It's only the MARK that sets the servo position, the SPACE doesn't seem to matter?. But usually you're making the SPACE longer, not shorter, but I see no reason why it won't work - but I should test it first!.
Why are you wanting to use a shorter SPACE anyway?.
running a servo at 250 hz is pushing it a bit - some servos will work at that high of a speed, others won't. Try it - and if it works monitor the temperature of the servo - if it's getting really hot then, obviously, you have a problem. Most servos will work up to around 200hz. Some will stop working well before that, unfortunately. I've heard some of the digital ones can go to 400hz - though I don't have any digitals so I am unable to verify this.
That's because you're trying to use the PWM module, it's not intended for servo's, and it's not really suitable - there's plenty of PIC servo examples on the net, google will find them for you.
This is not 100% relevant but I believe it may help a bit.
I was working with parallax cont. rotation servo working perfectly with the pulse that it required.
However there was another guy in the lab that was using the same motor with totally different timings but the servo worked just fine! The reason why his timings were wrong is because he messed up with the prescaler. However the "duty cyle" (percentage) of the pulse was correct even if the duration was wrong. Now the key to the whole situation is that those servos had a screw which you can use to adjust its stability. I believe this is why his servo also worked.