Peter, as I experiment with Puget sound sunlight (actually rain) for
exposing kallitypes, I have found that overcast sky still makes an
image---it takes longer, of course, but the image seems to have about a
half-paper grade more contrast. If the short-length UV (200-300nm) is
selectively absorbed by the clouds, but longer-wave (in the 350-500 nm
range, where ferric compounds are most 'absorptive'), this would explain
why the image is still formed, though the intensity of the light is
diminished. I have no idea why the contrast is different, perhaps one
of the expers on the list will explain.
--Ter