376 ACCOUNT OF THE ARCTIC llEGIONS. 



7. Before very heavy storms, wlien the barometer 

 falls uncommouly low, the mercury seems to get 

 below its natural level, and often rises two or three 

 tenths of an inch as soon as the predicted storm 

 commences ; hence this first rise of the mercury is 

 no indication whatever of an abatement of the 

 wind. 



8. On account of the different states of the ba- 

 rometer in west and east winds, the usual level of 

 the mercury with a moderate wind at west, not be- 

 ing much higher than with a gale at east, a change of 

 wind from one of these quarters to the opposite, 

 may be accompanied with the greatest alteration in 

 the strength of the wind, without producing any 

 effect on the barometer. Thus a storm at east 

 veering to west, may subside, or a gentle breeze at 

 west veering to east, may increase to a storm, with- 

 out any alteration taking place in the mercury of 

 the barometer. But in these cases, it has already 

 been observed, that the change which takes place in 

 the temperature, often compensates for the imper- 

 fect action of the barometer. 



