On the Seiches of the Lake of Geneva. 299 



diatcly rises half a line or even a line, without the existence of 

 any general cause to produce this rise ; and according to his cal- 

 culation (Essai IV. chap. iii. p. 476), he finds that the simple 

 refrigeration of three degrees, in the whole extent of a column 

 of air, suffices to explain a variation of 0.85 of a line in the ba- 

 rometer. 



But independent of these greater local variations, which occur 

 but seldom, there must exist lesser ones every moment, so to 

 speak, in the various regions of the atmosphere. And, in truth, 

 it is but seldom that we can suppose that the different columns 

 of which it is composed can have exactly the same temperature; 

 and whenever there is this variation, it is likewise impossible to 

 suppose that the liquid upon which they rest, shall be accurate- 

 ly on a level. 



Observation comes in here in support of reasoning ; for Mr 

 Sencbier, to whom we are indebted for so many valuable works, 

 and who has bestowed a great deal of attention upon every thing 

 which affects the variations of the thermometer, has often as- 

 sured me, that when he observed this instrument, at ten minute 

 intervals, it seldom happened that he found the mercury at ex- 

 actly the same height, and it almost continually was varying tlie 

 sixteenth of a line and more. He had the kindness to observe 

 this instrument at the time that I was marking the variations on 

 the lake. I may also mention, that whensoever, in a variable 

 atmosphere, I attentively observed a very admirable barometer 

 which Mr Jurine had the kindness to lend me, I have scarcely 

 ever failed to discover variations as considerable as those speci- 

 fied above. 



With the intention of carrying conviction on this point as far 

 as possible, I frequently transported this barometer to the shores 

 of the lake, to certify that its variations corresponded with 

 those on the surface of the water ; but a little reflection soon 

 convinced me that this experiment was not at all calculated to 

 supply the result I wished to obtain. In truth, when the water 

 was rising and falling at one spot, it was impossible to ascertain 

 whether the change was effected in virtue of the variation of that 

 column of the atmosphere which was perpendicular to it, or of 

 some of the neighbouring strata. Besides, there was nothing to 

 prevent the column which supported the barometer undergoing 

 a change, which change would not manifest itself on the surface 



