Comment on Effects of Disorder on Ferromagnetism in Diluted Magnetic Semiconductors

In a recent Letter, Berciu and Bhatt have presented a mean-field theory of ferromagnetism in III-V semiconductors doped with manganese, starting from an impurity band model. We show that this approach gives an unphysically broad impurity band and is thus not appropriate for (Ga,Mn)As containing 1-5% Mn. We also point out a microscopically unmotivated sign change in the overlap integrals in the Letter. Without this sign change, stable ferromagnetism is not obtained.