SUB-QUADRATIC BLOAT

If the programs within the population remain close to the ridge in the number of programs versus their shape and they increase their depth at a constant rate this leads to a prediction of sub-quadratic growth in their lengths'. (For modest size programs we expect size $\mbox{O(gens}^{1.3}\mbox{)}$rising to a limit of quadratic growth for $\vert\mbox{program}\vert \gg 1000$cf. [Flajolet and Oldyzko1982, Table II]. Up to generation 50 [Langdon2000, Table 5] reports good agreement on average.