its edge it is perfectly rigid, as such pressures tend to increase the interstices; when placed on its edge it sustains 2 cwt.
These experiments made on tangible matter fall short of what would be the dilatations if made with absolutely rigid uniform spheres. But after seeing the evidence that dilatation is the ruling property of granular matter, I venture to think that you will agree with me that without this clue the cause of gravitation would not have been discovered.
It now appears, however, as a complement to dilatation, in causing gravitation, we have the relative motion of the grains, which has already been shown to impart the elasticity necessary for the transmission of light, and this affords the complement for the complete explanation.
But, even so, the explanation is not what was expected; for gravitation is not the result of that dilatation which results from uniform parallel strains in the medium in normal piling, but results solely from those components of the dilatations caused by the space variation of the inward strains.
Thus, as long as the dilatation strains are parallel there is no attraction; but if there is curvature in the strains there will be efforts,
3—3
