Xpath reverse searching

Is there a way, when using DOM_Document Xpath to search in reverse (from the end of the page moving up instead of from the top down?) If so, how would I do this?

I am doind a scrape of a web site. (linked below). http://www.sturmfh.com/obit-display.jhtml?DB=update/obits/dbase&DO=display&ID=1189477693_24578

I only want to scrape the 3 obituary paragraphs. So i figured it'd be easiest to start at the end and move up.

Use:

(//p)[position() > count(//p) - 3]

This selects the last (up to three) p elements in the XML document.

XSLT - based verification:

<xsl:stylesheet version="1.0"
 xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
 <xsl:output omit-xml-declaration="yes" indent="yes"/>
 <xsl:strip-space elements="*"/>

 <xsl:template match="node()|@*">
     <xsl:copy-of select="(//p)[position() > count(//p) - 3]"/>
 </xsl:template>
</xsl:stylesheet>

When applied against the document, referenced in the question, this transformation evaluates the XPath expression and outputs the selected p elements.

The result is:

<p>
                If you would like to share your thoughts and memories,<br/> we will deliver your message to the family.<br/>
   <a href="mailto:staff@sturmfh.com?Subject=For%20the%20Family%20of%20Lyle%20Meier">Click</a>
   <a href="mailto:staff@sturmfh.com?Subject=For%20the%20Family%20of%20Lyle%20Meier">
      <img src="/images/email_condol.gif" alt="Logo" border="0" align="middle"/>
   </a>
   <a href="mailto:staff@sturmfh.com?Subject=For%20the%20Family%20of%20Lyle%20Meier">here</a>.
        </p>
<p>To Request a Tribute Folder
                <br/>
   <a href="./obit-foldreq.jhtml?fname=Lyle&amp;lname=Meier">Click</a>
   <a href="./obit-foldreq.jhtml?fname=Lyle&amp;lname=Meier">
      <img src="/images/email_condol.gif" border="0" alt="View" align="top"/>
   </a>
   <a href="./obit-foldreq.jhtml?fname=Lyle&amp;lname=Meier">here</a>
</p>