Eddy Fernandez [moderator of the ARF list, University of North Texas] rightly points out that when you create a lot of different primary reinforcer associations with the clicker, you strengthen the clicker's reinforcing power, and you mitigate the effects of satiation for any one reinforcer. I'd add to that the reinforcing effects of cues that were established with positive reinforcement, and the self-reinforcing aspects of such behaviors as tracking, running, jumping, succeeding at difficult tasks, etc. Eddy was hesitating, as I see it, over the possibility that using these very powerful reinforcers could allow you to develop animals that would work long past the humane level. He mentioned pigeons on long schedules that actually crack their beaks, pecking. [This of course was not developed by some superb sort of pigeon feed, but by shaping for longer and longer schedules, and probably in a food-deprived bird.]
Of course for thousands of years people have, with aversives, been working animals past their point of endurance, whether it is pack animals carrying loads, oxen grinding grain by turning a mill, dogs turning spits over the fire, or whatever. So that's always been a possibility. For me, the value of a panoply of reinforcers is that I can get what I need from my animals even in stressful circumstances. In a totally new environment among strangers my kitten will still immediately target, jump, and recall without hesitation (YAY MIMI!).
If on some rare occasion I DO need extraordinary endurance, I can ask for it and get more than I would have imagined possible. This has come up for me at photo shoots. Twice now my terrier Twitchett has worked continuously in excess of three hours, for tiny (not particularly favored) treats of beef jerky, learning new poses and actions and delivering them over and over and over for a camera crew until even the crew was wiped out.
I don't see that as manipulation or oppression, but as cooperation and delivery beyond the call of duty in a pinch; on a daily basis, I wouldn't ask for that. Of course we do ask for exhausting levels of unreinforced behavior for long periods of time from human beings; it's called graduate school.