Howe et al. have mistaken gene x environment correlations for environmental main effects. Thus, they believe that training would develop the same level of performance in anyone, when it would not. The heritability of talents indicates their dependence on variation in physiological (including neurological) capacities. Talents may be difficult to predict from early cues because tests are poorly designed, or because the skill requirements change at more advanced levels of performance. One twin study of training effects demonstrated greater heritability of physical skill after than before training. In summary, talents are real.