Daniell on a New Pi/rvmeter. 313 



of this end, and in the partition, were round holes which 

 just admitted the stem of the pyrometer. The instrument 

 was passed through these holes up to the shoulder, 

 the small division of the case was then tightly stuffed with 

 tow and lute, and the larger division was filled with mercury. 

 The whole length of the platinum bar was thus immersed in 

 that metal. The apparatus so arranged, with the needle 

 pointing to 0°, and the temperature of the air being about 60°, 

 was placed over two Argand lamps, or a small furnace, and 

 gradually heated. The index of the pyrometer moved slowly and 

 steadily forward from 0, and when it had reached 85° the mercury 

 boiled rapidly. It was kept in a state of ebullition for half an 

 hour, and the index remained stationary at that point. The 

 strong concussions, indeed, of the boiling metal caused the 

 needle to vibrate, but its motion was confined between the 

 degrees of 83 and 85. Now, if we assume the boiling point of 

 mercury, under the pressure of the atmosphere, to be 656° of 

 Fahrenheit's scale, as we are justified in doing from the best 

 authorities, and deduct 56° for the temperature of the air when 

 the pyrometer stood at 0°. We have 85° degrees of the py- 

 rometer equivalent to 600° of the thermometer, making each 

 degree of the former equal to about 7.0 of the latter. 



Having in this manner calculated the value of the degrees, I 

 proceeded to ascertain the equability of the expansion through- 

 out the thermometric scale. The same apparatus was made 

 use of with the addition of a thermometer plunged into the 

 mercury. The heat was applied very carefully and gradually, 

 and the progress of the two instruments was compared at every 

 50° of the thermometer, both in heating and cooling. The re- 

 sults of the experiment are contained in the following table. 

 The first column contains intervals of 50° of Fahrenheit's ther- 

 mometer ; the second shews what the corresponding degrees 

 of the pyrometer ought to have been, by calculation from 

 the former experiment; the third exhibits the actual ascent 

 of the index while heating ; and the fourth its descent when 

 cooling. 



