CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA: EVIDENCE FROM UPPER-CRUSTAL XENOLITHS IN KIMBERLITE DIATREMES

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY 125

IntroductionKimberlite diatremes contain a wealth of informationconcerning the composition of the mantle and lowercrust, in the form of xenoliths and xenocrysts (includingdiamonds), that have received extensive attention overthe years (Dawson, 1980; Nixon, 1987, Gurney, 1989).Due to the nature of diatreme formation – explosiveeruption, followed by fall-back of debris into the cavitycreated – they also contain a suite of xenoliths fromupper-crustal lithologies penetrated by the diatreme.Although these upper-crustal xenoliths dominate overtheir deeper counterparts, they have received lessattention in much of the kimberlite literature.

Upper-crustal xenoliths provide a record of the host-rock sequences into which the kimberlite diatremeswere emplaced, which may no longer occur in thevicinity of the diatremes as a result of post-emplacementerosion. The diatreme xenoliths are useful in estimating,for example, the original distribution of certainstratigraphic units within a sedimentary basin or thepalaeogeographic extent of volcanic units such as lavaflows. When kimberlite diatremes of varying ages occurin the same broad area, it may be possible to reconstructthe erosional history of the region. Certain xenoliths thatoccur in older diatremes may be absent in younger ones,due to an intervening period of erosion. The levels to

which kimberlites of different ages are eroded (crater-,diatreme-, and/or root-facies) may vary. As a result,estimates of rates of erosion with time, as well as thepalaeo-geomorphological evolution of an area, may alsobe revealed, assuming that kimberlite pipe morphologyis generally consistent.

The central part of South Africa (Figure 1) providesan ideal region to undertake such a study. This regioncontains several Group II kimberlites (approximate agesof 140 to 114 Ma) as well as younger Group I kimberlites(approximate ages of 100 to 75 Ma). These Cretaceouskimberlites were emplaced during a period of activeerosion in central South Africa, about which little isknown as there is virtually no preserved onshoresedimentary record. In addition, they straddle thewestern outcrop margin of the main Karoo Basin andmay provide evidence of the former distribution ofstratigraphic units within the Karoo Supergroup. In anattempt to improve our understanding of Cretaceouserosion in southern Africa, we have examined upper-crustal xenoliths from some 26 Group I and II kimberlitepipes in central South Africa (Figure 1, Table 2).Because these kimberlites were emplaced within themain Karoo Basin, this study has focused primarily onKaroo-age sandstone and basalt xenoliths. The resultsallow for estimations of erosion rates in central South

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA: EVIDENCEFROM UPPER-CRUSTAL XENOLITHS IN KIMBERLITE DIATREMES

E.K. HANSON, J.M. MOORE, E.M. BORDY, J.S. MARSH, G. HOWARTHGeology Department, Rhodes University, P.O. Box 94, Grahamstown, South Africaemail: [email protected]; [email protected]; [email protected]; [email protected];[email protected]

J.V.A. ROBEYDe Beers Group Services, P.O. Box 47, Kimberley, South Africaemail: [email protected]

© 2009 June Geological Society of South Africa

ABSTRACT

Twelve Group II and fourteen Group I kimberlite diatremes in central South Africa were examined for upper crustal xenoliths in

order to estimate the extent of various lithological units of the Karoo Supergroup in the main Karoo basin at times of kimberlite

eruption, the Cretaceous erosional history of the area, and the approximate vertical extent of the kimberlite diatremes prior to

erosion. Sandstone and amygdaloidal basaltic lava xenoliths from the Karoo Supergroup were specifically selected as their modal

mineralogies and geochemical compositions respectively can be attributed to specific stratigraphic positions within the Karoo

Supergroup. Results indicated that, at the time of Group II kimberlite eruption (~120 Ma), basaltic lavas of the Drakensberg Group

covered the entire area, but by the time of Group I kimberlite eruption (~85 Ma), they were restricted to the south-eastern half of

the study area. At the latter time, an escarpment is proposed to have existed at the basalt outcrop limit, some 180 km west of its

current position. Sandstones of the Stormberg Group had a restricted original distribution in the north and east of the study area,

whereas sandstones from all other Karoo groups occurred throughout the entire area. In the Kimberley area, approximately

500 m of erosion is estimated to have occurred from 120 to 85 Ma and 850 meters from 85 Ma to the present day at average rates

of approximately 15 m/Ma and 10 m/Ma respectively. Both Group I and II kimberlite diatremes had vertical extents of

approximately 1350 m at eruption. An inland scarp-retreat model is proposed for the Cretaceous erosion cycle in central

South Africa.

SOUTH AFRICAN JOURNAL OF GEOLOGY, 2009, VOLUME 112 PAGES 125-140

doi:10.2113/gssajg.112.2.125

Africa through the Cretaceous and constitute evidencefor an interior scarp-retreat model for the region. In addition, modifications to the classic Hawthorne andClement models for southern African kimberlite pipesare proposed (Hawthorne, 1975, Clement, 1982).

Regional GeologyAt the time of intrusion of the Cretaceous kimberlites,the upper-crustal geology of the central South Africanregion was dominated by rocks of the KarooSupergroup. The kimberlites sampled in this study aresituated in the north-western part of the main KarooBasin which is considered to have formed part of aretro-arc foreland system, generated by subduction of the palaeo-Pacific plate under the southern margin ofGondwana in the Late Paleozoic-Early Mesozoic(Veevers et al., 1994, Catuneanu et al., 1998). The KarooSupergroup is represented by a variety of clasticsedimentary rocks which originated under varyingclimatic and depositional settings (Table 1). Duringdeposition of the Dwyka and Ecca Groups, climateswere glacial to cold, and deposition was dominated byglacial diamictites and mudrocks respectively. Uplift insource regions to the south and warming conditions ledto the subsequent progradation of deltaic and fluvialsediments into the basin, containing increasingproportions of sandstone, during deposition of theBeaufort Group. The Stormberg Group (Molteno, Elliotand Clarens Formations) contains fluvial and aeolian

sandstones associated with warmer, more arid climates.The entire sequence represents a transition from cold-climate marine to warm-climate terrestrial deposits(Catuneanu et al., 1998, Johnson et al., 2006).

The progressive changes across the depositionalhistory of the Karoo Supergroup are reflected inmineralogical variations of the sandstones of the mainKaroo Basin. Sandstones from the basal portions of thesequence (Dwyka, Ecca, lower Beaufort Groups)contain significant lithic and feldspar componentswhereas those from the upper portions (upper Beaufortand Stormberg Groups) are dominated by mono-crystalline quartz (this study; Johnson et al., 2006). This mineralogical variation has been used to identifythe stratigraphic sources of sandstone xenolithsrecovered from the kimberlite diatremes.

The Karoo sedimentary sequence in the study areahas been extensively intruded by dolerite dykes and sills of the Karoo Igneous Province. Extrusiveequivalents of these dolerites, the Drakensberg Group,are present as an erosional remnant of continental floodbasalt to the immediate east of the study area (theLesotho Remnant). The presence of basalt xenolithsfrom the Drakensberg Group in many of the kimberlitediatremes in the study area clearly indicates a greaterformer extent to the flood basalts, covering much ofcentral South Africa.

Using geochemistry, particularly the ratios betweenimmobile incompatible elements, Marsh et al. (1997)

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA126

Figure 1. Locality map of the various Group I and II kimberlites sampled in this study (1-Melton Wold, 2-Lace, 3-Voorspoed, 4-Roberts Victor,

5-New Elands, 6-Phoenix/Star, 7-West End, 8-Wimbledon, 9-Finsch, 10-Markt, 11-Frank Smith, 12-Pampoenpoort, 13-Uintjiesberg,

14-Koffiefontein, 15-Monastery, 16-Kimberley, 17-Jagersfontein, 18-Kaal Vallei, 19-De Beers, 20-Bultfontein, 21-Kamfersdam,

22-Leicester/Balmoral, 23-Lushof, 24-Britstown, 25-Hebron, 26-Lovedale). Section A-B represents the profile in Figure 6. The positions are

shown of projected palaeo-escarpments at 85 Ma and 120 Ma.

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY

SOUTH AFRICAN JOURNAL OF GEOLOGY

127

demonstrated that the Karoo basalt sequence in theLesotho Remnant could be subdivided into a number ofchemostratigraphic units (Figure 2). In the lower part ofthe sequence, the Barkly East Formation comprisesbasaltic units which are geochemically diverse, havesmall volumes and have limited areal extent. In places,particularly in the south, these units are interbeddedwith fluvial sandstone and/or volcaniclastic deposits.The upper part of the sequence consists of the LesothoFormation characterised by large-volume basaltic unitswith a more limited geochemical range and whichextend right across the Lesotho Remnant. The distinctivecompositional signatures of the various basalt units have been used to identify both a Karoo provenanceand the specific stratigraphic source within the

Drakensberg Group for the basalt xenoliths sampled inthis study. In doing so, care has been taken to eliminatesamples that may be contaminated by kimberlite and emphasis has been placed on characteristic immobile incompatible element ratios which are leastlikely to be changed by low-temperature alterationprocesses.

