The complex [gamma(1,2)-SiW(10){Fe(OH(2))}(2)O(38)](6)(-) (1) has been reported to catalyze the much sought reductant-free selective O(2)-based epoxidation of alkenes (Nishiyama, Y.; Nakagawa, Y.; Mizuno, N. Angew. Chem. Int. Ed. 2001, 40, 3639-3641) in chlorocarbon-acetonitrile solution. The challenge of reproducing catalysis by 1 led us to examine this chemistry in detail. In H(2)O, a desirable solvent for catalysis, 1, does not exist in the proposed organic-medium form in which the two iron atoms are in the binding pocket defined by the equatorial oxygens and, importantly, by two oxygens bound to the central Si heteroatom. Instead, 1 in H(2)O initially forms an unusual trimer [{Fe(2)(OH)(3)(H(2)O)(2)}(3)(gamma-SiW(10)O(36))(3)](15)(-) (2). The X-ray structure of 2 shows that the Fe-O(Si) bonds are cleaved and new bonds (mu-hydroxo bridges) form between these Fe centers and those of the neighboring [gamma(1,2)-SiW(10)Fe(2)] units. Structural, physical, and computational evidence indicate that if the bonds between the d-electron center, M (Fe in the case of 1 and 2), and the terminal ligands on M are stronger than the M-O(x)() bonds, then the out-of-pocket form is more stable and is the one observed. Significantly, 2 in H(2)O forms an intermediate that catalyzes the effective aerobic oxidation of sulfur compounds (mercaptoethanol is oxidized to the corresponding disulfide by O(2) at ambient pressure and temperature). All experimental findings are consistent with dissociation of a gamma-SiW(10) Keggin unit from the trimer, 2, to form the catalytically active species.