Great. So the last elegans figure can be (a) log probability of alpha and beta, (b) prob of cost exponent, (c) the best cost exponent in the alpha -beta plane and (d) the 3 subfigures with o.5, 1 & 2 costs.
The last figure you give is also nice as it shows why for exponent 2 beta is larger than alpha.
I also like that there is no change in the best alpha and beta for 0.5 and 1. Had we found very different values (say something ever closer to pure wiring economy values for 1.0) I would have been confused. This way all point to 0.5 as the best value and we can even udnerstand that does calculations with exponent 2.0 we got higher values.
It would be interesting to use either teh bayes or pure wiring economy values and xi=0.5 and repeat our PNAS procedure. But that would be for your thesis, not for teh paper that is getting too big.
BTW, for the last subfigure (best alpha and beta for different exponets), could you add to it a standard deviation? I am curious to see whetehr the ‘ideal wiring economy’ values are within the standard deviation.