The Epis series, like most episcope lenses, is a Cooke triplet design. Usually in these (like also the TTH Cooke Series II), the aperture is positioned close to the negative center element, on its rear side. That's also true in descendents of the triplet design which have doublets for the front and/or back positives (Heliar, Saphir etc.).
Triplet designs are in general not completely symmetrical, and they are calculated in a way that the negative center element is slightly stronger than the front positive, so the beam is slightly diverging behind the center element.
Apertures are usually placed in the optical center of a lens, i.e. the point where the chief rays of all pencils of rays intersect. Doing it this way makes so that every pencil of rays is affected the same way by the aperture (and it doesn't cause vignetting). Placing the aperture outside the optical center usually isn't very much of a problem though. You will have two effects: introduce vignetting in pencils that come in from the extremest angles, and for out-of-focus objects "moving the cutout of the bokeh ball inwards or outwards from the axis", therefore affecting the distortion characteristics of the lens.
When only having the aperture slightly off center (i.e.: directly behind the negative element), the effect will be minimal. Historically though, aperture positioning has in fact been used artistically (cf.: Eder, Ausführliches Handbuch der Photographie 1,4: Die photographischen Objektive. - chapters on the simple lens and the Petzval).
https://digital.slub-dresden.de/werkansicht/dlf/449451/52?tx_dlf_navigation[controller]=Navigation&tx_dlf_tableofcontents[action]=main&tx_dlf_tableofcontents[controller]=TableOfContents&cHash=516a4004ec135b499b0cce24d375c4d5
For the lens in question here:
- you're probably quite lucky you have a barrel design that can be screwed into two parts. The two Epis 1000mm I have (an f/3.5 and an f/4.8) have the helicoid tube design, they can't be screwed apart (only taking out the elements by screwing out their retaining rings).
- the argument that "the opens we make in the waterhouse stops may not be accurate given the calculations we make" is bs. Simply write down a
ray transfer matrix analysis for that system. I.e.: Measure the focal length of the front element and the negative middle element roughly, the distance between front and middle, and the distance between middle and desired aperture position. Simplify the system matrix as: (free propagation between middle element and aperture)*(middle element as thin lens)*(free propagation between front and middle element)*(front element as thin lens). Then you're inputting a ray of (x0, 0), calculating what the x1 at the aperture is. Then plug in an aperture number N=f/x0 into that, and you find an equation for x1 (in terms of a diameter then) at any aperture number N. Not hard, just half an hour of actually writing down the math.
You could also simplify further, using just (free propagation between front and middle element)*(front element as thin lens) as matrix (would save you the hassle of measuring the negative element).
- because your barrel can be screwed apart: you could, instead of cutting the barrel for Waterhouse, go the route of washer stops instead (essentially: 3D print a set of half-cylinder/can-shaped parts that you can insert directly into the barrel that include different apertures)
- or: you could even experiment with just putting the aperture in front or behind the lens, depending on the format you want to shoot. Either just essentially cardboard pieces with aperture holes or 3D printed hollow cylinders, or these airflow aperture irisses like this that you can find at any hardware store:
https://www.amazon.de/-/en/Airflow-Regulator-Shutter-Stopper-Exhaust/dp/B084LJBKQ8?th=1
I think the last option would be the least hassle. If it's too much vignetting, you can still go the more
Much success,
Ferdinand