Temperatures — Autumn. 299 



are not caused by the warming or cooling of the water but by 

 the fluctuations in the level of the thermocline. These fluctua- 

 tions go on to a certain extent without an assignable cause, but 

 the larger movements, at the station where observations were 

 taken, are plainly due to the wind. 4. The upper layers of the 

 cool water become mingled by the action of the wind with the 

 lower part of the warm water above it and are taken into the 

 warm layer. Thus the thermocline moves constantly downward 

 ►during summer, while the water below it is little or not at all 

 •changed in temperature. 5. The water below the thermocline 

 is practically stagnant during the summer, and is cut off from 

 direct exposure to sun and air. As a result, it may become unfit 

 to support most forms of animal life, as is the case in lake Men- 

 dota. 6. The larger changes in temperature below the thermo- 

 <cline are due to currents caused by winds. 



Autumn temperatures. 



By the latter part of September the temperature of the sur- 

 face water has fallen so that it exceeds that of the bottom by 

 barely 5°. At this time also gales from the north are 

 -apt to occur whose effect is to break the thermocline and 

 render the lake homothermous. This result is reached at 

 different dates for different depths, but in both years the 

 lake became homothermous in its deepest parts about two or 

 three davs after the time when a similar condition was reached 

 at 18 meters. In each year the homothermous condition was 

 reached at a temperature not much exceeding 16°; and in 

 general the temperature for the 1st of October may be stated 

 .as about 16°. 



The breaking up of the thermocline is accompanied by a 

 marked rise in the temperature of the bottom water. In 1895 this 

 rise amounted to 2.8° from the 26th to the 23th of September; 

 .and in 1896, to about 1.5° in the same time. 



During October and November the temperature falls with 

 -singular uniformity, as indicated by the weekly averages, pass- 

 ing the temperature of the maximum density of water late in 

 November. The decline continues steadily until a temperature 

 is reached between 2° and 3°, after which the cooling goes 



