] 62 M. Langberg on the Determination of the Temperature 



however, as far as they have been made public, appear to me 

 to prove exactly the converse of what they are intended to do, 

 as the temperatures observed decrease in a much quicker ratio 

 than the geometrical progression would require, and the dif- 

 ference between the calculated and observed values is too con- 

 stant to admit of the supposition that it is solely due to errors 

 of observation. One source of the difference may probably 

 arise from the Newtonian law of cooling having been made 

 the basis upon which to found the mathematical deduction of 

 the law in question. According to Newton's law, a heated 

 body cools with a rapidity proportional to the degree of tem- 

 perature to which it has been raised above that of the air sur- 

 rounding it, and applies with correctness only to very slight 

 differences of temperature. [In the experiments alluded to 

 above, this difference amounts to 60° or 70° C] Again, it is 

 presumed that the power of conduction remains unchanged at 

 different temperatures, which is certainly not probable ; and 

 the theory further requires that the heated rod should be in- 

 finitely thin, or at least so thin that its temperature in every 

 part of a normal section should be exactly the same. 



Desprelz used in his experiments prismatical rods, the square 

 section of which was 21 millimetres in breadth ; holes were 

 bored in these at 10 millimetres distance from each other, 6 

 millimetres in diameter, and 14 millimetres deep. When the rod 

 had been brought into a horizontal position, these holes were 

 filled with mercury, and in each the bulb of a thermometer 

 was placed, the temperature of which, when it had become 

 stationary, was considered as that of the section of the rod 

 passing through the middle of the hole. As the breadth of 

 the holes amounted to nearly one-third of the whole breadth 

 of the rod, there is reason to fear that these large and frequent 

 interruptions in the continuity of the rod might cause a con- 

 siderable obstruction to the progress and distribution of the 

 heat. The results show that the method employed in the ex- 

 periments fulfilled very imperfectly the conditions required by 

 the theory, and it remains therefore still uncertain whether 

 the variations from the theoretical law which were observed are 

 to be attributed to an inaccurate method of observation, or to 

 an error in theory. 



The importance of the law, forming as it does the basis of 

 the mathematical theory of the phsenomena of heat, as well 

 as from its application for determining the conducting power 

 of solid bodies, appeared to me sufficiently great, and induced 

 me to seek a mode of observation not subject to the objections 

 which I have raised above. The first requisite was therefore 

 a mode of ascertaining correctly such slight differences be- 



