Photomacrographic lenses, part 1:
Zeiss Luminar 63 mm f/4.5
Leitz Milar 65 mm f/4.5
Olympus 20 mm f/3.5
Lomo 9x 0.20 microscope objective

In the 50’s, 60’s and 70’s, several manufacturers produced series of photomacrographic lenses, mainly designed for reproduction ratios above 1:1. Almost without exceptions, these lenses are designed to be mounted on bellows or extension tubes. Typically, these lenses are mounted in narrow and short barrels with manually operated diaphragms and no focusing helicoid. Most of them have a RMS threaded attachment similar to that of European standard microscope objectives. Adapters to connect RMS lenses to M42 (“Pentax” thread) equipment are relatively easy to obtain, as well as adapter rings to mount M42 lenses to other brands of cameras. Sometimes, it is even possible to find RMS adapter rings with other lens mounts, e.g., a Nikon bayonet.

Some of these lenses, however, (especially longer focal lengths like 100 mm and above, or lenses with unusually large apertures) use a wider, non-standard threaded mount and need proprietary adapters. If you own one of these lenses and/or adapters with odd mounts, epoxy can make miracles possible, like solidly connecting adapter rings with incompatible threads. There is even a type of epoxy with the consistency of modelling wax when not cured, which you may press by hand against proprietary threaded attachments to make an adapter that still allows screwing and unscrewing (albeit not as durable as a metal one) instead of gluing them permanently. For technical data and/or pictures of these lenses, see macrolenses.de/ and markerink.

From left to right, the above picture shows the following photomacrographic lenses:
Zeiss Luminar 63 mm f/4.5
Leitz Milar 6.5 cm f/4.5
Leitz Summar 10 cm f/4.5
Olympus 20 mm f/3.5
Lomo 9x 0.20. This is a microscope objective, not a photomacrographic lens. However, it is roughly equivalent to a 28 mm f/4.8, and it is unusual as a microscope lens in that it has a diaphragm mounted at the back of its optical elements, and operated by an aperture ring on the lens barrel. I included it in the present test because it can be used just like a photomacrographic lens, and is - potentially - a low-cost alternative to some of the other lenses shown above.

With a couple of exceptions, this type of lenses is no longer produced, and the second-hand market is the only available source. While camera shops tend to sell these lenses at high prices (especially the series regarded as best performers, and therefore most sought-after), eBay and other auction sites now and then sell specimens in good condition at quite reasonable prices (i.e., 1/3 to 1/2 the asking price for a new general-purpose macro lens of good quality).

I happen to have two such lenses of medium focal lengths, the Zeiss Luminar 63 mm f/4.5 (leftmost in the above picture) and the Leitz Milar 65 mm f/4.5 (second from the left). My specimen of the Luminar is not from the latest series (also called Blue Dot, which usually commands very high prices), but probably from the next-to-latest series, maybe produced in the 60’s or early 70’s. This series is usually cheaper, and said to be almost as good as the Blue Dot. The Milar lens is older (probably 40’s or 50’s) and much cheaper (1/3 of the Luminar, or less). It is peculiar in using a non-standard threaded attachment wider than the RMS. It came with an original adapter ring, which, however, also ends in a non-standard M40 thread, too big to fit a standard M39 attachment and too small for an M42. I epoxied it into an M42 to Nikon adapter, where it sits happily. The Luminar 63 mm is specified for a 2x-10x magnification range, the Milar 65 mm for 1.1x-9x (technical data mentioned on this page is mostly from Bracegirdle, B., 1995: Scientific photomacrography, Royal Microscopical Society/Bios Scientific publishers, Oxford).

I also happen to have a 100 mm f/4.5 Leitz Summar lens (centre in the above picture). I did not include it in the present test because its optical condition is not as good as the other lenses, and its focal length and optimum magnification range (0.8x to 5x) are also quite different.

Olympus 20 mm f/3.5 (second from the right) is probably the newest of my lenses in this category (I purchased mine in the mid-70’s - it was first marketed in 1972). It gives good results, but is very sensitive to internal flare, which produces a large, very obvious washed-out central spot in pictures taken against even a moderately light background. For this reason, I modified the front of the lens by attaching a short lens shade (a longer and narrower shade can be threaded into this if needed). This reduces the already short working distance (about 19mm), but at least makes the lens usable. It is optimized for a 5x-12x (according to some sources, 4x-12x) magnification range.

In general, a frequent recommendation in the choice of a photomacrographic lens is to buy the most modern model one can afford (compatibly with the available choices at the desired focal length). The Leitz Photar, Zeiss Luminar and Macro Nikkor series are consistently mentioned as the best in absolute terms. However, slightly older series like the Leitz Milar are probably available in larger numbers and at much lower prices although not always in prime condition). Therefore, one of the purposes of this test is to verify whether the Leitz Milar 65 mm is a usable alternative to the Zeiss Luminar 63 mm.

