SEICHES AND OTHER OSCILLATIONS 
41 
the middle is 4 to 6 inches. Suppose the greater part of the top of 
the tank to be covered over and a current of air to be blown along 
the surface. So long as the current is below a certain strength (0'45 
mile an hour), there is no disturbance of the mirror surface ; above 
that limit, and under a velocity of about 2 miles an hour, there is 
disturbance which is transient, i.e. does not long survive the disturb- 
ing causes (cats'-paws). For higher current velocities a regular train 
of so-called progressive waves is formed, which increase in height and 
in length as we go down the wind. In nature, the water is com- 
paratively calm at the windward end of the lake, but more — it may 
be very much — agitated at the other. Even at their greatest, these 
waves in such a tank as that above described are very short (say 
X = 0-25 foot); their period also is short (say T== 0-^2 sec). By 
watching a thin stream of red ink dropped from a pipette into the 
water in the tank, it may be seen that the oscillatory disturbance, of 
which these progressive waves are the manifestation, dies away on 
descending from the surface downwards, and is not appreciable at any 
great depth. Apart from the drift current near the surface set up by 
the wind, the motion of the individual water particles is in closed 
elliptic orbits. From the formulae 
it is easy to calculate the velocity of propagation and the period of 
these waves. 
For Loch Earn, common values would be about X = 20 feet, 
T = 2 sec, V = IO feet per sec. = 6*8 miles per hour. 
Generation of a Seiche by Horizontal Stirring at the Nodes. — 
Following a method due to the two young experimenters, Messrs 
White and Watson, to whose results I shall presently refer, it is 
easy to start a seiche in a long parabolic tank such as described 
above. This is done by stirring horizontally in the middle of the 
tank, with a period of about 5 sec. In a parabolic tank this period 
happens to be independent of the depth of the water ; in general 
it depends on the shape of the basin in which the liquid is confined 
and on gravity, but on nothing else. The result will be found to be 
a motion, quite different in kind from the two cases just described, 
called a pure uninodal seiche. It is a periodic wave motion, but the 
wave form does not travel as in the tM'o former cases. At first 
sight it would appear as if the surface particles merely moved 
vertically upwards and downwards, except at one point called the 
node, where there is little or no perceptible vertical motion of the 
surface. In reality the water particles describe rectilinear orbits of 
various lengths, inclined at various angles to the horizon. These are 