Central South Africa has been the focal point ofCretaceous kimberlite magmatism, both within theKaroo Basin and along its western margin. The magmatism takes the form of narrow intrusivedykes, rare sills and subvolcanic-to-volcanic pipes.Kimberlite pipes are typically subdivided into crater,diatreme and root zones (Clement, 1982) depending onmorphology and the textural nature of the associatedkimberlite. In the study area, kimberlite pipes are small(mostly <12 ha) and have a tendency to occur in clusterscontaining several pipes/dykes. Due to extensiveerosion, crater-zone kimberlites are not preserved. The kimberlites are divided into older Group II (~140 to114 Ma) and younger Group I (~100 to 75 Ma)occurrences (Table 1). Group II kimberlites areidentified by their phlogopitic nature, enrichments in K, Rb, Ba, Pb and depletions in Cr compared to GroupI kimberlites and by their enriched Sr and Nd isotopicsignatures. Group II kimberlites examined in this studyinclude diatreme-zone pipes (Finsch, West End, Lace,Voorspoed) and root-zone dyke/small-pipe associations(Phoenix/Star, Roberts Victor, New Elands, Wimbledon,Melton Wold). Group I kimberlites sampled arediatreme-zone pipes (Kimberley cluster, Koffiefontein,Jagersfontein, Kaal Vallei, Uintjiesberg) with Monasteryvirtually the sole root-zone dyke/small-pipe association.The Kimberley cluster contains the Kimberley, De Beers,Bultfontein, Dutoitspan, Wesselton and Kamfersdampipes.

Subsequent to the eruption of the Karoo floodbasalts around 183 Ma (Duncan et al., 1997), thegeological history of southern Africa has beendominated by the break-up of Gondwana and an

Figure 2. Schematic summary of geochemical stratigraphy in the

Drakensberg Group of Lesotho (from Marsh et al., 1997 and

unpublished data). Available data suggest that the different units of

the Barkly East Formation wedge out against an arch trending

roughly eastwest at about 30oS. The single palaeomagnetic reversal

within the basalt sequence (van Zijl et al., 1962) is consistently

located in the lower part of the Mafika Lisiu unit (Marsh et al.,

1997).

Table 1. Lithostratigraphy of the Karoo Supergroup in the northwestern part of the main Karoo basin. Thickness data and estimations from

this study, Winter and Venter (1970)*, Hawthorne (1975)**, Beukes (1970)***.

Lithostratigraphy Main rock types Age Climate/ Average unit thickness

depositional Monastery Kimberley Melton

setting area area Wold area

Drakensberg Group Basalt Early Jurassic Humid/ 1500 m 1445 m ?

(183 Ma) continental

‘Stormberg’ Group Sandstones, Late Triassic- Increasingly arid/fluvio- ~260 m*** (350 m)** 0 m

mudstones Early Jurassic lacustrine to aeolian

Beaufort Group Sandstones, Late Permian- Semi-arid ~700 m* 550 m** >1850 m*

mudstones Early Triassic fluvio-lacustrine

Ecca Group Mudstones, Early to Post-glacial/ ~480 m* 300 m** ~1100 m*

sandstones, coal Middle Permian fluvio-deltaic

Dwyka Group Diamictites, Late Glacio-fluvial ~40 m* <60 m** ~380 m*

varved shales, Carboniferous

conglomerates,

sandstones

extensive period of erosion (>100 Ma) – the Africancycle (Partridge and Maud, 1987). Due to significantuplift and erosion during subsequent post-AfricanMiocene/ Pliocene cycles in central South Africa anddevelopment of the modern Orange and Vaal drainages,virtually no evidence remains of the African cycle in thestudy area aside from certain etched surface remnants.Offshore sedimentation rates indicate Cretaceous peaksat 130 to 115 Ma and 86 to 78 Ma (Dingle et al., 1983)on the west coast and at 136 to 120 Ma and 93 to 67 Maon the south coast (Tinker et al., 2008b) that are relatedto increased rates of erosion, but in general, there hasbeen a decline in rates of sediment accumulation andthus erosion through the duration of the African andpost-African cycles. The two periods of increasedsedimentation have been linked to uplift associated withthe intrusion of the Group II and I kimberlitesrespectively (Scrutton and Dingle, 1976; Tinker et al.,2008a).

Previous studies of upper-crustal xenolithsXenoliths correlating with Karoo-Supergroupsedimentary unitsThe earliest descriptions of upper-crustal xenoliths inkimberlite diatremes (Matthews 1887; Du Toit 1906;Rogers and Du Toit 1909; Harger 1913; Wagner 1914;Williams 1932) used lithological criteria (e.g. colour,grain size) and fossils to correlate xenoliths withdifferent units within the Karoo Supergroup. On thisbasis, the presence of Beaufort Group xenoliths hasbeen reported from Melton Wold (Mambali, 1998),Voorspoed, Roberts Victor, Koffiefontein, Monastery,Jagersfontein, the Kimberley kimberlite cluster (Wagner,1914; Williams, 1932), and Finsch (Visser, 1972; Clement,1982). Geochemical characteristics of sandstonexenoliths from the Kimberley pipes (Rambula, 2005)suggest that they were mostly derived from the EccaGroup and lower Beaufort Group.

Xenoliths correlating with Karoo-SupergroupbasaltsPetrographic descriptions of basaltic xenoliths fromMelton Wold, Voorspoed, West End, Makganyene, FrankSmith, Uintjiesberg, Koffiefontein, Monastery andJagersfontein pipes led Du Toit (1906), Rogers and Du Toit (1909), Harger (1913), Wagner (1914) andWilliams (1932) to conclude that they are Karoo basalt.Vesicular basalt of presumed Karoo provenance isdescribed by Robey (1981) from the Lovedale No.1 pipe(not sampled in this study). Some of these observationshave subsequently been confirmed by geochemicalcomparisons between basalt xenoliths and DrakensbergGroup basalts (Roberts, 1997; Letsale, 1998; Mambali,1998).

A review of descriptions of igneous and volcanicxenoliths found during early mining operations atKimberley reveal that the original descriptions of thesexenoliths do not provide any unequivocal indication thatthey are Karoo basalts. While Matthews (1887), Becker

(1904), Lindgren (1905) and Wagner (1914) describebasalt, diabase, whinestone and melaphyre boulders andfragments in the pipes, none of these descriptions implythat they correlate with basalts of the DrakensbergGroup. Instead, the descriptions are more consistentwith mafic lavas of the Ventersdorp Supergroup and/orKaroo-age intrusive dolerites. Some authors describethese ‘igneous’ inclusions as being identical to thecountry rock of the kimberlite at depth, i.e. Ventersdorplava (Harger, 1909). These conclusions are in contrast toWilliams (1932) who is of the opinion that the lavaxenoliths are identifiable by their colour and thepresence of zeolite and calcite pipe amygdales as beingof Karoo origin. Samples collected during our surveythat had a visual similarity to Karoo basalts, all had bulkcompositions indicating that they were lavas of theVentersdorp Supergroup (Figure 4a).

Previous erosion estimatesPrevious estimates for post-Gondwana erosion in centralSouth Africa derive from geomorphological, geologicaland fission-track studies. Based on assumptions of pre-break-up Gondwana topography and subsequentpost-African uplift estimates, Partridge and Maud (1987)calculated about 1200 m of erosion in the Kimberleyarea during the African cycle. Brown et al. (1998)estimated 3.4 + 1.4 km of mid-to-late Cretaceousdenudation from fission-track results from the centralKaapvaal Craton, to the north of the study area, with apeak around 90 + 10 Ma. Tinker et al. (2008a) estimated2.5 to 3.5 km of mid-to-late Cretaceous (100 to 80 Ma)weathering to the south-west of the study area, togetherwith <1 km weathering since then (<80 Ma).

Geological studies have focused primarily on theGroup I kimberlite cluster at Kimberley where post-emplacement (~85 Ma) erosion estimates range from 800 m (Rambula, 2005) to >1400 m (Hawthorne, 1975).The most commonly cited estimate is that of Hawthorne(1975), which provides a minimum constraint of 900 m,based on the minimum permissible depth for formationof kimberlite sills intruding sedimentary rocks of theKaroo Supergroup, and a maximum constraint of 1900 mfrom depth of burial, based on the degree of host-rockdiagenesis of the country-rock mudstones. The averageof these, 1400 m, was adopted as the value for thethickness of strata eroded subsequent to emplacementof the kimberlite pipes. This value requires the presenceof basalts of the Drakensberg Group in the Kimberleyarea at the time of kimberlite emplacement, as it isunlikely that the sedimentary units of the KarooSupergroup achieved such a thickness (Johnson, 1976;1991). The model relies partially on observations in thehost rocks (diagenesis) that are unrelated to the timingof kimberlite emplacement which took place some 50 to100 Ma later. Hawthorne’s (1975) model further impliesthat a maximum of 500 m of erosion occurred in the first100 Ma subsequent to eruption of the Karoo floodbasalts (183 to 85 Ma), whereas 1400 m of erosionoccurred in the following 85 Ma (85 to 0 Ma).

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA128

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY

SOUTH AFRICAN JOURNAL OF GEOLOGY

129

This threefold increase in erosion rate is counter to thedeclining erosion-rate estimations from the offshoresedimentary record (Dingle et al., 1983).

In addition, Hawthorne (1975) allocated a 350 mthickness to the Stormberg Group, but sandstones of thisgroup have not been reported from the xenolith suite ofthe Kimberley cluster (Rambula, 2005; this study) or thesurrounding Group II kimberlites, and the StormbergGroup is not thought to have extended this far west.Noting the absence of Stormberg Group sandstone andDrakensberg Group basalt xenoliths in the Kimberleypipes and removing these stratigraphic units from theHawthorne (1975) model, Rambula (2005) concludedthat 800 m of erosion has occurred since Group Ikimberlite emplacement at ~85 Ma. This is similar to anearly estimate of 900 m (Wagner, 1914), which is basedon the presence of sandstones and fossil-bearingsedimentary rocks from the Beaufort Group in thexenolith suite. Other studies based on upper-crustalxenolith suites indicate approximately 1000 m of erosionat Melton Wold (Mambali, 1998); as little as 300 m, butmore likely 1.5 to 2 km of erosion at Voorspoed

(Roberts, 1997); and 1200 to 1700 m at Finsch (Visser,1972). Hawthorne (1975) estimated the total thickness ofthe Karoo Supergroup in the Koffiefontein area to be2200 m, based on country-rock diagenesis.

