I've been wondering for awhile now if camera exposures are
>repeatable, ie, if you set the thing for 1/125@f/11 and shoot 36 shots,
>will they be identical? This is what I was testing on those shots. The
>first run of diagonals was done on auto, and I just changed the f/stop.
>They should be identical, right? The next set were done on manual, with me
>changing the shutter speeds to known values. These should *really* be
>identical, assuming constant light. But they weren't identical, which leads
>me to believe you can figure out the exposure all you want but you better
>bracket (with slides) cuz they're so sensitive to small changes in exposure
>caused by who knows what that exposures are not even repeatable!

Hmmm... You are right about the last, but it has to do with the aperture,
especially in a zoom. The aperture the camera uses is a simulation
(variable resistor in body) of the actual aperture (marked lens aperture), which is a simulation of the ACTUAL aperture (where the leaves are supposed to be at the markings), which is a simulation of the A C T U A L aperture
(where the leaves actually are at one FL, which is a simulation of the
A * C * T * U * A * L aperture at the FL you have just shifted to,
causing the diaphragm to adjust.... So, an F3/FE/FM/F2A/F2AS with the AI
tab disengaged, using stop-down metering (with eyepiece shielded [F3
best here] gives the most accurate exposure.... [so much for the applique
of all those funny exposure features, like Matrix, spot, etc., since they are built on top of an erroneous system...]).
So we Luddites have our reasons....;-)