Chuck,
I think most of the optics industry has forgotten that when Köhler and Leinhos of Zeiss published their results they specifically stated that the formula usually referred to now as the 'twilight factor' only applied to the range of leight levels between 0.1 and 0.003 stilb or cd/m2 in SI units. That means moonlight to most of us. At higher light levels magnification is the predominant factor and below that exit pupil.I think most birders would be on their way home well befor the twilight factor formula applies.
Although they didn't go beyond 15x, they found that at levels above the TF light level range and what we might call birding twilight, a 3.3mm, 2.3mm and 1.6mm exit pupils made no difference to the average apparent acuity in their study. Certainly their work predicts that an 18x56 would show more detail than a 15x56 for our kind of use.
Berek from Leitz (Leica) took a rather different approach. He looked at the threshold of target detection. My understanding is that magnification is significantly more important at moonlight levels and beyond than for apparent acuity, but again in birding twilight magnification in the primary determinant. It's a very long time since I had a look at the US studies, but as I recall they came up with very similar findings to Zeiss but produced much more complex formulae.
There appear to be a couple of things going on here. In normal daylight the detail limiting factor is optical resolution of the lens complex of the eye. It seems the potential optical advantage of a dilated pupil is normally offset by increasing aberrations. So, on average, the acuity is more or less constant over much of the daylight (photopic) range. It is at the onset of twilight where the eye starts to switch between daylight colour vision and night monochromatic vision that the retinal illuminance increasingly determines not only the level of detail you can see but also the threshold of contrast detection and starts to explain Köhler and Berek's results. I should point out that all these studies were probably conducted on conscription age men and our results mat differ. Holger factored in age in his paper on the subject.
Just to add a little twist to this story. All those studies were done on fixed black and white inert targets, but is that how things always work? I've had a just a few attempts at trying to see how magnification and exit pupil work for me. I've seen enough to be satisfied that both the acuity and contrast approaches work more or less as predicted by those studies for me. However on a couple of occasions animals have appeared in the field. A cat hunting a mouse at the bottom of my garden on one occasion,and a fox hunting a rat on the far side of the field out the back on the other. It was absolutely obvious that both the threshold for detection and the level of visible detail increased dramatically, when I detected movement, and magnification was the main determinant deep at those moonlight light levels. We know our brains exercise some control over the way the eye functions and most certainly how the incoming information is processed. Just seems it might shift up a gear when predators or prey are involved. I've found no relevant publications on this subject, but at least in principle it might make evolutionary sense.
Cheers,
David