It is probable that the structure of rocks has also a greater, and not less, diversified influence upon the humidity of productive soil. Solid rocks, which are not traversed by numerous perpendicular fissures penetrating to a considerable depth, allow the water to remain in the soil; but columnar and schistose rocks, with perpendicular fissures, and strata declined from the horizontal position, draw off the water from the soil covering their surface, into lower places, where it often re-appears under the form of springs. In these circumstances, we find a partial explanation of the great difference between the humidity of soil covering a surface of solid granite, and that lying upon limestone, which is intersected by numerous fissures. Granitic mountains are often furnished with marshes, whereas, on the other hand, the dryness of the soil upon calcareous mountains is generally excessive[411], the cause of which phenomenon is, in a great measure, to be attributed to the circumstances above mentioned. Columella observes, that silex having a moderate covering of earth, preserves to the latter its humidity; and Palladius repeats the remark. In districts which consist of quartzose rocks, not less than of granitic ones, the surface is often covered with marshes. Porphyritic rocks, on the contrary, which have a remarkable segregation of parts, as well as columnar basalt, let off the water to lower places. Springs are very frequently found at the bottom of basaltic mountains; for the atmospheric waters penetrate by the perpendicular fissures to the strata on which the basalt rests, and appear at the place where the two rocks meet. The effect of different rocks upon the preservation and diminution of the moisture of fertile soil, influences vegetation in various degrees. The retentive power of the surface of rocks is of the greatest importance, where the soil consists chiefly of sand, through which the water percolates, and passes off entirely, unless it meets with a stratum of such a nature, as to obstruct its passage, or comes upon a surface of solid rock. The cause of the sterility of sandy plains is not merely their sandy nature, but also the great depth of the mass or rock capable of retaining the water. The same sand, when covering mountains consisting of sandstone, has a much less degree of sterility than in those plains, because the surface of the subjacent rock impedes the progress of the water, and consequently retains it in the soil[412]. It has been sufficiently proved by experiments, that plants can grow in pure sand, when furnished with the necessary quantity of water. A subjacent rocky surface has an entirely different effect upon soil which is very retentive of moisture, upon a clay soil for example, as, in that case, the humidity is increased to a prejudicial degree. In land of this nature, a substratum of rock having the property of drawing off the water would be useful.