Thanks to inbalance and sycamore for their informative comments
I think there's another nasty aspect of the Rotary, which is the effect of the rotating mass of the crankcase, cylinders, and prop. As I understand it, this caused the engine speed to react very slowly to changes in position of the air and fuel levers. This is worst on accelleration, where the engine could be flooded if the levers were opened too quickly. The way around this was to make small adjustments to the levers, and allow the engine RPM to catch up.
On the point of Castor Oil, most aircraft had an engine cowl with only a small open section around 6 o'clock. This was an attempt to collect the exhaust gases and direct them out under the fuselage, so the pilot's didn't get the s***s. I think this was only partially successfull.
We're finding out some of the reasons why the Rotary engine soon gave way to the 'static' radial.