The Bushmanland kimberlites and related rocks

7 THE BUSHMANLAND KIMBERLITES AND RELATED ROCKS By A. K. CORNELISSEN* and W. 1. VERWOEROt

ABSTRACT A cluster of several hundred pipe-like bodIes of Cretaceous to early TertIary age are descrIbed from an area where alluvIal diamonds occur. Most pipes are filled with crater deposits but drIlling has shown that this IS very oftell underlalll by kImberlItlc tuff and breccIa. Early reports of dlamondiferous kimberlite have not been substantIated. Plugs of olivine melIlitite are also present but their relationship with the other pipes is not clear. Emplacement of the pipes was controlled by joints. The volcanic activity of which they form part was of typical kratogenic type and appears to have been restrIcted to a 400 km wide upwarped zone parallel to the west coast Bore-hole samples show a deficIency of magnesIa III comparison with typical kimberlIte but pyrope-rich garnet, picroilmenite and chrome-dlOpside are found III some. It is concluded that dlamondlferous kimberlite may yet be dIscovered in tl1lS area.

INTRODUCTION

The Bushmanland plateau is situated south of the Orange River and approximately 100 km inland, adjoining a highly dissected portion of the Western Escarpment of South Africa. Kimberlite was first reported from this area by REUNING (1931) who sought to establish the source of the coastal diamonds at the mouth of the Buffels River. He claimed the discovery of a kimberlite dyke and seven pipes, including a diamondiferous one on the farm Burtonsputs. Subsequently Du TOIT (1939) stated that a small diamond had been found in a pipe on the boundary between the farms Gamoep and Koppieskraal. On chemical grounds he considered this occurrence to be intermediate between kimberlite and melilite basalt (i.e. olivine melilitite). Melilite-bearing rocks have long been known in the same area (RoGERs, 1911) and a genetic relationship with kimberlite was proposed by TALJAARO (1937). The pipes attracted little further attention from geologists until alluvial diamonds were also discovered at Galputs and Bosluispan along the disrupted eastward drainage of the Koa river. An intensive prospecting campaign followed during the years 1961-6. At least 151 pipe-like features were drilled to depths ranging from 20 m to 279 m and material resembling kimberlite was encountered in seventy-seven of them. It is the purpose of this paper to give a brief outline based on available data and to summarize the geological information brought to light by prospecting.

* 19 JoslIng Street.

Upington 8800. South Africa.

t UniverSIty of Stellenbosch. Stellenbosch 7600. South Africa. 71

72

A.

CORNELISSEN AND W.

K.

VERWOERD

J.

DISTRIBUTION OF THE PIPES In regional context the Bushmanland pipe swarm belongs to a zone of predominantly alkaline volcanic activity parallel to the west coast. Although the ages of some are still uncertain, this 4oo-km wide zone would include the Palaeocene trachyte and phonolitic trachyte intrusions on the Agulhas Bank (DINGLE and GENTLE, 1972); the melilite-bearing rocks of Heidelberg, Robertson, Sutherland, Bitterfontein, Pofadder and Pomona; several recently discovered carbonatites in South West Africa and the Salpeterkop and Brukkaros volcanoes; kimberlite pipes mentioned by JANSE (1959); diatremes in the Great Karasberg and Tiras mountains; the Klinghardt phonolites and the trachytes and phonolites between Windhoek and Rehoboth (Fig. 1). It seems reasonable to postulate that this volcanism was related to a continental warp axis parallel to the rift that eventually became the Atlantic Ocean.

.

WINDHOEK • . .I

'

A DIATR£ME

.: KIMBERLITE

I

.. CAR80NATlTE

SWA

o

OLIVINE MELILITITE

•

PHONOLITE

I

TRACHYTE

LOCAL ITIES

,

1 Auas Mts 2 GI bean 3 Brukkaros

'4

ras Mts 5 Porona

4 TI

• "'7

6 Kllnghardt Mts

7 Great Karas Hts 8 Pofadder 9 10 11 12 13 14

BUSHMANLAND PIPE SWARM Bltterfonteln Sutherland

Robertson Heidelberg Alphard Banks

SOUTH

AFRICA

FIG.!. Zone of kratogenic volcanism parallel to the west coast of South Africa, probably coinciding with Cretaceous upwarp preceding drift

From Platbakkies in the south to Aggeneys in the north some 270 pipe-like bodies are now known in an area of approximately 8400 km 2 • Without any doubt many more remain to be discovered as only about two-thirds of this area was held under option and properly prospected. In the'dissected country the pipes generally show up on aerial photographs as

