BY THE USE OF KITES OFF THE WEST COAST OF SCOTLAND, 1902. 
129 
To ascertain the variations of temperature and humidity with height, it is necessary 
to obtain simultaneous readings of all three elements from the curves. This presents 
some little difficulty. The time scale of the paper is 2 centims. for 50 minutes, 
or a millimetre for 2T minutes. Simultaneous readings could be easily obtained if 
the })ens were always so adjusted that the lengths of the levers for the several 
records Avere always equal, but the displacement of a pen by half a millimetre implies 
a^ difference of more than a ininute in time, and the winding in was occasionally 
sufficiently rajnd to give a vaiiation of height of upwards of 100 metres per minute 
(see fig. 1). From an examination of the curves it appeared that the setting of the 
temperature pen as compared with the height jjen might show a time difference of 
from a minute fast on the one side to six minutes slow on the other. Durinn the 
o 
ascents, if the Avind is strong, the levers are subject to considerable mechanical 
vibration, and the setting of the pen, Avhich is simply slid on to the lever and held by 
small clips, is not sufficiently secure. In any future ascents it aauII be desirable to 
pay special attention to this part of the apparatus, with a vieAv to securing really 
synchronous readings Avithout difficulty. In the present discussion an error of 
50 metres of lieight must be regarded as not improbable, on account of the uncertainty 
of the correspondence of the time scales. 
One might expect to obtain a series of truly synchronous readings from the curves 
as they stand by taking readings at salient jaoints, sucli as are given by pi'ojections 
from the height curve Avith corresponding projections on the temperature or humidity 
curve. Such a series of salient points is very conspicuous in the curves for August 11 
(fig. 1), but to proceed in this AA'ay first assumes that the instruments register-the 
maxima or minima of the separate elements simultaneously, though such an 
assumption is not necessarily justified, and, secondly, it prevents tlie use of a 
continuous cuinm to smooth the effects of isolated peculiarities Avhich may arise from 
the peculiar circumstances of the moment, or even from the instiaiments themselves. 
The process adopted Avith the traces under discussion AA’as to tabulate each curve 
according to its OAvn time scale for eacli tu-o minutes of time, and subsequently t<3 
determine by a careful examination of the original traces the proper equation of time 
betAA-een them. 
Curves of Amriation ol tenqjerature and height were plotted for each ascent. The 
cuiwe thus obtained for the ascent of August 11 (represeidmd in fig. 1) is giA'en in 
fig. 2, p. 130. The adjustment of tlie time scales aauis subjected to careful scrutiny, and 
to obtain a final curve of Amriation of temperature with height the observations Avere 
re-plotted, after the scrutiny, and a mean curve taken hetAveen the curves of ascent 
and descent; Avith a little practice it is easy to recognise in the })lotted curves 
the eccentricities due to Avant of synchronism, and in the correspondence finally 
adopted the observations are synchronous Avdthin a minute. The smoothed cuinm 
probably gives the relation hetAveen temperature and height Avell Avithin that limit. 
More accurate equivalence of the scales coidd ]>e obtained by closer taludation as 
VOL. CCJT.—A. 
s 
