Mg isotope ratios (²⁶Mg/²⁴Mg) are reported in soil pore-fluids, rain and seawater, grass and smectite from a 90 kyr old soil, developed on an uplifted marine terrace from Santa Cruz, California. Rain water has an invariant ²⁶Mg/²⁴Mg ratio (expressed as δ26Mg>δ26Mg) at −0.79 ± 0.05‰, identical to seawater δ26Mg>δ26Mg. Detrital smectite (from the base of the soil profile, and therefore unweathered) has a δ26Mg">δ26Mg value of 0.11‰, potentially enriched in ²⁶Mg by up to 0.3‰ compared to the bulk silicate Earth Mg isotope composition (although within the range of all terrestrial silicates). The soil pore-waters show a continuous profile with depth for δ26Mg>δ26Mg, ranging from −0.99‰ near the surface to −0.43‰ at the base of the profile. Shallow pore-waters (<1 m) have δ26Mg>δ26Mg values that are similar to, or slightly lower than the rain waters. This implies that the degree of biological cycling of Mg in the pore-waters is relatively small and is quantified as <32%, calculated using the average Mg isotope enrichment factor between grass and rain (δ26Mggrass-δ26Mgrain>δ26Mggrass-δ26Mgrain) of 0.21‰. The deep pore-waters (1–15 m deep) have δ26Mg>δ26Mg values that are intermediate between the smectite and rain, ranging from −0.76‰ to −0.43‰, and show a similar trend with depth compared to Sr isotope ratios. The similarity between Sr and Mg isotope ratios confirms that the Mg in the pore-waters can be explained by a mixture between rain and smectite derived Mg, despite the fact that Mg and Sr concentrations may be buffered by the exchangeable reservoir. However, whilst Sr isotope ratios in the pore-waters span almost the complete range between mineral and rain inputs, Mg isotopes compositions are much closer to the rain inputs. If Mg and Sr isotope ratios are controlled uniquely by a mixture, the data can be used to estimate the mineral weathering inputs to the pore-waters, by correcting for the rain inputs. This isotopic correction is compared to the commonly used chloride correction for precipitation inputs. A consistent interpretation is only possible if Mg isotope ratios are fractionated either by the precipitation of a secondary Mg bearing phase, not detected by conventional methods, or selective leaching of ²⁴Mg from smectite. There is therefore dual control on the Mg isotopic composition of the pore-waters, mixing of two inputs with distinct isotopic compositions, modified by fractionation. The data provide (1) further evidence for Mg isotope fractionation at the surface of the Earth and (2) the first field evidence of Mg isotope fractionation during uptake by natural plants. The coherent behaviour of Mg isotope ratios in soil environments is encouraging for the development of Mg isotope ratios as a quantitative tracer of both weathering inputs of Mg to waters, and the physicochemical processes that cycle Mg, a major cation linked to the carbon cycle, during continental weathering.