1825.]" Barometrical Measurement of Heights » Xl\ 



Constant error. 'lin. 'Sin. 'Sin. •4 in. 'Sin. 



30in. .. 3ft. .. 7 .. 10 .. 13 .. 17 





^ ins .. 6 .. 13 ..20 .. 26 .. 33feet. 



25 ..4 .. 8 .. 12 .. 16 .. 20 

 20 ..5 .. 10 .. 15 .. 20 .. 25 



The corrections are additive or subtractive according as the 

 pressures are observed in excess or defect.* ' 



An error of 1° F. in the difference of the indications of the 

 attached thermometers will render the calculated altitude incor- 

 rect by a quantity varying with the temperature of the air from 

 2*2 to 2*7 feet. To obtain the correct mean temperature of the 

 mercurial columns without considerable loss of time is one of 

 the greatest difficulties the observer has to encounter. In the 

 barometers of Newman, the cistern is furnished with a thermo- 

 meter having its bulb (of a tapering form) immersed in the 

 mercury. 



An erroneous estimate of the mean temperature of the air 

 amounting to 1° will vary in value from one foot in 500 at 32° F.. 

 to one in 480 at 90° F. Errors will occur when the arrange- 

 ment of the strata of the atmosphere of different temperatures 

 is such that the half sura of the thermometers does not represent 

 their correct mean temperature ; or when one or both thermo- 

 meters have been placed with so little discretion as to give the 

 temperature of some small portion of air affected by local heat, 

 in lieu of the general temperature of the surrounding air. Errors 

 may also be introduced when the lixed points of the thermome- 

 ter are placed too high or too low ; when the calibre of the tube 

 is unequal, and not allowed for in the graduation of the scale, 

 or when the height of the mercurial column has been noted as 

 affected by parallax. 



An inaccuracy of 1° in observing the dew point cannot possi- 

 bly give rise to an error of more than one foot in 2000. Most 

 of the other sources of error in estimating the mean hygrometric 

 state of the air will be similar to those occurring in ascertaining 

 the mean temperature. Whatever the possible state of the 

 atmosphere as to humidity, provided we add to the mean of the 

 detached thermometers, half the correction for saturated air 

 under the pressure of 30 inches, the error in altitude will not 

 exceed -rroth, but more generally it must fall short of half that 

 quantity. 



Were we in possession of a sufficient number of barometrical 

 observations made at the foot and summit of isolated mountains, 

 carefully levelled, together with the true mean temperatures and 

 dew points of the intercepted strata of the atmosphere, we could 



'* This is only an approximation ; for the half sum of the pressures we should sub- 

 stitute tile mean-density pressures. 



