It is one of the most important but rather paradoxical problems of the world-meteorology that the seasonal correlation changes sign suddenly and the reversed correlation becomes to exist in the new epoch as good as it did in the preceeding period. The present investigation will certainly throw some lights on this point, though it is not yet quite conclusive. The author examined the seasonal correlation for 40 years (1885-1924) between the summer temperature (June, July and August) in Tohoku district (Average temperature for Akita, Miyako, Yamagata, Ishinomaki and Niigata is taken) and the winter temperature at Nemuro. He found that the winter temperature to be correlated with must be that of January for a certain group of years and that of February in the other group. Moreover he found that the sense of correlation must be reversed in some years. All these, according to his idea, are ruled by the activity of the sunspot. The laws proposed by the author are as follows:- (I) According to that the annual Wolf-Wolfer's number for the previous year was above or below 5, January or February temperature must be taken as to be correlated with. When, however, the number was more than 50 and in the decreasing phase, then the January temperature must be referred to. (II) Generally the correlation is positive. It is reversed when (1) the sunspot number of the previous year was less than 10 and (a) increasing very slowly within the limit of 2 per year, or (b) decreasing very rapidly over the limit of 9 per year, or (2) when the number of the previous year was in a domain between 26 and 50, and in the decreasing phase. These laws are only applicable to the case above mentioned. The coefficient for the positive correlation was 0.79±0.04 in the case concerned, the number of years being 31 and the negative correlation was -0.71±0.11, the number of years being, 9.
The author compared the record of the rice crops in Okayama Prefecture for 37 years beginning in 1887 ending in 1923 with that of the sunspots activity in the same period expressed by the annual sum of the total sunspots area. The relation between them was not remarkable but the author pointed out the existence of the tendency that the crops decreases in the first and fourth year after the appearence of the sunspot minimum and increases gradually in the second and third year. The amount of decrease was 20% of the average value in the worst case.
By the snow fall of 30th January 1925, Tokyo and its vicinity were mantled in silverly white. On the next morning the sky was characteristically blue and air was almost still. We enjoyed bright sunshine. On a wooden fence in the compound of the Central Meteorological Observatory the author found a nicely developed snow garland as shown in Pl. 1. fig. 1. In his excellent book “Rain, ” Prof. T. Okada has made some remarks on this phenomenon and given the dimensions of one observed by Kassner in Berlin1). The present one was far larger than German one, which excited much the author's interest. He then found the second one which is shown in fig. 2. There were remains of many fallen garlands, some of them exceeding 6 feet in length. In one place the garlands hung outside, and in other place inside the fences, irrespective of the direction of fences-some run N-S and other E-W. It depends on the inclination of the upper surface of the fences. When it was inclined inwards accumulated snow slipped inside, and hang there forming snow garland, and vice versa. The slip took place when the snow began to melt from bottom. The dimenisions measured are as follows:- Maximum dip means the distance between the lowest point of the garland and the upper surface of the fence. The meteorological elements observed at the Observatory was as follows:- Densities of snow measured at 10h 25m of the 30th and 31st were respectively 0.058 and 0.0842). It is noteworthy that a torsional tendency was perceived on the snow garlands on the present occasion. (Feb. 3rd. 1925. Central Meteorological Observatory)