In the first example, their IQ of 120 is from a mix of factors and the author estimates some percentage that is due to the environment outside genetics, and tha

In the first example, their IQ of 120 is from a mix of factors and the author estimates some percentage that is due to the environment outside genetics, and that is what is lost in reverting to the mean of the environment. In the second example the author assumes that the environment of the desert island contributes nothing to genetics and so the offspring get the full 110 (heritability is 100%). It's odd to me though because it wasn't clear that the environment was meant to be affecting the parent and not the child too- the trick is the author treats the heritability parameter different for the parent and the child. Also, I don't think the desert island would actually have no effect on the 110 generation's IQ, or their upbringing outside of it, but I guess I can accept it as a thought-experiment-only place that does that by definition and not a literal desert island.