lme help

Does anyone know how to fit a linear mixed effects model when the within-group covariance is known!    I know how to fit an lme model when the within-group covariance STRUCTURE is known, but that doesn't help.    

Any help would be greatly appreciated!