In my 25-frame sequence of that boat crash the first frame is: Date/Time: 2004:06:08 11:21:25
The last (25th frame) is:
Date/Time: 2004:06:08 11:21:28
25 frames / 3 seconds comes to 8.333 FPS.
Just about what Phil claims. The camera, however, only times to the closest second. And hmmmm...8.5 or 8.3. Is it really THAT big a deal? Let's not start measurebating this please. That old 10D focus crap still gives me forum-nightmares.

Wouldn't the last frame be 24 frames later? Also with your times being only accurate to the nearest second surely the first shot could have been at 25.49 seconds and the last one at 27.51 seconds (over 11fps). Or maybe first frame at 24.51 and last at 28.49 (Only around 6 fps).
For an accurate test you need good light on a stopwatch that has hundredths of a second, take a burst and each frame should be within 12 hundredths of the previous one.
If someone could lend me a MkII I would be happy to perform some tests.
Mark: I/we have already addressed the fact that my test is not accurate:
Just about what Phil claims. The camera, however, only times to the closest second.
And we already discussed the stopwatch method:
So take pictures of a digital clock that shows hundredths of a second.
And Timmy already requested (anyone's) MKII:
Send me the body, i'll do it.
Did you have anything else? 

