264 PROCEEDINGS OP THE AMERICAN ACADEMY 



wooden pillars, about 20 cm. above the top of a long wooden table. 

 For the steady flow experiments two bars were placed in line with 

 their ends nearly touching, and both were heated at once by the 

 same Bunsen flame, which was protected from drafts of air by a tin 

 casing about 38 cm. tall and 13 cm. square, pierced on two sides to 

 admit the ends of the bars. The general temperature of the room 

 was observed by means of six ordinary Centigrade thermometers, 

 of which one was placed on the table near the heated bar, one about 

 1.25 cm. above the middle of the bar aud about 30 cm. from the 

 ceiling of the room, one near the north wall about 2.5 m. distant, 

 another about equally distant upon a south wall, another some 10 m. 

 away upon the western wall, and another about the same distance 

 toward the east. During the experiments on steady flow and on 

 rate of cooling, the thermometer lying on the table usually read 

 about 0°.5, and the one hanging above the bar about 2°, higher 

 than the mean of the six. The mean reading of these six ther- 

 mometers differed from the mean reading of the eight thermometers 

 used on the bars less than 0°.l, when all were placed together in 

 water near 20°. The mean reading of the six room, thermometers 

 during the actual experiments upon the bars usually varied only a 

 few tenths of a degree in an hour. 



Not wishing to devote a great deal of time to the work in hand, 

 I was content with a rather rough approximation to a state of 

 steady flow of heat. The method followed was to take several 

 rapid observations of all the bar thermometers at five-minute inter- 

 vals, after they had become reasonably constant. Upon the basis 

 of such observations calculation was easily made of the rate at 

 which heat was accumulating in the bar at any given instant. In 

 some cases this accumulation made a difference of several per cent 

 in the amount of heat entering into the calculation of the con- 

 ductivity. As an example of the method of observation, the fol- 

 lowing series made with the nickel bar will serve, the columns 

 headed 1, 2, 3, etc., being readings of the thermometers no. 1, 

 no. 2, no. 3, etc. : — 