To begin with, the above picture is a of a micropalaeontological subject (quite large fossil Foraminifera) taken at 1x with a Micro Nikkor 60 mm. Since I am more concerned about practical situations and subjects than theoretical resolution tests, the subject is a practical one, containing stacked and slightly oblique objects (pictures at higher magnifications are centred and focused on the large spines visible in the centre-right of the above picture). This subject also provides a high range of contrast, with a light subject against a mostly black background. This is likely to expose problems with flare and low contrast, as well as chromatic aberration. The same test subject, at different magnifications, is used throughout this test. If you take pictures of very flat subjects like IC chips, or other subjects with special properties, your tests may lead you to different conclusions than mine.

I used Nikon PB-6 bellows at maximum extension (209 mm flange-to-flange) for all photomacrographic lenses. I did take pictures at several apertures, but all tested lenses performed best with their diaphragms closed approximately half-way. This may well be a practical rule-of-thumb for using these lenses. Consequently, this page only shows pictures taken with the diaphragm closed half-way. Typically, pictures taken with the diaphragm fully open display flare, much reduced contrast and slightly reduced resolution. Pictures with the diaphragm fully closed display good contrast, but low resolution because of diffraction.

Throughout the test, an incandescent lamp was used for focusing, and a Nikon SB-800 flash in remote iTTL mode for the actual exposure (with manual exposure compensation if needed) . This eliminates vibration as a source of unsharpness. No lens shades were used, except those built into the lenses. It is entirely possible that using additional lens shades and/or placing the subject on a wide black background would improve contrast and reduce flare (my experience tells me that this is certainly the case, for instance, with the Olympus 20 mm).

As I argued here, photomacrographs do not show the same level of detail that you can expect in a landscape or close-up picture. Therefore, 1:1 crops of portions of these pictures are not very informative, and the picture as a whole is more indicative, even when reduced for publication of the Web. Therefore, here I provide only a reduced version of the whole frame. This page shows examples of 1:1 crops. Resolution in photomacrography is discussed on other pages of this site, both on theoretical grounds and as detailed tests.

The two above pictures were taken with the Zeiss Luminar 63 mm (A) and Leitz Milar 65 mm (B). There are some differences in colour (which could be an artifact of mixing incandescent and flash illumination) and, barely noticeably, in contrast, and the Milar shows a slight chromatic aberration if examined closely. The Luminar shows a higher depth-of-field because its diaphragm closes to a smaller minimum diameter, and therefore the half-way setting is not the same aperture in both lenses. However, as a whole, I would state the differences as minor, and certainly not as much as the difference in price would lead you to believe. If the price of a Luminar is too steep for you, you can get almost as good results (you will need to compare identical shots closely to see any differences) with a recent Milar in top condition. If you do this, watch out for molds or other contaminants on internal lens surfaces, for diaphragm rings hard to turn because of solidified grease, and for uncoated surfaces in really old Milar models. Also, many Milars have rusted screws that make it impossible to disassemble the lens for cleaning.

As tested here, the Luminar provides a magnification of 3.25x, the Milar 3.1x. Both lenses are being tested well within their design parameters, and probably quite close to their optimum magnification.

The two above pictures were taken with the Olympus 20 mm (A) and the Lomo 9x microscope objective (B). The Olympus provides a larger magnification (the Lomo has a higher focal length of about 28 mm). The Olympus lens provides an obviously higher resolution (focus was set on the long spine closest to the centre of the picture), while the Lomo produces relatively "fuzzy" images that display chromatic aberration even when reduced to a low size. However, the Lomo lens has a much higher contrast. Contrast can be increased in post-processing, while resolution can't (sharpening can be applied, but it cannot restore detail that is missing in the original image, while a RAW 24-bit image can have its contrast increased and then be converted to a 16-bit image, essentially without artefacts being introduced).

The Olympus lens has by far the lowest contrast of those tested here. This confirms my experience with this lens, which is extremely sensitive to flare. I bought a new specimen of this lens about 30 years ago, and it did have this problem right from the start. According to information on http://www.alanwood.net/photography/olympus/macro-lens-20-35.html, my specimen belongs to the very first series (single-coated). Multi-coated series became available later on, and they may have a better contrast.

The Lomo lens is the worst among those tested here, as far as resolution is concerned. However, considering that it is not designed as a photomacrographic lens and that it is quite cheap, it might be of some use in the lack of better equipment (as the saying goes, you can't take a picture with a lens you don't have).

As tested here, the Olympus provides a magnification of 16.25x, the Lomo 12.75x. It should be noted that the Olympus lens is being used here quite outside its optimum magnification range. The Lomo is also being pushed beyond its design limits, because it is designed for a microscope tube length (i.e., a distance from the mounting flange of the objective to the one of the ocular) of 160 mm. In addition, often microscope objectives by design do not correct all aberrations, and leave to lenses placed within the microscope tube or to the ocular the task of removing the remaining aberrations.

Conclusions

Some of the results of this test were expected, while others are both unexpected and interesting. Photomacrographic lenses, as a whole, perform better than equipment designed for other uses, including microscope objectives. This is hardly surprising. However, the fact that a relatively cheap and old Leitz Milar 65 mm performs, as a whole, like a more modern and much more expensive Luminar 63 mm (at least in this test) shows that there is still a practical use also for older lenses, at least when they are available in top condition.