THE BUSHMANLAND KIMBERLITES AND RELATED ROCKS

73

knolls or circular depressions among gneiss outcrops but towards the east they are covered by calcrete which reaches a thickness of 1-2 m on the pipes. Magnetometer traverses proved useful in locating some pipes and in delineating their boundaries. An aeromagnetic map by the Geological Survey shows hundreds of small circular anomalies but due to inadequate ground control these cannot be correlated with individual pipes. It seems likely that the Bushmanland swarm consists of several clusters in which pipes are more closely spaced; outliers which mayor may not belong here are found as far afield as Bitterfontein and Pofadder. LEGEND •

PiP"" Identified as klmberll tiC

o Sedlment-fl lied pipes

30°00'

0°

19°00'

FIG.

2. The Bushmanland Pipe Swarm. Country rock is 1000 m.y. old granite gneiss.

The distribution map (Fig. 2) shows that the pipe swarm as a whole has a north-northeasterly trend but no further structural control is apparent. Joint directions can only be established in the dissected area west of the plateau. The Riembreek-Kap Kap area was chosen for this purpose. According to JOUBERT'S (1971) detailed map several pipe lineaments can be recognized here and the area is traversed by a prominent set of north-westerly trending faults. The regional foliation of the gneisses appears to have exerted no influence

74

A. K. CORNELISSEN AND W. J. VERWOERD

on the emplacement of the pipes (Fig. 3). Remarkably enough, some of the minor joint directions (N. 25° E. and N. 47° E.) are found to have played as important a part as the major faults (Fig. 4).

,

I

I I

1

I I'

\

,

,\ I

\

I

1,

I

1

,

,\\' , , "I, ~

,

\\

"

........

~

\

\ ,

", , "

\

"\'

I ,\

~~

'~~

.,,'

't;

.

II

I

'\

I

.,

... , I

.\"

/ ,~:'.

\

,

", ,\

, , ,\

'\.'

'i

""

E3

\"

E--3

2

,

e ,9'

~.

\

@

E=3

4

I

0

...

\

,,",

I

/ 0

I,

\-

1"'" i" \

/, I' I

\

I

.\

"/ \

" e\ 4i .',',\:.,'. .. 'I I'

'I

IHAH 1

\II

\

'I.,

\'"

\\,:'

e , ,\ • \

".,'

:

\ I

')

i,'

,

(;)

,

\ ~ .<\

\

1\

'\

\

\

,.

'

,

\

" .\ \ '

I'~

\

\

0:»

e\

\,

,, " .,

e,

II .dI ., , •

' I\ \ , el \\ ,,'

\

\

..

\, .....,"\, \ , " I , . , ...." .. \ 1'\ \

\,

\

,' I

"

~, \.

I \ \ ,\ ,~ 'l,\'

\

"

'\ '\

...., .

"

\

I

\

,'

I

I

,'\

\'

\

I I

'", ", ' \"\1' , \

\ ..

,

E::3

6

I Km 8

FIG. 3. DIstribution of pipes (ornamented circular areas) in the Rlembreek-Kap Kap area (shown on Fig. 2) in relatIOn to faults (heavy dashed lines) and trend of foliation (dotted Imes).

_~o------------::::::ll:::c---

[al

--E

[bl

FIG, 4. OrientatIOn diagram of (a) 328 Joints and (b) eighteen pipe lineaments m the RlembreekKap Kap area.

THE BUSHMANLAND KIMBERLITES AND RELATED ROCKS

75

DESCRIPTION The pipes fall in three categories: (a) Olivine melilitites and olivine-nepheline melilitites which are being studied by A. E. Moore at the University of Cape Town. They sometimes form conspicuous brown rubble-strewn hills clearly distinguishable from the others. Here the rocks are perfectly fresh. However, on the calcrete and sand-covered plateau the melilitite-bearing plugs are as poorly exposed as the other pipes. (b) Sediment and breccia-filled diatremes. This description applies to the majority of occurrences. They range from 50 m to 500 m in diameter and are filled by shale, sandstone, grit, arkose and conglomerate showing graded bedding. The sediments are often disturbed by great blocks of country rock that collapsed into the crater. The latter may be adjoined by a volcanic neck (as on Riembreek and Kap Kap), consisting wholly of angular blocks of gneiss measuring up to several metres across and testifying to repeated explosive eruptions. More tranquil crater-filling conditions and a temperate climate are reflected elsewhere by deposits of carbonaceous shale, dysodile and mudstone with calcareous intercalations. These contain fossil frogs and dicotyledonous leaf impressions (HAUGHTON, 1931;

