I finally did a couple of tests. I took a 3 foot long strip of wood and nailed nails to it about an inch apart. I placed it about 3ft from the camera, focused on a nail in the middle of the strip and took a picture. Then I did the same with the strip about 10 ft away.

With the first of the above, the focusing appeared accurate. On the second, with the strip more distant, the nail I focused on was at the rear end of the depth of field.

I did some research, and I am led to believe the following:

1. My eyesight has nothing to do with what appears to be the best focus point on the ground glass screen that I am essentially viewing through the camera's viewfinder prism. The image is either in focus or not. Poor eyesight may make it harder to discern the best focus, but it could not cause an image to appear sharper when it is less sharp on the ground glass.

2. The point of focus in the split screen focusing aid in the center of the ground glass focusing screen must coincide with the best focus on the ground glass screen, absent manufacturing defect; the split screen prisms are embedded in the focusing screen. For more on how they work see:


3. Therefore, How accurately focus is depicted on the focusing screen should not vary with the distance of an object.

My conclusions: my focusing issues are either the result of operator error, it being harder to focus accurately on a smaller, more distant object; or there is a very small error in the camera that may cause focus to be off only a fraction of an inch up close but several inches at greater distance.

Curiously, I performed the same test on three different Nikkormats, with similar results. I also practiced with a D70 (sorry) with power off, then switched power on - each time the in-focus light was on, indicating that my focusing was accurate.

I can only think of one cause other than operator error, and that could be that over the years, wear could have caused the mirror to be off by a couple of thousandths of an inch. Rather than adjusting it and potentially making things worse, I will do a final test affixing some masking tape to the back of the mirror, to raise it slightly, and see what happens. But at this point I'm guessing that the problem is operator error.