Short answer: use
   (*//surname)[1]
instead of
   *//surname[1]

The long answer would be better handled by someone more expert. And
to whatever extent I could give a long answer at all, I couldn't do
it when I'm this tired. :-)

G'night.

> Thanks, Syd. You have identified the problem for me.  =20 Yes,
> there are some <biblStruct> entries in my data that contain=20 more
> than one <surname>. If I test this against entries that contain=20=

> only one <surname>, it works.  So I need to adjust my code such
> that it applies only to the *first* surname encountered.
> 
> Thus, I changed
> 
>   <xsl:sort select="*//surname"/>
> 
> to
> 
>   <xsl:sort select="*//surname[1]"/>
> 
> but this generates the same error message.
> 
> I need to find a way to select only the first occurrence of this 
> descendant. I am trying to figure this out now, but if anyone can 
> point me in the right direction, I'd be delighted!