- © 2003 Geological Society of South Africa
New whole rock chemical and Re-Os isotopic analyses of peridotite xenoliths from the Monastery kimberlite show the majority of the samples to have the low Fe, Al and Ca, and high MgO contents typical of magmaphile-element depleted cratonic lithospheric mantle. Two out of nine samples have elevated Al and lower Mg contents. Re-Os results show these two samples to be mid-Proterozoic in age indicating that additions of compositionally distinct material to this mantle section took place well after its initial assembly in the Archean. The remainder of the samples have the low Re/Os and 187Os/188Os characteristic of old cratonic peridotites, overlapping extensively with most other peridotite xenoliths contained in kimberlites erupted through the Kaapvaal craton. Excluding the two Proterozoic samples and two samples with anomalously old Re-Os model ages, the Monastery results give a mean Re-depletion model age (TRD) of 2.76 ± 0.41 Ga and mantle model age (TMA) of 3.25 ± 0.28 Ga. The oldest TRD observed at Monastery is 3.02 Ga. Compared to two well-studied localities from west of the Colesberg lineament that may mark a 2.9 Ga suture between the western and eastern blocks of the Kaapvaal craton, the Monastery mean TMA is older, but within uncertainty of, the mean TMA observed for Kimberley (2.93 Ga) and Newlands (3.18 Ga) peridotites. Similarly, the maximum TRD obtained at Kimberley (2.95 Ga) and Newlands (3.22 Ga) span the maximum TRD observed at Monastery. These results show conclusively that the most significant geochemical differentiation event affecting the Kaapvaal mantle occurred in the Archean. Compared to the results for peridotite xenoliths in kimberlites that penetrate the western Kaapvaal, the older mean TMA observed for Monastery peridotites is suggestive of older mantle beneath the older crust of the eastern Kaapvaal. However, the combination of the scatter observed in the TMA ages for samples from each locality and the uncertainty in the interpretation of Re-Os mantle model ages caused by the multistage, e.g. metasomatic, history of the samples does not allow conclusive resolution of this possible age difference.