7 
1907-8.] The Temperature Changes in Fresh-Water Lochs. 
Another method was also employed to arrive at the density distribution 
in the second phase, but it is not imitative of the natural conditions in a 
loch. A quantity of some soluble salt was put in the bottom of the trough, 
and fresh water carefully poured over it in such a way as to prevent mixing. 
The trough was then left for two or three days. The salt gradually dis- 
solved and diffused through the water, and after the lapse of a considerable 
time gave a density distribution such as was desired. The process was 
quickened by occasionally letting a wind-current blow along the surface, 
and also by starting ordinary seiches. For this experiment potassium 
bichromate was used, and was found very convenient to work with. 
The next problem was to pass from the second phase to the third phase 
experimentally. The change takes place at the time when the loch is just 
beginning to lose heat. The wind-current was continued blowing along 
the surface of the liquid in the trough. At first no change in the density 
distribution was noticeable ; but gradually the currents in the liquid pro- 
duced by the wind-current along the surface changed in character. To 
make these currents visible, a small quantity of black stain was intro- 
duced, and the motion of this black stain easily showed the trends of the 
currents. More will be said about these currents later, and they are only 
mentioned here as evidencing the progress of some change which eventually 
results in the formation of a discontinuity layer. The liquid gradually 
separates out into two distinct layers of different density, and the discon- 
tinuity becomes well marked. The transition from the second to the third 
phase has then been accomplished. 
This experimental imitation of the manner in which the discontinuity 
layer is produced is, I consider, one of the most remarkable results of these 
experiments. It indicates as a cause of the formation of the discontinuity 
layer that, when the surface water ceases to be of considerably less density 
than the lower layers, the return current is more localised, and takes place 
in shallower water. While the loch is gaining in temperature there must 
be on the surface a layer of water considerably warmer than the layers 
below. It is this warm layer which is carried along by the wind and 
accumulates at the end of the loch, the return current being, I imagine, 
slow and distributed through the whole depth of the loch. But when the 
surface loses instead of gains heat, the tendency for the water driven along 
by the wind will be to sink until it reaches a layer of water of equal 
density, where the return current will be set up. The mixing action of the 
waves, also, must have some effect in forming a layer at the surface of 
uniform temperature.* 
* See also p. 1 1 . 
