A breakthrough in the study of single-molecule magnets occurred with the discovery of zero-field slow magnetic relaxation and hysteresis for the linear iron(I) complex [Fe(C(SiMe3)3)2]- (1), which has one of the largest spin-reversal barriers among mononuclear transition-metal single-molecule magnets. Theoretical studies have suggested that the magnetic anisotropy in 1 is made possible by pronounced stabilization of the iron d z2 orbital due to 3d z2-4s mixing, an effect which is predicted to be less pronounced in the neutral iron(II) complex Fe(C(SiMe3)3)2 (2). However, experimental support for this interpretation has remained lacking. Here, we use high-resolution single-crystal X-ray diffraction data to generate multipole models of the electron density in these two complexes, which clearly show that the iron d z2 orbital is more populated in 1 than in 2. This result can be interpreted as arising from greater stabilization of the d z2 orbital in 1, thus offering an unprecedented experimental rationale for the origin of magnetic anisotropy in 1.