140 PRINCIPAL J. D. FORBES ON AN EXPERIMENTAL INQUIRY 



20. Thermometers were inserted in cylindrical holes, driUed in the upper side 

 of the bar. The holes were 0*28 inch diameter. Contact with the thermometer 

 was secured by mercury poured into the colder holes, and an amalgam or fusible 

 metal in a semifluid condition in the hotter ones* There were usually about 

 ten thermometers inserted in the bar, at distances varying from three inches to 

 eight feet from the zero point at the crucible. Those nearest to the source of 

 heat were defended from its action by two or three interposed polished metal 

 plates, which were found to act efficiently. 



21. When after several hours of exposure to steady heat, the bar attained 

 a normal temperature at its various points, the instant was to be seized, when 

 the casual fluctuations became inappreciable, not only on the thermometers 

 nearest to the source, but also in those at a distance. For though the source of 

 heat may, for a time, appear quite steady, the wave of temperature arising from 

 some antecedent irregularity may still be travelling along some remoter portion 

 of the bar. Experience, and the patient entry of a number of successive observa- 

 tions of all the thermometers, can alone secure the desired precision. These 

 experiments must never be made in a hurry. 



22. A modification of this mode of observing which occurred to me in the 

 course of these experiments, possesses important advantages, and may be used as 

 a check. I call it the method by stepping. One and the same thermometer was 

 transferred to the successive holes in the bar, beginning with the hottest and 

 going on to the coldest, and the temperatures were read in each case. In this mode 

 of operating, each hole was in the first instance provided with its mercury or 

 amalgam, and with its proper thermometer as before ; and the thermometer was 

 only withdrawn, and the stepjnng thermometer introduced, when the temperature 

 indicated by the latter (after being held in the hand to cool) reached as nearly 

 as possible the degree known to belong to the hole in which it was to be immersed. 

 It of course took the exact temperature almost instantly, without either heating 

 or cooling the mercury in the hole ; and so on to the end. It is an important 

 recommendation of this method, that a single reliable thermometer may be made 

 to perform the whole work, the others being merely used as rough indicators. 

 Nor is there any danger of error arising from slight changes in the temperature 

 of the Source of heat subsequent to the commencement of the readings; for expe- 

 rience shows, that the wave of disturbance of temperature advances much slower 

 along the bar than the stepping process can be perfectly well gone through. On 

 the whole, this is an immense facility for the extension of such experiments ; as 



* It was found tliat when mercury was used for these last, the surface became hotter by con- 

 vection than the central part of the hole, contrary to the law of the distribution of heat in a solid 

 bar, and consequently an undue (though perhaps hardly sensible) amount of heat was thereby dissi- 

 pated. I may add, that I ascertained by actual experiment, that the boring of several additional 

 holes between the extreme holes of a bar did not sensibly disturb the conduction of heat when the 

 intermediate holes had thermometers surrounded by mercury inserted in them. 