Field and Scott-Smith (1999) propose maximuminitiation depths for diatreme eruption of 1500 m forsouthern African kimberlite pipes. It follows thatapproximate post-emplacement erosion can beestimated by subtracting the depth from present-daysurface to the base of the diatreme. The base of diatremezones of the Kimberley pipes has been shown to occurat 400 to 600 m below the present land surface(Hawthorne, 1975; Clement, 1982; Clement et al., 1986)and this implies that a maximum of 1100 m of post-eruption erosion has occurred in the Kimberley area.Dawson (1971), however, indicates greater diatremeinitiation depths of 2.4 km.

Sample selection and methods of analysis The 26 kimberlites selected for the present study extendacross the central portion of the main Karoo Basin fromLesotho to Victoria West in South Africa (Figure 1).

Table 2. Kimberlites included in the present study; abbreviations used in subsequent tables and for sample numbering. Country-rock

information obtained from 1:250,000 geological maps, Council for Geoscience.

Kimberlite Pipes Age (Ma) Gp Surface Country Rock Xenoliths sampled

Sandstone Basalt

Melton Wold MW 143 ± 141 I/II Beaufort (basal) X

Lace LA 133.2 + 2.82 II Ecca (middle) X X

Voorspoed V 131.8 ± 1.72 II Ecca (middle) X X

Roberts Victor RV 127.33 II Ecca (lower) X X

New Elands NE 122.9 + 1.68 II Ecca (middle) X X

Phoenix/Star PH 1269 II Beaufort (lower) X X

West End WE 1184 II Transvaal Supergroup X

Wimbledon WI 1184 II Ecca (basal) X X

Finsch FN 118.4 ± 2.21 II Transvaal Supergroup X X

Markt MK 116.8 ± 15 II Ecca (basal) X

Frank Smith FS 113.6 ± 1.81 I/II Ecca (basal) X

Pampoenpoort PP 103.2 ± 0.75 I/II Ecca (upper) X X

Uintjiesberg UJ 100.7 ± 1.41 I Ecca (middle) X

Koffiefontein/Ebenhaezer KF 90.46 I Ecca (middle) X

Monastery M 88 ± 43 I Molteno (upper) X X

Kimberley/Big Hole* BH 87 ± 14 I Ecca/Dwyka contact X

Jagersfontein JG 85.6 ± 11 I Ecca (upper) X X

Kaal Vallei KV 84.9 ± 0.97 I Beaufort (lower) X

De Beers* DB 84.3 ± 33 I Ecca/Dwyka contact X

Bultfontein* BF 84 ± 0.97 I Ecca/Dwyka contact X

Kamfersdam* KD 86.96 I Ecca/Dwyka contact

Leicester/Balmoral LE 93.66 I Dwyka/Ventersdorp contact

Lushof LU 78.36 I Ecca (upper)

Britstown Cluster BC 74.4 ± 0.85 I Ecca (upper)

Hebron HE 744 I Beaufort (basal)

Lovedale LV 744 I Ecca (upper)

1Smith et al., 1985 4De Beers database (unpublished) 7Allsopp and Kramers, 1972Phillips et al., 1999 5Smith et al., 1994 8Allsopp et al., 19893Allsopp and Barnett, 1975 6Davis, 1977 9MacIntyre and Dawson, 1976*Member of the Group I kimberlite cluster in Kimberley (KC)

The complete analytical data-set may be obtained fromHanson (2007) and Howarth (2007) (the formeravailable electronically at http://eprints.ru.ac.za/855/).Table 2 presents details of the localities and the nature of investigations carried out at each. Xenoliths were obtained from dumps and oldweathering ‘floors’. The number of samples collecteddepended on how common the xenoliths were and howextensive the dumps were, and varied from a singlesample (Frank Smith) to suites of up to eight samples.Their representivity is therefore uncertain although theyrepresent a cross-section of available material. Basaltxenoliths were located and sampled at 16 of the pipesvisited and sandstone xenoliths suitable for point-countexamination were sampled at 15 pipes. In order to avoidintrusive dyke and sill samples, only amygdaloidalbasalts were collected. It is, however, conceded thatcertain shallow-level sills are known to have vescicularmargins and that other lavas were, as a result, excluded.

Sandstone samples range from very fine to verycoarse grained with dominant components beingmonocrystalline quartz, feldspar, lithic fragments as wellas polycrystalline quartz and minor accessory minerals.Modal proportions of these phases were determined bythe Gazzi-Dickinson point-counting method, asdescribed by Ingersoll et al. (1984). This methodovercomes the problem of grain-size differences,enabling comparison between samples with differentgrain-sizes. A 95% confidence level is achieved bycounting 300 grains per sample (Galehouse, 1971). The results have been normalized to 100% free ofaccessory minerals prior to plotting on the ternarydiagram (Figure 3) (see also Hanson, 2007; Howarth,2007). The plots were compared to composition fields ofsandstones from the different stratigraphic groups of theKaroo Supergroup compiled from previous studies(Kingsley, 1977; Eriksson, 1984; Johnson, 1991; Haycocket al., 1997; Fiedler and Adelmann, 1998; Hancox, 1998;Bordy et al., 2005) (Table 3).

The composition of basalt xenoliths was determinedby x-ray fluorescence spectrometry (XRF) on sparselyamygdaloidal samples. Major elements were determinedon fusion discs following the method of Norrish and

Hutton (1969). Trace elements (Zn, Cu, Ni, Co, Cr, V, Nb,Zr, Y, Sr, Rb, Ce, Nd and La) were determined onpressed powder pellets following the analyticalprocedure of Duncan et al. (1984). The compositions ofthe basalt xenoliths were compared (Figure 4) to thoseof geochemically defined stratigraphic units of theDrakensberg Group (Marsh et al., 1997; Marsh, 1998) aswell as to the Ventersdorp Supergroup (Bowen et al.,1986).

ResultsSandstone xenolithsThere is considerable overlap in the fields of sandstonesfrom various subgroups/formations (Figure 3) and, as aresult, the stratigraphic positions of xenoliths can onlybe broadly allocated to lower, middle and upper Karoo-Supergroup sources. Results of the sandstone analysesand correlations (Figure 3, Table 3) indicate thatxenoliths from the middle-Karoo Tarkastad Subgroup of the Beaufort Group are most abundant, presumablydue to the significant thicknesses and widespreaddistribution of these sandstones. These results areconsistent with visual observations of earlier workersciting Beaufort Group xenoliths in the Voorspoed,Roberts Victor, Finsch, Koffiefontein, Monastery,Jagersfontein and Kimberley-cluster pipes (Williams,1932; Visser, 1972; Clement, 1982). Xenoliths that maycorrelate with the lower-Karoo Adelaide Subgroup of theBeaufort Group and/or with the Ecca Group are alsocommon and widespread. Sandstone xenoliths that maybe correlated with the Dwyka Group are rare, havingonly been identified at Lace, Jagersfontein andBultfontein pipes. This is to be expected as the DwykaGroup in the study area consists primarily of very thinground-moraine and glacial-valley facies within whichsandstones are uncommon (Tables 2 and 3).

Sandstone xenoliths that may be correlated with theupper-Karoo Stormberg Group are more restricted,being confined to Voorspoed, Lace and Monastery pipes(Table 3). The Monastery xenoliths are to be expected asrocks of the Stormberg Group form the present-dayoutcrop there. The Voorspoed and Lace xenolithscoincide with an area of onlap to the east, of the

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA130

Table 3. Summary of stratigraphic positions of identified xenoliths within the examined kimberlites based on Figure 4c and Figure 3.

Kimberlite MW LA V RV NE PH WE WI FN MK FS PP UJ KF M JG KV KC

Drakensberg Evolved X

Group Basalts Oxbow X

Maloti/Senqu/Mothae X X X X X X X X X X

Mafika Lisiu X X X X X X X X X X X X X X

Barkly East X X X X X X X

Karoo Stormberg X X X

Supergroup Upper Beaufort X X X X X X X X X

Sediments L Beaufort/U Ecca X X X X

Lower Ecca X X X X X

Dwyka X X X

Ventersdorp X

Basalts

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY

SOUTH AFRICAN JOURNAL OF GEOLOGY

131

Stormberg Group over the Tarkastad Subgroup in thenorthern part of the main Karoo Basin. These reportedxenoliths are encouraging as they have logicalexplanations and point to the reliability of thesandstone-identification procedure. Previous visualobservations indicating that Stormberg Group xenolithsoccur in the Jagersfontein and Kimberley-cluster pipes(Wagner, 1914; Williams, 1932; Hawthorne, 1975) werenot supported by our study. In the case of Jagersfonteinthis may be the result of our limited sample suite.

Some sandstone xenoliths from the Finsch pipe mayalso correlate with the Molteno and/or Elliot Formationsof the Stormberg Group, but could equally be related tothe upper Beaufort Group (Figure 3). Stratigraphicequivalents of the Molteno and Elliot Formations arevery thin to absent in the south-west portion of theKalahari basin (Johnson et al., 1996; Catuneanu et al.,

2005) to the west of the Finsch pipe. With the exceptionof Voorspoed, Lace, Monastery and possibly Finschpipes, none of the kimberlites that contain DrakensbergGroup basalt xenoliths was found to contain StormbergGroup sandstones. This further supports the observationthat the original western extent of the main Karoo Basin during Stormberg times was more restricted (Cole,1992) and confined to an area east of the Koffiefontein,Roberts Victor and Phoenix/Star kimberlites. No samplesthat could be unequivocally correlated with the ClarensFormation were recorded.

