172 Mr. Nixon on the Theory of the [Sept. 



deviate from the perpendicular by the sa/we angular quantity, the 

 calculations may be correctly made on the supposition of the 

 mercurial columns having been truly vertical. It is merely to 

 imagine the mercury specifically too light in the ratio of the 

 secant of the angle of inclination to radius. Should the degree 

 of obliquity be considerable, the friction will be materially 

 increased, and the mercury will not settle so correctly as in a 

 vertical tube. 



To verify the adjustment of zero to the level of the cistern, 

 note the pressures (when very low) , with the instrument first in 

 a vertical, and afterwards in an inclined position. Then if the 

 pressure when inclined does not exceed the other in the ratio of 

 radius to secant of the angle of inclination, the zero dips within 

 the mercury, and the heights will be observed in excess. Or 

 we may find the correct pressure, and consequently the error of 

 adjustment, by measuring the height of the longer column of a 

 wheel barometer (freed of its pulleys) above a horizontal line 

 drawn from the summit of the shorter column. The mercury 

 in the two instruments must, however, be of the same specific 

 gravity.* 



Errors. — When the pressure is incorrectly measured at one of 

 the stations, which may arise from an inequality in the divisions 

 of the scale ; from parallax in reading off the vernier ; from 

 want of horizontal parallelism in the movement of the index ; 

 from the barometers differing in inclination ; from the varying 

 adhesion of the mercury to the sides of the tubes ; or from the 

 inexpertness of the observer in adjusting the notch, &c. of the 

 vernier in a tangent to the convex summit of the mercury, — the 

 value of the error when equal to "001 in. will be generally rather 

 more than one foot of altitude ; or more correctly to 



0-785 foot, the teaiperature being 0° F. | p^^^^^^^ 3j .^^^ 



1 ,747 • Q9 ^ 



i Aqo Q f Pressure 15 inches. 



When the zero of the scale is improperly placed, or the 

 allowance for capillarity is incorrect, so that the pressures are 

 always observed in excess or defect by some constant quantity, 

 the calculated altitude will be to the true elevation inverseh/ as 

 the half sum of the observed pressures is to that half sum aug- 

 mented or diminished as the pressures are in defect or excess by 

 the constant error. The subjoined scale exhibits the value in 

 feet of the error at different pressures, the altitude being 1000 

 feet. 



♦ An Englefield barometer in my possession (remarkable for the constancy with 

 which its mercurial columns would settle to the same height on being disturbed, and 

 there is a most material difterence in the instruments in this respect) had its zero placed 

 too high by nearly half an inch. 



