VI FLIGHT 235 
higher altitudes was to run the anemometer up the string 
of a kite by means of another string passed through a 
pulley ; as soon as it had reached the right altitude a 
third string, passed between the vanes and hanging 
loose, was to be pulled out, thus allowing the instru- 
ment to work. The altitude was to be measured by 
means of a sextant. Alternate hurricanes and calms 
put obstacles in the way of kite flying, and finally the 
anemometer struck work and refused to register. But 
results that are far from valueless were obtained by 
attaching the kite to a spring balance which measured 
the pull at different heights. 
Altitude measured by 
sextant. Pull of kite. 
420 feet ... vis sid sits _ 25 lbs. 
357 feet ... ae isk ee ite 19 lbs. 
350 feet ... oi a ais a 24 Ibs. 
204 feet ... a a te aikt 18 lbs. 
126 feet ... ist iiss ws ... 1§—20 lbs. 
60 feet ... i we a ..» 14—19 lbs. 
42 feet ... — via ene .. T1—20 Ibs. 
There is the question whether the greater length of 
string carried as the kite mounted higher may not have 
to some extent modified the results. Probably, however, 
the error due to this cause was but slight. At low 
levels, from 126 feet downwards, the jerking of the kite 
made an accurate measurement of the pull very diff- 
cult. But the series suggests, though it does not prove, 
that there is a steady though diminishing increase up 
to 420 feet. The pull at 357 feet measured during a 
temporary lull is the only false note in the scale. 
As a supplement to these kite-experiments, I tried 
the plan of letting go a number of small balloons, 
