Conditional extraction of data

It pays to be honest, so I'll be up-front about this.  The question arises
from an assignment on a course I'm doing.

However, I don't want to cheat, so I'll try to form the question in the most
general way that I can:

The purpose of the exercise is to convert some xml to html.  I have a
repeating element under the root - let's call it anElement.  Initially, the
requirement was to output all its contents, unchanged.  So I could do this:

<xsl:for-each select="anElement">
	<xsl:value-of select="." />
</xsl:for-each>

The anElement element has mixed content and would be described in a DTD as
follows:

<!ELEMENT anElement (#PCDATA | tag1 | tag2 | tag3 | tag4)* >

All the child tags contain #PCDATA only.

The new requirement is to print all the content as before, but to print the
contents of tag2 (say) in italics - i.e. surround its contents with <i></i>
html tags.

I've sweated over this all weekend and can't work out how to solve it.  If I
treat each tag individually in a for-each or an apply-templates/template
combination (if that makes sense), I can't see how to output the #PCDATA in
anElement.

Any help would be appreciated - especially a pointer to a resource that will
let me work it out for myself.  I've tried the obvious places like w3c
schools.

Thank for your attention


Peter