R help - Aug 2003 - bootstrapping nlme fits (was boot function)

(see archives for former discussion)

I probably didnt express myself very well in the last mail i wrote. The dataset
contains the variables subject, day, concentration    (measured). I would like
to bootstrap the variable subject. Now it is true that the subjects wont be
independent in the bootstrap samples but i still wanna conduct the bootstrapping
just for the sake of comparison with other results i got. Well, i succeeded to 
set up the program needed for bootstrapping. However, i get the error message:

Error: Unable to form Cholesky decomposition

or that convergence was not reached with the maximal interations.
What exactly does that mean? Is there possibly any connection to the fact that
the subjects are not independent?

Thanks a lot for answers

Dassy