Trans. Ceol. Soc. S.Afr ..

77,329-338

A GEOLOGICAL INVESTIGATION OF THE BUSHVELD COMPLEX IN

SEKHUKHUNELAND AND PART OF THE STEELPOORT VALLEY*

by

T. G. MOLYNEUX

ABSTRACT
The area mapped consists of approximately I 250 sq. km covering parts of the Main and Upper Zones
and the epicrustal phase of the Bushveld Complex in the Eastern Transvaal.
The oldest rocks are quartzite, feldspathic quartzite, hornfels and dolomite of the Pretoria Series.
These rocks are generally sandwiched between the Layered Mafic Sequence and the epicrustal and/or
granitic rocks of the roof of the Complex. Apart from the sediments, the rocks in the roof are
granophyre, leptite and granite. Coarse-grained granite is the dominant acid rock in the central and
northern parts of the area, where the underlying Layered Sequence is thickest; with lateral thinning
southwards of this sequence there is a corresponding transition in the roof from granite, through
granophyre at Paardekop, to Rooiberg felsite at Bothasberg which is south of the area. The present
work has yielded further evidence in support of Willemse's placing (1964, p. 97) of the epicrustal rocks
as older than the Layered Sequence which in turn predates at least some of the Bushveld Granite. There
is field evidence of granitisation of sediments and it is likely that some of the granophyre was derived by
that process. However, most of the granophyre (and leptite) was probably originally Rooiberg felsite
which was altered to granophyre by heat and volatiles emanating from the crystallising Layered
Sequence. Finally, at least some of the granite is likely to have been produced through anatexis and
palingenesis of granophyre.
The Main Zone of the Layered Sequence is approximately 2 860 m and the Upper Zone some I 790 m
thick. In these zones layering is remarkably persistent and uniform. The base of the Main Zone is at the
foot-wall of the Merensky Reef and the division between the Main and Upper Zones is taken as the base
of the mottled magnetite anorthosite 30 m below the lowest magnetite seam. Differentiation is most
pronounced in the Upper Zone and least developed in the middle of the Main Zone. Spotted anorthosite
is restricted to the Main Zone but the mottled variety occurs throughout the sequence. Iron-rich olivine
and widespread cumulus titanomagnetite are very largely confined to the Upper Zone. Cumulus
titanomagnetite co-exists with olivine in the compositional range F060- 0, orthopyroxene En 63 - 29 and
plagioclase A n60- 43 .
Plagioclase varies in composition from An70 at the Merensky Reef to An 43 in the quartz diorite at the
top of the Layered Sequence. In this interval orthopyroxene changes in composition from En78 to En 29
and cumulus olivine, which occurs almost exclusively in the Upper Zone, from F060 to Foo. Biotite is
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

prominent throughout the Upper Zone whereas apatite is restricted to the diorite constituting the top of
the Layered Sequence. Hornblende, which occurs only near the roof, may be a primary mineral in the
mafic rocks, but quartz in bulk, and orthoclase appear to have been formed only as a result of
hybridisation between the roof-rocks and magma of the Layered Sequence.
A surge of parent-magma was probably introduced into the chamber prior to the formation of the
Merensky Reef and the magma then embarked on a trend of differentiation which continued with little
interruption for about 2 200 m up to the foot-wall of the Pyroxenite Marker. Reversals in the
compositions of orthopyroxene and plagioclase indicate that there was a major influx of magma before
the formation of the Pyroxenite Marker. Differentiation then probably proceeded with little further
interruption, the end-product being the diorite constituting the top of the Upper Zone. Above the
Pyroxenite Marker some 750 m of layered rocks were deposited before the magma returned to its earlier
state and this interval may thus indicate the vertical dimension of magma introduced. After the
introduction of this probable final surge of magma the internal depth of the chamber is likely to have
been of the order of 2 400 m which is the present vertical distance between the Pyroxenite Marker and
the roof.

CONTENTS

Page
I. INTRODUCTION 330
II. FIELD GEOLOGY 330
A. General 330
B. Roof-rocks . . . 330
I. Sediments of the Pretoria Series 330
2. Epicrustal rocks 331
3. The Bushveld Granite 331
C. The Layered Sequence · 331
I. Gt:neral · 331
2. Main Zone · 332
3. Upper Zone · 334
D. Geological Structure · 335
I. Folding · 335
2. Faulting · 335
III. MINERALOGY OF THE LAYERED SEQUENCE · 335
A. General · 335
B. Plagioclase ............... . · 335
C. Olivine ............... . · 336
D. Orthopyroxene ............. . · 336
E. Distribution of other minerals in the Main and Upper Zones · 337
F. Suggested mechanism for differentiation in the Main and Upper Zones · 337
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . .338
* Extracted from a thesis submitted for the Degree of Doctor of Science at
the University of Pretoria, November 1970. Published in terms of the
Government Printer's copyright Concession No. 4518 of 5th July, 1971.
 330 TRANSACTIONS OF THE GEOLOGICAL SOCIETY OF SOUTH AFRICA

I. INTRODUCTION differentiation trends in the Layered Sequence especially

The area examined consists of part of the strike lengths those leading up to the formation of the titanomagnetite in
of the Main and Upper Zones and the epicrustal phase of the Upper Zone.
the Bushveld Complex in the Eastern Transvaal. It Without generous help from the following, this work
includes portions of the Steelpoort Valley, could not have been completed:
Sekhukhuneland, the Leolo Mountain-land and the The Anglo American Corporation of South Africa,
Pokwane Plateau (Fig. 1). Limited and Highveld Steel and Vanadium Corporation
Limited offered help and encouragement and also made
available official information.
The Bantu Affairs Department allowed access to
Sekhukhuneland.
The Geological Survey of South Africa provided an
honorarium and equipment to help with field-work.
The Oppenheimer Memorial Trust awarded an
honorarium.
Mr. P. Minckner did the draughting of Fig. 12.
Johannesburg Consolidated Investment Company,
Limited permitted inspection of bore-hole cores.
Lydenburg The late Professor J. Willemse offered much
o
encouragement in the initial stages of the study, which was
finished under Professor D. J. L. Visser. Drs. C. P.
Snyman, E. B. F6rtsch and G. von Gruenewaldt and other

1
i1r
members of staff of the University of Pretoria were at all
times interested and helpful. To them and numerous other
people who have helped in various ways, I offer my sincere
oBelfast I thanks.
II. FIELD GEOLOGY
SOkm A. General
The metamorphosed sedimentary roof-rocks of the
Pretoria Series are the oldest rocks in the area and in order
of decreasing abundance consist of quartzite, hornfels,
and dolomite. Some of these rock-types, particularly
hornfels, occur as xenoliths in the Upper Zone but most of
the sediments lie irregularly orientated between the diorite
at the top of the Layered Sequence, and the leptite,
granophyre and granite of the roof.
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

LEGEND Most of the work was devoted to the layered rocks of

the Main and Upper Zones, collectively referred to as the
~ Waterberg Sy. 1:::·:·:·.1 Lay. Maf ic R.ock Layered Sequence. The Merensky Reef is at the base of
~ Roof Rocks 0 Transvaal Syst. the Main Zone and some 1 200 m of rock above it is

E;3 Granite t:I Area mapped

differentiated into anorthosite, porphyritic gabbro and a
little pyroxenite (Fig. 14). Then follows approximately
Figurel
I 000 m of very poorly differentiated gabbroic rocks
Location Map of Sekhukhuneland Area. forming the bulk of the Leolo Mountains. At the top of the
Main Zone some 660 m of rocks are differentiated into
pyroxenite, anorthosite and gabbroic rocks.
Working for the Anglo American Corporation of South The division between the Main and Upper Zones is
Africa, Limited, T. G. Molyneux, accompanied by R. E. taken as the base of the mottled magnetite anorthosite
Jacob and K. B. McQuillin, spent most of 1961 and 1962 30 m below the lowest magnetite seam. Iron-rich olivine
mapping and prospecting the magnetite seams and plugs in and widespread cumulus titanomagnetite are very largely
the Upper Zone, from Stoffberg northwards to the Olifants confined to the Upper Zone. The uppermost 650 m of the
River, a distance of more than 100 km. Layered Sequence is largely dioritic.
The interest aroused by this work led to the study of the Post-Bushveld rocks include diabase, dolerite and the
magnetite seams at Magnet Heights (Molyneux, 1964). Spitzkop suite of alkaline rocks, and were not examined in
The present investigation was commenced in 1965, when any detail.
an area of some I 250 sq. km was mapped (Fig. 12, at end
B. Roof-rocks
of text).
1. Sediments of the Pretoria Series
Samples were collected along surveyed lines at vertical
Nearly all the sediments, which approach I 000 m in
intervals of on average 35 m in the Layered Sequence
thickness, are sandwiched between granodiorite at the top
from the M erensky Reef to the roof. A plane table and
telescopic alidade were used, and to obtain a complete of the Layered Sequence, and the over-lying granophyre,
leptite and granite of the roof. In the vicinity of Magnet
sequence it was necessary to traverse several lines (BCDE,
Heights (H5)* the general sequence upwards in the roof is
PQ, RS, TU, and XV, Fig. 12) as faulting and folding have
hornfels, feldspathic quartzite and quartzite. No
dislocated the sequence. Profiles were constructed (Fig.
petrographic study has been made of any of the sediments
13) and a stratigraphic column was compiled (Fig. 14, at
or other roof-rocks.
end of text).
In 1969, the oxide ores and some of the silicate minerals (aj Quartzite
were examined in the laboratory. Work on the silicates of The quartzite, which mostly contains some pink
the layered mafic rocks was confined to modal feldspar, forms conspicuous hills strewn with white talus.
determinations, microscopy and some refractive index and There is probably 600 m or more of quartzite constituting
X-ray studies. The results of investigations of the ore- Signal Hill (H4) and somewhat less in the mountains to the
minerals have already been published (Molyneux, 1970a). * Letter and numeral in parenthesis are for location purposes and refer to
In the present study an attempt is made to determine the Figure 12.
 GEOLOGICAL INVESTIGATION OF THE BUSHVELD COMPLEX 331

