『Abstract
　This study presents lithium (Li) and magnesium (Mg) isotope data from experiments designed to assess the effects of dissolution of primary phases and the formation of secondary minerals during the weathering of basalt. Basalt glass and olivine dissolution experiments were performed in mixed through-flow reactors under controlled equilibrium conditions, at low pH (2-4) in order to keep solutions undersaturated (i.e. far-from equilibrium) and inhibit the formation of secondary minerals. Combined dissolution-precipitation experiments were performed at high pH (10 and 11) increasing the saturation state of the solutions (moving the system closer to equilibrium) and thereby promoting the formation of secondary minerals
　At conditions far from equilibrium saturation state modelling and solution stoichiometry suggest that little secondary mineral formation has occurred. This is supported by the similarity of the dissolution rates of basalt glass and olivine obtained here compared to those of previous experiments. The δ⁷Li isotope composition of the experimental solution is indistinguishable from that of the initial basalt glass or olivine indicating that little fractionation has occurred. In contrast, the same experimental solutions have light Mg isotope compositions relative to the primary phases, and the solution becomes progressively lighter with time. In the absence of any evidence for secondary mineral formation the most likely explanation for these light Mg isotope compositions is that there has been preferential loss of light Mg during primary phase dissolution.
　For the experiments undertaken at close to equilibrium conditions the results of saturation state modelling and changes in solution chemistry suggest that secondary mineral formation has occurred. X-ray diffraction (XRD) measurements of the reacted mineral products from these experiments confirm that the principal secondary phase that has formed is chrysotile. Lithium isotope ratios of the experimental fluid become increasingly heavy with time, consistent with previous experimental work and natural data indicating that ⁶Li is preferentially incorporated into secondary minerals, leaving the solution enriched in ⁷Li. The behaviour of Mg isotopes is different from that anticipated or observed in natural systems. Similar to the far from equilibrium experiments initially light Mg is lost during olivine dissolution, but with time the δ²⁶Mg value of the solution becomes increasingly heavy. This suggests either preferential loss of light, and then heavy Mg from olivine, or that the secondary phase preferentially incorporates light Mg from solution. Assuming that the secondary phase is chrysotile, a Mg-silicate, the sense of Mg fractionation is opposite to that previously associated with silicate soils and implies that the fractionation of Mg isotopes during silicate precipitation may be mineral specific. If secondary silicates do preferentially remove light Mg from solution then this could be a possible mechanism for the relatively heavy δ²⁶Mg value of seawater. This study highlights the utility of experimental studies to quantify the effects of natural weathering reactions on the Li and Mg geochemical cycles.』

1. Introduction
2. Methods
　2.1. Experimental methods
　2.2. Analytical methods
　　2.2.1. Concentration measurements
　　2.2.2. Isotope chemistry
　　2.2.3. Isotope analyses
　　　2.2.3.1. Lithium
　　　2.2.3.2. Magnesium
3. Results
　3.1. Mineral phases
　3.2. Dissolution experiments
　　3.2.1. Dissolution rates
　　3.2.2. Lithium and magnesium isotopes
　3.3. Secondary phase precipitation experiments
　　3.3.1. Dissolution rates
　　3.3.2. Lithium and magnesium isotopes
4. Discussion
　4.1. Dissolution experiments
　　4.1.1. Stoichiometry
　　4.1.2. Dissolution rate
　　4.1.3. Mineral saturation states
　　4.1.4. Lithium isotope behaviour
　　4.1.5. Magnesium isotope behaviour
　4.2. Dissolution/precipitation experiments at high pH
　　4.2.1. Stoichiometry
　　4.2.2. Mineral saturation states
　　4.2.3. Dissolution rates
　　4.2.4. Secondary phase identification
　　4.2.5. Lithium isotope behaviour
　　4.2.6. Magnesium isotope behaviour
5. Conclusions
Acknowledgments
Appendix A. Supplementary data
References