Colombia hetioeen the years 1830 and 1830. 23 



Workin 



thermometer a barometrical scale. The product of 0°.974 of 

 Reaumur by 19 is 18.506, or, in round numbers IS. 5, i. e. 18°. 5 

 of Reaumur corresponds to 19 inches of the barometer. Then 

 measuring 18.5 from the summit, or 80° of Reaumur's scale, he 

 transferred it to the opposite side of the thermometer, dividing it 

 into 19 equal parts, or inches of the barometer, subdividing 

 these by a nonius into 24 each = half a line of the barometer. 

 In this manner the elevation of the thermometer by boiling 

 water indicates the corresponding elevation of the barometer 

 under the same atmospheric pressure. Caldas observes that 

 Humboldt, to whom he had communicated these ideas, when 

 they met in Popayan, objected the variability of the heat of 

 boiling water under the same atmospherical pressure ; to which 

 he replies : '•' Long practice has taught me its invariability in 

 this respect, using the requisite precautions in making the ex- 

 periment : otherwise, how could there be equal thermometers ? 

 Is not the invariability of the heat of boiling water under the 

 pressure of twenty-eight inches, the foundation of the superior 

 term of all thermometrical scales ? It is true tliat boiling water 

 does not immediately acquire its extreme heat, but pushing the 



Tnaxtm/umi 



p. 24. 



Caldas did not consider an invariable exponent possible, on ac- 

 count of the variability of atmospheric pressure. The want, 

 however, of a barometer induced me to make some experiments 

 to this effect, by way of rendering this method of measuring el- 

 evations still more simple, and of more general use. Is the va- 

 riability of atmospheric pressure such as to make any important 

 difference in these calculations? Does not water boil constantly 

 at 212° at the level of the sea? At Quito I found the same re- 

 sult as Caldas had several years before ; and several times the 

 same result in this and other parts of the Andes. The difference 

 then, is scarcely perceptible in the thermometer, and consequently 

 unimportant in the results of a calculation founded on the heat of 

 boiling water. The thermometer besides, immersed in boiling 

 water, is less liable to a variety of atmospheric influences to 

 which the mercury of the barometer is necessarily subject- 

 Hence the great differences in different barometrical measure- 

 ments of the same elevations, and the differences observed be- 

 twixt different thermometers exposed to the air in the same place, 