(A)

(B)

(c)

100

D

100

SiliCified

zon~

~Shale

D B

Arkose and grit Conglomerate

200

=

300 Metres

~BreCCla ~

o

C

Tuffaceous kimberlite Kimberlite Granite country rock

FIG. 5. (A) Geological map of Gamoep PIpe GPI illustratmg relatIOnship between kimberlIte dyke and crater deposit (B) One of the sediment-filled depressions on Kap Kap which probably overlies a kimberlite pipe (after H. Jansen's map). (C) Geological sectIOn through Koppieskraal Pipe K 5 drilled to a depth of 255 m.

76

A.

K.

CORNELISSEN AND W.

J.

VERWOERD

RENNIE, 1931; KIRCHHEIMER, 1934). Drilling has proved that such sediments extend down to depths of 250 m (Koppieskraal Pipe K5) and more than 266 m (Hoendernesvlei). In many cases, e.g. Burtonsputs, Gamoep GP 1 and Koppieskraal K5, the shales were found to have intercalations of blue-green tuffaceous kimberlite (cf. chemical analysis by REUNING, 1934) and eventually to overlie solid blue-ground resembling kimberlite breccia. The prospecting programme thus strengthened the opinion that the diatremes and sediment-filled depressions are surface expressions of deep-seated pipes with kimberlitic affinities. A peculiar feature of the sediment-filled pipes is the presence of late-stage silicification around their margins: ferruginous opal, veins of opaque white silicified kaolin and sometimes manganese oxide encrustations afford excellent indications ofthe proximity of a pipe. In a semi-arid area such as this the pipes are of considerable importance as aquifers. (c) A few bodies (some of them dyke-like rather than pipe-like) are known where weathered kimberlitic rock is exposed at the surface. The best-known occurrence is Gamoep GP 1 (referred to by Du Ton, 1939) where an off-shoot from the sediment-filled pipe consists of "hardebank". It contains nodules of an earlier kimberlite (?) in which chrome-diopside is concentrated. No undoubted eclogite or peridotite inclusions have been found. The only known occurrence with a large amount of biotite in a weathered serpentinous groundmass is located on the boundary between Papkuilsfontein and Couragiefontein.

AGE The Bushmanland pipes cannot be dated closely. The fossil evidence from the crater deposits indicate an Upper Cretaceous or Eocene age. Detailed investigation of fossils collected during the prospecting activities is still in progress.

CHEMICAL COMPOSITION The results of the prospecting were not only disappointing, but it actually failed to yield a single diamond from any of these pipes, despite the fact that several thousand loads were treated in washing plants. Consequently some doubt has arisen about the validity of their identification as kimberlites and the older reports have come to be regarded with skepticism. Figure 6A shows that it is possible to distinguish with a fair degree of assurance between kimberlite and the melilite-bearing rocks of South Africa on tl)e basis of partial chemical composition. There is only a small area of overlap when all the published analyses are plotted on an AI 2 0 r MgO-FeO diagram. The O'okiep Copper Company carried out analyses for Si0 2 , A1 2 0 3 , FeO, CaO and MgO on 512 samples of Bushmanland kimberlite. Approximately half of these were on near.surface samples and were rejected. According to Fig. 68 the rest are nearly all deficient in magnesia in comparison with classical kimberlite. The CaO and Si0 2 contents were not utilized because these constituents mainly depend on the calcite content which is not considered diagnostic. An interesting feature appears when the data of individual pipes are compared (Fig. 6c). In most cases the AI 2 0 3 /MgO/FeO ratio shows comparatively little variation in each pipe but considerable variation from one pipe to the next. Could this be a differentiation trend? The occurrences which are closest to a kimberlitic composition on this basis are Vermeulens Rust VR1, Klein Katvlei KK1 and Gamoep GP1 (Table 1).

THE BUSHMANLAND KIMBERLITES AND RELATED ROCKS

I

(A)

I ,:' .

(B)

(el

:'

/

f

~

.....

