56 



NATURE 



[May 2 1, 1908 



INVESTIGATION OF THE UPPER 

 ATMOSPHERE. 



THE work carried out in England in connection 

 with the investigation of the upper atmosphere | ment is returned 

 forms part of a wide and well-organised scheme. On j The record consists of fine lines, which remain 



styles on a sheet of copper. This constitutes an im- 

 portant improvement, for the ordinary smoked record 

 is not infrequently obliterated by the time the instru- 



certain days appointed by an international committee, 

 balloons and kites are sent up from some thirty 

 stations scattered all over the world. The kites are 

 used for the study of the lower layers, and free bal- 

 loons for greater altitudes. In a discussion bearing 

 on the isothermal layer we have therefore only to deal 

 with the latter. 



.Abroad, instruments of many types have been de- 

 signed and improved by Hergesell, Assmann, Teis- 

 serenc de Bort, and de Ouervain. The fact that nearly 

 twenty different modifications exist is sufficient excuse 



-SO - 25 



Teinperature in de^/rcs Centigrade. 



Variation of Atmospheric Temperature with height oti .\pril 2, 3 and 4 



for not undertalcing here an analysis of their respec- 

 tive merits. These instruments cost from \ol. to 12Z., 

 and their weight, which exceeds a pound, involves the 

 use of relatively large and expensive balloons. 



In Continental countries but few instruments are 

 permanently lost, and as very substantial Government 

 grants are in most cases available, the question of 

 working expenses is not of vital importance. In 

 England we are financially and geographically less 

 favourablv situated, and there is little doubt that but 

 for the energy and inventive genius of one man, little 

 work would have been attempted. 



The Dines balloon meteorograph costs one-tenth 

 and weighs one-eighth of the corresponding Con- 

 tinental instruments. The record, though microscopic, 

 is nearly iitdcstriictible, for it is engraved by two sharp 



NO. 2012, VOL. 78] 



parallel as long as the temperature remains constant. 

 The bimetallic thermograph is carried on a support 

 which is moved by the barograph. In ordinary 

 circumstances a curve is obtained for any point of 

 which the abscissa measures the deflection of the baro- 

 graph, and the ordinate minus a constant the deflec- 

 tion of the thermograph. 



The calibration of the thermograph is a simple 

 matter, the deflection being practically a linear func- 

 tion of the temperature. The calibration of the baro- 

 graph, on the other hand, offers considerable 

 difficulties. 



In order to increase the sensitiveness of 

 the instrument without adding to its weight, 

 it is necessary to leave a certain amount of 

 air in the aneroid box. This, as also the 

 change of elasticity of the metal, causes the 

 calibration of the instrument to vary con- 

 siderably with temperature, and adds 

 materially to the labour involved in reducing 

 the results. 



.\fter trying and abandoning various 

 methods of approximation, Mr. \V. \. Har- 

 wood, who is in charge of this branch of the 

 work at Manchester, now draws out for each 

 instrument a number of independent calibra- 

 tion curves, each one referring to one de- 

 finite temperature. These form a series of 

 nearly parallel curves, which cover the range 

 of temperature and pressure over which the 

 instrument is used. The calibration made 

 before the ascent is verified after the return of 

 the meteorograph. 



The accuracy obtainable of course decreases 

 with the density of the air; over the first few 

 kilometres a difference of level of less than a 

 hundred metres is easily measured, but when 

 the height attained is such that the prevailing 

 pressure is only one-tenth of an atmosphere, 

 the possible error becomes considerable. 



The diagram here given, which refers to 

 the last date on which balloons were sent up 

 on three consecutive days, is a typical ex- 

 ruiiple of the results obtained. The Man- 

 chester record for January 2 lies between the 

 Pvrton Hill curves for January :; and 4 up to 

 about c) km., above which height the tem- 

 perature remains constant for some distance. 



Two traces were inscribed by the instrument 

 sent up from Manchester on January 4. 

 1903. Probably in this case, as occasionally occurs, 



the balloon did not burst and fall at once on 

 attaining the maximum height, but drifted down 

 slowly as the gas diffused out. In such cases 

 the time to which the second trace applied cannot be 

 determined, and the rise of temperature might well 

 be due to the next day's solar radiation. Trust- 

 worthy or otherwise, thi's trace has its place on the 

 present diagram, the primary object of which is to 

 give a complete collection of the results obtained from 

 one set of ascents. The Pyrton Hill balloons were 

 sent up at 3 p.m., and the Manchester ones at 

 6 p.m. 



It is noticeable tliat during the three days the 

 average temperature gradients between 2 and 8 kilo- 

 metres (/.(■. above the disturbing influence of the 

 ground, but below the level of the isothermal layer) 

 remained practically constant, whereas a considerable 



d 



