But, if I'm looking at your second drawing correctly, there is only one return tube opening, right?

Well, my thought is, if you keep the pregnant females and fry in the tanks that are in the middle of the operation, they would be safest, correct?

And, since your water conditions would be consistant  due to the "shared" water system, you would have an easier time of moving them if necessary.

I'd get it set up and run it.  Fool around with the valves and returns.  Maybe add some fish food, or something similar, into the water, and see how the flow operation works.

Sometimes questions from the least experienced systems person might bring up necessary dumb questions.

The references alluded to below are by people much more experienced in this stuff than I. Just grist for the mill.

I'll bet a google search for "aquarium flow through systems" could have a person reading until at least next week. ;)

Would the water pump be overworked (burn out) if too much back pressure was put on it by partially blocking the flow? I've seen a smaller set-up where a power head feeds into a reservoir which in turn, by way of a bunch of smaller tubes fed into fry tanks. Through a small gage stainless steel mesh those tanks then drained down into the tank below the whole operation.

When the pressure got too great, a couple of extra hoses were set in the reservoir which bypassed the tanks and led to the overflow.

Another larger system had a water faucet type valve which could just bleed off extra water to be run through another pipe back to the central reservoir.

Your foam would be an effective fry block, but would it be too efficient as a mechanical filter and clog up? Would one have to seek a really course foam so holes would be bigger?

I recall an arrangement similar to your second graphic where the builder put an elbo in the pipe. As with the case about, they cut fine mesh stainless steel (available in the older hardware store sometimes), placed the mesh on the plastic pipe OUTDOORS in a very well ventilated place and applied just enough of a solvent that it melted a little of the pipe and as the temporarilly melted plastic reset, the mesh was tightly set in the end of the pipe.

Automatic siphons can also be contructed for tanks if the water supply is not always consistant. See
http://www.aka.org/pages/libary/flow_through.html

Here's another version of the sponge plug:
http://www.aka.org/pages/libary/fry_flow-through.html

Here a system using the steel mesh:
http://chika.aka.org/library/flowfry/flowfry.htm