An important but poorly understood factor that affects diffusion rates is the role of speciation during diffusion of a multi-species component. The diffusion of such a component is complicated by the different diffusion coefficient of each species and the interconversion reactions among the species. These complexities can be treated by a diffusion equation that incorporates the diffusive fluxes of all species contributing to the concentration of the component. The effects of speciation on the diffusion of the component can be investigated experimentally in some simple cases by measuring concentration profiles of all species developed during diffusion experiments or by studying some of their other consequences. Experimental data on water diffusion in rhyolitic glasses indicate that although dissolved water is present as two species, H2O molecules and OH groups, molecular H2O is the dominant diffusing species at very low to high water concentrations. This explains the apparently complex behavior of water diffusion. Experimental data on oxygen diffusion in some silicates using 18O tracers in the form of H2(18O) are consistent with the idea that 18O transport is dominated by diffusion of H2O molecules even at lower water contents (ppm or less). This explains why oxygen transport depends on the presence of water and generally depends on water fugacity linearly. For this mode of oxygen transport, there is a simple theoretical relationship between the effective total oxygen diffusion coefficient and the total water diffusion coefficient that is a function of only the water concentration of the silicate at low water content. This relationship appears to describe quantitatively the existing data over a wide range in water contents and diffusion coefficients in several phases.