This paper provides an alternative method to describe the water retention curve over a range of water contents from saturation to oven dryness. It makes two modifications to the standard Brooks and Corey <1964> (B-C) description, one at each end of the suction range. One expression proposed by Rossi and Nimmo <1994> is used in the high-suction range to a zero residual water content. (This Rossi-Nimmo modification to the Brooks-Corey model provides a more realistic description of the retention curve at low water contents.) Near zero suction the second modification eliminates the region where there is a change in suction with no change in water content. Tests on seven soil data sets, using three distinct analytical expressions for the high-, medium-, and low-suction ranges, show that the experimental water retention curves are well fitted by this composite procedure. The high-suction range of saturation contributes little to the maximum capillary drive, defined with a good approximation for a soil water and air system as HcM=∫0∞krwdhc, where krw is relative permeability (or conductivity) to water and hc is capillary suction, a positive quantity in unsaturated soils. As a result, the modification suggested to describe the high-suction range does not significantly affect the equivalence between Brooks-Corey (B-C) and van Genuchten <1980> parameters presented earlier. However, the shape of the retention curve near natural saturation has a significant impact on the value of the capillary drive. The estimate using the Brooks-Corey power law, extended to zero suction, will exceed that obtained with the new procedure by 25 to 30%. It is not possible to tell which procedure is appropriate. Tests on another data set, for which relative conductivity data are available, support the view of the authors that measurements of a retention curve coupled with a speculative curve of relative permeability as from a capillary model are not sufficient to accurately determine the (maximum) capillary drive. The capillary drive is a dynamic scalar, whereas the retention curve is of a static character. Only measurements of infiltration rates with time can determine the capillary drive with precision for a given soil. ¿ 1999 American Geophysical Union