Hi, I just wonder if R is still used for PK analysis. However, I have to use R for some purposes anyways. Here is a really strange problem I am running into and I hope to get some assistance from the group. The following code works fine in S-Plus (Population estimates and MAP estimates look reasonable ). cs.fit<-nlme ( CONC~phenoModel(CID,TIME, DOSE, lCl, lV), fixed = lCl+lV ~1, random = pdDiag(lCl+lV ~1), data = cs1grp, start = c(-5, 0), na.action=na.include, naPattern = ~ !is.na(CONC) ) But if I run the same problem in R, I get the following error: "Error in chol((value + t(value))/2) : the leading minor of order 1 is not positive definite" These are simulated data with a total of 100 subjects studied at one dose level and rich sampling scheme. (I can send the dataset, if required. I had some unsuccessful attempts to send by default). If it was a problem with parameterization of the model , I don't understand how come sparse sampling in generic "Phenobarb" data does not have a problem in R or S. BTW, the exact same code for "Phenobarb" runs in R and S-Plus w/o any modification needed. Thank you in anticipation. Pravin Pravin Jadhav [[alternative HTML version deleted]]