Oh, okay, you're talking column loading instead of beam loading. Still a fairly easy calculation (though involved) for one member. Once you get the full assembly and need to run the math on that, well, it's more complicated and that's where I let the computer do the crunching.
Check the following link out: . Once welded up into a structure, you will effectively have fixed-fixed end conditions, so your "effective length", Le, will be 0.5*L.
You'll need to work through that page, there are different formulas to use depending on the "slenderness ratio" versus a "critical slenderness ratio". When finding the critical slenderness ratio, σy is the yield stress of your column material. Note that I had an error in my formula for I, the area moment of inertia, in my last post, which I have gone back and fixed. You'll have to do a couple of iterative runs to get an idea of the load/length. My hunch is that you will end up as a "short column" setup, which uses Johnson's formula on that page, but you may be able to get away with Euler's after all.
I also found a decent-looking set of lecture notes on this topic that may be of help (it's been a while since I've looked at this stuff. Definitely a use it or lose it type of thing, though it's slowly coming back to me):
Or, yea, run the real-life test