north. South of Signal Hill there are only sporadic small Assuming that the granophyre is stratified and dips
lenses of quartzite. westwards conformably with the Layered Sequence, it
Impure quartzite has been feldspathised to a variable would be approximately 650 m thick at Paardekop. Units
extent and at places in the field the transition to what of granophyre dipping gently westwards and 60-150 m
appears to be granophyre is striking. This granitisation was thick have been mapped by photogeology on Dwars-in-
probably brought about by alkali material entering the die- Weg 137-JS. The nature of these units is not
roof from the differentiating gabbroic magma (Willemse, understood and no evidence was seen of chilling or
1964, p. 115). Groeneveld (1968, p. 56) also concluded anything indicating separate flows or intrusions.
that on Houtenbek 97-JT the gabbro mass was mainly The granophyre appears to be older than the Bushveld
responsible for the feldspathisation of impure quartzite. Granite. In the gorge of the Tsweneng River descending
from Spitskop (L3), dykes of granite cut the granophyre,
(b) Hornfels
and various stages in the alteration of granophyre to
There are two types distinguishable in the field: a fine- granite were seen along the escarpment west of the
grained banded rock consisting of small-scale alternations Steelpoort River (K4). Locally, contacts between granite
of hornfels and quartzite, and a coarse-grained massive and granophyre are very irregular but regionally the
variety, at places exhibiting current-bedding. Where granophyre seems to occur as remnant sheets in the
hornfels is abundant it tends to be the second variety, as is granite. At places the granite and the granophyre cannot
the case in the north of the area (A3). be mapped separately because they are intimately mixed.
Hornfels occurs as xenoliths in the Layered Sequence It is probable that a considerable amount of granophyre
though only upwards from the Main Magnetite Seam. On was altered to granite and the diminution in quantity of
Government Ground 846-KS (H5) a lens of hornfels some granophyre northwards may be due to its conversion to
10 m long is overlain by Upper Seam No.1; a xenolith granite. It is noteworthy that granite is the dominant acid
approximately 150 m in length lies some 50 m higher in the roof-rock where the underlying Layered Sequence is
sequence. In the south of the area (M4) a xenolith 10 km thickest (i.e. between Spitskop and the Olifants River)
long and mostly of hornfels lies approximately 500 m whereas southwards below Bothasberg (Fig. I) felsite
above the Main Seam. prevails above the much thinner mafic sequence.
(c) Dolomite Granophyre is abundant between these two extremes.
This rock-type occurs only as rare xenoliths in the The relationship of the granophyre to the Layered
Upper Zone. The well-known calcareous xenoliths west of Sequence is not clear as there is no mutual contact. On
Maila (E6) and at Magnet Heights are altered to a Dwars-in-die-Weg 137-1S, however, a gabbroic rock has
predominantly wollastonite-vesuvianite rock (Willemse intruded the granophyre. If this rock originated from the
and Bensch, 1964, p. 1). Bushveld magma chamber as it is thought that it has, it
Below Signal Hill there is a dolomite xenolith partly would prove the granophyre to be older than at least part
replaced by magnetite. The largest dolomite body, of the gabbro.
approximately 150 m across, recorded in the area is above
3. The Bushveld Granite
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

Upper Seam No.7 on Buffelskloof 141-JS (M4).

Most of the granite lies above the roof-sediments and is
2. Epicrustal rocks younger than these rocks and the granophyre. At least
Willemse (1964, p. 100) grouped leptite, felsite and some of it is also younger than the Layered Sequence into
granophyre in this category. which it is intrusive on Steelpoortpark 366-KT and
Magnet Heights 846-KS. The granite of the roof tends to
(a) Leptite be medium-grained at the base and upwards, over ap-
The leptite is fine-grained, brown or dark red in colour, proximately 300 m, becomes steadily coarser in texture.
and appears to be metamorphosed Rooiberg felsite which The Magnet Heights Granite intruded the gabbro as a
occurs in unaltered form only to the south of the area. large dyke some 9 km long and with a maximum width of
West of Sekhukhune (G6) and north of Jane Furse 1,5 km. A leptite xenolith approximately 3 km long and
Hospital (G4), dark green amphibolitic patches with the 200 m wide is well exposed in the Magnet Heights Granite
attitude of amygdales occur in dark red leptite and and if this leptite came from the roof, it would now be
indicate that the rock was originally lava. In this context some I 500 m below its original position.
Henriques (1966, p. 56) who investigated the Scandinavian Granitised sediments have been identified in the area
leptites, concluded that red leptite with phenocrysts was and it is possible that considerable amounts of sediments
derived from lava probably of rhyolitic composition. of suitable composition were converted to granite and that
The leptite contains numerous small patches of red only resistant varieties such as quartzite, hornfels and
granite and pegmatite which are thought to represent dolomite have survived. There are also indications in the
metamorphic segregations or, in the case of larger plugs field that a substantial amount of granite was produced
and dykes, mobilised leptite. through anatexis and palingenesis of granophyre.
Von Gruenewaldt (1968, p. 156) working in the vicinity Willemse (1969a, p. 12) suggested that a considerable
of the Blood River Valley which is south-west of the amount of alkalis may have diffused into the roof during
writer's area, concluded that leptite was produced through the crystallisation of the mafic rocks. These volatiles in
metamorphism of dark felsite. Willemse (1964, p. 100) conjunction with heat from the crystallising magma could
pointed out that though leptite is slightly coarser in texture have been responsible for extensive granitisation of the
than felsite, it is difficult to distinguish between the two roof-rocks.
rocks, and he also noted a gradation from leptite to both
granophyre and felsite. C. The Layered Sequence
( b) Granophyre 1. General
The granophyre mostly lies above the roof-sediments The thickness of layered igneous rocks from the
and below or adjacent to the Bushveld Granite. Merensky Reef to the roof is approximately 4 650 m. Lit-
Granophyre is most abundant in the south of the area and tle conspicuous thinning of the sequence or of individual
the largest mass is that constituting Paardekop (N I). bands was found along the 80 km of strike that was
Northwards, the amount of granophyre decreases, with a mapped. The rock-types are gabbro, hypersthene gabbro,
corresponding increase in the quantity of granite which is hyperite, norite, pyroxenite, anorthosite, magnetite seams
predominant in the roof in the central and northern parts and plugs, troctolite, diorite and granodiorite (Fig. 14 and
of the area. Table I).
 332 TRANSACTIONS OF THE GEOLOGICAL SOCIETY OF SOUTH AFRICA

Differentiation is most pronounced in the Upper Zone which is currently being exploited. The occurrence of
and least in the middle of the Main Zone. Spotted plugs in the vicinity of many small pegmatoid stringers
anorthosite is confined to the Main Zone but the mottled probably indicates a common origin. On Buffelshoek 368-
variety occurs throughout the sequence. Troctolite and KT which is some 10 km to the south, Hammerbeck
nearly all the cumulus titanomagnetite define and are (1970, p. 307) also mapped irregular dykes and seams of
restricted to the Upper Zone. magnetite at this stratigraphic level.
The sequence has been subdivided (Fig. 14) and is sum- The upper limit of Subzone A is taken as the top of the
marised as follows: "Upper Mottled Anorthosite" which is a composite unit
ROOF

