Estimate dispersion parameters under a general linear model. The statistical models cover a variety of possibilities like heterogeneous covariance components for residuals as well as other random effects, models with longitudinal data, random regression, and multi-environment analyzes. Additive and dominance relationships are implemented.