response to David Jacobson e-mail:

>>The Nikkor PC lenses (or any PC's) should be focused on a plain
>>(or plain plus grid lines) VF screen shifted to the taking
>>position - do not use focus "aids" (if you feel you must use
>>those dreadful "aids" :), center the lens first, but focus may
>>be slightly less accurate if the lens has field curvature).
>>Metering should be done with the lens centered on all bodies
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>except the F3, which will accurately meter through the shifted
> ^^^^^^^^^^^^^
>>lens at all but the most extreme rise in the vertical frame
>>position (this is beyond Nikon's recommended rise limits [That's
>>what those little numbers running from 7 to 11 on the outside
>>edge of the lens indicate: the max recommended shift at different
>>orientations of the lens - though they may be ignored at the
>>smallest stop or two.]).
>>Hope This Helps
>>
>
>Why is this? Don't the metering cells pick up light exactly the same
>as film does?

Hi-- No. Most bodies have the cells above the viewing screen, and
when the PC lens is shifted off center, the ground "glass" and Fresnel
are most unhappy, showing their displeasure to both the eye and the meter
by darkening one side of the screen (which does not appear on the film).
This can cause errors of up to two stops in the metering. The F3 meter
is under the mirror, and, though the meter area probably shifts as the
lens is shifted, the metering stays correct until the extreme shift
along the long axis is reached.
David Ruether