Subzone D Group 0 of magnetite seams,

granodiorite, diorite, mottled
anorthosite and troctolite 650 m

Subzone C Group C of magnetite seams,

UPPER gabbroic suite of rocks, mottled
ZONE anorthosite and troctolite 500 m

Subzone B Group B of magnetite seams,

gabbroic suite of rocks, mottled
anorthosite, troctolite and pyroxenite 450 m

Subzone A Group A of magnetite seams,

gabbroic suite of rocks and
mottled anorthosite 190m

Total 1790 m

Subzone C Gabbroic suite of rocks,

pyroxenite, spotted and
mottled anorthosite 660 m

MAIN Subzone B Slightly differentiated

ZONE gabbroic suite of rocks 1000m

Subzone A Gabbroic rocks, pyroxenite, spotted and

mottled anorthosite. 1200 m

Total 2860 m
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

2. Main Zone approximately 200 m thick consisting of spotted and

Wager and Brown (1968, p. 383) aptly described the mottled anorthosite and gabbroic rocks. This unit is easily
Main Zone as being composed of average gabbroic rocks mapped and can clearly be seen when looking from the
in that they generally contain three essential cumulus Maandagshoek Mission Station to more or less halfway up
minerals, i.e. two pyroxenes and plagioclase. No troc- the Leolo Mountains (C7).
tolites were detected in the present study though very (b) Subzone B
local, possibly pegmatoidal, patches of olivine magnetite This subzone consists of approximately 1 000 m of fairly
gabbro occur towards the top of Subzone B (Sample 35, uniform gabbroic rocks and is the least differentiated por-
G7) and warrant further investigation. Pyroxenite bands tion of the entire Layered Sequence. Locally, thin or poor-
occur at and near the base of Subzone A and C of the ly developed anorthosite bands were seen; for example, on
Main Zone. Fernkloof 639-KS (B6), a mottled anorthosite some two
metres thick is visible in a river cutting a few hundred
(a) Subzone A
metres north of the Djate road.
Rocks of this Subzone outcrop prominently along the
As no monomineralic or other markers could be es-
eastern slope of the Leolo Mountains which are the most
tablished in Subzone B, recourse was made during map-
beautiful and spectacular topographic feature in the
ping to differential erosion effects. Six sheets from 100-
Bushveld Complex. The Merensky Reef and the Bastard
200 m thick and varying little petrologically were dis-
Reef are intermittently visible at the foot of the mountains
tinguished (Fig. 12-14) using photogeology and by view-
and they are exceptionally well exposed in a stream some
3 km south of the Maandagshoek Mission Station (D8). ing the hills from different angles in the field. There are
particular areas where each is well displayed and
Approximately 300 m above the Merensky Reef is a band
elsewhere boundaries were inferred, mostly using the
containing needles of orthopyroxene (B8). This was
relief shown on the 1 : 50000 scale topocadastral maps.
pointed out by Mr. L. Davis, who worked in the area for
Unfortunately, the change from primary orthopyroxene to
the Johannesburg Consolidated Investment Company,
inverted pigeonite occurring in Subzone B was not
Limited. This unit is succeeded by some 200 m of gabbroic
detected in the field.
rocks containing conspicuous porphyritic crystals of
orthopyroxene (Willemse, 1969a, Fig. 9, p. 14). In the mid- Djate Sheet
dle of the subzone is the "Main Mottled Anorthosite" This is the first prominent scarp above the Upper
consisting of some 150 m of alternating bands of spotted Mottled Anorthosite of Subzone A. It constitutes most of
and mottled anorthosite and gabbroic rocks. the hill on Djate 249-KT (B6) and southwards, owing to
Above this marker-band is 300 m of slightly banded gab- poorer outcrops was detected only sporadically. At its
bro accommodating some magnetite plugs and numerous base it is banded and slightly differentiated.
magnetite-bearing pegmatoid stringers. These plugs were Soupiana Sheet
mapped on Nooitverwacht 324-KT (F8), De This is spectacularly exposed below Trigonometrical
Goedverwachting 322-KT (G8), and Kennedy's Vale 361- Beacon No. 131 on Houtbosch 323-KT (E8, Fig. 2). It con-
KT (H8). The Kennedy's Vale plug contains some two per sists of hard, massive gabbroic rocks and forms a precipice
cent V20 5 and hence constitutes a valuable ore-body up to 150 m high. It is identified as the second scarp above
 GEOLOGICAL INVESTIGATION OF THE BUSHVELO COMPLEX 333

Figure 4
Peneplain on Het Fort 329-KT (G7). Viewed from the north.

Het Fort 329-KT (G7, Fig. 4) and is one of the prominent

ridges where the Leolo Mountains drop towards the
Figure 2 Steelpoort River on Boschkloof 351-KT (H8).
Cliff of Soupiana Sheet in Main Zone below beacon 131, Hout-
bosch 323-KT (E8).

the Upper Mottled Anorthosite and forms the highest spur

on Nooitverwacht 324-KT (G8).
Garatouw Sheet
It is approximately 100 m thick and supports the
Garatouw Beacon No. 15 (C7). It is separated from the un-
derlying Soupiana Sheet by a poorly outcropping mottled'
anorthosite band some 2 m thick. The Garatouw Sheet
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

outcrops prominently with a conspicuous westerly joint

dip of approximately 25° and in the north contains striking
poikilitic crystals of inverted pigeonite up to 30 cm in
diameter. Its profile is clearly visible in the north of
Fernkloof 639-KS (A6) and southwards it forms a large
outlier on Hoogste Punt 290-KT (E7). On Boschkloof 351-
KT (H8), however, it is barely discernible.

FigJlre 5
Pillars of gabbro of Het Fort Sheet on Honoko Mountain west of
Genokakop 285-KT (D7).

Het Fort Sheet (Fig. 5)

This sheet is conspicuous on Boschkloof 331-KT, on the
eastern slopes ofThama Koosh Mountain (G7) and on the
sides of Honoko Mountain (D6). Only where there is
severe erosion is the boundary visible between the Outlier
and the Het Fort Sheets.

Figure 3
Banding in Subzone B of Main Zone on northern slope of
Honoko Mountain, Genokakop 285-KT (07).

Outlier Sheet ( Fig. 3)

This sheet is so named because outlying parts of it are
prominent on Grootvygenboom 284-KT (C7), on Hoogste
Punt 290-KT (E7), on Houtbosch 323-KT(E7) and on
Soupiana 325-KT (F7). The tendency to form outliers
seems to be due to pronounced vertical 'jointing which Figure 6
locally on scarp slopes has permitted the down-dip part of View looking westwards from Djate 249-KT towards part of the
the sheet to be removed. The sheet underlies the plain on Main Zone gabbro (A6).
 Fig~ TRANS, GEOL. SOc. S.AFR, , VOL. LXXVII IRREGULAR UNDULATING ROOF OF LAYERED SEQUENCE
54

-- -.-
4500 _---283
~281,
'282
_____ '
650 ~ ,

/' I'
SEAM 21 (9m)

I ,'279
2~ , ~ Anorthosit~and Ovicular Diorite~).
51 \ ,50

Anortho.~~nd
4000 ~274
273 ' ,
272, ,

seam13lmJ
48~'268
269( ~,
,L
267 ........~2 l2.t~~48
191t~'263 "
3500 260 ~!~ ~sC!!..Iroctolite Markeri!.Q..ml

Magnet
Heights ~255
1 '

Trave.rse ~854 ,
• _Zone
Upper Pvroxenite~'2 ni'
11~1!~45 45--~==========~~
_ _ ...:......L. _ _ _'I
14
43~--'44 Jt
_ _ _ _ _ -..........--..41 Oliv,

~284
3000
289'1-......,.290
Appearance of ~'291
------ 310
Pigeonite ~' 06
L'305
'~ Steelpoortdrift I
•I Percent Fosterite in Olivine
I I I
300,~'302
' " Traverse 80
I
'60
,I 1..,0 2,0 I
';
2500 299-"
298~
•
{38 ~236
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

CD
a: ~ ~236
.:
II)
296' , 235
295··~
c
CD ~94.........232 229 .
~ _N_o__ _________1
P~ig~e_O_nl_'te
:i 33;·~·23f"'··-'229
~ ------ ~228
o
~
32?'~
to 327~
! 2000 325L'223
221, ~. ,
i
2J
Appearance 0~.....-----~23 - Plagioclase Curve
Plagioclase

.5 Pigeonite 320]" 1'216 - . -Olivine Curve Orthoclase

Mooimeisjes Ojate Traverse
......... Orthopyroxene Curve Quartz
fontein 363· KT '213
~09
34'
345 Sample Number Hornblende
317' ~.206 x Intercumulus Material Apat,lte
MOOimeiSjeSfo~tein I Biotite
Traversel /.202
1500 Ti-magn. +11m
~~ 19 Appearance of Pigeonite
198·~14 90;;-Ojate 249 KS - Olivine
Orthopyroxene
/88
Clinopyroxene

y6 . . .
Upper Mottled Anorthosite Mark~r (s~e ~ig ..14)
1000 /'179
-- - -----
175

172 J 7~
'l;~

\169
Percent Anorthite in Plagioclase

Percent Enstatite in Orthopyroxene

40
I
60
SO

SO
I
60

40
I
70

30
I ------J-
Lower Mottled Anorthosite Marker(see Fig. 14 ) .

.~66
500
~te Traverse

,163

160" -----------

,7,58
,
O~____________________________________________~r-

;.~____ Values of Schwellnus(1956 p17) .,/ ,

____~M~ER~E~N~S~K~V~R~.F
o 20 ~ ~
Modal distribution of minerals
80

VARIATION IN COMPOSITION OF PLAGIOCLASE,ORTHOPYROXENE & OLIVINE. MODAL DISTRIBUTION OF MINERALS IN MAIN AND UPPER ZONES
IRANS. GEOL. soc. S.AFR., VOL. LXXVII MOLYNEUX , TABLE 1
Ck:
Ck:
w ww
lL MODES - PERCENTAGES BY VOLUME W
wW
lL MODES - PERCENTAGES BY VOLUME
m >w w CD
~ >w W x x ;i W
~
~ OCk: W 0 0
0
~
z· 00:: (!)
~ 0
:J
:J m> ~ Ck: (!) z
Z m> Ck: Ck: « Z
z
w
«~ --.J
0::: > « w
--.J COMPOSITIONAL W «?Jij ....J
u >
a.. ~ W ~ W
W
...J W N COMPOSITIONAL
t-(j') U
0 ~
a.. W ~ W CD --.J 0 0 m t-
t-
...J
a.. IZ 0 Z 0 t- a.. t-z 0 0 z
z
~
-l- - Ck:
~
(!)w
_Ck: ~
0 I > Z ~
Z
DETERM INATIONS ~
IUJ
~a:: ~ -
Z :I: :> «
Z
Ck:
t-
« « DETERMINATIONS
« wW
Z
:J
t- « 0
Ck:
0 ROCK TYPE « WUJ ~
t-
Ck: ....J
-t-
0 o a.. :::> ROCK TYPE
(j') :::t:~
...J
a.. u
Ck:
0
....J
0
t::
t- m I OF MINERALS (j') :::t:~
....J
a.. u 0 0 t- CD I « 0 OF MINERALS
291 291Sm 65,1 20,2 1,2 - 13,4 0,1 - AnS6 En63 Magnet ite gabbro 283 4S70m 54,6 13,6 0,3 8,7 3,2 0,7 11,9 0,5 5,5 An41 FoS Olivine diorite
312 2905 90,0 - tr - 10,0 tr - AnS9 Anorthosite 282 4510 62,2 10,3 6,0 7,9 4,2 2,5 3,9 3,0 tr An46 En29 F018 01 iv ine d ior ite
308 2860 93,0 - - - 7,0 tr tr AnS3 Anor thosite 281 4420 56,7 4,9 15,8 6,9 8,7 ',8 0,5 4,7 - En33 F032 Magnetite diorite
306 2830 En67 Hyperite 280 4330 42,3 6,3 15,5 13,1 9,2 7,3 0,7 5,6 - An47 F032 0\ ivine diorite
305 2 740 62,9 16,9 20,6 - tr 0,1 0,1 An64 En67 Hyperite 279 4220 70,5 10,1 5,0 4,3 6,4 3,3 - 0,4 tr An48 En46 F03S Diorite

