58 Proceedings of the Royal Society 
first, that the temperature above alluded to, 39*2° Fahr., was not 
observed at all at any depth between the ice and the bottom ; and 
second, that the curve representing the vertical distribution of tem- 
perature had a point of contrary flexure, showing that the actual 
distribution could never have resulted from the freezing of a thin 
layer on the surface of a large volume of water at a uniform tempera- 
ture, and the further cooling of it by conduction from the lower ice 
surface. The observed temperatures showed that, if the water had 
ever been at a uniform temperature throughout its depth, that 
temperature was certainly much below 39 *2°, and that while the 
water was being cooled by conduction downwards from the ice, it 
was being warmed by conduction and convection upwards from a 
source of heat at the bottoim 
These unexpected results induced me to repeat my visits to the 
loch, and the observations at Stations Nos. 2, 3, and 4 were made 
on the 9 th and 1 1th January. It will be seen that the results of 
them fully bear out the conclusions derived from the preliminary 
observations at Station No. 1. At Stations Nos. 2 and 4, situated 
in the deep western basin of the loch, the suspected heating sur- 
face is separated by a sufficiently thick stratum of water to enable 
its action to be studied separately. The observations at No. 4 are 
