THE FRESH- WATER LOCHS OF SCOTLAND. 421 



deeper waters. This return current, however, acts on the water below 

 the discontinuity layer just as the current of wind acts on the natural 

 surface of the loch, and a secondary current is produced at the surface of 

 discontinuity. This secondary surface current is much slower than the 

 surface current produced by the winds, but to take the place of the 

 water carried along by it there is a secondary return current at the 

 bottom of the lake. The secondary return current is very slow, and 

 its existence was first suggested to the writer by experiments carried 

 out in a glass trough, but observations support the view taken. The 

 current systems thus described are shown in Fig. 70. 



Another effect of the separation of the loch into two compartments 

 by the surface of discontinuity, is to render possible the temperature 

 seiche. The surface current produced by the wind transfers a large 

 quantity of warm water to the lee end of the loch, with the result that 

 the surface of discontinuity is deeper at the lee than at the windward 

 end of the loch. When the wind moderates or ceases a temperature 

 seiche is started, just as a seiche is started in a basin of water which has 

 been tilted. The temperature seiche was also studied experimentally, 

 and was made possible by superimposing a layer of paraffin on a layer 

 of water. By driving the paraffin to one end of the trough by a current 

 of air, the water, corresponding to the water below the surface of 

 discontinuity in a loch, received a tilt, and when the current of air 

 was stopped, a seiche started in the lower layer of water independently 

 of the upper layer of paraffin. 



The temperature seiche was first described by Mr. E. R. Watson in 

 the autumn of 1903, and a good deal of doubt was expressed as to the 

 accuracy of his views, but the theory of a temperature seiche was 

 established by the observations taken in 1904. For a considerable 

 period observations were taken at Fort Augustus every two hours, so 

 as to obtain a continuous record of temperature. Fig. 71 is drawn from 

 the observations taken in July and August, 1904, and shows the tem- 

 perature variations at Fort Augustus at the surface and at depths of 

 50, 100, 150, and 200 feet. It will be observed that in July changes 

 at the surface, which are chiefly produced by winds, are accompanied 

 by similar changes at all depths, but that in August, when the dis- 

 continuity layer has been formed, the temperature variations at the 

 surface are independent of the variations at a depth of 100 feet, where 

 the variations are principally due to the temperature seiche. 



Observations made at the two ends of the loch further support the 

 theory, as showing that the layer of discontinuity was in general rising 

 at one end when it was falling at the other end. Continuous records 

 obtained from the Callendar recorder are also easily explained by the 

 temperature seiche. Rough calculations were made of what should be 

 the period of this seiche, based on the assumption that the loch contained 

 two layers of water of different but uniform density. The observed 



