550 MULL. METEOROLOGY. BAROMETER. 



instruments were in the same room and the temperature 

 constant. There are certain causes already well known as 

 disturbing forces, of which some are capable of calcula- 

 tion and others are generally avoided in instruments of 

 even tolerable execution. These are chiefly, a diameter 

 of the tube too minute, varieties in the sp. gr. of the 

 mercury, and variations of temperature. But I need not 

 dwell on these and other circumstances already well 

 known ; since those which are not capable of being 

 accurately determined, are productive of but trifling errors, 

 and a ready mode of calculation is applicable to the 

 others. 



The error of chief moment, since it appears to be the 

 cause of the non-coincidence between proximate barome- 

 ters, and is also incapable of being estimated or corrected, 

 is one compounded of erroneous construction and varying 

 temperature. This has not been sufficiently noticed. It is 

 better known to the makers than to the purchasers of 

 barometers, that it is nearly impossible to make two act 

 together under all circumstances. It is not often that the 



o 



column is precisely of the same length in any number 

 when constructed, and to conceal this defect the scales 

 are adapted accordingly. These inequalities are diminished 

 by boiling, but it is scarcely possible even thus to expel 

 air and water from the tube ; besides which the imminent 

 hazard of breaking which it undergoes from the unequal 

 heating, prevents this from being almost ever effectually 

 done. In ordinary instruments it is not even attempted. 

 Even when the tubes have been sealed when hot, and 

 opened only at the moment of filling them with boiled 

 mercury, air will still insinuate itself. It is this air which 

 is the cause both of the unequal altitudes and the unequal 

 oscillations in question. The variations of temperature 

 above mentioned as affecting the movements of proximate 

 barometers, act by influencing this substance ; not by 

 altering the specific gravity of the mercurial column 