O"~~" Gl'IS

liP 14

C' '"

... qo,L------~---------------'F.O

FIG. 6. (A) TrIangular diagram showing a comparison of the AI 20 3 /MgO/FeO (total Iron) ratIOs of Southern African klmberlites and melilite basalts. Analyses from HALL (1938), MACKINLAY (1955), DAWSON (1962), TALJAARD (1937) and VISSER (1964). 1,2 = Average dyke and pipe kimberlite after DAWSON (1962). (B) The AI 2 0 3 /MgO/FeO (total Iron) ratios of 200 samples of Bushmanland kimberlite superImposed on the kImberlite and melilite basalt fields. (C) Diagram sho'wing the data of Fig. 6(B) pertaimng to individual pipes. VR = Vermeulens Rust, KK = Klem Katvlel. GP ~ Gamoep, KA = Kamiebees, KP = Koppieskraal.

TABLE I. PARTIAL ANALYSES OF SOME BUSHMANLAND KIMBER LITES ----------

-

-~-~-

-

Klein Katvlei KKl, BHl, 120'--140'

22.88 6.89 8.83 12.17 20.97

36.98 10.57 14.30 6.58 20.81

GamoepGPI BH4,320'-329'

------

-

SI02 Al 20 3 FeO (total Fe) CaO MgO

------

Vermeulens Rust VRI, BRS, 300'-320'

----

------ -

38.80 10.06 10.80 4.73 17.58

77

78

A.

K.

CORNELISSEN AND W.

J.

VERWOERD

MINERALOGY

In thin section and in hand specimen the Bushmanland rocks resemble kimberlite. However, the field designation of pipes as kimberlitic was based on the recognition of chrome diopside, ilmenite and deep red garnet in heavy concentrates from boreholes. In addition the following minerals have been found: orthopyroxene (En so _ 85), zircon, magnetite, biotite, olivine, perovskite, pyrite and barite. Pyrite is a common constituent of the black shale; fresh olivine and perovskite are typical constituents of olivine melilitite and could have been derived from this rock type rather than kimberlite. The opaque ore and the garnet from several pipes were further investigated. The opaque ore was found to show very variable magnetic properties. Twenty samples from sixteen pipes were analyzed. They fall in two groups, viz. (a) picroilmenite similar in composition to kimberlitic ilmenites according to DAWSON (1962) and (b) titaniferous magnetites (Fig. 7). From the fact that samples intermediate in composition between the two groups appear to be absent it may be concluded that ilmenite and magnetite are not closely associated. The pipes that contain kimberlitic ilmenites are Klein Katvlei KKl, Burtonsputs, Banke BK2 and BK3 and Koppieskraal KP8 and KP9. Indications are that the "low magnesia kimberlites" do not contain ilmenite and chrome diopside.

MgO

•

qf

••• • o

0

o

o

T i 02'~---''------->L._---''----",,------->'---_--->L._----'''---------' Fe 0 FIG, 7 ComposItIOnal variation of the opaque black ore from Bushmanland pipes (open circles) compared with picroilmenite from kimberlite accordmg to DAWSON (1962) (solid black dots),

The garnets are also found to fall in two groups (Fig. 8): (a) Eight out of twenty-three grains representing five different pipes are pyrope-rich and coincide with the field established by MATHIAS et al. (1970) for garnets from peridotitepyroxenite xenoliths in kimberlite. Despite considerable variations in colour (wine red, pink, purple, dark orange, light orange and brown) no garnets that could only have come from eclogite xenoliths have been identified. (b) The other garnets (often from a pipe which also contains pyrope-rich garnet) are almandine-rich antl were derived from the country rock. The pipes with kimberlitic affinities according to their garnet content are Vermeulens Rust VR 1, Klein Katvlei KKl, Gamoep GPl, Kamiebees, Papkuilsfontein C and Dikdoorn.

79

THE BUSHMANLAND KIMBERLITES AND RELATED ROCKS , 86

, 84

176

PYROPE 11 50

11 60

11 70

11 80

11 90

12 00

- - - a o l.8.lFIG 8 Graph of refractIve Index against unit cell edge length (ao) for garnets from borehole concentrates of Bushmanland pipes. Comparison with garnets from Namaqualand gneiss (starred) and from peridotite--pyroxemte xenoliths and eclogite xenolIths in kimberlite according to MATHIAS et al. (1970).