303 2 700 60,0 22,0 17,3 - 0,4 0,1 0,2 En74 Hypersthene gabbro 278 4180 67,6 3,0 0,5 16,0 9,0 1,0 - 2,9 - EnSl Fo42 Olivine diorite
302 2640 79,4 11,2 9,1 - tr 0,2 0,1 An64 En69 Hypersthene gabbro Sl 4 130 65,9 22,7 - - 10,3 1,1 - - - An46 Magnetite diorite
299 2500 59,1 24,1 16,3 - 0,2 0,3 - An67 En7S Hyper st hene gabbro 276 4 115 95,0 tr tr 1,0 An47 Anort hosite
296 234O 61,6 17,7 20,1 - tr tr 0,6 An68 En74 Hyperite 274 4020 66,9 4,7 - 17,6 6,5 0,2 - 4,1 - FO!.3 Troctolite
294 2270 9,3 12,9 77,7 - tr 0,1 - An66 En73 Pyroxenite 272 3870 60,3 26,6 6,5 - 5,5 1,1 - - - Anso EnSl Hypersthene gabbro
229 2200 10,0 6,2 83,8 - - - tr An64 En67 Pyroxenit~ Marker 269 3740 55,4 28,3 2,6 1,2 11,9 0,6 - - - An51 EnS5 F043 Magnetite gabbro
226 2130 69,1 20,6 10,3 - tr - - An62 H y perst hene gabbro 268 3690 52,6 27,4 9,0 - 9,3 1,6 - - - AnS3 EnS3 Hypersthene gabbro
224 2050 60,7 29,8 9,5 - tr - - An62 Hypersthene gabbro 48 3630 66,6 1,3 0,7 22,8 5,2 3,4 - - - AnsI. En62 FOS3 Troctol i te
221 1970 67,3 19,9 12,8 - tr - - An62 En66 Hypersthene gabbro 466 3 613 69,S - 2,5 - 23,7 4,3 tr - - Magnetite anorthosite
219 1920 71,4 14,1 14,5 - tr - - Hyperite 22 3610 75,1 2,9 0,5 14,4 5,6 1,5 - tr - F057 Upper Magnetite Sm.l1
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

216 1 820 56,3 29,3 14,4 - tr - - En64 Hypersthene gabbro 47 3580 95,0 An58 Anorthosite

213 1770 72,2 4,4 23,4 - - - - En65 Nori te 46 3500 57,8 0,4 1,1 31,1 8,6 1,0 - - - An55 F054 Troctolite, Sisal Marker
206 1615 54,7 15,0 30,2 - tr 0,1 - An62 En64 Hyperite 255 3365 61,4 11,' 21,6 - 5,1 0,5 - - - En61 Hyperite

202 1520 65,2 24,3 10,4 - tr 0,1 tr An62 En63 Hyperite 254 3310 57,0 9,0 17,5 - 15,2 1,3 - - - AnS6 En 62 Magnet i te hyper i te
189 1460 57,3 19,6 23,1 - - - - En68 Hyperite 18 3265 32,1 1,5 6,4 3,3 55,7 1,0 - - - AnS9 En62 Upper Magnetite Seam?
188 1330 62,3 35,S 2,2 - tr tr - En70 Gabbro 253 3240 70,4 11,0 10,2 - 7,5 0,9 - - tr En68 Hypersthene gabbro
186 1230 69,2 14,2 16,6 - tr tr - An64 En71 Hyperite 45 3205 22,9 52,6 21,7 - 2,3 0,5 - - - An57 En60 Pyroxenite
182 1060 95,0 5,0 tr - tr tr - An66 Anorthosite 242 3205 2,2 54,2 40,1 - 1,5 1,5 0,5 - - AnS6 En63 Pyroxenite

179 970 87,9 6,6 5,5 - tr tr - An68 En70 Anorthosite 44 3180 69,2 16,7 5,8 - 6,8 1,5 - - - AnS6 En62 Hypersthene gabbro
177 930 53,9 21,0 24,8 - tr 0,3 - An65 En72 Hyperite 43 3140 63,7 1, 1 35,2 - - - - - - En68 Nori te

172 660 95,0 2,0 3,0 - - tr - An67 En73 Anorthosite 41 3082 75,4 0,9 2,3 - 20,0 1,4 - - - En61 Magnetite anorthosite

166 50S 60,2 21,1 18,4 - tr 0,3 - An67 En73 Hyperst hene gabbro 8 3080 39,5 1,9 - - 56,4 2,2 - - - An60 Top of Main Magn.Seam
163 320 44,8 50,1 4,6 - 0,5 tr - An56 En68 Gabbro 6 3079 51,8 1,9 - 0,2 44,8 1,3 - - - An60 Feldsp.parting in M.M.S
161 305 An66 En76 Hyperite 3 3 078 79,0 some - some some some some - tr AnS9 F058 Anorthosite, fwall, M.S,
160 250 65,8 14,1 20,1 - tr tr - An67 En77 Hyperi tr 37 3070 56,4 31,4 0,5 - 11,~ 0,2 - - - An58 Magnetite gabbro
159 170 72,5 13,0 14,5 - tr - - An65 Hyperite 40 3050 74,7 0,6 1,8 22,1 0,2 0,6 - - tr Fo60 Troctol i te

157 110 68,0 2,5 29,S - tr - - An68 En78 Norite 288 3 010 67,1 8,7 0,2 13,9 9,9 0,2 - - -- An58 Olivine gabbro
154 25m 99,9 0,1 - - tr - - An71 Anort hosite 2 2 960 34,0 - 3,3 62,7 - - F011 Lowe r Mag n Seam 3

TABLE I Modal Compositions and Mineralogy of some rocks of the Main and Upper Zones
 334 TRANSACTIONS OF THE GEOLOGICAL SOCIETY OF SOUTH AFRICA

magnetite plugs occur near the base of the Upper Zone

and the prevalence of plugs in the vicinity of the lowest
seams probably reflects a build-up of the ferric iron con-
tent of the magma at that level. These plugs tend to be
coarsegrained which is an indication of their pegmatoid af-
finities.
Excellent exposures of the Upper Zone are found on
Government Ground (Magnet Heights) 846-Ks, Ironstone
847-KS (IS) and Steelpoortdrift 36S-KT. North of Magnet
Heights exposures are poor and in the Steelpoort Valley
outcrops are moderately good .
The Upper Zone has been divided somewhat arbitrarily
into four Subzones on the basis of the distribution and
composition of magnetite seams, anorthosite and troc-
tolite (Fig. 14). Subzones B, C and D start with a troctolite
band followed by anorthosite, magnetite seams and finally,
relatively uniform magnetite gabbro. Subzone A, however,
begins with magnetite anorthosite.
Figure 7 (a) Subzone A
Banded gabbro below Pyroxenite Marker in north of area, on This is best exposed on Steelpoortdrift 36S-KT (16) and
Thornhill 544-KS (B5).
the lowest unit is the previously mentioned mottled
magnetite anorthosite band about 10m thick. This
Thama Koosh Sheet (Fig. 6) marker-band is visible a few hundred metres upstream
This is the uppermost sheet in Subzone B and it from the drift three kilometres north-west of the Steel-
weathers to produce a jagged profile as seen beneath the poort bridge. The top of the subzone is taken as the base
Thama Koosh Trigonometrical Beacon No. 62 (F7). Near of the troctolite band some 30 m below the Main Mag-
the top is the Banded Gabbro (BS, Fig. 7) which is the first netite Seam.
useful marker-band above the Upper Mottled
(b) Subzone B
Anorthosite. The sheet forms a large outlier on Thama
The troctolite at the base is about five metres thick and
Koosh Mountain from which the Banded Gabbro has
was mapped on Steelpoortdrift 36S-KT, De Hoop 886-KS
already been removed. A locally occurring band some
(KS) and Uitvlugt 887-KS (LS). This troctolite band marks
three metres thick and containing cumulus titanomag-
the appearance in bulk of olivine in the sequence and it is
netite and some olivine (Sample 3S) was seen near the base
stratigraphically the lowest of several troctolite bands
of the Thama Koosh Sheet on Government Ground 846-
which are characteristic of the Upper Zone. Above the
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

