Are your timings based off of a single run each for Firefox and Chrome? The web is a very noisy place; to get good timing data you need to run many tests. WebPageTest (www.webpagetest.org) can be a good way to do this.

(I work on PageSpeed, and we need to run tests like this often.)