Just my two pence here for rads designs:
1. I noticed that some car rads (the bigger ones) have shaped tanks i.e. they are 3D wedge shaped. I believe that is done to balance the pressure gradients, or rather get rid of it to prevent a situation where different tubes get different flow rates through. That's single pass cross flow rad I am talking here.
2. Performance peaking and then a drop-off is due to breaking the turbulent flow Re number and therefore expotentially(dunno the exact functional relationship as I write this) increasing restrictivness compared to flow rate increase. That's just my theory on that one.
3. I think Cathar or was in Maxxracer hit the nail in the head when wrote about dT in single pass rads.
4. What I personally gathered from various radiator testing results of other people was that with current line of thinking in rads design water side of it is rather insignificant (flow rates have next to none impact on performance as used in PC water cooling) and the air side of the equation is of the utmost importance and that's why P.160s are so good there from noise/performance point of view and commercially of course