KS (G7). A cluster of magnetite plugs occurs in the sheet

troctolite there is approximately 10m of magnetite gabbro
west of Zwemkloof 283-KT (C6).
and shortly above this is the Main Magnetite Seam which
(c) Subzone C is the most conspicuous marker in the Complex.
This includes approximately 660 m of differentiated The Main Seam is easily traced northwards along the
rocks consisting of spotted and mottled anorthosite, Steelpoort Valley until it is truncated by the Steelpoort
equigranular and porphyritic gabbroic rocks and pyrox- Fault on De Hoop 886-KS (KS). North of this fault the
enite. At the base of the subzone is the distinctive Pyrox- Main Seam can be followed as far as the Magnet Heights
enite Marker, two metres thick, first recorded by Lom- Granite and then to near the Magnet Heights store, where
baard (1934, p. 17) on Chieftain's Plain 46-JT, which is to it disappears into the core of a faulted anticline (Fig. 13).
the south of the writer's area. There is a second pyroxenite Northwards it outcrops only in small fault-bounded blocks
one metre thick some 60 m above the marker band. Owing at Maila (E6, Profile FGH, Fig. 13) and at Sefelere (CS),
to dislocation of the Main Zone by faulting it was before reappearing on Malek's Kraal S09-KS, which is
necessary, in order to obtain a stratigraphically complete north of the area (Geol. Survey Sheet 2429B).
series of samples, to correlate the Pyroxenite Marker in A true norite band (Sample 43) was recorded 6S m
the north on Thornhill S44-KS (B6) with its occurrence above the Main Seam on Government Ground 846-KS.
south of the Magnet Heights Granite (17). What is particularly striking about this rock is the complete
Rocks of Subzone C outcrop on the western slopes of absence of ore-minerals, a situation virtually unique in the
the Leolo Mountains south of the Magnet Heights Upper Zone.
Granite, and also occur south of the Steelpoort Fault on Approximately 120 m above the Main Seam is the only
De Hoop 886-KT (KS). However, the Sekhukhune Fault pyroxenite (Samples 242, L4 and 4S, HS) recorded in the
has largely eliminated these rocks at surface in central Upper Zone. It is followed by Seams No.6 and 7 which are
Sekhukhuneland. persistent and useful in mapping. Above these are some
Within the subzone are several spotted and mottled 200 m of fairly uniform magnetite gabbro incorporating a
anorthosites up to 10 m thick and also a number of band of mottled anorthosite seen only on Government
transgressive magnetite plugs. The top of this subzone Ground 846-KS. The top of the subzone is taken as the
(also of the Main Zone) is taken as the base of a con- base of the prominent troctolite band (Sisal Marker)
spicuous mottled magnetite anorthosite some 30 m below which outcrops at the sisal hedge by the road about I,S km
the lowest magnetite seam (16). In discussion with the late north-west of Magnet Heights. It is possible that the
Professor Willemse, the Main Magnetite Seam was thickness calculated for this subzone is incorrect as the
designated as the top of the Main Zone (Willemse, 1969a, only exposed profile is at Magnet Heights where there
p. 8) but recent petrographic work makes it more sensible may be a strike-fault along a 10 m wide dolerite dyke (Fig.
to end the Main Zone with the abrupt appearance of inter- 13).
cumulus and cumulus titanomagnetite.
(c) Subzone C
3. Upper Zone The Sisal Troctolite Marker which is resistant and ap-
Magnetite seams are restricted to the Upper Zone and proximately five metres thick outcrops very well on
their field characteristics have already been described Government Ground 846-KS but in the Steelpoort Valley,
(Molyneux, 1964 and Willemse, 1969b). Most of the in common with other troctolite bands, was rarely iden-
 GEOLOGICAL INVESTIGATION OF THE BUSHVELD COMPLEX 335

tified owing to poorer exposures. There is another promi- Heights Granite. The throw of the fault reaches a max-
nent troctolite band some 30 m above Seam No. II. The imum of some 2 100 m at Sekhukhune (G6) where it is
uppermost 100 m of Subzone C overlying Seam No. 14 joined by a minor branch which follows a straight deep
consists of magnetite gabbro. valley southward across Mooimeisjesfontein 363-KT (17)
to the Steelpoort River. Northward from Sekhukhune the
(d) Subzone D fault has several branches but the main dislocation follows
The base consists of troctolite which contains the first
the Sehilwane River as far as Malagale (05). The blocks of
cumulus apatite in the sequence (Samples 274, 50), and the the Upper Zone in the vicinity of Maila (Profile FGH,
plagioclase is andesine which places the rock in the diorite Fig. 13) and constituting Sefelere Mountain (C5) lie
suite. Above the troctolite is some 10 m of mottled between components of the fault. From Malagale the
anorthosite followed at Magnet Heights by five metres of main fault swings north-eastward past Kwaane Mountain
ovicular magnetite diorite (Willemse, 1969b, Fig. 5, where the throw is some I 800 m. It then bends westward
p. 195), a useful marker which was traced for a strike dis- round Sefelere Mountain and resumes a northerly course.
tance of some 15 km.
Approximately 100 m higher in the sequence is a group
of six magnetite seams of which only No. 17 and No. 21 are III. MINERALOGY OF THE LA YERED SEQUENCE
useful in mapping. Seam No. 21 is the uppermost seam in A. General
the Layered Sequence and is normally some 10m thick. I n the field, samples were taken along surveyed lines at
However, in the north of the area it is poorly differentiated approximately 35 m vertical intervals in the sequence from
and of indefinite width possibly due to interference by the the M erensky Reefto the roof (Figs. 12-14). Some 150 thin
adjacent roof-sediments during its deposition. Beneath the sections were examined and determinations were made of
Signal Hill quartzite, which comes close to Seam 21, the the compositions of plagioclase, orthopyroxene and
latter is only about half its usual thickness. olivine; clinopyroxene was not investigated owing to shor-
Above Seam 21 is approximately 200 m of olivine diorite tage of time. However, Atkins (1969, p. 238) found that in
grading upwards into some 30 m of dark, pinkish the Bushveld Complex there is a close correlation between
granodiorite which is the topmost unit of the Layered Se- changes in composition of ortho- and clinopyroxene:
quence. This rock-type was probably produced through hence for the moment, one can draw tentative conclusions
hybridisation of diorite with the overlying roof-rocks. In on the behaviour of the magma from compositional
the north, the layered rocks of subzone 0 consist of imper- changes of orthopyroxene alone. Summaries of the results
sistent bands of magnetite, anorthosite and troctolite. of this work are set out in Table I and Fig. 8.
These local varieties may have been formed by differentia-
tion of residual magma isolated into cusps beneath the B. Plagioclase
undulating roof. The composition of plagioclase was determined optical-
On Oroogehoek 842-KS (14) a lens of granodiorite ly using a 4-axes Universal Stage mounted on a Carl Zeiss
(Sample 405) up to 70 m thick was emplaced between lep- microscope and three or more large, fresh unstrained
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

tite and granophyre. On Owars-in-die-Weg 137-JS (Nl) a cumulus crystals exhibiting Albite twinning were
lens of gabbro (Sample 459) occurs 300 m up from the base measured in each slide. Albite twinning is preponderant in
of the granophyre. In appearance it is quite unlike diabase Bushveld plagioclase though some Manebach twins were
found elsewhere in the area and it is thought to be related also encountered. Indicatrices were replotted normal to
to the Layered Sequence. It is medium-grained with a the 010 composition plane and the molecular composition
chilled margin some 30 cm wide and contains phenocrysts, was evaluated from Plate VIII of Burri, Parker and Wenk
up to I cm long, of plagioclase and olivine. It possibly (1967). Extinction angles and 2V measurements were used
represents gabbroic magma squeezed up from the main to check these results. The compositions determined are
magma reservoir into the overlying granophyre. of the inner unzoned parts of the crystals.
D. Geological Structure Zoning, where evident, is restricted to the outer portion
1. Folding of the crystals and is pronounced where a myrmekitic
Near the west side of the Magnet Heights Granite, type of intergrowth is found in margins of crystals (Wager
Seams No.6, 17 and 21 dip eastwards in contrast with the and Brown, 1968, p. 387). This intergrowth which was not
normal westerly dip of the Layered Sequence. Oown- investigated commonly occurs with, and shortly after, the
folded roof-sediments near Sekhukhune confirm the ex- appearance in the sequence of pigeonite which on cooling
istence of a major anticline. This anticline abuts against inverted to hypersthene.
the Magnet Heights Granite and was traced northward Working on plagioclase from the Merensky Reef on
over a distance of 30 km (Profile ABCOE, Fig. 13) and it Forest Hill 117-KT, a few kilometres north of the present
continues out of the area. There is a fault along the axis of area, Schwellnus (1956, p. 139) found that the plagioclase
the fold at Magnet Heights and probably northwards of this reef has a composition An 65 - 75 . Wager and Brown
beneath the plain. Bordering this fold to the east is the (1968, p. 351) present a composition at this level of An 76 .
Sekhukhune Fault. Low in the Main Zone on Winnaarshoek 250-KT (A 7) the
composition determined by the writer is on average An70
2. Faulting and this changes gradually upwards to An60 below the
The Steelpoort Fault Pyroxenite Marker. In the vicinity of this unit the com-
It has a south-west to north-west trend and is followed position of the plagioclase jumps to An 67 , a change that
along much of its length by the Steelpoort River. It has a was evident on both the Ojate 249-KTffhornhill 544-JS
maximum throw of some 750 m on De Hoop 886-KS (K5) (A5) and Mooimeisjesfontein 363-KT (17) traverses, which
where the Main Seam south of the fault abuts against are approximately 50 km apart.
Seam No. 12 to the north. The throw diminishes in both Upwards in the sequence from the Pyroxenite Marker
directions along the fault an'd is virtually nothing to the the anorthite content again decreases to An60 near the
north of Kennedy's Vale 361-KT (H8). It is noteworthy lowest magnetite seam. In approximately the lowest 300 m
that the throw of the Steelpoort Fault is a maximum where of the Upper Zone minute crystals probably of magnetite
it is approached by the south-western part of the (Groeneveld, 1968, p. 70), darken the labradorite. Taylor
Sekhukhune Fault. (1964, p. 25) also recorded these in plagioclase in part of
The Sekhukhune Fault the Layered Series of the Duluth Complex.
In the south the main fault is aligned along, and may Labradorite in the Upper Zone oscillates in the range
even have controlled the emplacement of the Magnet An 50 _60 before moving into the andesine field in Subzone D
 336 TRANSACTIONS OF THE GEOLOGICAL SOCIETY OF SOUTH AFRICA

