Abstract
Perdew et al. gave a proof that the highest occupied Kohn-Sham eigenvalue of an N-electron system represents the negative of the ionization energy (or electron affinity if N=Z+1). I argue that this cannot be the case.