Hmisc / Design question

All,

I have been reading Dr. Harrell's excellent 
"Regression Modeling Strategies" book and trying out
the exercises.  I understand that contrast( ) is used
to obtain contrasts between two variables for given 
levels of other nuisance variables; is there a way
to use contrast( ) to obtain, for example, Scheffe
confidence intervals / hypothesis tests for many
post hoc contrasts at once?

Any help would be much appreciated,

    Greg