of the Upper Zone. The uppermost quartz diorite (Sample This olivine associated with magnetite seams is anomalous
283) contains plagioclase of composition An 43 which com- in composition and does not follow the differentiation
pares closely with a composition of An42 obtained by Von curve of the cumulus olivine (Fig. 8).
Gruenewaldt (1966, p. 60) in diorite below the roof at In the Upper Zone there is phase-layering of cumulus
Bothasberg, 80 km to the south. Plagioclase in the hybrid olivine and its abundance appears to have a poorly defined
granodiorite against or in the roof has a composition inverse relationship with that of orthopyroxene, by which
An 24 - 3S • it may also be replaced. The iron content of the Bushveld
Plagioclase in anorthosite is almost invariably cumulus olivine increases rapidly from F040 at the base of
saussuritised though its composition could be determined the diorite to Foo 700 m higher against the roof.
optically in all except the band below the Sisal Troctolite
D. Orthopyroxene
Marker. Raal (1965, p. 19) working on bore-hole material
Orthopyroxene compositions were determined by
from the Western Bushveld, found extensive alteration of
indisputably unweathered anorthosite, so the saussuritisa- measuring 2Vx and nz . By cutting the sections normal to the
tion must be regarded as a magmatic phenomenon. It is in- strike of the igneous layering, a partly successful attempt was
teresting that in the ovicular diorite (Sample 51) the made to obtain crystals suitably orientated for
plagioclase in the anorthosite ovoids is highly altered measurements of2Vx which however are of no use fordeter-
whereas that in the surrounding diorite is fresh. Willemse minations in the compositional range En 40_60 .
(1969a, p. 16) pointed out that the development of To measure 2V x the lower aperture of the microscope
anorthosite reflected a trend towards alkali enrichment. was opened fully, as recommended by Munro (1963,
This is borne out by Hall's analysis (1932, p. 334) of p. 312). After correcting the angle on Von Fedorow's
anorthosite (from beneath the Main Seam at Magnet nomogram (Trager, 1959, p. 123), the molecular composi-
Heights) which is unusually rich in Na 20 (four per cent), tion was evaluated from the graph of Hess (1952, p. 180),
K20 (two per cent) and H 20 (three per cent). Saussuritisa- compiled mostly from a study of orthopyroxene in North
tion may thus have been brought about by a concentration American basic complexes. Bushveld orthopyroxenes
of alkalis and volatiles in the anorthosite bands. seem to have a minimum 2V x of approximately 50°. This
differs from the value of 46° deduced from Hess's graph
C. Olivine which unfortunately contains very few points in the in-
Olivine compositions were evaluated primarily by termediate compositional range and hence warrants
measuring ny using a mixture of methylene iodide and further refinement.
a-monobromonapthalene or a sulphur/phosphorus liquid. To measure nx the same technique was used as for
Fluid was mixed to equal the ny and the refractive index of olivine: nz, along the c-axis, was identified as the max-
that fluid was then determined using a Leitz-Jelley refrac- imum refractive index in (110) cleavage fragments. The ac-
tometer. As values for different grains vary slightly the curacy of the determinations is thought to be ± ,002 which
average measured ny is thought to be accurate to ± ,005 corresponds with a compositional range of ± three mole
which is equivalent to a compositional range of ± two
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

per cent (graph of Hess, 1952, p. 180). in the bronzite-

mole per cent. The molecular percentage was read from hypersthene range, refractive index determinations
the graph of Kennedy presented by Trager (1959, p. 37). generally gave results of one to two mole per cent higher
X-ray determinations were made using the forward in the enstatite molecule than values obtained by 2V x
reflections in a 114,6 mm AEG Guinier camera as measurements.
developed by Jagodzinski. The molecular composition was I n the Merensky Reef on Forest Hill 117-KT,
evaluated from duo (Yoder and Sahama, 1957, p. 486) and Schwellnus (1956, p. 139) found the orthopyroxene to be
the b lattice parameter (Jahanbagloo, 1969, p. 248). The bronzite, En 82 - 83 . On Winnaarshoek 250-KT (A 7), the
accuracy of the determinations is thought to be ± five composition immediately above the reef is En78 and this
mole per cent which agrees closely with the limits of changes gradually upwards in the sequence until the first
Yoder and Sahama of ± four mole per cent towards the pigeonite (inverted to hypersthene) is encountered some
fayalite end-member. 1 430 m higher in the sequence on Djate 249-KT. This
Some density measurements were also made by means primary phase change took place over a vertical interval of
of a pycnometer but complete separation of grains was im- approximately 100 m and at a composition of about En 69 .
possible in all but the "cleanest" troctolites. Optical axial However, higher in the sequence where the magma moved
angles yielded erratic compositional values, particularly in back into the field of crystallisation of primary
the case of iron-rich olivines. orthopyroxene, the latter reappeared at a composition of
The scatter of compositional values obtained by En 66 ·
different techniques indicates that existing data are not in On Mooimeisjesfontein the level at which pigeonite
agreement for naturally occurring fayalitic olivines. This appeared is 500 m closer to the Pyroxenite Marker than on
was also the experience of Heckroodt as recorded in his Thornhill (Fig. 8). These discrepancies, when fully studied,
thesis (1958a, p. 30). Discrepancies were such that he did may indicate lateral temperature and compositional
not publish results for his iron-rich olivine (1958b, variations in the crystallising magma.
pp. 377-386). Just below the Pyroxenite Marker, primary orthopyrox-
Olivine is absent in the Main Zone except in very rare ene reappears and persists almost to the top of the Main
patches which may have pegmatoid affinities. Olivine Zone. In the vicinity of the base of the Upper Zone in-
appears in the Layered Sequence above Lower Seam No. verted pigeonite (Fig. 9) again takes the place of primary
3, and approximately 30 m below the Main Seam is orthopyroxene and persists up to the roof, the only excep-
stratigraphically the lowest troctolite band (Sample 40), tion being the occurrence of some primary hypersthene in
some three metres thick and consisting almost exclusively the pyroxenite above the Main Seam (Sample 242). At
of olivine (F06o ) and plagioclase (An6o). Brown (1957, places in the Upper Zone hypersthene was seen to have
p. 514) likewise recorded that in the Skaergaard intrusion, partly replaced hyalosiderite grains.
olivine reappears as a primary precipitate mineral above The composition of orthopyroxene in the Upper Zone
the Middle Gabbro and marks the beginning of the varies erratically and this may be due to iron in bulk hav-
ferrogabbro in the Layered Series. ing been extracted periodically from the magma to
Some olivine occurs at the base of Seam No. II and of produce magnetite seams. The most rapid enrichment in
the Main Seam (Willemse, 1969b, Fig. 16, p. 200) and was iron is between Seam 21 and the roof, the increase over
found by him (1969b, p. 206) to be forsteritic (Foso- 8s ). some 210m being from En 4S to En 29 . Atkins (1969, p. 248)
Olivine interstitial in Seam No.4 is also poor in iron (F078 ). similarly found the most ferfuginous orthopyroxene in the
 GEOLOGICAL INV ESTIGATION OF THE BUSHVELD COMPLEX 337

Bushveld to have a composition, estimated optically as

Mg 29 Fe 7l"

Figure 11
Biotite (pale grey) enclosing titanomagnetite (white), pyroxene
(grey) and apatite (small black laths). Large black crystals are
plagioclase. (Sample 280.) Negative photo, X 9.
Poikilitic inverted pigeonlte (hypersthene) surrounding
titanomagnetite (black), Seam 7. Reaction rim is clinopyroxene;
pale crystals are plagioclase. Crossed nicols, K 12.
hornblende may be a primary mineral near the top of the
Layered Sequence, orthoclase and most of the quartz are
thought to be products of hybridisation of the uppermost
diorite with the overlying acid rocks.

F. Suggested mechanism for differentiation in the Main and

Upper Zones
Prior to the formation of the Merensky Reef, a surge of
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

parent-magma was possibly introduced into the chamber

(Willemse, 1969a, p. 8). Upwards from the reef the magma
embarked upon a gradual trend of differentiating in the
stability field of plagioclase and two pyroxenes for an in-
terval of about 2 200 m up to the level of the foot-wall of
the Pyroxenite Marker.
The present work and that of Von Gruenewaldt (1970,
p. 69) indicates that there was probably a major addition of
magma before the formation of the Pyroxenite Marker. At
the level of the Pyroxenite Marker, the magma probably
embarked on its final and uninterruped course of differen-
tiation, culminating in the formation of the diorite con-
Figure 10
Titanomagnetite crystals set in poikilitic plagioclase (grey and stituting the top of the Upper Zone.
white). Speckled grey crystal is olivine, Seam 21. Crossed nicols, As there is presently no evidence of magma influxes in
X25. the Upper Zone certain tentative conclusions can be
drawn. Above the Pyroxenite Marker some 750 m of
layered rocks were deposited before the magma returned
E. Distribution of other minerals in the Main and Upper to its earlier state, as evidenced by the composition curves
Zones of plagioclase and orthopyroxene (Fig. 8).
Cumulus titanomagnetite (Figs. 9 and 10) appears This interval may thus be the approximate minimum
abruptly at the level of Lower Seam I and constitutes thickness of magma emplaced. After the introduction of
some five per cent or more of virtually all the rocks of the this probable final surge of magma the internal depth of
Upper Zone. Cumulus titanomagnetite coexists with the chamber is likely to have been of the order of 2 400 m
olivine in the compositional range F0 6o...:o , otthopyroxene which is the present vertical distance between the Pyrox-
En 63 - 29 and plagioclase An 6o- 43 ' The norite (Sample 43),50 m enite Marker and the roof.
above the Main Seam contains no ore-minerals and this will This postulation gives rise to questions such as, where is
be referred to later in discussing the mode of formation of the chromium which would have been present in the influx
magnetite seams. It is also noteworthy that titanomagnetite of primary magma? A geochemical study carried out by
is extremely scarce in Samples 284 and 40, taken from 20 and Dr. D . R. Bowes at Glasgow University on the writer' s
30 m below the Main Seam. samples from the Main and Upper Zones yields the following
Biotite (Fig. II) constitutes some one per cent or more data: the magma had more or less rid itself of chromium at
of rocks of the Upper Zone and appears in this quantity at the level at which sample 177 was taken some 600 m above
the same level as olivine (F0 6o ) approximately 120 m below the Merensky Reef. Samples 294,296 and 299 just above the
the Main Seam. However, crystallisation of apatite com- Pyroxenite Marker again contain appreciable amounts of
menced at more or less at the same level as that of chromium (450-825 p.p .m.) and this may represent
andesine and hence is restricted to the diorite constituting chromium which was introduced with the possible heave of
subzone D of the Upper Zone . parent-magma. Most of the remainder of the chromium in
Hornblende, quartz and orthoclase occur as essential the magma appears to have been precipitated in the
minerals only within 100m of the roof. Though some magnetite seams.
TRANS. GEOL. SOC SAFR .. VOL. LXXVII MOLYNEUX. FIGURE 12

