etree.strip_tags() does not remove all instances of a defined tag

I have the following python code on lxml 2.2.2 with libxml 2.7.6 on FreeBSD 7.2:

from lxml import etree
html = """
<div>
  <div>
    I like <strong>beer</strong>.
    <br/>
    I like lots of <strong>beer</strong>.
    <br/>
    Click <a href="www.beer.com">here</a> for <a href="www.beer.com">this</a> beer.
    <br/>
  </div>
</div>
"""
element = etree.fromstring(html)
etree.strip_tags(element, 'a','br')
print etree.tostring(element)

which prints:

<div>
  <div>
    I like <strong>beer</strong>.

    I like lots of <strong>beer</strong>.

    Click here for <a href="www.beer.com">this</a> beer.
    <br/>
  </div>
</div>

I would expect *all* the "br" and "a" tags to be stripped.
Another example, use "etree.strip_tags(element, 'strong','br')", you get this output:

<div>
  <div>
    I like beer.
    <br/>
    I like lots of <strong>beer</strong>.
    <br/>
    Click <a href="www.beer.com">here</a> for <a href="www.beer.com">this</a> beer.
    <br/>
    </div>
</div>

Again i would expect all the defined tags to be stripped.
Thanks