mismatch on script tags by getElementsByTagName and XPath

Hi,
I'm experimenting a strange behaviour.

On that page,
http://www.dixonvince.com/

either using Firebug (on Firefox 3.6.17) or XPath I match 31 tags
script.

However, in my java code (xulrunner 1.9.2), if I retrieve the
nsIDOMDocument and getElementsByTagName I get only 20 tags.
Also, if I use the nsIDOMXPathEvaluator to evaluate the //script
expression, the result is again 20 nodes.
Finally, I tried to save the page with
nsIWebBrowserPersist.saveDocument and on within the file there are
indeed 20 occurrencies of script.

Similar behaviour is on the following pages
http://www.bonetts.co.uk/
http://www.thelettingline.co.uk/

Do you have any idea of what is happening here?

Any hint is appreciated
Thanks