Quote Originally Posted by MattKing View Post
For low light reciprocity failure, there isn't something magical about a long exposure that somehow affects the film in a mysterious way. What happens is that the light hitting the film has so little energy/intensity that the film doesn't respond as strongly to it - there is a threshold light intensity at the film plane. Below that threshold, the film responds much less to light than would otherwise be predicted from extrapolating the straight line portion of the film's curve. So with light at the film plane at these low levels, exposing the film, for example, for twice as long gives you a result that is actually twice a lesser result than what one would otherwise expect.
I came up with this analogy: Imagine that film is like a bucket, and light is like water. Normally, when you click the shutter, the right amount of water flows into the bucket. When the light is low, the water trickles or drips into the bucket, and its rate is so slow that some of that water evaporates out. So to get to the proper "full" mark, it takes more time to fill the bucket.

(And for making photographs like Michael Kenna, scene selection is important. One time I photographed a landscape by moonlight, and Kodak E100S looked like daylight after 15 minute exposure at f/2.8. I'm sure Mr. Kenna made lots of test exposures to get the appropriate reciprocity data for his needs.)