『Abstract
　We dissolved Boulder Creek Granodiorite in a plug flow reactor for 5794 h at pH = 1 and T = 25℃. The primary purpose of the experiment was to identify controls on dissolved δ^44/40Ca, δ^44/42Ca, and δ^26/24Mg values during granite weathering. Herein, we also examine the origin of Ca and Mg isotopic variability among minerals composing the Boulder Creek Granodiorite, and we constrain fundamental characteristics of granite weathering important for quantifying the elemental and isotopic geochemistry of the reactor output. Nine Ca-bearing minerals display an 8.80‰ range of δ^44/40Ca values and a 0.51‰ range of δ^44/42Ca values. Three Mg-bearing minerals display a 1.53‰ range of δ^26/24Mg values. These ranges expressed at the mineralogical scale are higher than the ranges thus far reported for bulk igneous rocks. Most of the δ^44/40Ca variability reflects ⁴⁰Ca enrichment in K-feldspar, and to a lesser extent, biotite, due to the radioactive decay of ⁴⁰K over the 1.7 Ga age of the rock, whereas the entire range of δ^44/42Ca values reflects mass-dependent isotope fractionation during igneous differentiation and crystallization. The range of δ^26/24Mg values may represent either fractionation during the chloritization of biotite or interaction of the Boulder Creek Granodiorite with Mg-rich metamorphic fluids having low δ^26/24Mg values.
　The elemental and isotopic composition of the reactor output varied substantially during the experiment. We synthesize the mineralogical and fluid data using coupled mass-conservation equations solved at non-steady-state. Model calculations reveal an intricate balance between increasing specific surface area and decreasing mineral concentrations. While surface area normalized dissolution rate constants were time-invariant, specific surface area increased as a power-law function of time through positive feedbacks between mechanical disaggregation, chemical dissolution, and mineral depletion. Variations in dissolved δ^44/40Ca, δ^44/42Ca, and δ^26/24Mg values reflect conservative mixing rather than fractionation. Apatite and calcite initially control δ^44/40Ca and δ^44/42Ca values, followed by biotite, titanite, epidote, hornblende, and plagioclase. The release of radiogenic ⁴⁰Ca clearly defines the period where biotite dissolution dominates. The brucite layer of chlorite initially controls δ^26/24Mg values, followed by biotite, the TOT layer of chlorite, and hornblende. Through direct isotopic tracking, these results demonstrate that trace minerals, such as apatite and calcite in the case of Ca and brucite in the case of Mg, dominate elemental release during the incipient stages of granite weathering. The results further show that biotite dissolution dominates the middle stages of granite weathering and that plagioclase dissolution only becomes important during relatively late stages. The Ca and Mg isotope variations associated with these stages are distinct and potentially resolvable in soil mineral weathering studies.』

1. Introduction
2. Minerals and methods
　2.1. Characterization and preparation of the granite and its minerals
　2.2. Plug flow reactor experiment
　2.3. Chemical and isotopic analyses
3. Results and discussion
　3.1. Mineralogy and petrology of the Boulder Creek Granodiorite
　3.2. General geochemistry of the output solutions
　3.3. Magnesium isotope geochemistry of the rock, minerals, and output solutions
　3.4. Calcium isotope geochemistry of the rock, minerals, and output solutions
　3.5. Description of numerical model for mineral dissolution
　　3.5.1. Coupled conservation equations for the solid and fluid
　　3.5.2. Specific surface area: A power-law function of time
　3.6. Solution of numerical model for mineral dissolution
　　3.6.1. The dissolution of Mg-bearing minerals
　　3.6.2. The dissolution of Ca-bearing minerals
　　3.6.3. The dissolution of K-feldspar and concentrations profiles for Al and Si
　3.7. Feedbacks between mechanical and chemical weathering
4. Conclusions and implications
Acknowledgements
Appendix A. Supplementary data
References