I had some of this same foam, so I looked at it with the BD epi system, with much more interference visible. First the pictures, then some discussion.
5x brightfield
5x darkfield and cross-eyed stereo
10x brightfield
10x darkfield
For the technical discussion, we have to start by understanding that open-cell foam is like soap bubbles that were "frozen solid", so are basically air bubbles with thin walls separating them. Of course, it isn't soap (a liquid crystal), but rather a polymer that solidified, with stresses, strains and stretches due to not completely uniform cooling and solidifying. I think they call it "open-cell" foam because it is mostly the "connective tissue", rather than a large number of cell walls that are present. Especially in 3D, you can see that there aren't too many walls intact.
The first thing you may notice is that certain regions have colors prominent in BF, while DF is in other places. Also, in many instances DF seems to be more vibrant. I'm pretty sure this has to do with the angles of incidence, with more reflectivity at higher angles, as well as simply having the light make it into the aperture at different illumination angles.
So, the question becomes, if we get interference fringes from the bubble walls, why does epi have apparently such an advantage in seeing the colors? Some may be attributable to Photoshop post-processing - I'm not sure what James did in that regard. But even with no difference there, the physics says episcopic illumination has an advantage here.
My little cartoon shows diascopic on the top, with the light coming from below, and episcopic on the bottom panel, with light coming from above. At first glance, they look fairly similar. The difference lies in the fringe contrast, which is maximum for two-beam interference when the intensities of the two beams are equal.
If we make the assumption that 5% is reflected and 95% transmitted (see Fresnel equations), what we have is approximately:
Diascopic: 5% loss at the bottom surface and 5% of 95% at the top surface. So, the black output beam is 90.25% of the incident light.
The red beam is 95% of 5% of 5% of 95% or 0.225625%. This is a terrible mismatch!
Episcopic: The black beam is simply 5%.
The red beam is 95% of 5% of 95%, or 4.5125%. This is almost a perfect match!
There are many other factors, angles of incidence, polarization, indices of refraction, diffraction and scattering, absorption, etc... but this gets the basic idea.
I'm guessing many people here already knew this, but I hope a couple of people were enlightened (and endarkened... interference after all
).I should also point out that I've now put the larger versions, including parallel stereo and rocking on my site.
Mike
