THE LAWS OF CONDUCTION OF HEAT IN BARS. 77 



room in the neighbourhood of the long bar, and similarly exposed, but without 

 artificial heat." (Art. 23.) The arrangement is shown in Plate I. fig. 2. 



52. The gas furnace was usually lighted about 8 a.m., when the lead in the crucible 

 was gradually melted. The readings of the thermometers were not recorded until 

 about four hours had elapsed, and the experiment seldom lasted altogether less 

 than eight hours, generally ten or eleven hours. It was difficult to keep the flame 

 of the gas furnace steady, the " regulator" used for the purpose being apparently 

 of little use. The lead in the crucible, after being quite fluid, sometimes solidified 

 a little over the interior flange which grasps the bar. I at length found the best 

 way of regulating the temperature to be, to keep the eye very constantly on the 

 first thermometer in order, and whenever the slightest rise commenced, to dip a 

 little cold lead into the crucible, and either let it melt or chill the mass by its 

 contact, or the gas might be cautiously lowered. If the temperature was seen to 

 be falling, the gas had to be raised. With experience I learned to keep the tem- 

 perature of the three-inch hole within a range of 2° Cent, under favourable 

 circumstances, the temperature being nearly 200° Cent. In some experiments in 

 which solder was employed instead of lead, I used a thermometer whose bulb 

 dipped into the crucible, where it stood about 460° Fahr.* My able assistant, 

 Mr James Lindsay, learned to regulate this with great nicety. 



53. When the temperature had been for a long time quite steady at the three- 

 inch (or hottest) hole, the thermometers, disposed, as has been explained, along the 

 bar, were successively read, and the readings recorded. This was done from left 

 to right along the bar with all deliberation, without regard to any possible change 

 during the process in the temperature at the hot end of the bar, since any such 

 change is comparatively slowly propagated along the bar. In like manner, in 

 " stepping" with one thermometer from point to point of the bar (Art. 2), a 

 slight change in the temperature of the source is immaterial, since the " stepping" 

 is performed faster than the wave of heat can follow. 



54. A careful examination of the whole record of simultaneous readings was 

 made, and those corresponding to the most stationary conditions of temperature 

 were selected ; and these being corrected for scale errors and temperature of column 

 (Arts. 42, 43) were entered, after the free temperature indicated by the little bar 

 (Art. 51) had been deducted, in the following tables as the Corrected Excesses 

 of Statical Temperatures in centigrade degrees. 



55. Looking cursorily over Table I., we may observe, Firsts that each day's 

 observations are comparable only amongst one another, as no exact coinci- 

 dence of the temperature of the crucible, or source of heat on different days, was 

 attempted. Secondly, the bracketed observations are made with independent 

 thermometers. Thirdly, comparing the first series for the bar covered with paper 



* By an oversight in the first part of this paper (Art. 19, note), it was stated that in this 

 instance the thermometer was dipped in melted lead. 



VOL. XXIV. PART I. Y 



