Several mathematical models have been developed to describe water uptake by plant root systems. These models usually incorporate one or more of the assumptions that water is proportional to rooting density, to soil hydraulic conductivity, and to the water potential difference between the root surface and that in bulk soil midway between two adjacent roots. An experiment was conducted with cotton (Gossypium hirsutum) at the Auburn rhizotron to test the validity of these three assumptions. Soil water potential, soil hydraulic conductivity, water uptake, and rooting density were measured or estimated as functions of soil depth and time. Average plant water potential for a water extraction period was estimated from pressure chamber and continuous stem diameter measurements. The assumption that water uptake is proportional to rooting density was valid. However, there appeared to be a large resistance in the pathway from root epidermis to root xylem. The other two assumptions must be modified to include this resistance. Therefore, water uptake was proportional to the water potential difference between root xylem and bulk soil and to the hydraulic conductivity of the combined soil-root pathway. There was no depth-of-roots effect on the validity of these relationships.