of the Movements of the Barometer. 19 



highest extreme which took place during a long and uninterrupted 

 frost. 



It is impossible to show more clearly than by these facts the 

 connexion between rain, wind from the S. or S.W., and a low state 

 of the barometer; I merely venture to suggest one mode of 

 accounting for this association of phenomena : — during a period 

 of atmospheric tranquillity, evaporation goes on from the surface 

 of the earth, and expands and forms a part of the atmosphere, to 

 the absolute weight of which it must necessarily add, and the 

 barometer accordingly takes its highest station ; at length, from 

 the collision of different winds induced by irregularities on the 

 earth's surface, this vapour is precipitated, on the principle of the 

 beautiful theory of Dr. Hutton*, viz. that, as the power of the air 

 to imbibe moisture increases in a higher ratio than the tempera- 

 ture, two portions of air of different temperatures, and saturated 

 or nearly so with moisture, cannot be mingled together without 

 its precipitation ; this precipitation is, in its turn, a fresh cause of 

 wind, of the fresh collision of different airs, and of the renewed 

 formation of clouds and rain, — and the contraction which thus 

 takes place in the atmosphere diminishes its elasticity, lowers the 

 barometer, and again becomes a cause of wind. 



According to this view of the subject, the barometer may be 

 regarded as an instrument measuring the actual weight of a column 

 of the atmosphere in serene weather, but in serene weather only. 

 On other occasions it measures the changes which take place in 

 the elasticity of the lower region of the atmosphere, the pressure 

 of the superincumbent part of which is cut off for a time until the 

 equilibrium be restored. From this view, too, we observe that it 

 is only in serene weather, and with a transparent and tranquil 

 atmosphere, that the barometer should be employed to measure 

 the heights of mountains. That the variations of the barometer 

 do, in fact, depend upon changes which occur in the lower regions 

 of the atmosphere, is proved by the observations of M. Laval, re- 

 corded in the Memoirs of the Royal Academy at Paris for 1709. 

 During ten days the barometer was observed to vary two lines and 



* Trans, of the R.S. of Edin. v. I. p. 41. 

 C2 



