On what scientific principle are you basing your assertion that it takes at least X units of oxygen to cause Y amount of damage to developer? The typical developer has several ingredients, some in quite small quantities. If the oxygen were to "attack" only the smallest essential ingredient, I see no reason it would take more than a little oxygen to kibosh the developer's effectiveness.
EricM, I think your challenge to me demonstrates the importance of me getting out the calculator before I start making theories. If I do a little figuring I now estimate that a half liter of air contains about 1.2 grams of oxygen whereas the half liter of developer would contain in the order of magnitude of 2 mg of dissolved oxygen. So it would seen that I am wrong on two counts.
Now if we take D-76 developer as an example, you are indeed correct as it would contain about 1 gram of Metal and 2.5 grams of Hydroquinone per half liter. So you are 100% correct that the developer would contain critical components in almost the same portion as the oxygen. I'm sure the molar quantities are somewhat different but the important thing is that they would be in the same order of magnitude. I think I'm not only wrong, but dead wrong. It could indeed be the air that is the problem.