404 Messrs. W. Kellner and W. H. Deering. [Mar. 7, 



From tlie results of these experiments it may be stated that 



1. The initial decomposition of ethylene by heat is very rapid, and 



requires but a short flow through a heated containing vessel, 

 such primary decomposition, however, being but slowly com- 

 pleted, owing to secondary reactions, which tend to re-form 

 ethylene. 



2. Dilution has but little effect in retarding the decomposition of 



ethylene, unless it be very large. 



3. Increase in rate of flow diminishes the amount of decomposition 



when the heated area is small, but rapidly diminishes in 

 effect as the length of flow through a heated area increases. 



4. The decomposition of ethylene is chiefly caused by radiant heat, 



the effect of which is very great as compared with the decom- 

 position due to contact with heated surfaces. 



I desire to acknowledge the valuable aid given me by Mr. F. B. 

 Grundy in this investigation. 



III. *' On the Measurement of Pressures by the Crusher-Gauge."" 

 By W. KELLNER and W. H. DEERING. Communicated by 

 Dr. W. ANDERSON, F.R.S. Received January 21, 1895. 



The object of the experiments, the results of which are given in 

 this paper, was to compare the indication of pressure (produced by 

 the firing in a closed vessel of a gunpowder) by the crusher-gauge 

 with a simultaneous estimation of the same pressure by another and 

 simple method. This latter consisted in ascertaining the weight 

 which was equal to the maximum pressure of the gases of the fired 

 cordite (the smokeless gunpowder used in these experiments) acting 

 on a valve closing gas-tight an opening in the explosion- vessel of 

 area small enough to allow of the use of weights of manageable 

 amount, not so small a fraction of the unit of surface (1 sq. in.} 

 as to have to use a very large multiplier. Within the explosion- 

 vessel, on the top of which were the valve and weights, was placed 

 the crusher-gauge, the firing of the charge of cordite giving the 

 two indications of pressure. In the " weights " method r a series of 

 short steps was taken, weights in excess of that required being 

 placed on the valve, the cordite fired, and it being observed whether 

 the gases of explosion were blown out of, or kept in, the vessel. 

 The amount of compression of the copper-crusher was then, 

 measured, and the corresponding pressure ascertained from the 

 existing tables. In the next experiment the weights were lightened,, 

 and so on, until the gas blew out of the vessel, the mean of the 

 pressure at the blow-out point and of that in the last experiment in. 



