Inadvertently, the results calculated using the VASP-PAW method and
presented in figures 3 and 4 were obtained for the theoretical
equilibrium lattice constant. The correct figures with the values at
the experimental lattice constant are displayed below. The
conclusions of the paper are not modified by these changes.

Figure 3. The variation of the energy difference between the SDW and the
commensurate AFM1 state of Cr as a function of the wavevector. Full
dots: LMTO-ASA results; empty circles: VASP-PAW results; squares:
FLAPW results of reference [1]. All calculations use the
GGA.

Figure 4. The variation of the amplitude m0
of the SDW as a function of
the wavevector q as calculated in the GGA. Full dots: LMTO-ASA
results; empty circles: VASP-PAW results; squares: FLAPW results
(reference [1]).
We are grateful to S Cottenier, B De Vries, J Meersschaut and M
Rots for bringing this mistake to our attention.
References
[1] Bihlmayer G, Asada T and Blügel S 2000 Phys.
Rev. B 62 11937