Basalt xenolithsBasalt xenoliths with Drakensberg Group affinity occurin all Group II diatremes throughout the study area andin the Group I kimberlites that lie closest to the currentoutcrops of the Drakensberg Group in the Lesotho

Figure 3. Ternary diagram showing point-counting results (Qm-monocrystalline quartz; F-feldspar; Lt-total lithic fragments) and

classification of Karoo-Supergroup sandstone xenoliths. Sample numbers as for Figure 1. Classification fields based on: Dwyka Group

(Johnson, 1991); Ecca Group (Kingsley, 1977; Johnson, 1991; Fiedler and Adelmann, 1998); Adelaide Subgroup (Johnson, 1991; Haycock

et al., 1997); Tarkastad Subgroup (Johnson, 1991; Hancox, 1998); Molteno Formation (Eriksson, 1984; Johnson, 1991; Hancox, 1998); Elliot

Formation (Eriksson, 1984; Johnson, 1991; Bordy et al., 2004) and Clarens Formation (Eriksson, 1984; Johnson, 1991).

Remnant, namely Jagersfontein, Monastery and KaalVallei. No Karoo basalt xenoliths were found in Group Ikimberlites at Koffiefontein/Ebenhaezer, the Kimberleycluster or Leicester/Balmoral (Figure 1). Basalt or ‘lava’fragments that may be correlated with the DrakensbergGroup have also been reported in diatremes northwest of the study area, from Prieska andMakganyene (Du Toit, 1954; Hawthorne, 1975).

Minor/trace element ratios were used to discriminatebetween basalt samples derived from the VentersdorpSupergroup and Drakensberg Group (Figure 4a), as wellas samples showing probable kimberlite contamination.A suite of samples from the Group I Kimberley cluster,that bore superficial visual similarities to theDrakensberg lavas, was in this way identified as comingfrom the Ventersdorp Supergroup. Several samples

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA132

Figure 4. Bivariate plots showing various geochemical discriminants for Drakensberg-Group basaltic xenoliths from Group I (black

diamond) and Group II (open diamond) kimberlites and Ventersdorp Supergroup xenoliths (star) from the Group I Kimberley cluster; (a)

Zr/Y vs P/Zr plot showing fields for general Drakensberg-Group basalts (Marsh et al., 1997) and Allanridge Formation of the Ventersdorp

Supergroup (Bowen et al., 1986); (b) Zr/Y vs P/Zr plot showing fields for various stratigraphic units from the Barkly East Formation

(Wonderkop, Moshesh’s Ford, Golden Gate, Roma) and Lesotho Formation of the Drakensberg Group (Marsh et al., 1997); (c) P/Zr vs FeO

plot showing similar fields (W-Wonderkop; MF-Moshesh’s Ford) as well as the field for evolved Oxbow dykes; (d) P/Zr vs Mg# plot showing

similar fields and the field of evolved basalts from the Markt pipe.

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY

SOUTH AFRICAN JOURNAL OF GEOLOGY

133

showed distinct enrichments in the alkali elements Naand K over unaltered basalts (Figure 5). As amygdaloidallavas were selectively sampled, it is conceivable thatsome alkali enrichment came from secondary vesicle-fillminerals which are predominantly zeolites, quartz andcalcite (Dunlevey et al., 1993). However there was noassociated Ca enrichment to support this contention.Samples that were strongly enriched in K cameexclusively from Group II kimberlites that arethemselves K-rich. Low-temperature post-magmatic andnear-surface ground-water-related alteration hasoccurred in many kimberlites with the formation of Na-and K-rich secondary minerals such as natrolite,analcime, pectolite, apophyllite and vermiculite reportedassociated with kimberlite xenoliths (Berg, 1968; Kruger,1980, 1982).The nature of the Na and K enrichments hasnot been clearly identified in this study, but is likelyrelated to the latter processes.

Analysis and correlation of the basalt xenoliths withthe Drakensberg Group, using FeO, Mg number (Mg#)and immobile minor/trace elements P, Zr, Y (Figure 4),indicate that the Lesotho Formation covered the entirestudy area at the time of emplacement of the Group IIkimberlites (140 to 100 Ma) (Table 3). This is notunexpected as the Lesotho Formation represents thethicker and more widely spread of the two formationswithin the Lesotho Remnant. Here the DrakensbergGroup attains measured thicknesses of 1500 m, with thelower Barkly East Formation averaging less than 200 mand the Lesotho Formation in excess of 1000 m (Marshet al., 1997). Most pipes contain basalt xenolithscorrelated with the lower Mafika Lisui unit of the

Lesotho Formation and/or with the upper Maloti, Senquand Mothae units that are difficult to separatelithogeochemically (Table 3, Figure 4c). These findingsare in agreement with previous geochemical studies atVoorspoed, Melton Wold and Uintjiesberg kimberlites(Roberts, 1997; Letsale, 1998; Mambali, 1998). Xenolithsamples that could possibly be correlated with theBarkly East Formation were collected at Jagersfonteinand Monastery Group I pipes and from Roberts Victor,New Elands, West End and Melton Wold Group II pipes.Basalt xenoliths from the Markt pipe, in the extremenorth-west part of the study area, have low Mg# (Figure4d), are enriched in FeO and incompatible elementssuch as Ti and Zr and appear to be evolved Karoo lavas,possibly from the uppermost Lesotho Formation(Oxbow unit) or from an upper unit not preserved in theLesotho Remnant.

At Koffiefontein, Karoo-correlated basalt xenolithswere observed by Harger (1913) and Wagner (1914),where confusion with Ventersdorp Supergroup lavas isnot possible, but were not noted by Naidoo et al. (2004)or in the present study. No Karoo basalt xenoliths havebeen observed in the Group I Kimberley mine dumpsover the last 25 years (JVA Robey, pers. observation) andthe only documentation of such xenoliths is presentedby Williams (1932). Furthermore, Karoo basalt xenolithswere not directly described by Hawthorne (1975),Clement (1982) or Clement et al. (1986), nor were theyfound by Rambula (2005). Whilst it is conceivable thatbasalt xenoliths were overlooked during the more recentstudies, these pipes do not host the abundance of basaltxenoliths that is found at other Group I kimberlites, suchas Jagersfontein, Monastery and Kaal Vallei.

DiscussionImplications for kimberlite emplacementOn the assumption that xenoliths of lithologies sampledduring kimberlite emplacement are distributed uniformlythroughout the diatreme, lack of wallrock xenoliths fromcertain stratigraphic units would indicate that those stratawere absent at the time of kimberlite emplacement.However it is conceivable that variable processes withinthe lower diatremes of certain kimberlites, such as lessexplosive eruption, excessive host-rock fall-back ormultiple eruptive events, might result in a non-uniformdistribution of xenoliths with depth in the diatreme.Thus the presence or absence of xenoliths found in anyone kimberlite might be a function of the presenterosion level or mined portion within that pipe.

Karoo-correlated basalt xenoliths were not observedin certain Group I kimberlites that are eroded down totheir middle-to-lower diatreme zones. The Leicester-Balmoral pipes outcrop at the Karoo Supergroup/basement contact and are dominated by shale xenolithspresumably from the Ecca and lower Beaufort Groups.The Kimberley and Koffiefontein pipes outcrop in thelower-to-middle Ecca Group and contain sandstonexenoliths from the upper Beaufort Group. The Jagersfontein, Kaal Vallei and Monastery pipes

Figure 5. Bivariate plot of Na2O vs K2O for Drakensberg-Group

basaltic xenoliths from Group I (black diamond) and Group II

(open diamond) kimberlite pipes with the general field for the

Drakensberg Group in the Lesotho Remnant (Marsh et al., 1997).

outcrop in the upper Ecca and Beaufort Groups andcontain basalt xenoliths from the Drakensberg Group.This may indicate a broad ‘stratification’ to the xenolithsuites in Group I kimberlite diatremes. The absence ofKaroo basalt xenoliths in the preserved portions of thediatremes would therefore not necessarily imply that the Drakensberg Group had eroded away prior toeruption of the Group I kimberlites.

In contrast, xenoliths of basalt from the DrakensbergGroup and sandstones from the Beaufort Group arecommon at many Group II kimberlites, several of whichare eroded to their root zones (e.g. Wimbledon,Phoenix/Star, New Elands, Roberts Victor), indicatingthat xenoliths from across the entire Karoo stratigraphyare down-rafted to the very base of the diatreme inGroup II kimberlites. Large rafts of Drakensberg Groupbasalt are reported from Voorspoed and Finsch pipes atdepths significantly below their projected stratigraphicoccurrence (Clement, 1982; Roberts, 1997). Theseobservations indicate that the processes of formation ofGroup II kimberlite diatremes involved the evacuationof the diatreme at some stage, with subsequent fall-backof material containing a significant component of near-surface material.

Drakensberg Group basalt xenoliths are numerous atJagersfontein and occur at Kaal Vallei, Uintjiesberg andMonastery pipes, indicating that similar eruptiveprocesses to Group II kimberlites were probably activeat these Group I pipes. The present-day outcroppingMonastery pipe occurs near the diatreme/root zoneboundary (Whitelock, 1973) and contains significantnumbers of basalt xenoliths. The Lovedale No. 1 (Robey,

1981) and Uintjiesberg Group I pipes would also appearto be eroded to levels close to the root zone and contain amygdaloidal basalt xenoliths. If Group I and IIkimberlite diatremes form by similar mechanisms, then itis likely that Group I diatremes were also evacuatedduring their formation and that the absence of basaltxenoliths of the Drakensberg Group indicates that Karoobasalts were absent or very rare (i.e. eroded away) in theLeicester, Kimberley and Koffiefontein areas at the timeof Group I kimberlite eruption. Diatremes of Group Iand II kimberlites have similar dimensions (length,shape, wallrock-interface angle) and general rock types(tuffisitic kimberlite breccia) indicating that similarphysical forces operated during their formation (Rice,1999).

