> No. The "order" of an axis only "rules" for the step that it is in, ie
> it affects the meaning of position() and last() in predicates 
> that occur
> in that step. It doesn't affect the constructed node set.

And the last step has the say on which way the nodes are used
for position() and last() in any predicate?

A/B/C[last()] then should mean that whatever axis is used in step C
determines if document order or reverse order is 'in play' for the
predicates?
Not forgetting any shorthand expansion :-)

DaveP

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list