(1) The equation of motion of the needle of Benndorf's self-recording electrometer was solved when the potential of the needle was kept constant, the potential of the “Referenzpunkt” kept constant and the latter varied periodically. (2) Constants ω and κ were determined from the observation of the motion of “Zeiger.” The relation of κ and the density of sulphuric acid, derived from the increase of its volume, is illustrated in Fig. 1. The value of κ also varies according to the position of the mica-vane in the liquid. It is noticeable that κ becomes larger when the upper rim of the mica-vane approaches the liquid surface. (3) A method of determining the leaking-rate of the whole electric system under the working state of the collector was proposed. Which involves the observation of the motion of “Zeiger” when the potential of the needle increases from zero by the action of the collector. (4) The position of the “Zeiger” after one or two minutes in the damped oscillation which was initially caused by the kick of the lever-work was calculated with the observed κ and ω. (5) The decrease of amplitude and the phase retardation of the oscillation of the needle, when the potential of the “Referenzpunkt” varied periodically, was calculated with observed κ, ω and α+β for some given periods.
This problem has been attacked by many authors, such as J. Hann, Max Margules, W. Trabert, Paul Jaerish, H. Takaya, H. Arakawa and S. Idubuti. The author of this paper generalized Takaya's result and obtained the following formula; where p=po(1+ε), T=To(1+γ), γ=Ae-sz sin(pt+mx), φ: latitude, Lo=4×109, Θ=8.64×104, To: mean temperature, and s>10-7, α is very small, and we can put _??_1. The author calculated the variation of pressure when the change of temperature of one degree occured on the earth's surface, giving some adequate values of s and the heights of neutral plane of pressure and wind velocity.
As the measure of mental laxness the ratio of the number of articles left in the city-ears to that of the total passengers was calculated for each day. From January to December 1933, the daily passengers amounted to about one million and a half, white the daily number of articles left was some two hundred in Tokyo. The correlations of the ratio with the meteorological elements were studied and the relative humidity was found to be the most effective. The coefficient of correlation was computed: in Tokyo r=0.64 and w=0.027 (January-December, 1933) and in Yokohama, r=0.42 and w=0.030 (April 1932-March 1933) where In addition, it was ascertained that the sudden increase of humidity and the decrease of the atmospheric pressure generally precede the increase of the above-mentioned ratio. The data from the city of Osaka were also brought under the discussion.
For the purpose of p edicting mean air temperature in January at Chee-foo, China, the author investigated the correlation between the conditions of the continental high in October and the above mentioned element. As the arguments to express the condition of the high, two quantities were taken: the one is the mean air pressure at a certain station, and the other the position (latitude) of the centre of the high area in the isobaric chart drawn from mean air pressure in October at many observatories and stations in the Far-east. Tientsin, Chinan, and Hankow were examined taking them as the pressurestation, and it was found that Tientsin was the best. The regression equation was given as: Δt=-1.20Δp+0.348l+0.925 where Δt is the temperature deviation from normal in January at Chee-foo, Δp the pressure deviation from normal in October at Tientsin, and l the latitude in degrees counted from the parallel of 40° N and taking positive towards north. To my great disappointment, this result has three exceptional years in the whole period under the present inquiry, 1906-1932. They are 1906, 1913, and 1931 in which the sunspot-numbers in October, are all less than 10. Although such result must admittedly be valued less, in the remaining years other than the three, the correlation is so striking that it makes the author believe that the above equation may be used, with caution, for the year except those of sunspot minimum.
In this paper the author studied about a method of forecasting of the winter temperature by means of pressure gradients in November, in each directions about Gihu. The results are Y=0.24-0.12X, where Y is winter temperature and X denoted by the following equations: X=F•A+B•D when G>0, X=F•A+L•K when G<0 In the above equations A, B, D, K, L and G are numerical values calculated from the pressure gradients by suitable formulae under the consideration of following conditions: a) It is very cold in winter when the continental high pressure has developed in November. b) or when the high pressure has developed at central China. c) It is very cold in winter when the pressure has fallen in northeastern parts to Gihu districts in November. d) The winter takes moderate coldness when the pressure has fallen in southern part to Gihu districts in November. The correlation coefficient between X and Y was calculated to be -0.75.
As the result of examination of Margules' theory on the diurnal barometric oscillations using the observed values of air temperature and atmospheric pressure in the upper air at Lindenberg, we found out that the constancy of the equilibrium temperature for different altitudes assumed by him is not correct.
We analysed fifty-three microbarometric waves (period 5 min.-40 min.) observed at Tokyo in 1933 by the statoscopes and the Yosida's microbarograph at five different places. Most waves are found to be progressive like three examples reported by Messrs. T. Kita and H. Huzita last year. In addition, the waves seem to be topographically conditioned, since the direction of their propagation is almost to N, NE or to E., and some waves have local natura. Moreover, their propagating velocities (ca 0.2km/min to ca 4.0km/min) diminish as they advance. Daily changes of the wave occurrence are shown in § 5.