Physico-chemical models for kimberlite-diatremeformation are highly contested and not well constrainedor quantified, with both magmatic (Clement and Reid,1989; Skinner and Marsh, 2004; Sparks et al., 2006) andphreatomagmatic (Lorenz, 1975; Lorenz and Kurszlaukis;2003) processes advocated. In phreatomagmatic modelsinvolving groundwater-magma interaction, the processesare likely to be similar for Group I and II kimberlites andcommonly involve total initial evacuation of thediatreme (Lorenz and Kurszlaukis, 2003). Magmaticmodels generally involve the separation of a magmaticvapour phase from the rising magma that causesrupturing to surface once the lithostatic pressure isexceeded by the vapour pressure. Again, the physicalprocesses of diatreme formation would seem to besimilar for both Group I and II kimberlites.

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA134

Figure 6. Cross-section through the study area (see Figure 1), showing projected surface profiles at 180 Ma, 120 Ma, 85 Ma and 65 Ma as

well as the present-day land surface. The surface profiles were partly constructed based on an assumed 1350 m depth profile for the

kimberlites. The 120 Ma and 85 Ma escarpments were based on projections from outside the cross-section.

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY

SOUTH AFRICAN JOURNAL OF GEOLOGY

135

Implications for Cretaceous landscape evolutionThe variation in upper-crustal xenolith composition ofthe Group I kimberlites is therefore probably related to erosional/topographic differences at the eruptionsites, with the Kaal Vallei, Monastery, Jagersfontein,Uintjiesberg and Lovedale Group I kimberlites eruptingthrough the entire Karoo stratigraphy, whereas theKoffiefontein, Kimberley-cluster and Leicester/BalmoralGroup I kimberlites erupted through a truncatedsequence, devoid of Drakensberg Group basalts andStormberg Group sediments. The former are absent dueto active Cretaceous weathering and the latter due tobasin-margin pinch-out. The abundance of Karoo basaltxenoliths at Jagersfontein and their absence or rarity atthe nearby contemporary Koffiefontein pipe wouldappear to indicate the presence of an escarpment,similar to the modern-day Lesotho-Remnant escarpment,separating the two pipes (Figure 1).

The distribution of Karoo-basalt xenoliths within theGroup I kimberlites reveals that the Drakensberg Groupbasalts probably extended only as far west as Kaal Vallei,Jagersfontein, and Lovedale by 85 Ma. This pattern canbe explained by the presence of an eastward-retreatinginland escarpment (Figure 6) that occupied a positionsome 180-200 km west of the present-day outcrop of theLesotho Remnant. The Monastery pipe has been erodedto the diatreme/root zone contact (present surface area0.6 ha), whereas the other Group I pipes (Koffiefontein,Kimberley cluster) are all only eroded to the mid-diatreme level (present surface areas between 5 and 12 ha). The Monastery pipe therefore has experiencedsignificantly greater post-emplacement erosion(estimated at ~1350 m) than the other pipes (~850 m),supporting arguments for the presence of thisescarpment. The Jagersfontein and Kaal Vallei pipes aresituated close to the proposed escarpment and have anintermediate signature, having relatively large diatremes,but containing common Drakensberg Group basaltxenoliths. Over the past 85 Ma, the inland erosion fronthas continued to retreat eastwards, at an average rate ofjust over 2 km/Ma, to its present position at the westernmargin of the Lesotho Remnant.

Evolved-basalt xenoliths from the upper LesothoFormation come mostly from the adjacent Voorspoedand Lace pipes in the extreme north of the study area.Further north in the Springbok Flats Karoo sub-basin,the Drakensberg Group basalt stratigraphy indicatestruncation of the lower Lesotho Formation equivalentsand greater thicknesses of the more evolved upperLesotho Formation equivalents (Marsh et al., 1997). This appears to be the case in the Voorspoed-Lace areaas well, which would explain the dominance of upperLesotho Formation evolved xenoliths in these pipes.

Scarp-retreat erosion was also probably proceedingduring the eruption of the Group II kimberlite pipes(140 to 114 Ma). The Voorspoed and Lace kimberlites inthe north of the study area (Figure 1) are diatremesexposed in the Ecca Group with their root-zoneboundaries some 300 m below surface near the

Karoo Supergroup/basement boundary (G. Howarth,pers. observation), whereas the Phoenix/Star kimberlites70 km to the south are dykes with small pipes, at thediatreme/root-zone boundary, exposed in the overlyingBeaufort Group. An explanation for this root-zoneelevation difference (approx. 500 m) would be that anescarpment existed between the two areas (Figures 1and 6).

Such a scarp-retreat model indicates that atremendous volume of material was removed from thecentral part of South Africa in the last 140 Ma. An inlandDrakensberg-Group basalt escarpment has retreated atleast 200 kilometers eastward to the present Lesotho-Remnant outcrop during this period. This represents asimilar distance to that estimated for external scarpretreat from the east coast of South Africa to the presentoutcrop on the Great Escarpment over the same 140 Maperiod (Partridge and Maud, 1987). Rates of erosionabove the escarpment in the Lesotho Remnant havebeen an order of magnitude less than those below(minimum of 300 to 400 m compared to ~2300 m over 180 Ma; Dunlevey et al., 1993; this study). The largesurface areas of certain kimberlite pipes in the Lesothohighlands outcropping above 2500 m altitude (Letseng leTerai ~15.9 ha; Kao 19.8 ha; Mothae ~8.8 ha; Liqhobong~8.5 ha) tend to support this argument. The presence ofthis inland escarpment has been overlooked in most ofthe geomorphological models for southern Africa. Suchlarge-volume erosion in central South Africa wouldrequire significant contemporaneous uplift in the samearea (see Brown et al., 1998; 2000).

The Cretaceous palaeo-surfaces (Figure 6) all drainnorth-westward away from the north-east strikingpalaeo-escarpment and indicate internal drainages incentral South Africa throughout the Cretaceous (Mooreand Moore, 2004). This continued until the end-Cretaceous (~65 Ma) as evidenced by the north-westward palaeo-drainage at Mahura Muthla whichcontains transported fossil wood, sandstone clasts andagates of Karoo Supergroup origin together with in-situlate Cretaceous fossil wood (Ward et al., 2004). The Vaaland Harts drainages are shown to be post-Cretaceousfeatures that developed from exhumation of south-west-draining basal Karoo glacial valleys that effectivelybeheaded the north-west Cretaceous drainages.

Implications for kimberlite erosionErosion estimates are characterised by large errormargins that result from uncertainties about the originalthickness of the Karoo-Supergroup units in the studyarea. Borehole data from the eastern edge of the study area near Monastery pipe indicate thicknesses ofapproximately 40 m for the Dwyka Group, 500 m for theEcca Group, 700 m for the Beaufort Group and 250 mfor the Stormberg Group (Beukes, 1970; Winter andVenter, 1970) (Table 2). The Drakensberg Groupaverages about 1500 m in the adjacent Lesotho Remnant(Marsh et al., 1997) and is estimated to have beenbetween 1600 and 1800 m thick, based on zeolite

thermal stability (Dunlevey et al., 1993), amounting to atotal thickness of approximately 3200 m for the KarooSupergroup at this locality. In the south-west near theMelton Wold, Pampoenpoort and Uintjiesberg pipes,borehole data indicate significantly increasedthicknesses for the Dwyka, Ecca and Beaufort Groups(Winter and Venter, 1970) (Table 2).

Thicknesses in the Kimberley area are unknown andhave to be extrapolated using recorded thinning rates inremnants of the Karoo Supergroup sedimentary units(Johnson, 1976; Visser, 1984; Johnson, 1991; Johnson,1994; Veevers et al., 1994). It is generally assumed thatmost units thin towards the western margin of the mainKaroo Basin. However the basal part of the Karoo

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA136

Figure 7. Profile showing the relative positions of Group I and II kimberlites in relation to the stratigraphy of the Karoo Supergroup and

various palaeo-surfaces in the Kimberley area. No horizontal scale is provided as the pipes have variable dimensions (between 5 and 12 ha

at present-day surface in the Kimberley Group I cluster; Clement et al., 1986).

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY

SOUTH AFRICAN JOURNAL OF GEOLOGY

137

Supergroup is highly variable here, with the TierbergFormation of the Ecca Group showing maximumthicknesses of 700 m along this margin (Johnson et al.,2006) and deep glacial valleys existing locally in theDwyka Group. Uncertainty also arises concerning thetotal thickness of dolerite sills present in the stratigraphy.Hawthorne (1975) includes a total thickness of ~200 mof dolerite sills in his estimation. Winter and Venter(1970), however, indicate that the dolerite-sediment ratioreaches levels of 0.3 and greater over significant portionsof the study area.

If the Hawthorne (1975) model is reconsideredsimply with Stormberg and Drakensberg Groupsuccessions omitted for reasons discussed above, thepost-eruption loss due to erosion of the Kimberleycluster would be closer to 850 m (Figure 7) than the1400 m proposed. Hawthorne (1975) allocated 400 m tothe Dwyka and Ecca Groups, including 100 m of doleritesills. This may be an under-estimation given thepotential for increased thickness of Tierberg Formationin this area. Five hundred and fifty meters is allocated tothe Beaufort Group, including 100 m of dolerite sills.Given the north-westerly current directions and southernsource for the Beaufort Group (Johnson, 1991; Johnsonet al., 2006), thinning of the group into the Kimberleyarea, compared to the borehole near Monastery pipe(700 m thickness), is to be expected. However, theamount is unknown. For comparative purposes, thefollowing discussion assumes that the thicknessesproposed by Hawthorne (1975) for the Ecca andBeaufort Groups are correct, although in reality there isa high degree of uncertainty on the estimate.

