November 4, 1922] 



NA TURE 



615 



that tropical cyclones are essentially convectional 

 phenomena. Observations in the free air in and 

 around a tropical cyclone are not available. Dr. V. 

 Bjerknes, in his theory of the polar front, has 

 recently given a new conception of the part that 

 local wind systems play in the formation of cyclones. 

 The author states that it would seem that the cause 

 of the origin of the tropical cyclone may be found 

 in the counter current theory as to initiation of the 

 cyclone centre, while the convective theory accounts 

 for its maintenance after having started. Much 

 information is given on the movement of hurricanes, 

 and there are numerous charts showing the travelling 

 centres in association with the surrounding distribu- 

 tion of atmospheric pressure. 



Local or Heat Thunderstorms. — The U.S. 

 Monthly Weather Review for June gives an interesting 

 and instructive account of the development of 

 thunderstorms by Prof. C. F. Brooks, of the Clark 

 University, which was presented before the American 

 Meteorological Society in April last. The supply and 

 action of the ascending and descending currents 

 of air are explained, as well as the formation and 

 effect of rain in the development of the storm. The 

 physical make-up of the thunderstorm is said to 

 develop quickly into a central descending and out- 

 flaring current of cold air, surrounded by a cone of 

 rising warm air, and still farther out by a zone of 

 descending air. A thunderstorm is described as the 

 result of relatively large streams of air in violent 

 convection attended by abundant condensation of 

 moisture. With reference to the rapid rising of air 

 in cumulus clouds that are growing into cumulo- 

 nimbus, the author remarks that on different occa- 

 sions his rough measurements have shown upward 

 motions of 3, 4, and 7 metres per second in the tops 

 of cumulus clouds. Aviators and aeronauts who 

 have been within active portions of cumulo-nimbus 

 clouds have experienced great bumpiness owing to 

 the strong up-and-down currents. Violent convection 

 is said to be caused by the instability accompanying 

 a large lapse rate in temperature. Abundant con- 

 densation of moisture is essential to the start of a 

 thunderstorm. The gist of the communication is 

 the predicting of local thunderstorms, and certain 

 questions are formulated for the forecaster relative 

 to streams of air, convection, and condensation. 

 It is suggested that the conditions be tabulated and 

 that use be made of a + or - answer, the summing 

 up of which will indicate whether local thunderstorms 

 are probable. Important information is given as to 

 where local thunderstorms originate. 



The Sphere-gap Voltmeter. — When it is neces- 

 sary to measure the maximum or peak voltage of an 

 alternating current from a transformer or induction 

 coil the sphere - gap voltmeter is often used, as its 

 indications are independent of the humidity of the 

 air and of the form of the voltage wave. The 

 following particulars of such an instrument at the 

 National Physical Laboratory, furnished by Dr. E. 

 A. Owen in the October issue of the Journal of the 

 Rontgen Society, will prove useful. The spheres, 7-62 

 cm. in diameter, are mounted on ebonite pillars 21 

 cm. long, with sulphur rings 5 cm. long let into them 

 for additional insulation. One sphere is fixed and 

 the other supported on a slide which can be moved 

 towards the fixed sphere by means of a screw. A 

 scale on the slide gives the distance apart of the 

 spheres at their nearest points. The spheres are 

 connected to the supply and are moved slowly towards 

 each other till a spark passes. The peak voltage is 

 then deduced from the distance apart by the following 



NO. 2766, VOL. I IO] 



data : 1 cm. 32-7 kilovolts ; 2, 60 ; 3, 86 ; 4, 196 ; 

 5, 124 ; 6 cm. 141 kilovolts. 



Ozone. — Prof. E. H. Riesenfeld, of Berlin, has 

 recently described (Chemiker Zeitung, October 7) the 

 preparation and properties of pure ozone. Ozonised 

 oxygen containing 10-15 P er cent, of ozone was lique- 

 fied in exhausted glass bulbs by cooling in liquid air. 

 The deep blue liquid, on exposure to reduced pressure, 

 gave off mainly oxygen, and at a certain composition 

 separated into two layers : the upper, dark blue, layer 

 was a solution of ozone in liquid oxygen ; the lower, 

 deep violet-black, layer was a solution of oxygen in 

 liquid ozone. The lower layer, formerly considered to 

 be pure ozone, contains about 30 per cent, of oxygen at 

 - 183 C. The oxygen was pumped off from it, and 

 pure liquid ozone (B.P. - ii2-4°C.) obtained. The 

 vapour density of 48 (O a ) was found by Dumas' method. 

 On cooling in liquid hydrogen, solid ozone, in violet- 

 black crystals, M.P. - 249-7° C. was formed. The 

 gas, deep blue in colour, is, in the absence of all 

 catalysts, remarkably stable. Pure gaseous ozone 

 can be exploded by an electric spark, but some re- 

 mains unchanged. This would be expected from the 

 endothermic character of the substance. The critical 

 temperature is - 5 C. No evidence whatever of the 

 existence of higher polymers of oxygen was obtained : 

 both in the liquid and gaseous states the formula is 

 3 . This work is of great interest, and, apart from 

 the determination of the physical properties of ozone, 

 it removes the last doubt as to the simple character 

 of ozone — " oxozon " does not exist. 



Differential Gas Analysis. — Mention has already 

 been made in Nature of a method devised by Dr. 

 G. A. Shakespear of Birmingham University for 

 measuring differences in composition of similar gas 

 mixtures. The method, which has proved itself 

 valuable for controlling the purity of hydrogen, the 

 safety of atmospheres in balloon sheds, and many 

 other purposes, depends on the differences of thermal 

 conductivity of gases. Two identical spirals of 

 platinum wire are enclosed in separate cells in a 

 metal block, each spiral forming one arm of a Wheat- 

 stone Bridge circuit, the other two arms being of 

 manganin. An electric current flowing through the 

 bridge thereby heats the two spirals, which lose heat 

 to the walls of the cells. If the two cells contain 

 gases of different thermal conductivities the spirals 

 will cool at different rates, and one spiral will there- 

 fore be maintained at a higher temperature than the 

 other. The difference in temperature of the two 

 wires thus causes a deflection of the galvanometer, 

 the extent of which depends on the difference in 

 conductivity of the two gases. The construction is 

 such that changes in the temperature of the gases 

 affect both sides of the bridge equally. If, therefore, 

 one cell contains a pure gas, and the other cell the 

 same gas mixed with some other constituent, the 

 extent of the deflection will indicate the proportion 

 of the second gas present, and the galvanometer can 

 be calibrated to show directly the percentage com- 

 position of the mixture. The difference in conduc- 

 tivity between air and carbon dioxide enables the 

 method to be used to determine the percentage of 

 carbon dioxide in flue-gases. The other constituents 

 of flue-gases either have thermal conductivities differ- 

 ing but little from'those of nitrogen, or are negligible 

 in amount, while the effect of the water vapour can 

 be counteracted by keeping the gases in both cells 

 saturated. By attention to certain details the method 

 may be then applied to follow the change in carbon 

 dioxide content of the flue-gases in fuel-consuming 

 installations. The instrument is made by the Cam- 

 bridge Scientific Instrument Company. 



