Theory of Explosions. 169 



was fastened horizontally to a wooden support, and a brass rod 

 moved piston-wise through the holes in the ends ; the rod was 

 about six inches long. Inside the tube was coiled a spiral 

 spring of hard-brass wire ; one end of the spring rested 

 against the inner surface of one of the caps, and the other 

 was pinned to the brass rod. The result was, that the rod 

 was always held by the spring in one position. One end of 

 the rod was furnished with a brass disk about three inches in 

 diameter, the other end abutted on the bob of a lever pendu- 

 lum so suspended as always to press slightly on it. The 

 pendulum was supported by a small shaft passing through it 

 and through holes in two brass brackets. The distance from 

 the bob (which weighed about six ounces) to the fulcrum was 

 one inch, and the pendulum was extended upwards about 20 

 inches by a pointer of sheet zinc; the end of this pointer 

 moved over a paper scale. When properly supported in the 

 tank the gauges occupied much the same position as the ones 

 previously employed. The flat disks of the gauges were 

 placed .so as to face towards the centre of explosion. These 

 gauges seemed to leave little to be desired either as to con- 

 venience of reading or otherwise, and, both being exactly alike, 

 gave tolerably consistent readings. 



By firing some dozen charges, arranged as symmetrically 

 as possible, I managed to find out that the indications of the 

 gauges were proportional, in about the ratio of 1*3 to 1. I 

 did not require to be very accurate, because all I was looking 

 for was to find whether the streams of debris formerly observed 

 through the window were also streams of explosive effect. 



To decide this I made the explosion cartridges purposely 

 unsymmetrical, either by having the glass too thick on one 

 side, or by turning up the ends of the covered wires so that 

 they entered the bulb horizontally and facing one of the 

 gauges. The effects now became more puzzling, but on the 

 whole there can be no question that the gauge towards which 

 the bulb was turned suffered most. In fact the direction taken 

 by the streams- of explosive energy appeared to coincide with 

 the directions of projection of debris, and with the direction 

 foretold from the initial conditions. 



The experiments were repeated at various distances and in 

 various manners with more or less compressed charges, and 

 with variations in the position of the firing-point. The pen- 

 dulum readings were on the whole certainly proportional to 

 the direction of explosion as foretold from the initial condi- 

 tions. Of course, in some few cases, there were unexpected 

 actions on the gauges ; but this was hardly avoidable, since I 

 had seen by the previous experiments how small a 'change m 



