Well, I think I've read that the hp to turn a given prop within nominal conditions increases as the cube of the rpm? Not sure but if the lift goes up with V^2 it is at least that. 37/31^3 is 70% more, 37/31^2 is 42% more.
But assumptions and arithmetic aside, the bench prop is a very light drag load and I think it essentially lets the engine easily get on resonance. So the rpm certainly represents a major hp increase, but I don't know whether it represents the actual ratio between peak hp, or the engines/pipe "governed" out at those rpm due to tuning lengths and all that and there is more to it than that. Now we're getting to my fuzzy areas (in my brain, thank you).