The nongenetic differences in personality (that is, the amount of variance not accounted for by genes) are approximately the same for identical twins, fraternal twins, ordinary siblings, and adoptive siblings, even though the correlation between them varies widely -- from .50 for identical twins down to zero for adoptive siblings. How can this be? It's confusing, because we expect the correlation between siblings to equal the amount of variance explained by genes, and it doesn't. Here's why: the correlation reflects only similarities between siblings, whereas the amount of variance explained by genes includes both similarities and differences. Some of the differences between siblings who aren't identical twins are due to the fact that they have different genes. Fraternal twins sometimes differ in eye color, but genes nevertheless account for all of the variance in eye color. The correlation is less than 1.00 but the amount of variance accounted for by genes is 100%.

In regard to personality, genes account for all of the similarities between identical twins and none of the differences between them (genes can't account for any of the differences between them, since they have identical genes). For fraternal twins and ordinary siblings, genes account for all of the similarities between them and some of the differences. And for adoptive siblings, genes account for none of the similarities between them (there are no similarities to account for, because adoptive siblings are no more alike in personality than two people plucked from the street at random) and some of the differences. The amount of variance not explained by genes is roughly the same -- about 50 percent -- in each case.