*) 
») 
BY THE USE OF KITES OFF THE WEST COAST OF SCOTLAND, 1902. 
regards time, but such a proceeding would entail an amount of labour quite out of 
proportion to the increased accuracy, having regard to the character of tlie traces 
obtained. j.t has been already mentioned that the mechanical vibration of the leveis 
was considerable in high winds, and this gave rise to a thickening of the lines of the 
traces, sometimes to almost undecipherable smudges. The stability of the aneroid' 
lever is considerably greater than that of 
the others, so that the height trace is 
generally sharp and distinct, while the 
others are relatively blurred, so much so in 
some cases that the reading cannot he 
relied upon to within a division of the 
paper scale or a degree of temj^eratiire. 
The difficulty that may arise from this 
cause is illustrated by the record for 
August 23, the highest obtained, which is 
reproduced in fig. 3. 
As a degree of temperature C. could in 
no case correspond to less than a hundred 
metres of height, the system of tabulation 
and adjustment of time scales may be 
regarded as sufficiently accurate in com¬ 
parison with the other uncertainties of the 
experiments. But the method of recording 
is a matter which requires very careful 
attention with a view to increased accuracy. 
The results as they are here presented 
must be regarded as giving the general 
relation between height and temperature; 
any local variations, due to differences in 
parts of the same horizontal layer at short 
intervals, have been smoothed out by the 
process of reduction. Such variations form 
a separate subject of study, wdiich requires apparatus specially adapted for this 
purpose. 
It has been pointed out already that the humidity traces are subject to special 
disadvantages, so that as a rule only rough approximation can be regarded as secured. 
On two occasions, however, very interesting traces were obtained, namely, on July 31 
and August 12, when the air became dry at some distance above the surface (see 
diagram, fig. 4). July 7 also gave an interesting record of humidity (see Plate 12). 
In a number of cases the humidity trace was either lost or undecipherable, owing to 
the relative instability of the pen lever. 
S 2 
