> For some reason I could not get David Carlisle's method to work.

perhaps because, depending on how charitable you're feeling, my answer
was unduly cryptic, or wrong.

The general model is to do what I posted [position() mod 5 = 1] but I
failed to observe that your existing selection wasn't a single step
foo/bar[position() mod 5 = 1]
selects  every  5th element child of each foo, whereas you wanted every
5th item in the sequence so that's

(foo/bar)[position() mod 5 = 1]

but it's probably simpler to use for-each-group anyway, as (because the
initial items are mot selected with a simple step) you'd have to use a
correspondingly more complicated expression to select the following ones
(or did as I posted in 2nd reply) use the compartively more expensive
following axis rather than following-sibling.

David

________________________________________________________________________
The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs. 
________________________________________________________________________