ON THE SEICHES OF LOCH EARN. 463 



Thei'e is still greater uncertainty regarding the periods in Tables XI.-XIIL, roost of 

 which rest only on a single series. Possibly T = 2'88 is the sextinodal period. It 

 must, however, be borne in mind that the smaller the period the greater is the danger 

 of confusion with progressive wave disturbances, with possible transversal seiches, 

 or even with secondary local oscillations due to indentures in the shore of the lake. 



Determination of the Nodes. 



The difficulties anticipated * in determining the nodes by direct observation were 

 more than realised in practice. When the range of the seiche is large, there is nearly 

 always a great deal of wind-embroidery of an irregular character, which it is impossible to 

 eliminate either by damping the limnograph or by residuating the limnogram. Also, 

 where the amplitude is small, there is almost always an aperiodic variation of the lake 

 level, probably due to the heaping up of the water on the shallow shore, an effect 

 which will vary with the slope of the beach. The varying slope also affects the range 

 of the seiche to an extent which it would be difficult to calculate with any degree of 

 accuracy. Both these causes introduce uncertainty in the method of observing with 

 index limnographs on two sides of the node where the seiche is found in opposite 

 phases, and then deducing its position by interpolation. A mere null method would 

 scarcely lead to a satisfactory result, unless under exceptional circumstances which did 

 not occur during our observations. Of the many attempts made, only a few led to 

 limnograms which could be utilised ; and in every case the curves had to be purified 

 by residuation. 



Uninode, — The two best pairs of observations gave almost exactly the same 

 position for the southern end of the uninode, and led to the conclusion that it lies 

 about 105 yards west of the position given in my paper on the Calculation of the 

 Periods and Nodes of Lochs Earn and Treig. t This is precisely what was expected, 

 as the actual normal curve (C, p. 825) rises above the assumed biparabolic curve on the 

 west and falls below it on the east of the calculated position of the uninode. It would 

 be useless to calculate what the amount of divergence ought to be; because the 

 uncertainty of one of these determinations, as shown by the observations themselves, 

 is ±65 yards, and of the other ±129 yards, the latter being more than the divergence 

 itself. The exact agreement of the two determinations is probably an accident. 



Eastern Binode. — Two determinations agreed almost exactly in placing the 

 Eastern Binode about 117 yards west of the calculated position; but the uncertainty 

 of these determinations was ±94 yards in the one case and ±59 yards in the other. 



Western Binode. — The best pair of observations gave a position for the southern 

 end of the Western Binodal line 305 yards west of the calculated position. A 

 divergence in this direction was to be expected from the shape of the true normal curve 

 in the neighbourhood ; but the amount is somewhat surprising. There can be little 



* Trans. R.S.E., vol. xli., p. 850. + Henceforth referred to as " C." 



TRANS. ROY. SOC. EDIN., VOL. XLVI. PART III. (NO. 20). 71 