CONCLUSIONS 1. The presence of typical "indicator minerals" in at least some of the Bushmanland pipes would seem to confirm the existence of true kimberlite in this province. 2. Hydrothermally altered melilite-bearing rocks may also be present and would perhaps account for the low magnesia content of some pipes that have been confused with kimberlite. 3. The source of the alluvial diamonds along the ·lower and upper reaches of the Buffels River to the west and the Koa River to the east has not yet been found. In Tanzania over 200 kimberlite occurrences were discovered during a R4 million prospecting campaign, yet nearly all of them are barren and the Mwadui mine remains the only one of economic importance (EDWARDS and HOWKINS, 1966). Thus there is still a possibility that a diamondiferous pipe may be discovered in Bushmanland. The two provinces have many features in common. 4. The former existence of craters 250 m deep and their subsequent filling by lacustrine sediments represent aspects of the emplacement of kimberlite that are unknown elsewhere in South Africa. Does this imply a difference in erosion level, age or eruptive mechanism? 5. The age-old question whether kimberlite and olivine melilitite are genetically related may possibly be answered here. It has been claimed that both rock types are found together in the pipes on Tauseb and Klein Katvlei and that the kimberlite pipe on Kamiebees

80

A. K. CORNELISSEN AND W. J. VERWOERD

contains inclusions of olivine melilitite. Geological relationships are not sufficiently clear to confirm this. From the above it should be clear that a detailed study of the Bushmanland pipes promises to elucidate many intriguing problems of kimberlite geology.

ACKNOWLEDGEMENTS

The three mining companies that prospected for diamonds in Bushmanland during the years 1961-6 (De Beers, Rand Mines and the O'okiep Copper Company) generously provided access to plans, logs and reports which were formerly considered as confidential. Much of the field work on which this review is based was done by geologists H. Jenner Clark, H. Jansen, F. 1. G. Schreuder, K. Whitelock and P. Joubert. P. Gresse did the laboratory work on the garnets.

REFERENCES DAWSON. J. B. (1962) Bastuland Klmberlites. Bull. Geol. Soc. Amer. 73. 545-60 DINGLE. R. V. and GENTLE. R. I. (1972) Early Tertiary VolcanIc rocks on the Agulhas Bank. South African Continental Shelf. Geol. Mag. 109. 127-36. Du TOIT. A. L. (1939) The Geology of South Africa. 2nd ed. Ohver & Boyd. EDWARDS. C. B. and HOWKINS. J. B. (1966) Kimberhtes m Tanganyika with speCial reference to the Mwadui occurrence. Econ. Geol. 61. 537-54. HALL. A. L. (1938) Analyses of Rocks. Ores. Coal. Solis and Waters. Geol. Surv. S. Afr. MemOIr 32 HAUGHTON. S. H. (1931) On a collection offossil frogs from the clays at Banke. Trans Roy. Soc. S. Afr. 19. 233-49 JANSE. A. J. A. (1969) Gross Brukkaros. a probable carbonatite volcano in the Nama Plateau of South-West Africa Geol Soc. Amer. Bull. 80. 573-86. JOUBERT. P. (1971) The regional tectonism of the gneisses of part of Namaqualand. Precambrian Research Unit. Univ. Cape Town. Bull. 10 KIRCHHEIMER. F (1934) On pollen from the Upper Cretaceous dysodil of Banke. Namaqualand. Trans Roy. Soc. S Afr. 21.41-50. MACKINLAY. A. C M. (1955) Kimberlite intrusions cuttmg Karroo sediments m the Ruhuhu depressIOn of South-Western Tanganyika. Rec. Geol. Surv Tanganyika 5.63-80. MATHIAS. M .• SIEBERT. J. C. and RICKWOOD, P C. (1970) Some aspects of the mineralogy and petrology of ultramafic xenoliths in kimberlite. Contr. Mineral Petrol. 26. 75-123. RENNIE. J. V L. (1931) Note on fossil leaves from the Banke clays. Trans. Roy. Soc S. Afr. 19. 251-3. REUNING. E. (1931) A contribution to the geology and paleontology of the western edge of the Bushmanland Plateau. Trans. Roy. Soc. S. Afr 19.215-32. REUNING. E. (1934) The compositIOn of the deeper sediments of the pipe at Banke. Namaqualand and their relation to kimberhte. Trans Roy. Soc. S. Afr. 21. 33-39. ROGERS. A. W. (1911) Geological Survey of parts of Van Rhyn's Dorp and Namaqualand DiVISIOns Ann. Rep. Geol Comm. Cape of Good Hope 16. 7-84 TALJAARD. M S. (1937) South African melihte basalts and their relations. Trans. Geol. Soc S Afr. 37. 281-316. VISSER. J N J (1964) Analyses of Rocks. Mmerals and Ores. Geol. Surv. S Afr. Handbook 5