“…Furthermore, the sharp maximum at the bond midpoint has been argued to be inconsistent with previous theoretical (e.g. Wang & Klein, 1981;Yin & Cohen, 1983) and experimental (Spackman, 1986) valence densities. In contrast, the MEM density of Be based on single-crystal data (Larsen & Hansen, 1984) is quite smooth and such fine features do not appear (Takata, Sakata, Kumazawa, Larsen & Iversen, 1994;Iversen, Larsen, Souhassou & Takata, 1995).…”