Thank you for your quick response. You've confirmed what I pretty much
suspected. What I'm trying to accomplish is to identify paragraphs in generic
papers that might have "run-in" heads (that is, the head is part of the
paragraph, not on a line by itself). These heads do not have a standard
format but do generally follow a rough pattern. I've converted everything to
flat text by converting all the formatting to tagged markup. The wide variety
of possibilities in the tagged markup are what cause Word's wild card option
to fall short in this case. The regex identifies the paras with run-in heads
and places a marker at the end of the head text for later processing. The
placement of the marker is the problem. Of course, that's the time when the
string text is used for substitution.

I am intrigued by your last suggestion. That's exactly the kind of solution
I was contemplating. Your SymbolToUnicode macro supplies the hints on how to
do this. I plan on trying your 2 macros first before I get too involved with
writing the "character to tag" macro. It seems, however, that in order to
prevent "surprises" from other symbol fonts, that kind of solution will need
to be implemented.

Thanks again.