OF ARTS AND SCIENCES. 265 



From these observations the following estimate may be made : — 



Time 1. 2. 3- 4. 5. 6. 7. 8. 



This was an unusually bad case. 



Each bar was tested twice in a condition of approximate steady 

 flow, that end which was heated in one test being the cool end in 

 the other test. This trial of both ends in turn was probably sug- 

 gested by the fact that the ends of the nickel bar were somewhat 

 different in appearance, owing to that too rapid cooling of the 

 molten metal which has already been mentioned. 



For determining the rate of emission of heat each bar was twice 

 heated and cooled as a whole. As only eight thermometers were 

 available for this purpose, the middle hole was left unoccupied in 

 all the rate-of-cooling experiments. The order of experiments was 

 as follows : — 



Cast Gun-Iron. Southern Cast-Iron Nickel. 



May 30, rate of cooling. May 31, rate of cooling. June 10, rate of cooling. 



June 2, steady flow. June 2, steady flow. " 11, steady flow. 



" 7, rate of cooling. " 4, rate of cooling. " 11, " " 



" 11, steady flow. " 11, steady flow. " 12, rate of cooling. 



The bars were generally, if not in all cases, rubbed somewhat 

 to brighten their surfaces just before being used. In experiments 

 on the rate of cooling the whole length of the bar was first heated 

 to about 115 or 120°, and as it cooled observations of all the ther- 

 mometers upon it were made at regular intervals, — two-minute 

 intervals at first and longer ones afterward, — until the bar was 

 perhaps not more than 15° (in the earlier cases about 20° and in 

 the later about 10°) above the general temperature of the room. 

 The accordance of the two sets of cooling observations made with 

 each bar was, on the whole, very satisfactory. From each set a 

 curve was plotted. The following values taken from these curves 

 will show how close the agreement was : — 