Kimberlite pipe models generally involve thepresence of a crater facies within the uppermost part ofthe diatreme. Hawthorne (1975) allocated approximately200 m to the crater facies, although the crater facies atthe Group I Orapa pipe in Botswana is in excess of 400 m thick (Field et al., 1997). As the epiclastic rocks inany crater facies are superimposed on the underlyingpyroclastic/volcaniclastic kimberlites of the diatremefacies, they are not thought to materially affect theestimations of amounts of erosion. They are, in fact,products of the initial erosional processes themselves.However, should significant tuff rings or cones developat the eruption site, these may affect the estimations ofrates of erosion.

The Kimberley area is unique in that it hosts bothGroup II (Wimbledon) and Group I (Kimberley cluster)kimberlites within close proximity that can be comparedas a means of erosion estimation. The current erosionlevels of the Wimbledon and Kimberley-clusterdiatremes can be utilized to assess the erosion rates thatexisted in the time period between Group II and GroupI kimberlite eruption (in this case, 120 to 85 Ma).Wimbledon is eroded down to the root-zone/diatremeinterface, while this interface occurs at depths ofapproximately 400 to 600 m below present-day surfacein the Kimberley cluster (Hawthorne, 1975; Clement,1982; Clement et al., 1986). If Group I and II kimberlites

are broadly of similar size and shape in the study area(Clement and Reid, 1989), then Wimbledon had to haveexperienced ~500 m more erosion than the Kimberley-cluster pipes (Figure 7). Post-emplacement erosion forWimbledon, therefore, would have been approximately1350 m based on the model proposed above.

Erosion rates can also be estimated based on thedifferences in erosion levels between the Wimbledonand Kimberley-cluster kimberlites. From 120 Ma to 85 Ma, ~500 m of material, predominantly DrakensbergGroup basalts, was removed from the Kimberley area atan average rate of ~15 m/Ma. In the past 85 Ma,however, denudation rates have slowed to an average of10 m/Ma resulting in the erosion of 850 m of Beaufortand Ecca Group rocks and the development of themodern-day land surface. This apparent decrease inerosion rates since the late Cretaceous also applies on aregional scale (Tinker et al., 2008a). According toPartridge and Maud (1987), post-Gondwana landscapeevolution is characterized by variable denudation ratesof up to 15 m/Ma with resulting off-shore sedimentationpeaking in the early stages, from Late Jurassic to Mid-Cretaceous, and slowing through the Late Cretaceous tominimum levels by Tertiary times as landscape planationwas achieved. Crater-facies sediments are preserved inthe late Cretaceous/Tertiary kimberlite pipes atStompoor (Smith, 1986) and Banke, supporting thisargument. In the southern African region, fission-trackstudies indicate that peaks in erosion are evident in theearly (140 to 120 Ma) and mid-Cretaceous (100 to 80 Ma)(Brown et al., 2000; Tinker et al., 2008a). These peakscomplement the measured offshore sedimentation rates(Dingle et al., 1983; Tinker et al., 2008b).

Based on an eruption age of 183 Ma for theDrakensberg Group basalts (Duncan et al., 1997), andassuming a similar erosion rate to that of the 120 to 85Ma period (i.e. 15m/Ma), it is estimated that some 945 m of material, again predominantly DrakensbergGroup basalts, were removed subsequent to eruption ofthe Drakensberg Group lavas, but prior to the extrusionof the Group II kimberlites (Figure 7). This wouldrepresent a total thickness of 2295 m that has beenremoved in the Kimberley area subsequent to the Karooflood basalt eruption. It also implies that some 1445 mof Drakensberg Group flood basalts were present in theKimberley area, compared to the 1600 to 1800 mestimate in the Lesotho Fragment (Dunlevey et al., 1993).

An approximate 250 m thickness of Dwyka and EccaGroup sediments has survived erosion at the 90 MaKoffiefontein pipe (Naidoo et al., 2004) compared toabout 100 m at the Kimberley cluster (Hawthorne 1975,Clement et al., 1986). Assuming a similar erosion rate tothe Kimberley area of 10 m/Ma, some 900 m of erosionwould have occurred post-eruption. The absence ofDrakensberg-Group lava xenoliths would imply a totalthickness of Karoo-Supergroup sediments of at least1150 m, compared to 950 m in the Kimberley area. The 85 Ma Jagersfontein and Kaal Vallei pipes outcropclose to the Ecca-Beaufort contact in the lower Beaufort

Group. Assuming 10 m/Ma erosion rates, some 850 m ofpost-eruption erosion would have occurred. Both pipescontain Drakensberg lava xenoliths indicating that thecombined thickness of Beaufort and Stormberg Groupsediments was less than 850 m at these localities.

Taking erosional levels of 850 m in the Kimberleyarea, with an average of ~500 m of diatreme-facieskimberlite unexposed below surface, the mean extent ofthe diatreme facies of Group I kimberlites is estimated tobe 1350 m, which is close to that of Field and Scott-Smith (1500 m, 1999). Extending this to the Group IIkimberlites would indicate that the upper ~40% of thediatreme facies occurred within flood basalts of theDrakensberg Group (Figure 7). This would explain whybasalt xenoliths are so dominant in Group II kimberlitepipes.

Older Group I and Group II kimberlite pipes in thewest of the study area (Uintjiesberg, Melton Wold,Pampoenpoort) have been eroded down to similardiatreme/root-zone levels to Group II pipes in theKimberley area, implying similar post-eruption erosionrates (~1350 m since 120 Ma). However younger GroupI pipes in the Britstown area are either hypabyssal innature (Lushof, Britstown) or occur at the diatreme/rootzone boundary (Lovedale) (Robey, 1981). This impliesgreater erosion rates (~1350 m since 75 Ma), similar tothe situation at Monastery, indicating the probablepresence of an escarpment to the north-west ofBritstown prior to 75 Ma (Figure 1). Supporting evidenceis provided by the presence of basaltic xenoliths in theGroup I Lovedale pipe (Robey, 1981).

All these estimates measure absolute amounts oferosion and are unaffected by any differential uplift,either temporal or spatial, that might have occurred inthe study area. As a consequence, the palaeo-profilesshown in Figure 6 do not represent true profiles, but arebased on the present-day surface as a fixed datum.Proposed uplift events in the Miocene and Pliocene(Partridge and Maud, 1987) occurred along axes outsidethe study area and postdate most of the erosion historyoutlined above. Scarp-retreat models involve significantuplift in the fore-scarp regions (Gilchrist et al., 1994)which would cover much of the study area. Based onFigure 6, the maximum volumes of material removed,and hence the areas of potential maximum uplift, have migrated progressively eastwards with time fromthe western Finsch-pipe region for the period 183 Ma to 120 Ma, to the eastern Monastery-pipe region post-85 Ma.

ConclusionsThe presence or absence of xenoliths of KarooSupergroup rocks in the kimberlite diatremes in centralSouth Africa generally reflects respectively the presenceor absence of the specific stratigraphic units at the siteof eruption of the kimberlite at the time of eruption.

With regard to the original distribution of variousstratigraphic units within the main Karoo Basin, it isconcluded that the Stormberg Group had a restricted

distribution to the west of its current outcrop butextended further northwards into the Voorspoed/Lacearea. Basalts of the Drakensberg Group, mainly from theLesotho Formation, originally extended across the entirestudy area and had thicknesses of ~1500 m in theKimberley area at the time of eruption (183 Ma).

By 140 to 120 Ma, the entire surface area was stillcovered predominantly by rocks of the DrakensbergGroup, despite ~1000 m of erosion. An escarpment haddeveloped between the Voorspoed-Lace and Phoenix-Star kimberlite clusters. By 90 to 85 Ma, a further ~500 m of erosion had occurred at a rate of ~15 m/Maand an eastward-progressing escarpment comprisinglavas of the Drakensberg Group had advanced to a lineto the west of the Jagersfontein and Kaal Vallei pipessome 180 km west of its present position. From 85 Mato the present day, a further ~850 m of erosion havesubsequently taken place at a slower rate of ~10 m/yr,with an eastward scarp-retreat rate of ~2 km/Ma.

Models for kimberlite pipes in the Kimberley clustershould exclude the presence of rocks of theDrakensberg and Stormberg Groups. This reduces the extent of the diatreme/crater facies from 1900 m, asproposed by Hawthorne (1975), to 1350 m. Reductionsof this nature will lower significantly estimations of thediamond budget provided by the Cretaceous kimberlitepipes to the younger alluvial and marine deposits.

AcknowledgmentsE. K. Hanson acknowledges financial support receivedfrom the Research Committee of De Beers GroupServices Exploration Division for her MSc thesis study.We would like to thank various mining companies andfarmers for providing access to the sampling locations.Mike Skinner is thanked for fruitful discussions.Improvements to the manuscript resulted from reviewsby Craig Smith and Mike de Wit.

References Allsopp, H.L. and Barnett, D.R. (1975). Rb-Sr age determinations on South

African kimberlite pipes. Physics and Chemistry of the Earth, 9, 605–617.

Allsopp, H.L., Bristow, J.W., Smith, C.B., Brown, R., Gleadow, A.J.W.,

Kramers, J.D. and Garvie, O.G. (1989). A summary of radiometric dating

methods applicable to kimberlites and related rocks. In: J. Ross, (Editor).

Kimberlites and related rocks, Proceedings of the Fourth International

Kimberlite Conference. Geological Society of Australia Special Publication,

14, 343–357.

Allsopp, H.L. and Kramers, J.D. (1977). Rb-Sr and U-Pb age determinations

on Southern African kimberlite pipes. Extended Abstracts, 2nd

International Kimberlite Conference, Carnegie Institute, Washington,

United States of America.

