Hi, I have data with negative spatial autocorrelation at short lags. It is well described by a spherical model, but flipped upside from what is normally seen. It seems that gnls will work with this correlated residual type IF one parameter in the spherical model can be negative (it is usually positive). Does anyone know if this parameter is constrained to be positive? -seth -- View this message in context: http://n4.nabble.com/gnls-for-negative-autocorrelation-tp1838023p1838023.html Sent from the R help mailing list archive at Nabble.com.