The idea of this section is to try and give you an idea about what has been done in the field. It may help with deciding what to do.
Cost is not an issue - the trinocular head
The trinocular head replaces the standard binocular with a more complex beam-splitter that creates a third optical path that feeds the image to a camera sturdily mounted 'somewhere at the back'. If you are working at the microscope all day, and need instant photography ready for when the hoped-for animal suddenly swims into your field of view, there is really no other choice. The entire set-up is bought from the best manufacturer in the world, and yes, buyers are high corporate, high academic, or eccentric millionaires. Unless you really have tuned things so the camera synchronises focus with your eyepieces, the camera, out of reach at the back, has to be focused. If it is a digital camera, the image is transmitted as high speed, high cost video to a monitor screen, probably via a dedicated computer with the manufacturer's interface card inside. Without getting up, you optimise focus on the monitor, and press the camera button when the hoped-for animal smiles.
The downside to this charmed environment can be seen at present in a number of these incredible installations that are on the secondhand market about ten years after purchase. The mechanics and optics are still state of the art, but the camera is film-based, possibly even Polaroid, which is foundering, and there is no way to update the camera. If you buy a mechanical/optical/electronic mix, be prepared to do it again in ten years time when the electronics are out of date. Cost really has to be no object to take this path. Note you will also require a large, dedicated workbench to carry the gear, and possibly a temperature-stabilised room to put the workbench in. Space considerations apply equally if ordinary mortals take out a bank loan and buy one of these installations secondhand. Remember you can't put it in a garage, and you never buy a bargain from a dealer. What you might get in a big package, are exotic objectives, that may be unbuyable by any other means.
Cost is fairly high - the manufacturer's camera adaptor
The 25 year old Wild M20 that we use all day, every day, came third hand of course, with a Wild SLR adaptor. The binocular housing clamps on with an inverted cone fastening, and is removed to drop in the inverted cone of the camera adaptor. Things go downhill from there, with a strange shutter, and a 'spreader lens' to size the image out so it fills a 35 mm film frame properly. An SLR adaptor, unfortunately not for Pentax, is variable for height to set up for the SLR in question. Even set up properly, with an old Miranda which appears to focus in the right place, you are twisting your neck, peering into the near dark of the image, and hoping all will be right with the focus and exposure. It really is a blood sport, but the thing that impressed was the absolute rigidity of the inverted cone mounting. Our camera is not adjustable for height, but fixes to an adaptor to replace a given binocular head via a custom-machined inverted cone for precise centring. The adaptor height puts the sensor image plane 160mm above the microscope objective, for the one model for which the adaptor was designed. The adaptor costs over two hundred dollars Australian and cost varies with the trickiness of the job.
If the hoped-for animal comes into view, some fast footwork is required to swap the camera for the binocular head, but systems of this sort still give you a chance.
Cost is minor - eyepiece tube and eyepiece adaptors
If you like to get things going fast and cheaply, and run uphill for the rest of your life, a manufacturer's eyepiece adaptor is for you. The eyepiece tube, and the twin eyepiece distance varying mechanism, are not designed to carry the weight of a camera. They will probably not break, but they may bend and will certainly wiggle, as well as never being in the same place twice. Polaroid did a cute little camera that actually slid in to replace an eyepiece, at a couple of thousand US dollars. I'm no artist, but I'd rather draw pictures of what I see than try and get consistent working out of devices of this ilk. Kodak now has an adaptor that does not look to me a great deal better, with a Kodak digital camera (lens left in) followed by an additional lens to allow the image to focus, followed by a tube adaptor. It may be OK.
Buy a 'Scientific' microscope camera and trust the manufacturer.
I did a Google search almost a year ago looking for possible affordable scientific digital microscope cameras. Prices then started at 5000 US dollars for the 'cheap one', going up to about $15000. Cameras of this variety terminate in a C-mount thread, so all you have to do is find someone who will adapt them to your microscope, or buy a pre-made adaptor from the camera maker. Apart from the previously mentioned Pixera, using a webcam VGA chip in this sort of price range, genuine 1.3 Megapixel chips were offered. Exotic circuitry, hence the cost, took out high speed video into an industrial-strength video card (priced extra at about $1500) and out to a computer monitor hooked to your own fast computer. Assuming you have used an inverted-cone adaptor (see above) and your sensor is at the right place, you can do comfortable, serious work. The fast video link allows good viewing and focusing. Nice if you can afford it, but we could not.
You form your own opinions, but I will set down here the conclusions that we came to.
Domestic space considerations, as well as the soaring cost, ruled out the digital camera/video link/computer/monitor arrangement. You try keeping that ready for a month.
An inverted-cone adaptor replacing the binocular head was the only system rigid enough for reproducible use. For a camera-only system, a really swish adaptor could be machined up to give equal rigidity to an older monocular microscope head. For use on either, the camera would have to weigh not much more than a binocular head.
The camera would use the internals of a cheap genuine 1.3 Megapixel camera, and would be optically enhanced so viewing and focusing on the inbuilt LCD display would be fast and positive in the hands of a skilled user. Camera memory would be raised to allow at least 100 high-resolution JPEGs to be stored, and single use or rechargeable batteries would allow at least a continuous afternoon's shooting. Used in conjunction with the ledlyt, the system should need no Mains supply.
The camera should allow dust cleaning from the now exposed imaging chip. The microscope adaptor should be easily removable, leaving a simply-machined interface that could take all manner of lenses, telescope adaptors etc, that could be used for technical photography in conjunction with the hot mirror filter. The camera case should itself be boltable to any handy object.
The camera should be saleable, all-up, at less than 1500 dollars Australian.