PRESENT STATE OF OUR KNOWLEDGE OF THE UPPER ATMOSPHERE. 79 
_ apparatus was employed in the Atlantic expeditions of Rotch and 
Teisserenc de Bort, and of Hergesell in 1902-05.' Rotch made the 
first series of registering balloon ascents in America at St. Louis in 1904.2 
In 1907 the International Committee at Milan, adopting the suggestion 
of Teisserenc de Bort, determined to carry out the observations on a 
much more extended scale in the northern hemisphere. The work was 
extended to Africa and India, and several stations in Great Britain 
began to take part regularly in the ascents. Almost all the countries 
of Europe had previously taken part in the monthly international 
ascents, made since 1901 on the first Thursday in each month, and 
these countries continued to participate in the extended series, which 
included ascents of several balloons on successive days at stated 
periods. The results are collected and published by the International 
Committee. In addition, special ascents have occasionally been made, 
such as those at Milan during the month of September 1906 and at 
Manchester, June 2 and 3, 1909. On the last occasion twenty-five 
balloons were liberated in twenty-four hours, and during the same 
period four balloons were sent up at intervals of six hours from most 
of the Continental stations. 
III. (a2) Apparatus and Instruments employed in ascents of Balloons 
and Kites. 
The increasing use of captive balloons, which were subject to sudden 
shocks and jars, of ballons-sondes, and of kites, gave a strong impetus 
to the work of designing really satisfactory self-recording instruments. 
The light self-recording aneroid barometers, Bourdon tube thermometers 
and hair hygrometers of Richards Fréres, came to be considerably used 
with kites and ballons-sondes. They recorded through levers and metal 
styles on smoked paper, wrapped round a revolving clockwork drum. 
They were used with kites at the Blue Hill Observatory, U.S.A., alongside 
a meteorograph designed by Fergusson, which included also an anemo- 
meter, and by Hermite and Besancon with ballons-sondes in 1893-98. 
In 1891* Assmann described a new form of aspirated psychrometer, 
which was so far independent of shocks and jars as to be suitable for use 
with captive balloons.‘ In the following year, in a review of the results 
of tests and observations made in Germany by balloons and captive 
balloons, he stated that the aneroid barometers and aspirated thermo- 
meters which had been employed were satisfactory, the aspiration being 
absolutely necessary in order to obtain consistent and comparable 
results. The self-recording instruments used registered temperature by 
means of a bent Bourdon tube filled with alcohol, humidity by means 
of a bundle of hairs, and pressure by an aneroid barometer, the whole 
being enclosed in an aspirated space.> At the second meeting of the 
International Committee in 1898° Teisserenc de Bort exhibited a self- 
recording thermometer consisting of a blade of German silver fixed in a 
frame of Guillaume steel, which had small thermal inertia (requiring only 
15 seconds to indicate a sudden change of temperature of 9° C.),and which 
was not affected by shocks. Cailletet showed an instrument for photo- 
graphing simultaneously the face of the aneroid and the ground in order 
' See above, p. 77. 2 Ann. Harvard Obs. vol. \xviii. part 1. 
: Nature, vol. xliv. p. 502. ‘ Ibid. ai 
> Ibid., vol. xlv. p. 168. ® Ibid., vol. lviii. p. 380. 
