After further research -
A : the drawing showing the cooling airflow is wrong , it is pushed through and out just as it is described in the text .
B : Apparently , The A was fitted with a fighting compartment heater by stealing some heated air from one or both radiators on the left , with the fan blowing DOWN .( so the hot air is coming from the rectangular-grated volume (s ) )
Evidently , the ducting for this heated air interfered with the normal cooling air flow on that side of the engine where it had to squeeze past - within an already super-sealed , over-cramped and overheating environment -- so they shrouded that sides exhaust piping and took some of the supply air from that fan and pumped it around and out the back with the exhaust .
The introduction of the cooling tower for the G with necessary reversal of the fan to blow UP and I ASSUME the change in ducting / piping made the forced air cooling of that side's manifold unnecessary .
The engine compartment and the fan/radiator compartments are water-tight , fully separated bays - so that's the only way a crew heater using discharge air from the radiators could have an impact on the temperature of the exhaust manifold . ducting .
The fans always blew DOWN except at some point early , with the D for sure , when one fan at least was reversed to blow up to suck flammable vapors out of the engine compartment .
Then they fixed their fuel leaks and returned to the nominal both fans blowing DOWN for the A and the G - until the G was fitted with the heater tower ( or "lump " as Robin calls it ) and then that side had it's fan reversed to blow UP .
.. . and the ducting from the new tower didn't interfere with the manifold cooling on that side OR they did something in addition . ???
Does anyone have any info on the A's crew heater ?