~s
24°30'+,~~~~~~~--------~--··------~~----------------
A
prug, Dyke or
56am 21;- -

KANAAN 783-KS

GflUK'S

STER~SPFWIT 807- KS

E
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

HOEGELfG[N 809-KS

VERGfLEGEN 819-K5

MORGENZON 949 - KS

LEEUWKRAAL 877-" KS

J
LEGEND

c=J ALLUVIUM RECENT

IT:] ALKALI INTRUSIVES

} SPITSKOP COMPLEX
~ CARBONATITE

~ DOLERITE a DIABASE POST-BU$HVELD AGE

1+ +1 COARSE GRANITE It' PEGMATITE

} 8USHYElD GRANITE
~ MEDIUM 8. FINE GRANITE

GRANOPHYRE
K LEPTITE (ALTERED FELSITE)
} EPICRU$TAL ROCKS

DIORITE (.:. OVICULAR VARIETY)

366-KT
MAGNETITE NOR ITE -GABBRO (UPPER ZONE)

NORITE-GABBRO SUITE - MAIN ZONE

BANDED GABBRO
M~~~~~~TEE _ ;EMA~~I~E~~R~16PlUG BuSH VELD
SPOTTED ANORTHOSITE MAIN e.
UPPER ZONES
LAYERED SEQUENCE
MOTTLED ANORTHOSITE

~ TROCTOLITE
~ PYROXENITE

I·~·>: -:1 PORPHYRITIC GABBRO - NOR I TE

25°00'
25°00'
~ MERENSKY REEF /Pd PEGMATOID
L
~ DOLOMITE

~ HORNFELS } TRANSVAAL SYSTEM

30°00' ~ QUARTZITE

FOLO AXIS
'::"5~ FAULT 8: SHEAR lONE
__ 5° DIP DIRECTION
45° INVERTED STRATUM
OBSERVED BOUNDARY
INFERRED A"<L ,J" r:':ROP BOUNDARY
DOUBTFUL BOUNDARY.
~- STREAM
... TRIGONOMETRICAl SURVEY BEACON
@ SAMPLE NUMBER

M
FIG. 12

THE GEOLOGY OF PARTS OF SEKHUKHUNELAND

AND OF
THE STEELPOORT VALLEY
145-J$

N
Scale- 1'100 000

FOR DETAILED MAP OF SPlTSKOP SEE STRAUSS a TRUTER,1950

AND OF MAGNET HEIGHTS AREA, MOLYNEUX, 1964

2 3 4 5 6 7 8
TRANS. GEOL SOC S.AFR., VOL LXXVlI MOLYNEUX, FIGURE 13

PHEPANE MTN.
B LEOLO MOUNTAINS
APPEARANCE OF
INVERTED PlGEONITE c o
A
SEFELERE MTN.

! E

, -; REEf
MEREi'lSK
CRITICAL ZONE

F G LEOLO

--
MOUNTAINS

NTSWETETAU MTN.
SEKHUKHUNE do
FAULT
do
do
H

MERENSKY REEf
CRITICAL ZONE

Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)
x y
SIGNAL HILL APPEARANCE OF

POSSIBLE DISPLACEMENT
ON DOLERITE DYKE.
+ +
BRAZES OF SEKHUKHUNE FAULT\ LE~~~OU_:;;;'W~EE~T-::.,_I_NV_'"_TE_l_."'_EO_N_ITE_'.
+ + + + STEELPOORT
FAULT
KENNEDY'S VALE
MAGNETITE PLUG J
tM~.Fn<;ER' . --'--OuTIJERSHEET' _ _
' __._~~_,___
+ MAGNET
+ , p-;ROXENITE " f . - - ' -~~JA~UwS~-'-~--' - - , - - , '-'::~=--:=-===:=-==:'--:--~rJ;;;;;::;;::;P!~~-:--
+ HEIGHTS
+ + +
GRANITE ~ .-
, , SH
'THAMA.J5,°.9---,
- - RT SHEET - .-
SHEET_'--'-f:~;;"""-;;,-'~;;'
HET.-!2......-;",,--"ET '~~
- --~, ;S~O~U~P_I~A~NAl'~S~H~E~E_T~:~''C-~-rrp.'::r:;l-~-Ji-:n:::::g:~~c>:-~'-:c-tJ:C'~~'="":=~=~=~'1i'r~~JO~~~~~Fo~~~~~~~~1
'DJATE SHEET' , • , . , . , . . . . . HOS\TE .
+ + + J ,_-~TLI~:R!~~W_S_HE~ i UPPER, MOTTLED, ANOR,THOSI:rE , r:lATuM LINE _' SEA LEVEL i ' MAIN, MOTTLED ~NORT, ,

LEGEND:
FIG. 13 1+ +1 BUSHVELD GRANITE ITTTTT I TROCTOLITE

GRANOPHYRE

Sections across Main and Upper

MAGNETITE NORITE - GABBRO ,UPPER ZONE
M

--
LEPTITE POSSIBLE LAYERING
NORITE - GABBRO SUITE t MAIN ZONE

Zones, Sekhukhuneland and the QUARTZITE DOLERITE

HORNFELS FAULT
Steelpoort Valley.
~ DIORITE

f2J MAGNETITE SEAMS a PYROXENITE

SCALE
[2] MOTTLED ANORTHOSITE
o 2 Km.
I
~
1

SPOTTED ANORTHOSITE
~

DRAWN: J,Quintao-1973
TRANS. Gl:.:OL. SOc. S.AFR .. VOL. LXXVII MOLYNEUX FIGURE 14

I
- UPPER MAGNETITE
+ +
GARATOUW
206- 18- 0,12,OJ280,6m ~-....j SEAM 7
1HEET
253-
tI 17-
252-
O,258Q,27m ~---4SEAM 6
1,2m PYROXENITE
45,242 - >-
I x
2~-
4» I 44- .,
lit
~ I~m
SOUPIANA
SHEET
.,...
III
f----
N N N
> 43- ~ o 2m SEAM 5
I
.=

i
200- 14- ~45m SEAM 4
189- I u 15m SEAM 3

428~~;+
I 0,75m SEAM 2
198-
I 3,486 °lg~ SEAM I
, MAIN MAGNETITE SEAM
314-
I 2iI= Sub-Zone B
00000
T
---- + 40- T

•I 286-
:::>
I-
Sub Zone A G
405 -
ROOF OF
LAYERED
188- --G-- I 288- ., 3000m G
G
G
SEQUENCE
289- ~

187-
G G DJhTE
1----- SHEET
2-
290-
~ I ,5m tm~~ LOWER SEAM 3
54 8 55 - UPPER ZONE Y

I .=
0 0,3 m T.
G G
r, LOWER SEAM 2
283-
I
291- -I~
I 312- 0,35m G G LOWER SEAM I
186-
I Sub Zone A
282-
UPPER ZONE
Sub-Zone 8 t MAIN ZONE
Sub Zone A f Sub-Zone C
UPPER 306-
MOTTLED
ANORTHOSITE
2 m I:m:az::l~ I 281-
....... t
305-

182- 280-
181-
303- >-
313- ., 10m' . ·U· •
III 650- X
4; G2556600-
IOOOm
>
o
.,
179-
302-
.= .,~>
o
300-
.=
177- CONSPICUOUS 279- c
o
t------t LAYERING
278- .,
III

