Wendell,

Thanks for the extra information on this problem.  I like
the fact that the position is being explicitly calculated
via count(preceding-sibling::Cell), as indeed it would be
more robust, particularly if the sequence which was the context
when the template was invoked was not what I am obviously
expecting (ie all the Cell elements in a Row) it to be below.

The fact that Cell elements can be missing is a bit of an
issue for this approach, but as I am prototyping the stylesheet
using Excel xml input, I can make sure the input never has this
problem.

I have to say this list and its citizenry is amazing.

Cheers,
Peter