We're not as far apart as it seems. The 55/2.8 MicroNikkor reversed and shot at f/4 is indeed competitive with the 63/4.5 shot wide open, its best aperture. With the 40 and 25 Luminars at their recommended magnifications shot wide open, maybe, but at the cost of considerably more extension. Not competitive with the 16/2.5, which benefits from having a larger numerical aperture.
Originally Posted by Sparky
Note the qualifications. The 55 MicroNikkor's major advantage over shorter Luminars, if the extension can be obtained and managed, is somewhat greater working distance. Its major advantage over all of them is cost.
Um, er, ah, about tests that don't have features at a range of spacings, well, they're really acceptance tests. The lens can resolve the features in front of it or it can't, but the test tells nothing about whether it can separate more closely spaced features. Most of the tests I've seen on the 'net were done at arbitrarily selected magnifications, not stepping over a range of magnifications. Klaus Schmitt and I had a strong disagreement about this some years ago. He tested every lens he had by shooting a microchip at 10x. This is a fine test for deciding which lens to use at 10x for subjects with features spaced no closer than the chip's closest features but isn't very informative about which lens to use at other magnifications with subjects whose key features are more finely spaced. If you want to know what a macro lens will do, shoot the USAF 1951 on glass at a range of magnifications. This is what I did after I got organized and this is why I insist that most macro lenses are optimized for a range of magnifications and have an optimum aperture that sometimes varies with magnification. It all comes down to asking the right question in an appropriate way.