EECENT EXPERIMENTS ON FOG-SIGNALS. 317 



To the circumstance that the broadsides of the slabs 

 faced the sea is probably to be ascribed the remarkable 

 fact observed on March 23, that in two directions, not 

 far removed from the line of fire, the gun-cotton 

 detonated in the open had a slight advantage over the 

 new gun. 



Theoretic considerations rendered it probable that 

 the shape and size of the exploding mass would affect 

 the constitution of the wave of sound. I did not think 

 large rectangular slabs the most favourable shape, and 

 accordingly proposed cutting a large slab into fragments 

 of different sizes, and pitting them against each other 

 The differences between the sounds were by no means 

 so great as the differences in the quantities of explosive 

 material might lead one to expect. The mean values 

 of eighteen series of observations made on board the 

 ' Gralatea,' at distances varying from 1! mile to 4-8 miles, 

 were as follows: 



Weights . . 4 oz. 6 oz. 9 oz. 12 oz. 

 Value of sound . 3-12 3-34 4-0 4-03 



These charges were cut from a slab of dry gun-cotton 

 about 1J inch thick: they were squares and rectangles 

 of the following dimensions: 4 oz., 2 inches by 2 

 inches; 6 oz., 2 inches by 3 inches; 9 oz., 3 inches by 

 3 inches; 12 oz., 2 inches by 6 inches. 



The numbers under the respective weights express 

 the recorded value of the sounds. They must be simply 

 taken as a ready means of expressing the approximate 

 relative intensity of the sounds as estimated by the ear. 

 When we find a 9-oz. charge marked 4, and a 1 2-oz. 

 charge marked 4*03, the two sounds may be regarded 

 as practically equal in intensity, thus proving that an 

 addition of 30 per cent, in the larger charges produces 

 no sensible difference in the sound. Were the sounds 

 estimated by some physical means, instead of by the ear, 