Becker, G.F. (1904). Determination of rocks surrounding the Diamond pipes

of Kimberley, South Africa. In: Reports on Determination of Rocks,

Kimberley Diamond Mines, De Beers Internal Report, Kimberley, South

Africa, 1902–1905.

Berg, G.W. (1968). Secondary alteration in eclogites from kimberlite pipes.

American Mineralogist, 53, 1336–1346.

Beukes, N.J. (1970). Stratigraphy and sedimentology of the Cave Sandstone

Stage, Karroo System. In: S.H. Haughton (Editor). Proceedings 2nd IUGS

Symposium on Gondwana Stratigraphy and Palaeontology. CSIR,

Pretoria, 321–342.

Bordy, E.M., Hancox, P.J. and Rubidge, B.S. (2005). The contact of the

Molteno and Elliot formations through the main Karoo Basin, South

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA138

E.K. HANSON, J .M. MOORE, E.M. BORDY, J .S . MARSH, G. HOWARTH AND J.V.A. ROBEY

SOUTH AFRICAN JOURNAL OF GEOLOGY

139

Africa: a second order-sequence boundary. South African Journal of

Geology, 108, 351–364.

Bowen, T.B., Marsh, J.S., Bowen, M.P. and Eales, H.V. (1986). Volcanic rocks

of the Witwatersrand Triad, South Africa I: description, classification and

geochemical stratigraphy. Precambrian Research, 31, 297–324.

Brown, R.W., Gallagher, K., Griffin, W.L., Ryan, C.G., de Wit, M.C.J., Belton,

D.X. and Harman, R. (1998). Kimberlites, accelerated erosion and

evolution of the lithospheric mantle beneath the Kaapvaal craton during

the mid-Cretaceous. 7th International Kimberlite Conference - Extended

Abstracts, 105–107.

Brown, R.W., Gallagher, K., Gleadow, A.J.W. and Summerfield, M.A. (2000).

Morphotectonic evolution of the South Atlantic margins of Africa and

South America. In: M.A. Summerfield (Editor). Geomorphology and

Global Tectonics. John Wiley and Sons, Chichester, United Kingdom,

255–281.

Catuneanu, O., Hancox, P.J. and Rubidge, B.S. (1998). Reciprocal flexural

behaviour and contrasting stratigraphies: a new basin development model

for the Karoo retroarc foreland system, South Africa. Basin Research,

10, 417–439.

Catuneanu, O., Wopfner, H., Eriksson, P.G., Cairncross, B., Rubidge, B.S.,

Smith, R.M.H. and Hancox, P.J. (2005). The Karoo basins of south-central

Africa. Journal of African Earth Sciences, 43, 211–253.

Clement, C.R., (1982). A comparative geological study of some major

kimberlite pipes in northern Cape and Orange Free State. Unpublished

PhD Thesis, University of Cape Town, South Africa, 432pp.

Clement, C.R. and Reid, A.M. (1989). The origin of kimberlite pipes: an

interpretation based on a synthesis of geological features displayed by

southern African occurrences. Geological Society of Australia Special

Publication, 14, 632–646.

Clement, C.R., Harris, J.W., Robinson, D.N. and Hawthorne, J.B. (1986). The

De Beers kimberlite pipe – a historic South African diamond mine. In: C.R.

Anhaeusser and S.Maske (Editors). Mineral Deposits of Southern Africa,

Geological Society of South Africa, 2193–2214.

Cole, D.I. (1992). Evolution and development of the Karoo Basin. In: M.J. de

Wit and I.G.D. Ransome (Editors). Inversion Tectonics of the Cape Fold

Belt, Karoo and Cretaceous Basins of Southern Africa. A.A. Balkema,

Rotterdam, The Netherlands, 87–89.

Davis, G.L. (1977). The ages and uranium contents of zircons from

kimberlites and associated rocks. Annual Report of the Director

Geophysical Laboratory, Carnegie Institution, 613-635.

Dawson, J.B. (1971). Advances in kimberlite geology, Earth Science Reviews,

7, 187–214.

Dawson, J.B. (1980). Kimberlites and their xenoliths. Springer-Verlag, Berlin,

Germany, 148-183.

Dingle, R.V., Siesser, W.G. and Newton, A.R. (1983). Mesozoic and Tertiary

Geology of Southern Africa. A.A. Balkema, Rotterdam, The Netherlands,

375pp.

Duncan, A.R., Erlank, A.J. and Betton, P.J. (1984). Appendix I: Analytical

techniques and database descriptions. Geological Society of South Africa,

Special Publication, 13, 389–395.

Duncan, R.A., Hooper, P.R., Rehacek, J., Marsh, J.S. and Duncan, A.R. (1997).

The timing and duration of the Karoo igneous event, southern Gondwana.

Journal of Geophysical Research, 102, 127–138.

Dunlevey, J.N., Ramluckan, V.R. and Mitchell, A.A. (1993). Secondary mineral

zonation in the Drakensberg Basalt Formation, South Africa. South

African Journal of Geology, 96, 215–220.

Du Toit, A.L. (1906). Geological Survey of the Eastern Portion of Griqualand

West. 11th Report of the Geological Commission, Cape Colony, 89-176.

Du Toit, A.L. (1954). The Geology of South Africa, 3rd Editon, Oliver and

Boyd, Edinburgh, United Kingdom, 611pp.

Eriksson, P.G. (1984). A palaeoenvironmental study of the Molteno, Elliot

and Clarens Formations in the Natal Drakensberg and northeastern

Orange Free State. Unpublished PhD Thesis, University of Natal, South

Africa, 209pp.

Fiedler, K. and Adelmann, D. (1998). Provenance of Late Permian sediments

of the Laingsburg subbasin (SW Karoo Basin, South Africa). Journal of

African Earth Sciences, 27, Addendum, 3–4.

Field, M., Gibson, J.G., Wilkes, T.S., Gababotse, J. and Khujwe, P. (1997). The

geology of the Orapa A/K1 kimberlite, Botswana: further insight into the

emplacement of kimberlite pipes. In: Proceedings of the Sixth

International Kimberlite Conference, Kimberlites, Related Rocks and

Mantle Xenoliths, Russian Geology and Geophysics, 38, 24–39.

Field, M. and Scott-Smith, B.H. (1999). Contrasting geology and near-surface

emplacement of kimberlite pipes in southern Africa and Canada,

Proceedings of the 7th International Kimberlite Conference, 1, 217–237.

Galehouse, J.S. (1971). Point counting. In: R.E. Carver (Editor). Procedures

in Sedimentary Petrology. Wiley Interscience, New York, United States of

America, 385-407.

Gilchrist, A.R., Kooi, H. and Beaumont, C. (1994). Post-Gondwana

geomorphic evolution of southwestern Africa: implications for the

controls on landscape development from observations and numerical

experiments. Journal of Geophysical Research, 99, 12,211–12,228.

Gurney, J.J. (1989). Diamonds. Geological Society of Australia Special

Publication, 14, 935–965.

Hancox, P.J. (1998). Stratigraphic, sedimentological and palaeoenvironmental

synthesis of the Beaufort-Molteno contact in the Karoo Basin.

Unpublished PhD Thesis, University of the Witwatersrand, South Africa,

380pp.

Hanson, E.K. (2007). Estimating erosion of Cretaceous-aged kimberlites in

the Republic of South Africa through the examination of upper-crustal

xenoliths. Unpublished MSc Thesis, Rhodes University, South Africa, 138pp.

Harger, H.S. (1909). The occurrence of diamonds in Dwyka conglomerate

and amygdaloidal lavas; and the origin of Vaal River diamonds.

Transactions of the Geological Society of South Africa, 12, 139–158.

Harger, H.S. (1913). Some Features associated with the Denudation of the

South African Continent. Annual Proceedings of the Geological Society of

South Africa, XXII–XXXIX.

Hawthorne, J.B. (1975). Model of a kimberlite pipe. In: Ahrens, L. et al.,

Proceedings of the 1st International Kimberlite Conference, Physics and

Chemistry of the Earth, 9, 1–15.

Haycock, C.A., Mason, T.R. and Watkeys, M.K. (1997). Early Triassic

palaeoenvironements in the eastern Karoo Foreland Basin, South Africa.

Journal of African Earth Sciences, 24, 79–94.

Howarth, G.H. (2007). Estimating erosion of Group II kimberlite pipes and

palaeo-distribution of the Karoo Supergroup, in the Free State province of

South Africa, using upper-crustal xenoliths. Unpublished BSc Honours

thesis, Rhodes University, South Africa, 50pp.

Ingersoll, R.V., Bullard, T.F., Ford, R.I., Grimm, J.P., Pickle, J.D. and Sares,

S.W. (1984). The effect of grain size on detrital modes: a test of the Gazzi-

Dickinson point-counting method. Journal of Sedimentary Petrology,

54, 103–116.

Johnson, M.R. (1976). Stratigraphy and sedimentology of the Cape and Karoo

sequences in the Eastern Cape Province. Unpublished PhD Thesis, Rhodes

University, South Africa, 276pp.

Johnson, M.R. (1991). Sandstone petrography, provenance and plate tectonic

setting in Gondwana context of the south-eastern Cape – Karoo Basin.

South African Journal of Geology, 94, 137–145.

Johnson, M.R. (Editor) (1994). The Lexicon of South African Stratigraphy. Part

1: Phanerozoic Units. Geological Survey of South Africa Handbook, 56pp.

Johnson, M.R., van Vuuren, C.J., Hegenberger, W.F., Key, R. and Shoko, U.

(1996). Stratigraphy of the Karoo Supergroup in southern Africa: an

overview. Journal of African Earth Sciences, 23, 3–15

Johnson, M.R., van Vauuren, C.J., Visser, J.N.J., Cole, D.I., Wickens, H. de V.,

Christie, A.D.M., Roberts, D.L. and Brandl, G. (2006). Sedimentary rocks

of the Karoo Supergroup. In: M.R. Johnson, C.R. Anhaeusser, and R.J.

