76 REPORTS ON THE STATE OF SCIENCE. 
introduction of sufficiently rigid self-recording barometers and reliable 
recording thermometers and hygrometers, however, rendered captive 
balloon observations far more practicable.' 
The kite-balloon of Siegsfeld and Parseval, a more elaborate apparatus 
than that of Archibald, was first used to raise meteorological instruments 
in 1898 at Strassburg, and has since been used regularly at the Prussian 
Meteorological Institute, Lindenberg, to obtain observations in calm or 
nearly calm weather. At most other stations ordinary captive balloons 
have been used, and in weather when both kites and captive balloons 
are useless, small pilot balloons have been employed to determine the 
direction and velocity of the wind. 
Pilot Balloons—The use of small free balloons was first suggested by 
Le Verrier in 1874.2 In 1877 M. Secretan of Paris, under the direction of 
M. W. de Fonvielle, sent up a series of small indiarubber balloons in order 
to investigate the changes of wind direction with altitude and to determine 
the heights of clouds.* 
The method was quickly adopted in America, and before the end of 
1877 it was decided to use these small pilot balloons regularly in Arctic 
work.! They were also employed in 1879 by the French Academy in 
preliminary ascents to determine the paths which manned balloons would 
take.° 
In continuation of the investigation of the variation of wind with 
height, M. Bonvallet in 1891 despatched ninety-seven paper balloons from 
Amiens, and sixty of the cards attached to the balloons were returned. 
The experiments were continued by Hermite during the period 1893-1898, 
and about half of the balloons sent up from Paris were returned from 
within a radius of 100 miles. 
Subsequently these pilot balloons have been employed regularly with 
theodolites in determining the direction and velocity of the wind at 
various heights, and to continue the observations when kites could not 
be flown owing to calm weather, or when an opposing current prevented the 
further rise of the kite.* They have, too, the advantage of reaching greater 
heights than kites. 
Kites —The first use of kites for scientific purposes was made by 
Alexander Wilson and his pupil Thomas Melville at Glasgow in 1749.’ 
In these experiments thermometers were raised to considerable heights. 
Three years later Franklin performed his famous experiment of collecting 
electricity with kites.‘ In 1822-23 the Rev. George Fisher and Captain 
Sir Edward Parry, using self-registering thermometers, obtained tem- 
peratures by means of kites at different heights in Arctic regions.° Some 
time later, in 1840, Espy, an American meteorologist, employed kites 
to verify his calculations of the heights of clouds from measurements of 
humidity.'° The experiments also extended to England, for W. R. Birt 
of the Kew Observatory flew kites in 1847 with the hope of obtaining 
the changes of temperature, humidity, and wind with height. In 1883-85 
KE. D. Archibald used kites with steel piano-wire to obtain the wind 
* Nature, vol. xlv. p. 168. 
? Ibid., vol. xlviii. p. 160. 8 Tbid., vol. xv. p. 458. 
4 Tbid., vol, xvii. p. 171. 5 Thid., vol. xx. p. 401. 
° P. 244; Compt. rend., 141, pp. 605-608, October 9, 1905; 142, pp. 918-921, 
April 9, 1906. 
" Trans. Roy. Soc. Edin., vol. x., part ii. pp. 284-286. 
® Sparks’ Works of Franklin, vol. v. p. 295. LE: 
® Symon’s Meteorolog. Mag., April 1897. 
%° Espy, Philosophy of Storms, 1841, p. 75. 
