ON THE SEICHES OF LOCH EARN. 469 



Observers are now agreed that the development of seiches usually accompanies local 

 disturbances of the barometric pressure whose duration if they are transitory, or period 

 if they are periodic, does not differ greatly from the period of the seiche in question. 

 Our observations on Earn fully bear out this conclusion. Disturbance on the micro- 

 barogram is always accompanied by disturbance on the limnogram, although the 

 magnitudes do not always correspond. Sometimes a violent disturbance on the micro- 

 barogram is accompanied by a moderate or slight disturbance on the limnogram ; and 

 occasionally the disturbance on the limnogram is much greater than might at first sight 

 be expected from the microbaric disturbance. The mathematical theory (and indeed 

 common sense apart from recondite theory) indicates the reason for this. If an increase 

 of pressure operates on one half of a symmetric parabolic lake during half the period of 

 the uninodal seiche while the water in that half is falling, it will evidently work the 

 whole time towards increasing the amplitude of the seiche. Also, if there were to be 

 increase of pressure for half-periods alternately on the two sides of the uninode, always 

 tending to drive the water in the direction in which it was going, it is obvious that a 

 very small increase might end by producing a very large seiche. As a matter of fact, 

 we do find occasionally a considerable rise of seiche when the microbarogram is 

 comparatively smooth ; but in such cases a closer examination usually shows a faint 

 undulation with a period not very different from that of the seiche which is generated. 



On the other hand, if we suppose our increase of pressure to act on one half of the 

 parabolic lake during the whole period of the uninodal seiche, or if it is distributed 

 equally on both sides of the uninode, it is easy to see that the final result in altering 

 the range of the seiche will be nil, however long the increase of pressure may act. 



Absence of microbaric disturbance is accompanied by absence of seiche disturbance ; 

 that is to say, either there is no seiche at all, or an existing seiche continues unaltered. 

 Under these circumstances the limnograms from Earn are of great beauty. As an 

 example, we may mention a record taken by the converted Sarasin near the binode 

 from 23rd to 27th August. This shows a regular uninodal seiche with an average range 

 of 6 to 7 millimetres, which continued for over 89 hours. During all that time the 

 microbarogram shows only very slight disturbance — faint undulations, occasionally 

 periodic. The range of the seiche is not absolutely constant, but sometimes rises 

 and sometimes falls gradually, the minimum being, say, 4 '5 mm. and the maximum 

 8 mm. (corresponding to 77 mm. and 13 '6 mm. at St Fillans). There is nowhere any 

 sudden change of phase. 



Examination of the limnograms shows that seiches may be generated " suddenly," 

 i.e. attain their full range in one or two oscillations, or may be generated " gradually," 

 i.e. the full range may be attained only after a considerable number of oscillations. 



Among the causes that might generate seiches suddenly we may consider the 

 following ; — 



1. The sudden release of a static denivellation of the whole lake-surface, due to 

 the progression of the general system of the atmospheric isobars. 



