382 THE SEICHES OF LOCH EARN. 



In fig. 18 are given tracings of an actual seiche record taken on Locli Earn on 

 the 14th September 1905. In taking the curve 2, an access tube 12 feet long 

 of ^g--inch diameter was used ; in taking the curves 1 and 3, 6 feet of 5-inch diameter. 

 In the interval between 12 h 55 m and 14 h 25 m the state of the lake had altered very 

 little. The lower trace is interesting, as it gives an automatic record of the finer 

 denivellations — or vibrations, as Forel calls them — which could not be accurately 

 registered by the most skilful eye observations, and which would be almost invisible 

 on one of the large self-registering limnographs running at the ordinary speed of 1 mm. 

 per minute. These vibrations are also beautifully shown in figs. 10 and 11. 



If the instrument were to be constructed specially with a view to seiche observa- 

 tions, the above theory suggests several obvious modifications which it is needless to 

 discuss here. 



On the Reduction of Limnographic Measurements, the Method of 



Residuation, ETC. 



The deduction of the periods, not to speak of the nodes, of the various seiches of 

 a lake from limnograms taken at various points, and under various conditions, is by no 

 means so simple a problem as appears at first sight. 



It has been proposed to apply the harmonic analyser to the discussion of limno- 

 grams ; but this is not possible, at least directly, for the harmonic analyser simply 

 enables us to deduce the amplitudes of harmonic components whose periods are known 

 or assumed a priori. But the periods are precisely what we have in the first instance 

 to determine from the limnogram itself. 



If only one seiche appeared at a time, unaccompanied by any other denivellation, 

 it would simply be necessary to measure the time between two distant maxima or 

 minima and divide by the number of oscillations between. It is, however, very rare 

 that a pure seiche prevails for more than a very short time. During iVugust and 

 September 1905, the new limnograph at St Fillans on Loch Earn may be said to 

 have never drawn a straight line ; yet in all that time it gave scarcely a single certain 

 example of a pure seiche. 



Use of the Configuration Period. — The most commonly occurring type of seiche 

 on Loch Earn is a more or less pure dicrote, in which the uninodal and binodal are 

 the predominant components. Figs. 3 and 4 are reproductions of good examples. 

 Such seiches are occasionally very persistent on Earn. When the new limnograph 

 was at Lochearnhead, a dicrote seiche was traced continuously, a]though with varying 

 amplitudes and varying embroidery, from about 8 h 30 m on 16th October to 20 h 15 m on 

 22nd October 1905. 



Owing to the accurate or approximate commensurability of the periods T 1 and T 2 

 of the uninodal and binodal, the limnogram of a protracted seiche of the kind just 

 alluded to consists of a repetition of a certain portion, which we may call the "Con- 



