This is purely for the number-crunchers out there, let me take Szen's formula another step. Some people like to work with shutter speeds, instead of apertures, so here's how.
If you take the aperture correction factor inside the bracket, i.e. (1 + M/p) and squared it, the resulting amount would be factor you need to multiply the effective shutter speed to get your actual shutter speed.
For example, let's say all you had was a bellows with a normal lens and a handheld meter, so no TTL. The handheld meter gives you a shutter speed of 1/125 sec and an aperture of f/11. You need a magnification of 3. Since it's a normal lens, p=1. The correction factor (1 + M/p) would then be (1 + 3/1) or 4. Square that to 16. Multiply 1/125 sec by 16 and you approximate 1/8 sec. You then need to dial in 1/8 sec on the shutter speed dial, retaining f/11 on the lens aperture.
If you had taken Szen's original formula, you would have divided 4 into f/11, to get f/2.8, the actual aperture you need to set on the lens to get the effective aperture of f/11 (what the handheld meter suggests). Note that either way would have yielded the same result: if you noticed, there are 4 steps between 1/125 and 1/8 sec, as there are between f/11 and f/2.8.
Of course, all these are pretty moot, if you're using a modern body with a AF-micro lens, since all these calculations are done internally for you.