A lot may depend on factors like the lenses you actually have in hand for the formats you are shooting, and I wouldn't discount grain as a factor in what the final print is going to look like.
One thing to be aware of, if you've got a really good macro lens for 35mm is that it will cover larger formats at larger magnification, if you can find a way to mount it physically to the camera. At the same time, it may be the case that the lens is better optimized for the magnification ratios likely encountered with the primary format than with the format the lens has been adapted to (say 1x-5x on 35mm could be better than 8x-40x on 8x10").
Just run some tests and make some prints with what you have, and you'll see what you like. When I've had a particular macro project, that's what I've done, trying different lenses, reversed lenses, enlarging lenses, etc. at different reproduction ratios, and seeing what works. If you have a real macro lens for 35mm, it may look better than a non-macro lens for 4x5", at macro magnifications. You may find, for instance, that a lens that isn't a dedicated macro lens for the magnification ratio at hand introduces an unexpected distortion or an internal reflection or flare (adequately shading a reversed lens can be tricky) that produces a hot spot, negating any other advantage of the lens.
So yes, shooting at larger magnifications is better (but - depending on the difference in film format, of course - only marginally).
It means, however, you have to use more difficult to handle larger format equipment.
And that's what behind the advice to use 35 mm format.
But it is a clash of the two approaches i mentioned before: the first is the compositional/format driven one, which comes with the benefit of better image quality when you step up in format. The second is the magnification driven approach, in which (as long as the subject fits in the format at the desired magnification) favours the smaller format (for greater ease of use with the same image quality).
So what to do is something that has to be decided on a case by case basis.
On top of that, you get all those considerations David mentioned.
I'm trying to think if you'll have 8x narrower DOF at 8:1 or more than that. If it is 8x narrower will it be less obvious in a contact print than the required 8x enlargement of a 1:1 image? Is there a nonlinear factor which won't come out in the wash?
Grain and tonality will be better on the larger piece of film. If the other factors can be practically overcome then the larger format should win but can you buy or make an 8:1 setup for an 8x10 camera?
There isn't enough DoF anyway, so an 8x reduction of it may sound pretty impressive, 1/8th of nothing isn't significantly less.
You typically use very short lenses (i use 16 and 25 mm lenses mostly) when in-camera magnification gets beyond 3 or 4x life-size.
It's not just grain and tonality, but also that - because of the larger in-camera magnification - you resolve more subject detail. Enlarging an image recorded on a bit of film will not increase the amount of detail you see in the subject, just enlarges what you have captured on it at that lower magnification. Get to the same size magnification on film, and there will be more detail in the captured image.
May just have been 35mm defenders but the old adage was always that 8x10 lenses didn't need to be that sharp as they weren't enlarged 10-20x on printing like 35mm needed to be for anything more than the 4x6" snapshot developing plus prints option at the lab.
Which 16 and 25mm lenses cover 8x10" sheets and have good close focus characteristics?
Zeiss Luminars, Olympus bellow head lenses (they were/are 20 mm and 38 mm in focal length), Leitz Photar lenses. I'm sure Nikon and others make the same type 'micro'-lenses as well.
Mind you, they only cover large formats at high(ish) magnifications. But then, that's what they are made for.
Different focal lengths are not chosen to get a different field of view, but are optimized for different magnifications (or rather, ranges of magnifications), so are chosen to match the magnification you're after.
As a rule, the higher the magnification you are after, the shorter the lens that's best suited. (For instance: the 16 mm Zeiss Luminar is best for 10 - 40 x, for a bit less, 6 - 25 x, the 25 mm Luminar would be better.)
Same as enlarger lenses. Thanks.
Well, just to split my half hair, if the spheric seed is opaque, you cannot see from any angle more than half of it, so you only need half inch of depth of field :whistling:
The question of the OP I think can be generalized beyond the macro field. If you use a 200 mm lens, on whichever camera, to take a shot of a landscape, and there is a certain bell tower in this landscape, the absolute size of the bell tower on the film will be the same regardless of the dimension of the photogram. On an APS, 135, 120, or LF film, given a certain focal lenght, the dimension of the object is the same, and the DOF is the same.
We have the "impression" 200mm is more a tele in a small format because we don't get, and don't print, all what would have been "around" the bell tower.
So if you use a 200 mm with a 4"x5" and with a 135, if you only print a 24x36 portion of your 4"x5" film, you have the exact same results.
The same applies to macro photography. So, strictly speaking you need a larger format than 24 x 36 only if you need a 1:1 reproduction, on film, of a subject that is "bigger than 24 x 36". (Or you need LF because you need movements and it is easier with LF).
At the end of the day, you can always use LF with movements, take a LF picture, and only print a 24x36 portion of it, but you will have used the Scheimpflug law.
Post #3 seems to suggest that maybe the OP deems that the actual f/value of 44 has an influence on the DOF. If this is the case, I'd like to say that to my (quite imperfect) knowledge, it has not. The light fall caused by extension tubes, bellows, or teleconverters, is not accompanied by a correspondent DOF increase, nor by a correspondent diffraction increase, because DOF and diffraction, as far as I know, depend on diaphragm aperture.