Moisture removal rates have been used for many years to estimate root distribution in the soil. Repeated observations have indicated that in the field, water is removed more rapidly from shallow depths when the soil is uniformly moist throughout the profile, but that as the surface soil dries out, the largest amount of water is removed each day from successively lower depths. After an irrigation, water is again removed more rapidly from the surface and a similar cycle repeated. It does not seem probable that root distributions would change in this manner.
A laboratory experiment was conducted in which a constant vacuum was used to simulate the suction created by the leaves of a plant. Absorbers made of Coors No. 762, porosity 5, filters were placed in the soil in a way to represent the root distribution in a soil. The water removed was measured in burettes.
With uniform suction on all absorbers (representing roots), the amount of water extracted per unit time was roughly proportional to the area of the absorbing surface when tensions were uniform. The tension increased more rapidly in the soil with greatest concentration of absorbers after which more water per unit area of absorbers was extracted from the wetter soil with fewer absorbers.
The amount of water extracted per day from a soil decreased continually with time when the suction was constant. If total water extracted per day were to be maintained anywhere near constant, the applied suction would have to increase with time.
It appears that a soil moisture tension gradient, characteristic of the amount of absorbing surface, would be established in the soil at the point where water would be extracted at the same rate from each volume of soil regardless of the surface areas of absorber.