102 Mr. Hcrapath on True Temperature, and the [Aug. 



"blish one hypothesis hy another, the very probability of which 

 remains to be sho n ; namely, the hypothesis of " specific heat." 

 Without, however, stopping now to discuss this subject, I will 

 endeavour to show that a very trifling error in MM. Lavoisier 

 and Laplace's experiments with the calorimeter, from which 

 M. Biot drew his conclusion, is sufficient to make these experi- 

 ments coincide with my theory. By a formula which I shall 

 hereafter demonstrate, if W be a given weight of water, and T 

 its true temperature, the weight ofice it will meltin the calorimeter 



is -7T7T . Therefore, 100 ounces of water at .212° 



Fahr. will melt 126*57 ounces ; and hence 100 ounces of mer- 



29 



cury at the same temperature ought to melt only -^— x 126*57 



= 3-67 ounces. Consequently 100 ounces at 118*6°* should 

 melt only 1*835 ounce ; and at 122° only a 27th part more, or 

 1*903 ounce. Thus in operating with 100 ounces of mercury, 



an error of only - of the whole ice melted, or indeed of half of 



1 34 



it - , or of — — , oz. would be sufficient to overturn the inference 



of M. Biot. And if Lavoisier and Laplace operated with a less 

 quantity of mercury, 50 ounces for instance, an error of 

 1 34 



54' or TToo oz * wou '°- De enough to reduce the law of temperature 

 from the simple to the subduplicate ratio. But if so trifling an 

 error in the quantity of melted ice can produce so enormous a 

 difference in the results, to what a degree of accuracy ought the 

 experiments to be made to establish any law of temperature 

 whatever ! 



Thus the experiments with the calorimeter are exposed 

 to difficulties and errors against which it is doubtful whether 

 even the abilities of Biot, Lavoisier, and Laplace, could 

 protect them. Mercury, on account of the smailness of its 

 numeratom, or heating power, is certainly one of the worst 

 fluids that could have been ~ chosen for determining the 

 accuracy of the thermometric scale with the calorimeter. 

 Water would be much preferable if the condensation of its 

 vapour could be avoided. This, in experiments with the 

 calorimeter, will always make in favour of the old theory, 

 and against the new. By a formula which I have investigated 

 from my theory, and an experiment related by Dr. Thomson, in 

 his System of Chemistry, I find that one ounce of steam at 212° 

 Fahr. condensed on 100 ounces of water at 201°, would be suffi- 

 cient to raise it to 212°. Consequently, if a quarter of an ounce 

 of steam, or a quantity equal to ] -400th part of the weight of 

 water, were introduced with the water, it would be enough to 



* By mistake, I assumed the true mean temperature to correspond with 118-6°, 

 instead of 1 18-4°' ; but the difference is so trifling that I think it useless to repeat the 

 calculations, even simple as they are. 



