a loess/NaN question

[Paul Schlobohm <pschlobo@dri.edu>]

I have a dataset with missing values or NaNs for which I would like to
generate a loess fit for the entire dataset, including the NaNs.  I can
successfully perform a loess() routine and a plot.loess() routine on
this dataset with na.action = na.exclude.  However, the routine
loess.smooth(), which would provide the fitted values I am interested
in, does not want to run with my missing values.  Can you suggest a
solution?

Thanks in advance
Paul