On the TJicrmal Expansion of Marhlc. 6T 



tution of matter, the same particles, when irregularly congre- 

 gated, may follow a very different one. Such an analogy 

 might have given rise to conjecture, and might have incited to 

 experiment ; but it should surely never have been regarded as 

 justificative of an inference which militates against all experience, 

 and gives to white marble so unique a place among solid bodies. 



Tlie appeal to direct observation can alone set the matter at 

 rest. Induced by the high importance of the subject, we have 

 made this appeal, and now propose to give an account of the 

 results of our experiments. 



The difficulty of making any very accurate measurements of 

 the expansion of bodies by heat is best felt by those who have 

 essayed them. When the rate of expansion of one substance is 

 known, it is not a very difficult matter thence to determine the 

 expansibilities of others: the first determination is that which i$ 

 attended with the greatest difficulty. No substauce is free from 

 expansion by heat, so that we can have no permanent standard 

 for measuring the magnitude of that submitted to observation, 

 unless by keeping some slightly expansive material at a fixed 

 temperature. Heat, however, is communicated with such rapi- 

 dity through even the most slowly conducting media, that either 

 the standard or the variable body changes temperature before the 

 measurements can be effected. Berthoud's plan of lifting a heated 

 bar and placing it upon the plate of a pyrometer, is altogether 

 unfit for accurate purposes, nor is it easy to point out any me- 

 thod that may be entirely free from objections. The heating of 

 the substance, as well as the retaining of it for a considerable 

 time at a fixed temperature, is also attended with inconvenience ; 

 in fact, the sources of minute error are so involved, that no ob- 

 server can be certain of the accuracy of his Jesuits, and that it 

 becomes imperative on him to detail all the grounds on which he 

 has proceeded, and all the precautions which he has taken to 

 avoid error. 



As we have already hinted, there are two methods according" 

 to which we may proceed in determining the expansion of a sub- 

 stance : we may either compare its lengths, when differently 

 heated, with that of a body kept at a uniform temperature, or 

 otherwise with those of a body of known expansion subjected to 

 the same changes of temperature. The latter method is attended 



