tide seems to be scarcely different from the theo retical value by the equilibrium law. (4) In the data of the tidal stations where observations have been made for longer periods, the 10-11 yearly periodicity can be found pre-eminent, and it is quite noticeable that this is almost corresponding to the periodicity of the sun spots.