I've seen SBC irrigation pump engines with exhaust manifolds made from heavy square box tubing. They run at a fixed speed, maybe 3500 rpm?

Kevin

Wow, maybe because, like you've hinted, the working rpm is relatively low ?
I'm pretty sure if there was any benefit to running square sectioned exhausts the top racing teams would using them now. I know someone who used a guy who made exhausts for F1 cars, they were all round tube Inconel stuff. Real nice work but a bit expensive for me
Intake manifold design could be different, but currently F1 uses round tube runners from the plenum to the head. They do flare out to an elongated shape at the port due to the head being 4v / cylinder but surely if there was any sort of power advantage with a square design they would be using it ? They spend a lot of money developing stuff to get an edge I'm sure they would have tried it at some point.

i too, have seen square tubing used on exhaust "headers", only this time, on an emergency generator at the machine shop i worked at for almost 50 years in my working career as a machinist/midnight shift supervisor.
this was an engine that ran on natural gas, [i forget what brand the engine was] and had a section of square tubing bolted on the side of the engine where the exhaust ports were. it was an inline, but i don't remember if it was a 4 or a 6 cylinder.
the square tubing had a round section of what looked like black pipe that went to a threaded 90 degree elbow that pointed up, then about two feet above the elbow, there was a large diameter muffler [maybe three feet long or longer] then another section of straight pipe that went through the roof of the shop, capped by a "flapper valve" like used on a farm tractor.

The hard to find Max Wedge exhaust manifolds are square. Street Hemi are square, I think some Ford and Pontiac are too, 50s and 60s stuff.

Yes, that is more apples to apples. With a non-compressible fluid, round wins out by simplify having jess wetted surface area, ie friction. With a gas, all the other questions mentioned earlier and more come into play, IMO.
Taking this consideration further, can a rectangle ever have equal flow or exceed flow of a square with equal cross-sectional area, based on liquids, No.