『Abstract
　This paper presents an investigation into the degree and nature of chemical weathering during soil formation on a volcanic (phonolite) substrate on the southern slopes of Mt. Kilimanjaro in northern Tanzania. The high field strength elements Nb and Ta were used to estimate enrichments and depletions relative to the bedrock. The degree of weathering was found to increase with depth in the soil profile. At depths greater than 200 cm, Si, Na, K, Ca, and Mg have been depleted by nearly 100％ while Al has been enriched, resulting in a highly aluminous soil residue (40-50 wt.％ Al2O3). At depths shallower than 200 cm, the soil is also depleted in Si, Na, K, Ca and Mg though not to the extents seen at depths greater than 200 cm. The lower degrees of weathering in the upper 200 cm are also evidenced by the fact that layer above 200 cm is characterized by slight positive Eu anomalies relative to other rare earth elements whereas the deeper layer exhibits no Eu anomalies. The rare earth element systematics are consistent with preferential weathering of the glassy matrix in the upper 200 cm, leaving behind plagioclase phenocrysts, which are enriched in Eu. In the deeper layer, weathering appears to be so extensive that both Eu-rich plagioclase phenocrysts and Eu-poor glass/ash have largely weathered away. These observations collectively show that the upper and lower layers of the weathering profile have undergone different weathering histories. Four scenarios may explain the apparent inverted weathering profile: re-precipitation followed by erosion, Aeolian deposition, a buried paleosol, and enhanced weathering due to lateral subsurface water flow. The first hypothesis fails to explain the massive loesses of Si, Na, K, ca, and Mg below 200 cm. The Aeolian deposition hypothesis is also untenable because it contradicts the trace element and REE's behaviors. The latter two hypotheses are both reasonable; however, the buried paleosol model is inconsistent with some physical and geochemical observations and the subsurface flow model requires the influence of hydraulic conditions not tested in this study. It is concluded here that either the buried paleosol model or the subsurface flow model can explain the formation of the Machame soils with the latter having novel implications for the transport of dissolved cations to the ocean.

Keywords: Soil formation; Subsurface water; Chemical weathering; Tanzania』

1. Introduction
2. Study area and description of weathering profile
3. Methods
4. Results
5. Estimating mass changes during soil formation
　5.1. Calculating chemical weathering and element mobility
　5.2. Identification of an appropriate immobile element tracer
　5.3. Estimating mass fluxes
6. Possible explanations for the inverted weathering profile
　6.1. Re-precipitation followed by erosion
　6.2. Aeolian dust deposition
　6.3. Buried Paleosol
　6.4. Lateral subsurface water flow
7. Implications of lateral subsurface water flow
8. Conclusions
Acknowledgments
References