『Abstract
　The H/C mass ratio of the Earth's exosphere, which consists of the fluid envelopes plus the crust, is 1.95±0.15. In contrast, the H/C ratios of undegassed oceanic basalts are significantly lower, ranging from 1.2 down to 0.05. Reconstruction of source H/C ratios by accounting for H/C fractionation during partial melting and addition of carbon-enriched low-degree partial melts suggests that the source regions of MORB have H/C ratios in the range of 0.75±0.25 and those of OIB have ratios in the interval 0.5±0.3. Combining these estimates with plausible limits on the relative proportions of the OIB and MORB sources indicates that the total H inventory of the mantle is equivalent to between 0.2 and 1.6 times the H in the exosphere, assuming that there are no significant hidden reservoirs unsampled by oceanic basalts. Combining the H contents and H/C ratios of the mantle and the exosphere suggests that the H/C ratio of the bulk silicate Earth, (H/C)^BSE, is 0.99±0.42, significantly greater than the H/C ratio of chondrites, which have H/C ratios no greater than 0.55. The superchondritic (H/C)^BSE ratio likely results from preferential sequestration of C in the core, though it may also partly reflect a cometary origin for some portion of the BSE volatile inventory. The high (H/C)^BSE ratio, combined with a D/H ratio similar to chondrites, argues strongly that the BSE volatile inventory is not chiefly derived from a late veneer. The large difference in H/C ratio between the exosphere and the mantle could reflect early Earth processes such as preferential retention of C in a crystallizing magma ocean in reduced phases such as diamond, or selective loss of a massive CO2-rich atmosphere. alternatively, it may have arisen by enhanced subduction of carbon relative to hydrogen. If the latter is the case, carbon in the mantle is likely dominantly recycled.

Keywords: Deep Earth water cycle; Deep Earth carbon cycle; Mantle volatiles; Magmatic volatiles; Origin of the atmosphere; Late veneer』

1. Introduction
2. Inventories of H and C in the exosphere and in the mantle
3. Discussion
　3.1. The C contents of mantle domains
　3.2. H/C ration of the bulk silicate Earth
　　3.2.1. Constraints on (H/C)^BSE and comparison to chondrites
　　3.2.2. Extraterrestrial origin of high (H/C)^BSE?
　　3.2.3. Origin of high (H/C)^BSE or high (H/C)^Exosphere by removal of a thick CO2 atmosphere?
　　3.2.4. Origin of high (H/C)^BSE by C sequestration in the core?
　　3.2.5. A role for a late veneer?
　3.3. Mantle-exosphere fractionation of H/C
　　3.3.1. Retention of C in a crystallizing magma ocean
　　3.3.2. Preferential degassing of H during basalt extraction
　　3.3.3. Preferential subduction of carbon
　　3.3.4. Primitive mantle-normalized concentration of exosphere volatiles compared to continental crust: the key role of C subduction
　　3.3.5. Secular variation in the exosphere carbon budget?
　　3.3.6. A hidden reservoir could change everything
4. Concluding remarks
Acknowledgments
References