The allometric relationship between plant volume and phytomass of crested wheatgrass was studied for the 1981, 1983, and 1984 growing seasons in west-central Utah. Basal diameters, canopy diameters, and standing plant heights were measured for individual plants. Three models of volume (basal elliptical cylinder, canopy elliptical cylinder, and elliptical cone section) were tested as predictors of plant phytomass using nonlinear regression. Elliptical cone section produced the highest R2 and lowest SEE, but requires measurement of canopy diameters which may be subject to excessive measurement error. Basal elliptical cylinder produced R2 and SEE values nearly comparable to those of the elliptical cone section; moreover, this model does not require measurement of canopy diameters, making it the practical choice. Nonlinear regressions for plants by size class (small, medium and large) were produced using 1983 data. Predictive ability of size class-specific equations was compared to that of the equation for all size classes combined. When phytomass of only small or medium size class plants was predicted, the SEE of size class-specific equations was slightly lower than the SEE of the equation for all size classes combined. When phytomass of plants from all size classes was predicted, however, the equation for all size classes combined produced the lowest combined SEE for new data (i.e., data not used to generate the equation). There were substantial year-to-year differences between equations, which indicates the necessity of producing new equations each year.