Thomas (Editors). The Geology of South Africa. Geological Society of

South Africa/Council for Geoscience, 461–500.

Kingsley, C.S. (1977). Stratigraphy and Sedimentology of the Ecca Group in

the Eastern Cape Province, South Africa. Unpublished PhD Thesis,

University of Port Elizabeth, South Africa, 240pp.

Kruger, F.J. (1980). The occurrence of cebollite in kimberlite and included

zeolitized crustal xenoliths. Mineralogical Magazine, 43, 583–586.

Kruger, F.J. (1982). The occurrence of cebollite in kimberlite and included

zeolitized crustal xenoliths – a correction and discussion of the occurrence

of pectolite. Mineralogical Magazine, 46, 274–275.

Letsale, B.M. (1998). The geology of the Uintjiesberg kimberlite and its

xenoliths. Unpublished BSc Honours Thesis, Rhodes University, South

Africa, 64pp.

Lindgren, W. (1905). Determination of rocks from Kimberley, South Africa.

In: Reports on determination of rocks, Kimberley diamond mines,

De Beers Internal Report, Kimberley, South Africa, 1902–1905.

Lorenz, V. (1975). Formation of phreatomagmatic maar-diatreme volcanoes

and its relevance to kimberlite diatremes. Physics and Chemistry of the

Earth, 9, 17–27.

Lorenz, V. and Kurszlaukis, S. (2003). Kimberlite pipes; growth models and

resulting implications for diamond exploration. 8th International

Kimberlite Conference – Extended Abstracts, Victoria, British Columbia,

Canada, 5pp.

MacIntyre, R.M. and Dawson, J.B. (1976). Age and significance of some

South African kimberlites. Abstract, Fourth European Colloquium on

Geochronology, Cosmochronology, Isotope Geology, Amsterdam, 66.

Mambali, A.F. (1998). Melton Wold kimberlite: Petrography, geochemistry,

xenoliths and model of genesis. Unpublished BSc Honours Thesis, Rhodes

University, South Africa, 53pp.

Marsh, J.S. and Skilling, I.P. (1998). Field Excursion A3: Karoo volcanic and

intrusive rocks, Eastern Cape. International Volcanological Congree:

Magmatic Diversity; Volcanoes and their roots. Cape Town. 48pp.

Marsh, J.S. (1998). Geochemical stratigraphy in basalts of the Mohale Dam –

Katse dam areas, Lesotho. Report on Contract LHDA 1009, Mohale Tunnel:

Sampling and testing of cores. Lesotho Highlands Tunnel Partnership

(Mohale) (unpublished). Rhodes University, South Africa, 21pp.

Marsh, J.S., Hooper, P.R., Rehacek, J., Duncan, R,A and Duncan, A.R. (1997).

Stratigraphy and age of Karoo basalts of Lesotho and implications for

correlations within the Karoo Igneous Province. In: J.J. Mahoney and M.F.

Coffin, (Editors). Large Igneous Provinces: Continental, Oceanic, and

Planetary Flood Volcanism. Geophysical Monograph 100. American

Geophysical Union, Washington DC, 247–272.

Matthews, J.W. (1887). Incwadi Yami, or Twenty Years’ Personal Experiences

in South Africa. Rogers and Sherwood, New York, United States of America,

542pp.

Moore, J.M. and Moore, A.E. (2004). The roles of primary kimberlitic and

secondary Dwyka glacial sources in the development of alluvial and

marine diamond deposits in Southern Africa. Journal of African Earth

Sciences, 38, 115–134.

Naidoo, P., Stiefenhofer, J., Field, M. and Dobbe, R. (2004). Recent advances

in the geology of Koffiefontein Mine, Free State Province, South Africa.

Lithos, 76, 161–182.

Nixon, P.H. (1987). Mantle Xenoliths. Wiley, New York, 844pp.

Norrish, K. and Hutton. J.T. (1969). An accurate X-ray spectrographic method

for the analysis of a wide range of geological samples. Geochemica et

Cosmochimica Acta, 33, 431–453.

Partridge, T.C. and Maud, R.R. (1987). Geomorphic evolution of southern

Africa since the Mesozoic. South African Journal of Geology, 90, 179-208.

Phillips, D., Kiviets, G.B., Barton, E.S., Smith, C.B., Viljoen, K.S. and Fourie,

L.F. (1999). 40Ar/39Ar dating of kimberlites and related rocks: problems and

solutions. Proceedings of the 7th International Kimberlite Conference,

2, 677–687.

Rambula, T. (2005). Karoo-age xenoliths in the Kimberley kimberlites –

implications for post-Cretaceous erosion in South Africa. Unpublished BSc

Honours Thesis, University of the Witwatersrand, South Africa, 34pp.

Rice, A. (1999). Can the blasting excavation engineering sciences provide

insight into the processes of kimberlite emplacement and eruption?

Proceedings of the 7th International Kimberlite Conference, 2, 699–708.

Roberts, M.A. (1997). The geology of the Voorspoed kimberlite and its

xenoliths. Unpublished BSc Honours Thesis, Rhodes University, South

Africa, 105pp.

Robey, J.V.A. (1981). Kimberlites of the central Cape Province, Republic of

South Africa, Unpublished Ph.D Thesis, University of Cape Town, South

Africa, 261pp.

Rogers, A.W. and Du Toit, A.L. (1909). Kimberlite and Allied Pipes and

Fissures in Carnarvon and Victoria West. 14th Annual Report of the

Geological Commission, Cape Colony, 98-108.

Scrutton, R.A. and Dingle, R.V. (1976). Observations on the process of

sedimentary basin formation at the margins of southern Africa.

Tectonophysics, 36, 143–156.

Skinner, E.M.W. and Marsh, J.S. (2004). Distinct kimberlite pipe classes with

contrasting eruption processes. Lithos, 76, 183–200.

Smith, C.B., Allsopp, H.L., Kramers, J.D., Hutchinson, G. and Roddick, J.C.

(1985). Emplacement ages of Jurassic-Cretaceous South African

kimberlites by the Rb-Sr method on phlogopite and whole rock samples.

Transactions of the Geological Society of South Africa, 88, 149–169.

Smith, C.B., Clark, T.C., Barton, E.S. and Bristow, J.W. (1994). Emplacement

ages of kimberlite occurrences in the Prieska region, southwest border of

the Kaapvaal Craton, South Africa. Chemical Geology (Isotopic Geosciences

Section), 113, 149–169.

Smith, R.M.H. (1986). Sedimentation and palaeoenvironments of late

Cretaceous crater lake deposits in Bushmanland, South Africa.

Sedimentology, 33, 369–386.

Sparks, R.S.J., Baker, L., Brown, R. Field, M., Schumaker, J., Stripp, G. and

Walters, A. (2006) Dynamical constraints on kimberlite volcanism. Journal

of Volcanology and Geothermal Research, 155, 18–48.

Tinker, J., de Wit, M. and Brown, R. (2008a). Mesozoic exhumation of the

southern Cape, South Africa, quantified using apatite fission track

thermochronology. Tectonophysics, 455, 77–93.

Tinker, J., de Wit, M. and Brown, R. (2008b). Linking source and sink:

Evaluating the balance between onshore erosion and offshore sediment

accumulation since Gondwana break-up, South Africa. Tectonophysics,

455, 94–103.

Van Zijl, J.S.V., Graham, K.W.T. and Hales, A.L. (1962). The palaeomagnetism

of the Stormberg lavas of South Africa, I: Evidence for a genuine reversal

of the Earth’s field in Triassic-Jurassic times. Geophysical Journal of the

Royal Astronomical Society, 7, 23–29.

Veevers, J.J., Cole, D.I. and Cowan, E.J. (1994). Southern Africa: Karoo basin

and Cape Fold Belt. In: J.J. Veevers, and C.McA. Powell (Editors).

Permian-Triassic Pangean Basins and Foldbelts along the Panthalassan

Margin of Gondwanaland. Geological Society of America, Boulder,

Colorado, Memoir, 184, 223–279.

Visser, J.N.J. (1972). Sedimentêre insluitsels van Karoo-ouderdom in

kimberliet van die Finsch-diamantmyn. Tydskrif vir Natuurwetenskappe,

32–36.

Visser, J.N.J. (1984). A review of the Stormberg Group and Drakensberg

volcanics in southern Africa. Palaeontologia Africana, 25, 5–27.

Wagner, P.A. (1914). The diamond fields of South Africa. Transvaal Leader

Johannesburg, South Africa. 347pp.

Ward, J.D., de Wit, M.C.J., Bamford, M.K. and Roberts, M. (2004). Gravels,

woods and diamonds of the Late Cretaceous Mahura Muthla alluvial

deposit, Northern Province, South Africa: Palaeo-drainage implications.

Geoscience Africa 2004, Abstract volume, University of the Witwatersrand,

689pp.

Whitelock, T.K. (1973). The Monastery mine kimberlite pipe. In: P.H. Nixon

(Editor). Lesotho Kimberlites. Lesotho National Development Corporation,

Maseru, 214–218.

Williams, A.F. (1932). The Genesis of Diamonds. Ernest Benn Ltd., London,

United Kingdom, 636pp.

Winter, H. de la R. and Venter, J.J. (1970). Lithostratigraphic correlations of

recent deep boreholes in the Karoo – Cape sequence. In: S.H. Haughton

(Editor). Proceedings 2nd IUGS Symposium on Gondwana Stratigraphy

and Palaeontology. CSIR, Pretoria, 395–408.

Editorial handling: J.M. Barton

SOUTH AFRICAN JOURNAL OF GEOLOGY

CRETACEOUS EROSION IN CENTRAL SOUTH AFRICA140