﻿202 DE. H. C. SOEBY 0^ THE APPLICATION OF [May I908, 



average volume of the interspaces was 49*4 per cent., and when 

 well shaken 43*5 per cent. In the case of Calais sand, the 

 grains of which had a mean diameter of about '01 inch, the inter- 

 spaces as above were respectively 45-8 and 38'7 per cent. The 

 means of all my experiments with both kinds of sand were, for 

 not shaken 47*0 per cent., and for well shaken 40*0, which agree 

 remarkably well with 47*2 and 40*0 found in the case o£ shot. 

 In no case with sand of fairly -uniform grains did shaking reduce 

 the volume of the interspaces to anything like the theoretical 

 minimum, namely 25" 95 per cent. In the case of finely-powdered 

 sand, having grains on an average -^-^ inch in diameter, when 

 filled in quickly the interspaces were 47 per cent., and when well 

 shaken 34 : this much smaller amount being probably due to the 

 greater relative range in the size of the grains. 



Sand with Grains or Extremely Variable Size. 



Theoretically, the admixture of fine sand with coarse ought 

 greatly to reduce the amount of the interspaces. For example, 

 if the spaces between the coarse grains were filled with fine sand, 

 the interspaces should be 40 per cent, of 40 per cent. : that 

 is, only 16 per cent. ; but in no case have I obtained so low a 

 result by experiment. Thus, on mixing the coarse Millstone-Grit 

 sand with an equal volume of Calais sand, I found the interspaces, 

 when filled in quickly, to amount to 38 per cent., and when well 

 shaken, 32 per cent. When equal quantities of these two sands 

 and of the powdered sand were mixed, the interspaces when shaken 

 were further reduced to 28*9 per cent. This, however, is far 

 greater than it would be theoretically, if arranged with intelligent 

 design ; it might probably be the result, if the sands were exposed 

 to pressure and vibration for an indefinitely-long period. The facts, 

 nevertheless, clearly show that, as might have been expected, the 

 proportion of interspaces in such sand as would be deposited from 

 deep water, and not sorted by bottom-currents, would be much 

 less than in well-washed sand. Hence, other things being equal, 

 the amount of interspaces gives some indication of the relative 

 depth of the water, which agrees well with many general facts. 

 Calais sand mixed with half its weight of pipe-clay, when gently 

 compressed contained 36*7 per cent, of interspaces, which was re- 

 duced to about 18 per cent, by a pressure roughly estimated at 

 about equal to 60,000 feet of superimposed rock. In the case of 

 clay alone it was only 14-6 per cent. We thus see that the effect 

 of clay is great. 



Small Flakes of Mica. 



When fine particles of mica were put into a brass tube, the 

 amount of interspaces was about 74 per cent., and by moderate 

 pressure this was reduced to 55 per cent. When the mica was well 

 shaken before pressure the amount was about 57 per cent., and when 



