1904.] The Decomposition of Ammonia ~by Heat. 113 



minus the total pressure at the instant of observation. The experi- 

 mental data furnish values of P and 2p , and 2p Q - P has been tabulated 

 giving the pressure of ammonia at the time of observation. From 

 these numbers corresponding values of AP/A have been calculated by 

 taking the mean of the two consecutive differences and dividing by the 

 time interval ; but AP/A = dP/dt approximately within the limits of 

 experimental error, and dP/dt = dpi/dt, and therefore we have the rate 

 of change of the pressure of the ammonia at definite pressures, they 

 have been plotted in two series of curves. 



Results. The rate of decomposition was found to be much influenced 

 by the state of the globe ; it invariably increased after the globe had 

 been used once or twice, owing probably to the action of ammonia or 

 hydrogen upon the porcelain ; consequently we were not able to trace the 

 effect of temperature on the rate of decomposition. In the first experi- 

 ments the projecting stem of the globe was allowed to become too hot, 

 with the result that a trace of the alloy at the joint C found its way 

 into the globe, and by its catalytic action increased enormously the 

 rate of decomposition of the ammonia. The alloy contained a small 

 proportion of mercury, and a film of this metal was found on the 

 interior of the globe when it accidentally broke. In the later experi- 

 ments great precautions were taken to maintain the stem of the globe 

 at a low temperature. The experiments are here tabulated to serve as 

 a key to the curves. 



Greatest 



Mean variation of 



temperature. temperature. Remarks. 



791 0'2 Traces of mercury present. 



750 0-6 



841 0'2 New clean globe, pressure fell 



only 20 mm. in 2 hours. 



894 0'6 Globe as from last experiment. 



833 1-1 



1111 0-35 Globe freshly cleaned. 



902 0-8 



1061 0-6 Globe as from last experiment. 



815 0-4 



797 1-0 Iron wire in globe. 



1025 0-4 Pt. black in globe. 



The measurements made at 1111 and 1025 are given in full ; they 

 may serve as samples. 



l 2 



