A Little Help with Hpricot Parsing

I need to scrape out the name, address, city, state, zip, etc. from
site. Due to their wonderful coding, they didn't put lines of text in
<p> tags. So Hpricot doesn't see the address (et al) as an element,
but it does see the <br> as empty elements.

So how do I pick out the address and so forth if they are not in
elements?

Here is what I have scraped so for, down to my target table:
<table id="vbsSearch_pnlSortByOrg">
  <tr>
    <td>
      <h3>
        <span class="orgName">Bank of America</span>
      </h3>
      405 N. 3rd Street<br>
      Phoenix, Arizona&nbsp;60606<br>
      Distance: N/A<br>
      800-555-1212<br>
      <ul>
        <li>
          <strong>June 15 - 20 2008</strong><br>
          6:00 PM&nbsp;-&nbsp;9:00 PM<br>
          Event Theme Name<br>
        </li>
      </ul>
      <hr align="left" width="100%">
      <h3>
        <span class="orgName">Washington Mutual</span>
      </h3>
      3705 Beaver Creek Rd<br>
      Austin, Texas&nbsp;60606<br>
      Distance: N/A<br>
      800-555-1212<br>
      <ul>
        <li>
          <strong>July 07 - 11 2008</strong><br>
          9:00 AM&nbsp;-&nbsp;12:00 AM<br>
          Event Theme Name<br>
        </li>
      </ul>
      <hr align="left" width="100%">
    </td>
  </tr>
</table>

So how do I pick out the address and so forth if they are not in
elements?

Use that fact to your advantage. The only non-elements in the table
cell are your addresses.

doc.search('/table/tr/td/*').each do |node|
  print node unless node.elem?
end

This will print out the addresses only.