6 
Proceedings of the Royal Society of Edinburgh. [Sess. 
gerate the differences in density which in a loch are due to temperature, 
and so to make the experiments more manageable, and the effect of conduc- 
tion is thus eliminated. For the most part, brine solutions were used — a 
dense solution representing the coldest water in the loch, and water of less 
salinity representing the warmer and lighter layers. 
To represent the first temperature phase, when all the water in the loch 
is cold and of uniform temperature and density, the trough was filled to 
within about 2 inches of the top with brine ; and, as there is nearly always 
Fig. 3. 
a wind blowing at some part of a loch, a current of wind was driven along 
the surface of the brine. 
To imitate the gradual gain of heat in a loch, fresh water was very care- 
fully and gradually allowed to flow over the surface of the brine in such 
a way that, but for the action of the wind-current, there would have been 
little mixing, and a separate layer of fresh water would have formed on the 
surface of the brine. A syphon overflow was provided to compensate for 
the inflow of fresh water. The agitating influence of the wind-current, 
however, mixes this inflow of fresh water with the brine, and gradually 
the distribution of density becomes similar to the distribution of density 
due to temperature in the second period of the temperature cycle. 
