On the Height* of some ef the Cotswold J Jills. 39 



quested, and has since most obligingly supplied the approximate 

 heights above the mean level of the sea of sixteen remarkable 

 points in our vicinity which I shall presently read to you, together 

 with other data which I have myself obtained by the aid of the 

 aneroid barometer lately invented in France, and much vaunted 

 as applicable to the measurements of heights. 1 then procured 

 one of these instruments from Dent, with his pamphlet upon it, 

 and will now give the results of its comparison with the mea- 

 surements received from Capt. Yolland. 



It may be as well however first to make a few remarks on this 

 new instrument, with a view to show how far it may be appli- 

 cable in its present state to the purpose of measuring altitudes. 

 It is probably known to most of you, that in carrying a mercu- 

 rial barometer to the top of a high mountain, the mercury sinks 

 from two causes, the one purely barometric, the other therino- 

 metric. Whilst for every 850 feet of perpendicular ascent the 

 weight of the air decreases so as to show a fall, in its counterpoise 

 the quicksilver, of about an inch — for every 300 feet of ascent there 

 is also a decrease in the temperature of 1° Fahrenheit, occasioning 

 a proportional contraction in the quicksilver in the tube, making 

 it stand so much lower than it ought to do were its descent due 

 to the diminished pressure of the air alone. To calculate there- 

 fore correctly the height indicated by the mercurial barometer, 

 allowance is always made for decreasing temperature, and tables 

 have been compiled for this purpose from the known rate at 

 which mercury contracts by cold. 



The same double effect is doubtless produced in the aneroid 

 barometer, which Mr. Dent says is compensated by means of gas 

 in the "vacuum-vase" of the instrument. This however is, I 

 believe, a mistake*. 



In its present form, then, I conclude that a correction for 

 temperature is needed for the exact measurement of heights. 

 There are also two palpable defects, one of which is that the band 

 or index is frequently so far from the face of the dial, that its 

 parallax leads to error in reading off the scale, which may easily 

 amount to 20 feet in height. This however may be somewhat 

 corrected by bending the hand so as to make it nearly touch the 

 face of the dial. The other fault is that the inch is subdivided 

 into only forty parts, one of which corresponds to 22 feet in 

 height. It would be better to have it graduated to hundredth* 

 — so that the actual reading off should tally at once with the 

 barometric tables now in use — or if the size of the dial will not 

 admit of this, to subdivide the inch into fifty instead of forty parts, 



* I have since ascertained it to be one. M. Villi himself informed me in 

 November last, that although he at one time made some experiments on 

 the use of gas in the " vacuum vase M — he has now rejected it altogether. 



f2 



