Ekholm's, Besson's, Fineman's, Marvin's, Vettin's, Finkes, Strachey's, Galton's, Clement Lay's and some other's nephoscopes have respective superiority and defect. The present instrument is comparatively good for routine observations. It consists of a plane mirror mounting on an azimuth circle, which carries a bridge with a movable eye piece, as shown in fig. 1. The eye piece slides along the horizontal beam of bridge with a constant height above the surface of mirror. The equatíon of cloud motion for this instrument is V/H=m/t, where V is the true velocity of a cloud mass, H its height, m the distance on the mirror travelled by the image of cloud, and t is the time of the travelling, by taking t. suitably, i.e. t=10 seconds for H=1000 metres t=20 seconds for H=2000 m and so on, we have the formula V=m. The scale of m on the mirror shows at once the velocity in metre-per-second. In these respects the nephoscope is just convenient as that of Marvin's. Besides, our instrument has a scale along the bridge showing the altitude of cloud above borizon, hence we can very easily read off the azimuth and altitude of the cloud. We use such sheets of paper in which azimuth lines and altitude circles are printed with blue ink. The radii of altitude circles is proportional to their tangents and hence the point entered in this diagram represents at once the position at the level of cloud. If we enter the points for successive readings of the cloud, then we can trace a line (nearly always straight) which shows the motion of cloud and the locality of fiamament through which the cloud passed. Fig. 2 and fig. 3 show examples of such diagrams before and after a storm. We can get some idea on the local character of the upper current by constructing such a diagram every day. For clouds passing very near the zenith we use an incl_??_ned small mirror with angle of elevation of 15°. But in practice we scarecely find any use of this auxilialy mirror. Very easily and accurately we can measure the angles for halo, corona, rainbow and alied phenomena, with this instrument.
The author has special interest on the relation between the sunspot activity and the motion of upper atmosphere. He already pointed out the increase of westward component of cirrus movement in southern and central Japan.(1) As the semi-diurnal term in the barometric daily oscilation concerns mainly to the upper atmcsphere, he, for the present, compares the secular variation of the amplitude and phase of the said term in Tokyo (obtained by harmonic analysis) with sunspots activity and mean annual direction of motion of upper and middle clouds. The curves are shown in fig. l. p. 47. Numerical values are tabulated in the table in p. 48. With these we can see that between these elements a certain correlation, though not perfect, exists-positive correlation between spots and semidiurnal oscilation and negative one between these and the eastward compornent of upper cloud motion.
The author draws attention on the peculiarty of the variation of wind velocity near the ground measured at Kobe with pilot balloon, as shown in the table in p. 00. The velocity measured a little above the ground is smaller than that measured at the ground. The observatory stands on a hill. The stream lines most closely assemble at the hill top. The author left the problem. unsolved but his opinion is that the above peculiarity must due either to the stream effect or to the instrumental error of anemometer at the ground.
The records of spell of dry days in Tokyo since 1911 are tabulated in the table p. 52. As dry days the days without precipitation and those with precipitation less than 0, 7mm and not occur on successive two days are taken. The longest spell occurred on 1917 and it continued 39 days. The next longest happened this year and it was 34 days. Fine days prevail in this district when the general distribution of pressure belongs to so called West-high East-low type as already named by Prof. Okada, and also when the high pressure belt covers Japan proper from West to East or when the belt situates a little south of Japan Proper so that South or Southwest breezes prevail in the district. The West-high East-low occurs cyclically and hence in normal winters fine days is interrupted by rain or snow-fall nearly once a week or, once in 5 or 6 days. But when the high pressure belt takes position mentioned above, it persists rather long, during which the weather is kept dry. The former cyclic phenomenon superposes with this, so that we have cloudiness or week shower once in 5, 6 or 7 days even in this period. To know the cause of the movement of high pressure zone is most important for long range forecast, but it is not yet discovered. (S. F.).
Literature (1) T. Leconte: Observations on a remarkable exadation of ice from the stems of vegetables and on a singular protrusion of icy columns from certain kind of earth during frosty Weather. Philos. Mag. (3) XXXVI. p. 329. Fortsch. d. Phya. 1850/51, p. 264. (2) B. Schwalb: Ueber Eisfilamente. Met. Zeit. 1885. p. 185 Nature XXIX pp. 472. 496. (3) Woodd Smith: Nature XXXI. p. 5 and discussions of many others in the same Vol. pp. 29, 81, 193, 216, 264. (4) W. Prinz: Production des filaments de glace à la surface du sol. Ciel et. Terre 1885. pp. 208-210. Met. Zeit. 1885 p. 92 (145). (5) R. Assmann: Ueber Eisfilamente. Meteorologie und Luftschiffahrt p. 14. Met. Zeit. 1890 Literature bericht. 20. (6) Cleveland Abbe: Ice-columns in gravelly soil. Amer. Meteorol. Journ. Vol. IX. 1893 p. 523. (7) T. Okada: Monthly Weather Review. (8) M. Goto and L Inagaki. Report on the investigation in the Eiscolumns. Journ. Met. Soc. Japan. 18 year. (1897) p. 285. 351, 413. The last paper accentuate the action of bubbles suspending in the thin film of water between the bottom of the ice columns and the earth surface. But there still remain some doubts and further investigation is recommended. Fig. 1, 2, 3, 4 and 5 show the general feature of the phenomenon, which is very common in Japan.