Thanks again for your most recent reply sophiecentaur. It has actually been a help - I spent some time this evening thinking it through. I feel that you have some contradictory statements in there, but I suspect that just arises from my own limited understanding of some of the detail. Every time I dig into this stuff I uncover another layer of complexity and pretty soon I get well out of my depth...
Anyways, here's a couple of things that I thought seemed inconsistent.
You suggest that in fact there IS a downward force for any slice of atmosphere larger than a very thin slice. That implies a variation in air pressure over small scales, yet in the earlier thread, yoshtov claimed that there is almost no difference in air pressure over distances of 1 or 2 metres.
What is air pressure?
I guess 'almost' might be the operative word here? But 1 metre is a lot more than an 'infinitely thin slice'.
regardless, I still don't understand where this downwards force is coming from if the pressure responds to PV=nRT. A very small slice, or a larger slice, has a particular volume. According to my limited take on that equation, pressure is fixed on each surface, real or virtual. I still can't see weight in there anywhere.
I can see that gravity is pulling molecules down, hence on average density decreases with height. My 'n' is getting smaller, hence pressure reduces with altitude, or in reverse, pressure increases as we get lower. Again, this seems to be purely a function of volume, number of molecules and temperature. Gravity might on average increase my number of molecules, but where is this force we can call weight?
Next you suggest that for anything more than a very small slice of atmosphere, the molecules going down will go faster than those going up. That seems at odds with the idea of all molecules exerting the same pressure on a given surface area. Take my sealed container - the molecules respond to the fixed pressure (if n, V and T remain fixed) by bounding around energetically. If the ones going down have a greater force due to gravity, that suggests that a barometer in there will detect that effect. Yet it does not.
But then you seem to contradict yourself in the next sentence where you argue that for a thermally insulated container, the temperature will even out and all molecules will now go at the same speed over the whole column. But what happened to gravity? It was accelerating the molecules down a moment ago, but now with an even temperature profile it has no effect? What exactly did you mean here?
Your next paragraph actually gets right to the nub of my conceptual difficulty. You describe the effect of my container in space - a homogenous distribution of gas and pressure. On return to earth, we now have a pressure gradient. But this is what I said earlier - the effect of gravity is to increase density at lower levels, which introduces a pressure gradient. But that's not weight, at least it isn't what I understand by the term 'weight'.
Of course the molecules have 'weight', so too the parcel of air in our container. I am not arguing against that. But I mean weight in a Newtonian sense - that is, something I can measure as weight.
I am not convinced that the pressure we are measuring in my container is weight.