480 



MR E. M. WEDDERBURN 



different directions, and the surface waters were evidently in a very agitated condition, 

 although the surface of the water was quite calm. Everything goes to show that heat 

 is communicated from one part of the loch to another by convection more than in any 

 other way (leaving out of consideration the effects of winds, which have already been 

 discussed). Another fact which strengthens this view is that there are frequently 

 lesser Sprung schichts at different depths, and in the neighbourhood of all these there 



II Ji'oorj ) I 



/ P.M.. 



5 " 



4 ,f 



N? Z. 



/V ? J. 



W 3 



1 



7 



8 



9 

 lO 



llMid.-Mgkt 



I A.M. 



7 n 



# " 

 6 .. 



Fig. 14. 1 



must be convection currents set up. The behaviour of water in the neighbourhood of a 

 Sprungschicht is very similar to the behaviour of water near the surface. Before this 

 was understood, the observers were often puzzled when working with mercury thermo- 

 meters by obtaining readings which seemed to show the presence of very marked inverse 

 stratification, which always disappeared in a very illusory fashion when any attempt 

 was made to follow it up. An explanation was offered by the platinum thermometers. 

 The curves obtained by means of the recorder with the thermometer at various depths 



* No. 1. Thermometer at 5 feet, 22-23 May 1904. No. 2. Thermometer at 5 feet, 23-24 May. No. 3. Tliermo- 

 m eter at surface, 24-25 May. See also footnote, page 423. 



