140 MESSRS. N. SHAIV AND H. DINES: METEOROLOGICAL OBSERVATIONS 
surface. The temperature on the mountain is therefore not likely to differ much 
from the temperature at the surface of the sea diminished by the adiabatic decrease 
due to the altitude; or to put it in another way, the potential temperatures on the 
mountain and at the sea surface are not likely to be very different. At the place 
where the kites were flown, the stream lines, apart from convection currents, are 
probably horizontal, since the disturbance produced by the mountains cannot extend 
very far to windward. The temperature gradient has been shown to be generally 
less than the adiabatic gradient in undisturbed air. In any fluid that expands with 
heat the potentially coldest layers will be at the bottom. The opposite condition 
can only be transient, otherwise it would involve the continuance of a state of 
unstable equilibrium. Admitting therefore that these two suppositions are correct, 
namely, that the temperature gradient over the Atlantic is generally less than the 
adiabatic one, and that the air rises from the sea level up the slopes of the mountain 
and in its ascent follows the adiabatic la\v of cooling, the temperature on the 
mountain must be less than that in the free air over the sea at the same level. 
The adiabatic fall of temperature for the height of Ben Nevis would be 13°‘5 C. for 
dry air, and 7° for saturated air, with an initial temperature of 12° C. The observed 
differences between the temperatures of the Ben and at sea level near Crinan are 
between these limits, except on July 7. The peculiar conditions of the atmosj^here 
on that occasion are well illustrated by the diagram, Plate 12. The differences may 
therefore be accounted for by circumstances which tend to produce an adiabatic 
gradient in tlie air, as it is found near the sea level. For example, on the 25th 
August, the day of the greatest diflerence, the wind was due west, and though the 
weather was showery, the cloud level is set at 2000 metres over the sea, and the 
surface temperature, 16° C., is the higliest recorded during the experiments. The 
mass of the mountain itself, therefore, was probably relatively cold, and the drift of 
surface air over the land might give a temperature diflerence even exceeding that of 
the adiabatic gradient for dry air, but the diflerence of 11° on July 18, when there 
was a northerly wind, be., blowing from the mountains towards Crinan, with cloud 
over the sea at 750 metres, cannot be explained in a similar way. No other single 
suggestion is more fortunate. But the fact that the Ben Nevis—Fort William 
average gradient in July and August, though it agrees witli the kite gradient of the 
United States Weather Bureau, exceeds the Berlin balloon gradient and the average 
nKnmtain gradient, points to an explanation depending on some differences of effect 
of a high level station in that position as compared with a layer of free air at the 
same height. 
It will be noticed, on comparing the readings of temperature at Fort illiam 
(Table C) with the initial temperature of the kite ascents (Plate 12), that there is no 
corresponding difference between the sea-level temperatures at the two positions. 
Tliere is another fact which supports the idea that the Ben Nevis temperatures are 
below those of the free air at the same altitude. I lie level at which the kites 
