342 BABCOCK. 



twenty minutes, and when the container was empty, for fifteen min- 

 utes. At the end of this time the heater was cut off. The tempera- 

 ture of the mercury continued to rise for about twelve seconds and 

 then started to drop. This "coast" of the mercury temperature is 

 due, no doubt, to the lack of equalization of temperature between the 

 thermometer on one side of the calorimeter, and the heater on the 

 other side. The heater is the hottest part of the calorimetric system 

 while the current is on, and some time is required to allow the heat 

 to flow arovmd to the thermometer after the current is shut off. In 

 other words, the temperatures recorded on the thermometer during 

 the heating period are not exactly mercury temperatures, but are 

 slightly below the true equalized mercury temperature. However, 

 the amount of this lag is so small that no correction was made for it. 



The drop of temperature occurring after this "coast" takes place, 

 is, however, far from negligible, and is very noticeable. This drop is 

 explained by the fact that the container and its contents are lagging 

 behind the mercury during the heating period, so that after the heater 

 is cut off, the mercury is losing heat both to the surroundings and to 

 the container, which accounts for the rapid drop in its temperature. 

 As the temperature differences between container and mercury get 

 less and less, the curve drops less steeply, and after a time, a point is 

 reached where the curve becomes sensibly a straight line, or at most a 

 line with very slight curvature. Where this occurs, it is assumed that 

 the calorimeter has equalized its temperature differences and that the 

 curve of temperature against time from that point on is due entirely 

 to stirring and heat transfer to the surroundings, as during the initial 

 period. This point of equalization may be located fairly accurately 

 by plotting the curvature, {(Fd/df), as a function of the time. After 

 the point of equalization is passed, the curvature is practically zero, 

 while before this it is changing very rapidly. While the slope of the 

 curvature function has no actual discontinuity, a "corner" may be 

 determined with some degree of certainty. The final line may slope 

 either up or down depending on the difference of temperature between 

 bath and calorimeter, and no attempt was made to have this slope 

 the same in each observation. This final line became sensibly straight 

 after about ten minutes in the case of the ammonia measurements, and 

 after about fifteen minutes in the case of the calibration measure- 

 ments. This straight line is projected back to the time ordinate corre- 

 sponding to the instant at which the heater was cut off, and the value of the 

 temperature so obtained is taken to be the equalized temperature of the 

 calorimeter at the instant the switch was cut off. 



