Whoa! This is counterintuitive. I've recently had a renewed interest in divided developers owing to their ability to linearize the sensitometric curve at the low negative density end. This seems to clearly provide for greater discrimination of detail in the shadow region of the print. Maybe I've not grasped the suttleness of your comment above. Perhaps you are implying that the lower gamma in the tail of the sensitometric curve (the first derivative of that curve function) when taken a second time yields a steep slope and thus a greater shadow discrimination than would be intuitively expected. Interesting thought but I'm not sure I'm grasping your intent very clearly.
Nate Potter, Austin TX.