Ah, so it's a challenge to a belief that nobody actually holds. No big deal, then.
Miniscule and exponentially manifest organismic response to environmental pressures is only predicted in a system where the organism strives to survive.
"Natural selection", as explained by Dawkins and as initially suggested by Darwin to explain the obvious commonality he saw, requires that the organism cares whether it survives. It also requires that there be an empirically demonstrable acquisition of new code (or activation of dormant code) that is beneficial to a species in order to be proved.
Margulis' theory utterly solves the problem of "new code". But this is, in my opinion, merely incidental to the true beauty of her theory and its potential to unify our comprehension of several fields of science which have been relatively bad bedfellows to date.