Q.
E
299- 2500m 51- ~
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

276-
m ~== PYROXENITE
0,12mp: , 238-
50-
174- Detail from I 298- T T
J.C.I. borehole I T
Mkwanestad - I
East Tvl. MAIN 236-
T T
274-
~o~oU:o~o~ol;f MOTT LED
Sub-Zone 0 ~,T ~'(
ANORTHOSITE
296-
235- 4 OOOm- '\~TI~\
I 3m .•...•.
I G
o 0 273 - Su b-Zone C
172- o 0 0 t 295- 233- --G-- G G
232-
294 ....;~ I G G G
I~ • • PYROXENITE
231- 0,180,1 m ~G__G--4SEAM 14
PURE -~
272-
0,3m F====I ANORTHOSITE 335- 230- ~ub-Zone C
169-
~~~= ~n= ____~:::~~~PYROXENITE
~ub-Zone B
MKR.
PUfE ANORTHOSITE
49-
I 63- 10m~~........!
166- 500m above 27- 0,35m G (oj SEAM 13
Merensky Reef.
329-
226- I
I 269-
G
.t=:====1 'BANDED GABBRO'
G G
I
I G
165- I 24- 0,2280,35m 1---..-01 SEAM 12
327- 224-
THAMA 268-
W
KOOSH 465-
1,2 m prJGlO;z:f 223-0 SHEET
325- o I
2000m 267-
~___
NEEDLES OF
PYROXENE 221-
m I 48- T T T

u:
-- t
CI
163- G
35- MAGNETITE 22:~~= 0,5m SEAM II
_<L_G_ o f------ GABBRO 263- 0,22m ......-.....-1 SEAM 10
w
o 2m .••••.
0.. 323- 219- ., t 47-
I
160- 0 3,5 m ~c:::I::z::q .,~> HET FORT
19 -
260-
0,22 m I-~
G-G~ SEAM 9
0,17m G G SEAM 8
o SHEET
.= I 62- Sub-zo~e5g0m
•f>
216- T T 'SISAL MARKER'
I 46- ___ •.....;..,;.~ ;;';'_I-~....!....-I

----- + Sub-Zone B
159- ~

158- g
10m •.•...
215-

213-
I
----- OUTLIER
I
•

W
157- : N N N
Q. - 1------ SHEET
E 320- I
o 255- POSSIBLE BREAK
CJ)
I
154- I
Sub-Zone A 0
t g:::fi: 0 00 BASTARD REEF
318-
t
GARATOUW
IN SUCCESSION
ON TRAVERSE X Y

317- SHEET 254-

153- MAIN ZONE
CRITICAL ZONE
MERENSKY REEF
~
Legend:
E3 BUSHVELD GRANITE r;-l
~
OVICULAR DIORITE

FIG. 14 r~~;~ GRANOPHYRE ~ MOTTLED ANORTHOSITE

Geological Column of Main ~ LEPTITE ~ SPOTTED ANORTHOSITE

~ QUARTZITE ~ TROCTOLITE
and Upper Zones in ~ HORNFELS o HYPERSTHENE GABBRO 6 HYPERITE

Sekhukhuneland. ~ GRANODIORITE ~ GABBRO

_ DIORITE ~ NORITE

IZJ MAGNETITE SEAMS AND PYROXENITE o PORPHYRITIC GABBROIC ROCK

 338 TRANSACTIONS OF THE GEOLOGICAL SOCIETY OF SOUTH AFRICA

During the formation of the rocks in the interval of Hamilton, W. (1960). Form of the Sudbury Lopolith. Canad.
some 500 m below the level of the Pyroxenite Marker the Miner., 6,437-447, .
Hammerbeck, E. C. I. (1970). The Steelpoort Park ~~anlt.e,
chemistry of the magma appears to have been sympathetic eastern part of the Bushveld Complex, and the magnetItItes In
to the onset of crystallisation of cumulus titanomagnetite the gabbroic country. Geol. Soc. S. Afr., Spec. Publ. 1,
and iron-rich olivine. However, after the probable influx 299-311.
of parent-magma below the level of the Pyroxenite Heckroodt, R. O. (1958a). Die platinumd:aende du.n~etpy'p of'
Driekop (Oos-Transvaal) en die sames!e1lmg. van oilVlen m die
Marker, precipitation of titanomagnetite ceased until Bosveldstollings Kompleks. M.Sc. thesIs, Unlv. Pret. (unpubl.)
crystallisation reached the level of the base of the Upper 38 pp. . . f
Zone at which time the magma again moved into the ____ (1958b). An X-ray Method for the DetermInatIOn 0
stability field of titanomagnetite and iron-rich olivine. Olivine. Trans. Geol. Soc. S. Afr., 61, 377-386. . .
Henriques, A. (1966). Geology of ores of the Ammeberg DIstrIct
In considering the mode of formation of cumulus (Zinkgruvan), Sweden. Arkiv. Miner. Geol., 4, 3-238. .
titanomagnetite in pasic complexes the conclusions of Hess, H. H. (1952). OrthopY.r0xe~es of ~he Bl!shveld type, 10~
Osborn (1962, pp. 215-220) are drawn upon. He found in substitutions and changes In Unlt cell dImenSIOns. Amer. J. SCI.
studying the system CaSi03-Mg2Si04-Si02-Fe304 that (Bowen Volume), 180, 173-187.
- - - - (1960). The Stillwater Complex, Mountana. Mem. Geol.
when the oxygen pressure is low, the effect of fractional Soc. Amer., 80,230 pp.
crystallisation is to produce successively mo~e iron-rich Jackson, E. D. (1961). Primary textures and mineral associations
crystalline silicate aggregates. In constrast with this, a con- in the Ultramafic Zone of the Stillwater Complex, Montana.
stant high oxygen pressure produces primary magnetite U.S. Geol Surv. Prof Pap., 358, 106 pp. . . .
J ahanbagloo, C. (1969). X-ray diffraction study of olIVIne solId
and prevents the crystallising liquid from moving towards solution series. Amer. Miner., 54,246-250.
a higher iron content. Hence he drew up two discontin- Lombaard, B. V. (1934). On differentiation and relationships of
uous series depending upon the oxygen pressure: when the rocks of the Bushveld Complex. Trans. Geol. Soc. S. Afr.,
olivine crystallises it takes Si0 2 from the magma and leads 37,5-52.
Molyneux, T. G. (1964). The geology and the structure oft.he area.in
to a build-up in FeO, whereas when magnetite crystallises, the vicinity of Magnet Heights, Eastern Traf1:svaal,. with speczal
the reverse occurs. reference to the magnetic iron ore. M.Sc. theSIS, Unlv. Pret. (un-
In the Upper Zone of the Bushveld Complex, publ.) 112 pp. . . . .
hyalosiderite (Fo 60 ) appears in the sequence approximate- - - - - (1970a). The geology of the area In the VIClnlty of
Magnet Heights, Eastern Transvaal, with special reference to
ly 100m higher than cumulus titanomagnetite, and alter- the magnetic iron ore. Geol. Soc. S. Afr., Sp. Publ. 1,228-241.
nate or simultaneous crystallisation of titanomagnetite and - - - - (1970b). A Geological Investigation of the Bushveld
hyalosiderite may have exerted a buffering effect which Complex in Sekhukhuneland and part of the Steelpoort Valley,
maintained the oxygen pressure within certain limits. Eastern Transvaal with particular reference to the oxide minerals.
D.Sc. thesis, Univ. Pret. 125 pp.
The complete absence of magnetite shortly above the Munro, M. (1963). Errors in measurement of 2V with the
Main Seam (Sample 43) and its virtual absence shortly Universal Stage. Amer. Miner., 48,308-323.
below (Samples 40 and 284) strongly implies crystallisation Osborn, E. F. (1962). Reaction series. Amer. Miner., 47, 211-225.
near the floor, as suggested for the Stillwater Complex by Raal, F. (1965). The transition between the Main and Upper Zones .of
the Bushveld Complex in the Western Transvaal. M.Sc. theSIS,
Reproduced by Sabinet Gateway under licence granted by the Publisher (dated 2010)

Jackson (1961, pp. 94-98). This is because one would not Univ. Pret. (unpubl.), 59 pp.
expect such complete separation of phases had the crystals Schwellnus, J. S. I. (1956). The basal portion of the Bush veld Igneous
had to descend from near the roof through probably more Complex and the adjoining metamorphosed sediments in the
than a kilometre of magma. North Eastern Transvaal. D.Sc. thesis, Univ. Pret. (unpubl.),
206 pp. .
It is thus believed that cyclic crystallisation near the Strauss C. A., and Truter, F. C. (1950). The Alkali Complex at
floor of the magma chamber aided by magmatic currents Spitzkop, Sekhukhuneland, Eastern Transvaal. Trans. Geol.
of a presently unknown nature were mostly responsible for Soc. S. Afr., 53, 81-125.
producing the diversity of rocks in the Upper Zone. Taylor, R. B. (1964). Geology of the Duluth Gabbro Complex,
near Duluth, Minnesota. Minnesota Geol. Surv., 44, 63 pp.
Troger, W. E. (\959). Optische Bestimmung der gesteinsbildenden
Minerale. Tei! I. Schweitzerbart, Stuttgart., 147 pp.
REFERENCES Wager, L. R., and Deer, W. A. (1939). Geological investigation in
Atkins, F. B. (\969). Pyroxenes of the Bushveld Intrusion. South Eastern Greenland. Pt. III. Petrology of the Skaergaard Intru-
Africa. J. Pet. Oxford, 10,222-249. sion, Kangerdlugssuag, East Greenland. Medd. Grlmland, 105,.
Brown, G. M. (1957). Pyroxenes from the early and middle stages No.4, 352 pp.
of fractionation of the Skaergaard Intrusion, East Greenland. - - - - , and Brown, G. M. (1968). Layered Igneous Rocks.
Miner. Mag., 31, 511-543. Oliver and Boyd, Edinburgh and London, 588 pp.
Burri, C., Parker, R. L., and Wenk, E. (1967). Die optische Willemse, J. (1964). A brief outline of the geology of the Bushveld
Orientierung der Plagioklase. Birkhauser Verlag., Basel, 334 Igneous Complex. In Haughton, S. H., Ed., The Geology of
pp. Some are Deposits in Southern Africa, II. " Geol. Soc. S. Afr.,
Groeneveld, D. (1968). The Bushveld Igneous Complex in the 91-128.
Stoffberg area, Eastern Transvaal, with special reference to the " - - - - (\969a). The geology of the Bushveld Igneous Complex,
magnetitite seams. D.Sc. thesis, Univ. Pret. (unpubl.) 169 pp. the largest repository of magmatic ore deposits in the world.
Gruenewaldt, G. von (1966). The geology of the Bushveld Igneous Econ. Geol. Monog., 4, 1-22.
Complex east of the Kruis River cobalt occurrence, North of - - - - (l969b). The vanadiferous magmatic iron-ore of the
Middelburg, Transvaal. M.Sc. thesis, Univ. Pret. (unpubl.) 106 Bushveld Igneous Complex. Econ. Geol. Monog., 4, 187-208.
pp. - - - - , and Bensch, J. J. (1964). Inclusions of original
- - - - (1968). The Rooiberg felsite North of Middelburg and carbonate rocks in gabbro and norite of the Eastern part of the
its relation to the Layered Sequence of the Bushveld Bushveld Complex. Trans. Geol. Soc. S~Afr., 67, 1-87.
Complex. Trans. Geol. Soc. S. Afr., 71, 153-172. Yoder, H. S., and Sahama, G. (1957). Olivine X-ray determinative
- - - - (1970). On the phase-change orthopyroxene-pigeonite curves. Amer. Miner., 42,475-491.
and the resulting textures in the Main and Upper Zones of the
Bushveld Complex in the Eastern Transvaal. Geol. Soc. S. Afr., Geology Department,
Sp. Publ. 1,67-73. Anglo American Corporation,
Hall, A. L. (1932). The B ushveld Igneous Complex of the Central Johannesburg.
Transvaal. Mem. