a i 
PRESENT STATE OF OUR KNOWLEDGE OF THE UPPER ATMOSPHERE. 85 
The correction is thus :— 
Ah=0:01h + 32-4(T, —T,). 
Ah=correction in metres. 
h=height indicated. 
T,=temperature at ground level in °C. 
T, =temperature at height h+0-01h. 
The thermometers are tested by immersing them in a bath of alcohol 
cooled by carbonic-acid snow, the temperature being indicated by a 
standard pentane thermometer. 
The hair hygrometers are compared on the Continent with the 
aspiration psychrometer, and in England with the ordinary wet and dry 
bulb, the humidity being varied by means of sulphuric acid of different 
degrees of concentration or by other like means. 100 per cent. is obtained 
by wetting the walls of the enclosure. 
The Robinson and Assmann anemometers are compared with the 
indications of a standard instrument, the two being exposed together, 
or are placed in an artificial air-current of known velocity produced by a 
Scirocco fan. The Dines’ anemometer is calibrated by hanging various 
weights on the thread of the instrument, each weight corresponding 
to the pressure of a definite wind-velocity on the spherical balls, prede- 
termined by Mr. Dines. For example :— 
Test of Dines’ Anemometer. 
Weight Deflection | wind Velocity ae 
gm. grs. | 
8 123 | 14 / £5 
10 154 LF 70 
20 308 22 112 
30 463 27 13:9 
50 772 37 18:2 
70 1080 47 215 
80 1235 52 23°1 
90 1389 57 247 
Ballons-sondes Instruments.—The second meeting of the International 
Committee at Strassburg (April 1898) recommended that the ballons- 
sondes instruments should be tested as nearly as possible under the same 
conditions as those encountered during the ascent, and, if possible, to 
temperatures and pressures lower than those actually experienced. 
In the method of testing adopted by Teisserenc de Bort, the whole 
instrument was placed under the receiver of an air-pump and the pressure 
was lowered in stages down to about 50 mm. of mercury. The indications 
of the instrument at various pressures and atmospheric temperatures were 
thus obtained. 
To test the thermometer the Bourdon tube, or compound strip, was 
immersed in a bath of alcohol, the rest of the instrument being above the 
level of the liquid. The alcohol was then cooled by means of carbonic- 
acid snow to various temperatures down to —75° C. Teisserenc de Bort 
found that the Bourdon tube alcohol thermometers were satisfactory. 
The aneroid barometers were less accurate, the elasticity of the boxes 
_ being imperfect and giving rise to a considerable lag and to changes of 
zero after being subjected to low pressures.' The effect of temperature 
1 Compt. rend., July 11, 1898. 
