1826.] Measurement of Heights hi/ One Barometer. ^ 



density of the air approaching more nearly to that of the 

 mercury. 



In the second place, as the pressures are very irregular about 

 sunrise, and sunset, the observations should not commence 

 before the sun has risen an hour or two, and must termi- 

 nate when practicable at nearly an equal distance of time 

 previous to his setting. 



As to temperature, cold weather would appear to be most 

 ehgible, the variable correction foV humidity being unimportant, 

 and the density of the air differing less from that of mercury 

 than at more elevated temperatures. On the other hand, as the 

 barometrical experiments whence the principal coefficients of 

 the formula have been derived were made at temperatures some 

 httle above a mean, it will be most prudent to avail ourselves of 

 calm cloudy days, with the thermometer steadily between 50° 

 and 60° F. 



The observer will consequently avoid as much as possible 

 low, fluctuating, or rapidly varying pressures ; high and espe- 

 cially changeable winds ; also extremely hot weather, particu- 

 larly if extraordinarily dry or damp. Cold winds, with a hot 

 sun, are decidedly objectionable ; the difficulty of ascertaining 

 the true temperature of the mercury, or even that of the air, 

 being almost insuperable. Extremes of temperature occurring 

 in the course of the observations are very untoward. 



At the level of the sea the heights of the barometer will 

 remain unchanged under every vicissitude of temperature of the 

 atmosphere. Stationed on an eminence the mercury will rise 

 and fall with every accession and diminution of heat affecting 

 the subjacent stratum of air. Hence the impossibility of ascer- 

 taining in elevated situations the correct rate of the variation of 

 the pressure, and the absolute necessity of placing the inferior 

 barometer, whatever may be the differences of level required to 

 be measured, in the very lowest situation in the vicinity, ia 

 order to obtain this indispensable datum with the requisite 

 accuracy. In making the consequent calculations, it will also 

 be necessary t» compute, in the first instance, the altitudes of 

 the different superior stations above this inferior or reference 

 station, whence their relative heights may be subsequently 

 obtained by the mere subtraction of the one from the other. 



When the reference station cannot be conveniently revisited 

 subsequent to the ascent, a second observation should follow the 

 first after the lapse of at least half the time required to gain the 

 summit. Again, on reaching the base of the mountain in some 

 other direction, or on arriving at some eligible low situation, the 

 rate of the variation of the pressure may be deduced from two 

 or more observations made at intervals of an hour or two. By 

 taking a proper mean of the rate at the two places, the pressure 

 at the reference station at any given time may be inferred with 



i>2 



