Rick, I thought I covered it, but maybe I got sidetracked in another post. Anyway, here goes. We were talking about forcing dormancy(downard trending) and reducing moisture, temperature and light and CO2 was an uncontrolled constant value. When shortening the light (daytime) period, my practice is to always cut the daytime period short at the end of that particular day. My thinking is that this gives the plant plenty enough nightime period to complete it's functions of utilizing all of the food and energy it has made and stored that 'last long day' as well as maybe do some self cleansing of occluded or trapped unused CO2 and other impurities. The start up time for the next daytime period remains the same to start the new daytime-nightime schedule of decreased light hours. If I were to lengthen a light (daytime) period, such as with seedlings growing bigger (upward trending), I would add light to the beginning of the first longer day allowing the plants more time to produce food and energy to begin that new regiman. I'm not talking about light-dark reactions of the Calvin-Benson cycle--only daytime/nightime periods.
I remember you saying you thought it shouldn't make a difference what end light was removed or added, and it most probably doesn't. But that's how and why I do it--if nothing else, for consistency. Della mentioned adding supplimental light in the morning, then natural light during the day with more supplimental light at the end of the day. I don't think I like that idea. I would add it all in the morning only and then at some point in the future go to all supplimental lighting.