The exact direction that a population will take is not predictable. However, the central tendency of the potential directions of a population given known starting conditions is predictable, within the rules of chaos theory.

Of the 20 strains in the experiment, one developed the two particular precursors that had no evolutionary advantage which, in turn allowed the mutation that provided an advantage to occur. Given the first 2, the third became likely. That third change may well have happened in other populations, but without the precursors, it conveyed no advantage, and therefore was not expressed.

Because the first two gave no advantage, it took about 30,000 generations to have those first two available in one population when the change that conferred the advantage happened. The advantaged genes then propagated more vigorously and became obvious, expressed. They restarted cultures from the saved generation in which that pair occurred, and ran that culture forward a number of times. Sometimes, the advantage then occurred again. Not always.

This experiment vindicates Gould, not disproves. There is only a statistical chance of evolution following any particular path given identical germ material and conditions. But any step is contingent on certain other steps having been followed before, ie: Historical Contingency. That was Gould’s conjecture; his hypothesis. Now demonstrated conclusively.

This was a terrific experiment and you have to admire the extent to which it traced evolution for such an incredibly long period of time. But I’m also fuzzy on whether this result truly disproves S.J. Gould. I’m assuming that Gould made a strong claim that what happened in the experiment wouldn’t happen–or am I misreading Gould?