First of all, you misread the figures. The correlation between identical twins reared together is 0.85, while adoptive parent and child is 0.19.

Second of all, correlation coefficients do not scale linearly. A correlation of r=0.85 means r^2 explains 0.72 of the variance, while a correlation of 0.19 explains 0.03 of the variance. In other words, adoptive parents have 0.03/0.72=4% of the effect on IQ as being an identical twin does.