TEMPERATURE OF SCOTTISH LAKES 
103 
(4) Convection. — Radiation and conduction are the principal 
ultimate sources of gain or loss of heat. Currents produced by 
convection and by wind are the principal causes of the temperature 
distribution found in lakes. In spring and summer, when the surface- 
water is being rapidly heated up, there forms at the surface a layer of 
water of much higher temperature than the layers immediately below 
it, and convection currents are set up which equalise the temperature 
near the surface. I have observed at Dores, on Loch Ness, in perfectly 
calm weather and out of reach of river influences, the surface 
temperature to change as much as 6° Fahr. in two minutes. Again, 
in calm frosty weather there is a similar action of convection currents, 
the surface-water is rapidly cooled down, and the resulting difference 
of temperature between the water at the surface and the lower layers 
is equalised or minimised by convection currents. These currents 
were clearly demonstrated by the electrical recorder used on Loch 
Ness. Even the cooling which takes place in the course of cold spring 
evenings is followed by rapid convection currents, which were at 
times shown in a startling manner by the recorder. 
Convection currents are, however, not limited to the surface. 
When the discontinuity layer (see page 117) and the temperature 
seiche are in evidence, there are convection currents set up in the 
neighbourhood of the discontinuity. It is inevitable that there 
should be such currents where there are two layers of water of 
widely different temperatures in contact, especially as the tempera- 
ture seiche and wind currents cause relative motion between the 
two layers. 
(5) Wind. — The effect of wind in determining the temperature 
changes in lakes has never, I think, been given its due place. To the 
effect of winds more than to any other thing I attribute not only the 
cycle of changes which occurs in a lake, but also the absolute quantity 
of heat which becomes stored up in it. If one could imagine a lake 
which is never troubled with winds, and the waters of which are ever 
calm, the problem of its temperature changes would be very easy of 
solution — there would be no wind currents to carry warm water to 
its depths and no temperature seiche. There would be no sudden 
differences in temperature from top to bottom, but a gradual change 
throughout. The transference of heat from top to bottom would take 
place almost entirely by conduction, and, as has already been seen, the 
transference of heat by conduction is very slow. 
The detailed examination of the effect of winds will be more con- 
veniently dealt with after an explanation of some of the phenomena 
which appear. Mention is only made of it at this point for the sake 
of completeness. 
