Hi Paul

Thank you for the fuller explanation.

However, I’ve understood all that already, and that’s what I was doing. I had a mail merge output document, open as you describe in your point (1), containing the text that needs to go into the body of each mail; and I had a ready-merged directory document consisting of a two-column table with e-mail address in the first column and the path to the attachment in the second column, as you describe in your point (2). This latter was not open – as required. The records in the two documents were sorted in the same order, since they were created using the same original data source. This was the exact scenario that yielded the results I described in my second post, above.

The problem is that I want *the attachment itself* – to which, for each record, the path in the second column of the directory document points – to be a merged Word document, in which the content is different for each recipient, populated using fields in the original data source. This is the bit that I can’t achieve using the process described, because when you merge to letters in Word you get one long document and not separate files for each recipient. If the path in the second column of the directory document points to an unmerged Word document, the process doesn’t merge it – it just attaches it as it is.

Do you know of a solution that achieves what I describe?

Thanks,

Martin