Afocal photomicrography projects the microscope image through the camera's lens before reaching the film or more recently the digital camera sensor. The technique is popular with fixed lens digital cameras that have been found to be well matched to a microscope eyepiece, notably the now dated Nikon 9xx Coolpix series (which rather handily had a thread that matched some Leitz microscope eyepieces), but other 'small bore' fixed lens digicams have been found to work well e.g. the Sony W7, P200 and some Canon models.
It isn't usually the first choice for cameras with removable lenses e.g. film/digital SLRs or the recent mirrorless digital cameras, as at best the additional glass elements of a lens (typically six for a 50mm prime) may have no detectable deleterious effect but at worst can degrade the image or be a bad match causing severe vignetting. But using the camera body only may have its own problems—the projection eyepiece has to ideally closely match the sensor / film size as well as still fully meeting the requirements of the objective if it requires correction at the eyepiece. Suitable eyepieces are becoming highly sought after and attracting prices sometimes equalling or well beyond a camera body price, notably the Olympus NFK 1.25X which is reported to be a good choice for full frame DSLRs.
Left above: Older setup on a Zeiss Photomicroscope III using the thinner vertical tube.
Right above: Newer setup using the more robust focussing photo tube. Both with camera body only, no lens.
Each uses the same third party 35mm photomicro' adapter (this example a Meiji). These are lensless two part adaptors and has a Zeiss 10X Kpl W eyepiece (not a projection design) on a short collar ca. 5mm to project a real image and is exactly parfocal with bino' head view. The adaptors are potentially very robust as clamp to eyepiece tube, but typically have the internal metal 'C' collars secured by single screws. These fixings may flex no matter how tightly screwed, but careful shimming (Sellotape often sufficient) makes them very stiff. The current camera used, a Canon 600D DSLR body, has Electronic First Curtain Shutter so no mechanical movement occurs at beginning of and during exposure in Live View mode.
My preferred method of photomicrography with the setup shown above uses the normal microscope eyepiece raised on a short collar to give a real image exactly parfocal with the visual field. The collar is short enough to have no practical effect on image quality caused by tube length change (supported by Loveland's detailed studies in his book 'Photomicrography: A Treatise') and gives a high quality flatfield aberration free image. The potential downside of this setup is that only about 2/3rds of the field of a perfectly matched relay lens is captured, although partly a benefit as lateral chromatic aberrations does show if extend from an APS sensor to full frame. Most of the time I don't find the smaller field a problem and useful for small subjects that don't fill the eyepiece view e.g. small pond critters but occasionally capturing a subject that nearly fills the field is awkward e.g. circular plant stems or 
insects parts, without image stitching.
I've been keen to try the afocal technique as a potential route for widening the captured field when required as some enthusiasts have reported good results with removable lens DSLRs. A recent thread in the Yahoo microscopes forum highlighted a cost effective adaptor on eBay UK (seller ID 'joy8899'); it's actually sold for 1.25" telescope eyepieces but many microscope eyepieces can approach this in size, particularly modern examples or those for a stereo.
The adapter is sold in seven filter thread sizes from 28 to 58mm inclusive for £9.99. I opted for 52mm as most of my Nikkor lenses or camera adaptors have this thread. It is shown right fitted to a Zeiss Kpl 8X eyepiece and has a three point screw fixing. To bulk out the adaptor to clamp it to this or other eyepieces tried, one or two split collars of thick card from a Tesco aluminium foil roll were used. This also had the benefit of protecting the eyepiece from damage by the bare metal flat-ended screws.
The adapter potentially presents the eyepiece very close to the front element of the lens in use so also added a foam annular pad to all eyepieces tried just in case it slipped.
Fixed lens cameras that didn't work
All cameras available in the Walker household were tried, except a Sony P200 which is a small bore lens camera with no filter thread.
Sony S75 (left above) One of the older 3.3 Mpixel digicams but with a high quality Zeiss 3X zoom lens. I've home converted the camera to near IR only by removing the IR filter above sensor. This style of 'wider bore' fixed lens digicam I have never found easy to use on a microscope, a reversed 50mm lens or a dedicated LM-Scope adapter being best but both still give vignetting. The Sony lens adapter (shown fitted above) is for 52mm filters so suited the afocal adaptor. Howver, using the adapter with either a Zeiss 8X or 10X Kpl eyepiece gave marked vignetting on the microscope at all zoom levels.
Fujifilm X20 (right above) One of the new breed of Fujfilm's superbly crafted 'retro' style cameras with a very high spec. This belongs to my brother Ian and it's a delight for normal work with a fixed 28-112mm equivalent zoom lens which is fast throughout the zoom range at f2-2.8. With a new design of non-Bayer RGB array ('X-Trans CMOS II') and no anti-aliasing filter on the large (for a compact) 2/3rd inch sensor, we had high hopes for it for photomicrography.
The lens adapter which screws on lens front is for 52mm filters so was also suited to the afocal adapter to present the eyepiece (scarily) close to the front element which does not move on zooming. But despite the fast lens, wide zoom range, manual focus to try both infinity and super macro with both the Zeiss 8x and 10x Kpl eyepieces, vignetting was too severe on the microscope to be practical.
Eyepieces that were suited mechanically
Zeiss 8X Kpl (O.D. 25mm, eyepoint ca. ) - two card collars (each 1mm thick) required.
