26 Proceedings of Royal Society of Edinburgh. [sess. 
The binodal seiche, whose period is about 15 ‘3 minutes, is 
usually very well marked. It is the commonest type, and lasts 
longer than the uninodal seiche. The node is probably some- 
where in the neighbourhood of Inverfarigaig, but has not yet been 
accurately determined. It is also interesting because its period is 
less than half the period of the uninodal seiche, although, accord- 
ing to Du Boys, it ought always to be greater than half ; and in 
most lochs it is so, the most notable exception being Lake Geneva.. 
The basin of Loch Ness is so regular that it is difficult to explain 
it, as was attempted in the case of Lake Geneva, by assuming an 
oscillation of part of the loch. 
The polynodal seiche, whose period is 8*8 minutes, is always of 
small amplitude, but sometimes very regular. There are also' 
oscillations of shorter period, but they do not occur regularly 
enough to allow of their measurement with any degree of accuracy. 
On one or two occasions there were embroideries on the curve,, 
which may have been due to transverse seiches. Owing to the 
narrowness of the loch, the period of such a seiche would only be 
about 1 minute. These embroideries may be due to a variety of 
causes, such as the wash of steamers, the opening of the lock gates 
in the canal, etc. It will only be possible to determine whether 
they are vibrations or transverse seiches by simultaneous observa- 
tions at the two sides of the loch. 
The range of atmospheric conditions at Fort Augustus included 
thunderstorms and earthquakes, but these had no very marked 
influence on the loch. It seems probable that the cause of seiches 
is sudden local variations of atmospheric pressure ; and this view 
is supported by the records of a barograph at Fort Augustus. The 
polynodal seiches, and perhaps the uninodal and binodal seiches 
also, may be started by sudden gusts of wind. The wind blows 
down the various glens in strong, almost vertical gusts, and this 
may be sufficient to start the oscillation. 
All the speculations, however, regarding the causes of seiches 
can only be satisfactorily tested by quantitative measurements of 
the forces at work, and it is hoped that something will be done in 
this direction in the summer of 1904. 
( Issued separately January 15 , 1904 .) 
