On the Cold Caves of the Monte Testaccio at Rome. 209 



on the dryness of the air when at the external temperature, 

 and not when it is reduced to the mean temperature of the 

 place in the interior of the hill, because there it will always be 

 saturated. What the mean temperature is, has, by the errors 

 both of the writer and the printer of the article before us, been 

 most erroneously stated. " A quantity of loose materials," it 

 is said, '' whose mean temperature cannot be less than 46°." 

 And in a note : " The mean temperature of Rome is, accord- 

 ing to Humboldt, 45°.9 in winter, and 55^.2 in summer." 

 Puzzled by such a mass of mistakes, I referred to the original 

 passage in the Edinburgh Enc^/clopcedia, Article Physical 

 Geography, where I found the mean temperature still called 

 46°; but Humboldt's summer average, 75°.2 instead of 55°2, 

 giving the real mean temperature 60^.6, which I know from 

 experiment to be very exact ; this, therefore, is the real mean 

 temperature of the materials. In Saussure's most extreme ex- 

 ample, the case is, that air of a certain degree of dryness, and 

 at the temperature of 78", moistened to humidity with moisture 

 of the temperature of 60°, produced a cold air of 44°. This 

 is a case too complex to be solved by the common hygrometric 

 formulae ; I therefore undertook the following experiments, not 

 with the hope of a complete solution of these delicate inquiries, 

 but to obtain such general facts, as to give to the case in ques- 

 tion an approximate estimation of the truth. The experiments 

 themselves are in some measure similar to those of Saussure, 

 (though I had not them in my eye at the time,) but are con- 

 ducted upon more distinct principles, and by more precise data, 

 I. The room in which my first experiment was made had, 

 at the beginning of my operations, a temperature of 59J, but 

 at the end had risen nearly to 62°, and in both cases the tem- 

 perature of a thermometer with a moistened bulb was 55^. 

 The air was both hotter and drier, therefore, at the conclusion ; 

 but we shall take the temperature at 61 as an average. We then 



have for the dryness of the apartment, * y ' =ft — T^ZZTrB' 



in this case, supposing the barometer at 30 inches, as it was 



nearly ; /61 — ^|^q_3 > we obtain .0021893 grains in a 



* Ed, JEncy, Art. Hycjkombtry, 585. 



