It is true that the extent of the filtration is in proportion as the soil is more or less penetrable; but if we examine the standing pools with sandy bottoms, we shall perceive the water confined in the small compass it had hollowed itself, and the moisture spread but a very few inches; even in vegetable earth it has no great extent, which must be more porous than sand or hard soil. It is a certain fact, that in a garden we may almost inundate one bed without those nearly adjoining feeling any moisture from it[65:A]. I have examined pieces of garden ground, eight or ten feet thick, which had not been stirred for many years, and whose surface was nearly level, and found that the rain water never penetrated deeper than three or four feet; and on turning it up in the spring, after a wet winter, I found it as dry as when first heaped together. I made the same observation on earth which had laid in ridges two hundred years; below three or four feet it was as dry as dust; from which it is plain that water does not extend so far by filtration as has been generally imagined. By this means, therefore, the internal part of the earth can be supplied with a very small part; but water by its own weight descends from the surface to the greatest depths; it sinks through natural conduits, or penetrates small passages for itself; it follows the roots of trees, the cracks in rocks, the interstices in the earth, and divides and extends on all sides into an infinity of small branches and rills, always descending until its passage is opposed by clay or some solid body, where it continues collecting, and at length breaks out in form of springs upon the surface. It would be very difficult to make an exact calculation of the quantity of subterraneous waters which have no apparent vent. Many have pretended that it greatly surpasses all the waters that are on the surface of the earth.