I have just run a very simple JavaScript performance test (don't ask why). The test declares a variable, but doesn't assign anything to it:
var x;
It then compares the speed of comparing the value variable to null, and to undefined, in other words:
var y = (x == null); and var y = (x == undefined);. 
I was expecting the comparison with undefined to be the fasted. In fact it was nowhere near. The comparison with null was far and away the fastest, around 80% faster.
The results I've described above come from running the tests in Chrome (version 13). Running them in Firefox produces results far closer to what I would have expected (the comparison with undefined is faster than with null, albeit very marginally).
So, my question is what could the cause of this be? Why does Chrome seem to favour the comparison with null so greatly?
For quick reference, here's a screenshot of the results:

 
     
     
     
     
     
     
    