In this report a statistical t h eory of prediction that can be applied to the forecasting based on limited data or a small sample is outlined. The exact sampling prediction theory is framed on two foundation stones; the one is a two sample theory in which the multivariate regression theory is introduced, and the other is a theory of time series in which a sort of conditional independency is introduced for some types of processes that may be sufficiently general for practice. From the fundamental theory some special prediction formulas are derived; for instance, the joint prediction of future changes, the prediction of the pattern of distribution of a meteorological element and so forth. In order to express the uncertainty of a prediction an idea of stochastic prediction is introduced, which is defined by a sort of relative conditional probability distribution.
An idea of efficiency of a p rediction is introduced that is defined on the conception of risk function in which the effect of a prediction to the human behaviour or the social activity and the protection cost are considered. How to use the stochastic prediction introduced by the author is explained by the risk function and, from a view point of the efficiency, it is shown that the stochastic prediction is the most reasonable one. The calculation method of efficiency is explained by two examples.
The magnitude and the date of the next maximum of Wolf's sunspot numbers are stochastically predicted. The most probable values of them are 118.4 and Jan.1958 respectively and the uncertainty of them is indicated individually and jointly by a sort of probability distribution.
The effect of topography and influence of the vorticity flow near 35°N lat. on the height change of isobaric surface along 50°N latitude circle are discussed, which are, among others, necessary for the prediction of the height change of 500-mb surface over the northern hemisphere by harmonic coefficients computed along every latitude circle. It was found 1) that topography has a considerable effect in low latitudes, but, on the whole, little in our case, and 2) that we must take the area from the pole to 20°N lat. at least as a prediction area.
In the present paper we calculated the horizontal transfer of angular momentum to carry on our investigation of the atmospheric disturbances, aiming at getting an idea of the model suitable for investigation of the general circulation, that is the role of horizontal motion, or to determine the critical wave number in Fourier Series. Besides we estimated the amount of eddy transfer of angular momentum, horizontal conservation of angular momentum and some suggestions about the energy exchange between the disturbances and the basic flow in the atmosphere.
A cup-anemometer model is oscillated in the current of wind tunnel to simulate ship installation in pitching and rolling motion. The mean rotational speed of the anemometer is found larger by a certain amount during the oscillation than in the steady case. This result is contrary to the case of combination anemometer and wind vane, where a certain under-speed is observed in its mean value.
An integrating actinometer for ultraviolet radiation was constructed. It is operated by a 100 V A.C. power supply. The principle of operation is charging up the photo-electric current due to the incidient radiant energy in a condenser and then discharging and forming an impulse to operate a relay counter. Experiments on the linearity of the response, errors due to power supply fluctuations and the absolute sensitivity, were carried out. The general performance has been very satisfactory. Above all, the stability against the fluctuations of power supply is the essential merit of this instrument.
In considering the process of the decay of foam layer of sea water, the author has introduced the same quantity “A” as the activation energy in chemical changes. Thus by using “A” he discussed the phenomenon. The quantity “A” increases with increasing concentration of sea water, up to a certain maximum, and then decreases with further increase of concentration. And also he discussed the frequency distribution of the size of bubbles in the foam layer by means of statistical methods. The theoretical values of the frequencies calculated by statistical methods agreed fairly well with those obtained by experiments; thus it appears that it is possible to deal with the phenomenon of foaming by means of statistical methods or mathematical probability.
Though the author had once made a new current meter using a phototube  and observed many interesting facts with it already  , it is very expensive to use it. Now the author has made another and simpler current meter which will record current speed fluctuations more minutely but directions more roughly than the former. With it he observed ocean currents very well and discussed about them.1. Introduction
By the radioactivity in the air and sea water caused by the US hydrogen bomb tests in the area of Bikini-Eniwetok Atolls,1954. The radioactivity contamination to fishes and fishing boats, were brought about. The observation of the actual state of the radioactivity around Bikini-Eniwetok Atolls was carried out aboard the survey boat “ Shunkotsu-maru”. The maximum radioactivity in sea water was 7025 cpm/l at 570 km west of Bikini and the influence of the radioactivity was recognized as far as about 2000 km from Bikini Atoll. Some oceanographic discussions on the horizontal as well as vertical distributions of the radioactivity in the sea are given.