Zeiss 10X Kpl W (O.D. 28mm, eyepoint ca. ) - one card collar required.
Meiji 10X (O.D. ca. 33mm, for EMZ1 stereo) - one card collar to prevent eyepiece damage.
Eyepieces that were not suited mechanically
Leica 10X/23 (O.D. 32.3mm focussing, cat. no.10447131) - the standard eyepiece on my Leica S8 stereo. The O.D. is OK but does not extend down far enough into the adapter to be securely clamped. The eyepiece tubes are also sprung and unsuited for any weight on them.
Removable lens cameras that did work
Nikkormat FT3 35mm SLR film camera (left above with Nikkor 50mm f2 lens). Potentially works. This was duly rescued from the bowels of a cupboard as it's the only full frame 35mm camera I have (funds haven't yet permitted a digital full frame DSLR and don't have a suitable projection eyepiece for one that can afford). I didn't run a film through it but did quickly assess with piece of tracing paper at the film plane, shown right using the Nikkor 50mm f2 lens. This is what I regard as acceptable vignetting caused by edge of eyepiece field stop, much of the film area is used and in the x-plane it allows a crop of an image right to the edge of eyepiece visual field. The demanding micrometer as a subject looked aberration free and flat but would need proper tests running film through to confirm.
As with many older 35mm film cameras this is a heavy body and with 50/2 lens weighs 920g. I wasn't very comfortable having all this weight resting on a friction three point fixing even in a vertical photo port. The unit wobbles on the small supporting area of eyepiece in a vertical tube so great care would be needed to avoid camera vibration—especially with film camera with mirror and mechanical shutter. The FT3 does have a mirror lock but the coarse viewfinder is not ideally suited for critical focus (the clear central split image does give some area for focus).
Canon 600D digital SLR, (right above). This has an APS sensor and I use with my legacy Nikkormat lenses using a Canon-Nikon adapter.
Best lens: The Nikkor 50mm f2 AI shown on camera was compared to the Nikkor Series E 50mm f1.8 (shown to right of Canon DSLR above). The former has the front element well recessed into the mount, the latter is more forward facing. Both gave vignette free imaging and good field coverage but the series E had a noticeable hotspot in centre of image which can be tricky to remove. The 50mm f2 did not show this.
Other lenses tried: A Nikkor Series E 100mm f2.8. This gave marked vignetting. I do have a Nikkor Series E 75-150mm zoom lens but regarded this to be too heavy and optically complex cf. a lighter simpler designed prime lens. Many of the modern prime and zoom lenses are much lighter than the old lenses so may be worth trying if owned.
All lenses were set at infinity focus and full aperture.
Vertical eyepiece use
The camera / lens / adapter is most safely used on a vertical camera port to avoid off-axis strain. The system of choice the Canon 600D / Nikkor 50mm f2 lens was 780g and the card collars I used were not ideal with the screws needed regular checking and retightening. A stiff non-deformable plastic collar may be better.
Angled eyepiece use
I wasn't at all confident on using the adapter with either a digital or a film camera with lens on an angled eyepiece tube e.g. the Zeiss bino' head or a Meiji EMZ1 stereo bino head. The weight of the body and lens combination may strain the eyepiece tube and the card collar I used, even with screws very tight, did not reliably hold the system on axis. A stiffer and better internal padding from plastic or metal by the DIY'er may make it more suitable if the bino' head tubes are robust enough.
Conclusions on my set-up
The afocal adapter is a very cost effective adapter if find a good combination as I did with the Canon DSLR / Nikkor 50mm f2 lens / Zeiss Kpl W 10X eyepiece. But the slight increase in field of view over my present system was negated by its lack of rigidity and safety cf the third party 35mm photomicro' adapter currently used which securely clamps to the eyepiece tube not the eyepiece. The latter adaptors are widely available and cost effective allowing an eyepiece of choice to be used with the camera body only. The typical 'C' collars and single clamp screws used in the latter adaptor, although in their own way flawed, can be carefully shimmed for a very rigid system, as have done with mine.
I would be wary of using the afocal adapter for film work with an old heavy camera with the present crude card collars, it is not rigid enough. Better designed collars possibly tailor made to a given eyepiece may improve this, but the camera / lens system still relies on the rigidity of an eyepiece friction fitted into an eyepiece tube.
Afocal adapter Pros:
- very good value and well made
- with the right lens / camera / eyepiece combination, a potentially effective way of capturing a good area of visual field and with good quality
- if above satisfied, potentially save expense on a dedicated relay lens
- correct microscope eyepiece maintained to ensure full correction of microscope optics
Cons:
- design does not give the high rigidity often needed for critical work, extra care needed for vibration reduction
- design best suited for a vertical tube if a heavy camera body / lens combination
- three point fixings are metal and may damage microscope tube / eyepiece if firmly clamped, internal collar required
- great care needed to avoid the eyepiece slipping and damaging the front of camera lens, protection e.g. sponge pad on eyepiece advisable just in case
- (of afocal as a technique): very hit and miss, 'try it and see' only certain way of assessing if a given combination of camera / lens / eyepiece will work
Comments to the author David Walker are welcomed.
Footnote: I have no affiliation with the eBay UK seller of the adapter. Other sources of comparable adapters may be available.