smooth quantile lines

[FCollin@genelogic.com]

Adding a lowess line to a Residual vs Predicted plot gives me an indication
of any predicted value dependent systematic effects.  I would also like to
gage the dependency of variability on predicted value.  A lowess line
through the scatter of abs(Residual) vs Predicted (or Residual^2 vs
Predcicted) gives some indication of trends in variability.  I wouls also
like to look at the dependency of the tails of the distribution of errors
on Predicted value.  Any suggestion on how to do this would be appreciated.

Thanks in advance.

-francois collin