Using approximations based on the methods and theories of linear least squares, a procedure is presented for estimating the variance of the parameter estimates of a nonlinear time dependent growth model before sampling for growth. To use the procedure, prior knowledge is needed of the approximate model parameter values and the variance about the regression. An example is given where the relative standard error of the asymptote of the Chapman-Richards function is estimated for increasing growth series lengths. Forest Sci. 30:836-841.