Deep-stuck Ti-rich carbonatites: A link to the composition of primitive carbonatite melts?

Carbonatitic melts are subject to different processes during their ascent (e.g., fractional crystallization and crustal contamination), which may cause a strong change in their composition. Their original composition has not been definitively determined. In order to find indications of the primitive composition, it appears reasonable to investigate the deepest known carbonatite occurrences. One of the deepest known carbonatites is the Palabora complex in South Africa. Most of the complex is represented by varieties of pyroxenite, while the center of the complex comprises a multiple calcite carbonatite intrusion (called Loolekop). A new discovery of a second carbonatite in the southern part of the complex, reveals a head section of a stuck carbonatite intrusion, which is reflected by isolated veins on the surface and a more extensive abundance with depth. This carbonatite shows a strong enrichment in Ti (>10 wt.% TiO₂) with up to 20 modal% ilmenite. Fenitising fluids exsolved from the carbonatitic melt have even titanitised the surrounding pyroxenite. Additionally, previous investigations of the Loolekop show that the Ti content of the carbonatite increases systematically from the center of the intrusion to the margin. This could indicate a noticeably higher Ti content in the more primitive melt that crystallized in the marginal areas in comparison to a slightly more evolved melt that formed the center. Could this be an indication that primitive carbonatite melts are rich in titanium and lose the titanium rapidly during their evolution?