Such dew-formation, however, cannot happen to any appreciable extent under the conditions maintained in the writer’s work, viz, absorption within a thick-walled (two-inch) wooden box of the dimensions given above, and sunk in the ground in a cellar in which the temperature varies only a few tenths of a degree during 24 hours. The soil layer of one millimeter thickness being put down in the morning, the 7 hour absorption period falls at the time of slightly rising temperature, as an additional precaution against dew-deposition. Mitscherlich fails, moreover, to show that this source of error produces any wide or serious discrepancies except under such long absorption periods as he finds it necessary to use because of the great thickness of his soil layers. It is doubtful whether the limits of errors in soil sampling do not greatly exceed any of those involved in the writer’s method, and whether such accuracy as is attempted by Mitscherlich is of any practical significance. Mayer’s experiments relied upon to prove the uselessness of hygroscopic moisture to plant growth, were carried out in flower-pots, in which it was plainly shown that the plants wilted before even the visible liquid (capillary) moisture of the earth was entirely exhausted. But this simply proves that under such artificial conditions, plants cannot withdraw moisture from the soil _rapidly_ enough for their needs. In _nature_, and notably in the arid regions, the chief supply of water is received through the deep-going main roots, while the bulk of the active feeding roots of the plant may be surrounded by almost air-dry soil; under which conditions, as Henrici (Henneberg’s Journ., 1863, p. 280) has shown, slow growth and nutrition occurs even in such plants as the raspberry, a native of humid climates. But in the arid region this is the normal condition of the native vegetation through most of the rainless summer. That a higher moisture-coefficient does not necessarily imply that a larger amount of moisture can be withdrawn from the soil by the plants, is undoubtedly true in some, but not in all cases; for in soils rich in humus, the moisture is more freely shared with the roots than in non-humous, clay lands. The higher moisture-absorption is however of the most unquestionable service in the case of the occurrence of the hot, dry winds that so frequently threaten the entire crops of some regions. In this case the soil containing the greater amount of moisture requires a much longer time to be dried, and heated up to the point of injury to the roots, than in the case of sandy soils of low absorptive power, whose store is exhausted in a few hours and then permits the surface to be heated up to the scalding point, searing the stems and root crowns. That such injury occurs much sooner in sandy lands than in well-cultivated clay soils, is a matter of common note in the arid region.