The author found that at the great part of the meteorological stations in Japan Proper the seculer variation of the monthly mean air temperature of February and that of December correlated as follows;
i) the positive correlation for 1905_??_1918 (or 1919),
ii) the negative correlation for 1918 (or 1919)_??_1933 (or 1934),
that is, the correlation turned into negative from positive in the year 1918 or 1919.
The distribution of this turning year resembles the winter path of “Kuro-sio” along the Proper coast and the year 1917 is the most maximum year of Wolf's number of solar spot in a good long period.
Thus the turn of this correlation is likely to have a relation with Kuro-sio along the Proper coast. Indeed it has been reported by Mr. Uda that the abnormality of Kuro-sio at the southern Offing of “Kisyü” had caused for 1905_??_1906, 1916_??_1918, and 1934_??_.
Since about the year 1934 each seculer variation is at random one another, but the positive correlation will begin about that maximum year of solar spot number.