re: Crossed Polars
Light can be described as two sine waves orthogonal to one another and 180 deg out of phase.
Once the initial experiment is properly tuned, one lens removes the horizontal polarization plane vector, the second lens removes the vertical component. Adding a third, randomly oriented lens in between the two should make no difference, it doesn't add anything to the wave, it can only remove something from it.
I don't doubt what you're saying is true; I've performed the experiment myself, I just can't figure out how it works.