SEICHES AND OTHER OSCILLATIONS 
67 
of ordinary seiches in the variations of the general system of isobars 
shown in the daily weather charts. 
2. Effect of Wind Denivellation. — It is well established by the 
researches of Sir John Murray that a wind which has prevailed for 
some time causes transport of the water of a lake in the direction in 
which the wind is blowing, and the observations of von Cholnoky on 
Lake Balaton show that in shallow lakes this wind denivellation may 
be considerable, and that its sudden release may give rise to seiches. 
After a long and careful examination of our limnograms, we have 
arrived at the conclusion that this kind of denivellation is very small 
on Loch Earn under ordinary circumstances, and is rarely an effective 
cause of seiches. It is, however, not easy to judge of this matter. 
When the wind is light, the effect is very small, and cannot be 
separated from the denivellations due to precipitation and evapora- 
tion, and to variations in the barometric gradient. When the wind 
is high it is usually accompanied by considerable fluctuations of the 
barometric pressure, or by rainfall, or by both ; and again the 
difficulty of separating the causes arises. That wind denivellation 
should be small on Loch Earn is not surprising, for, looking at the 
ratio of its depth to its length, we must classify it as a deep lake, and 
in such lakes, as is now well known, the return under-current readily 
forms, and prevents the accumulation of wind denivellation. 
The seiche of 3rd September 1905 (fig. 23) is interesting in the 
present connection, and also because it was accompanied by the 
strongest gale experienced during the two months of observation. 
For some hours before midnight the wind had been very light, 
and at it was practically calm. About 2^ 37™ the wind began to 
rise, and in an hour it had reached a mean velocity of about 15 
(mile/hour). The velocity fluctuated between 6 and 15 till when 
a very sudden rise began. By 7^ 30"^ the average velocity had risen to 
35, with extremes of 45 to 50. About 8^' 30™ there was a sudden drop 
to about 25, then a more gradual drop to 10 at 9^ 20™. After that 
the gale rose again to a mean velocity of 35 to 40, with extremes 
occasionally reaching 53. After lasting four hours, the gale began 
to abate about 15^, and then fell more or less uniformly to calm 
about 20^, there being two rather sudden lulls at 17^ and 19^' 20™. 
Throughout the whole of this time the microbarogram is much 
disturbed. During the strongest parts of the gale it shows the 
characteristic wind blurring, and throughout there are fluctuations 
of various periods: e.g. 7-2' at 2'^ 5*6' at 4'^ 30™, 13-6' at 8^ 30™, 
IT at 16^. 
Till about midnight there had been a fairly regular UB-seiche 
with a small trinodal component, the total range of the whole being 
about 31 mm. Soon after midnight, that is, more than two and a 
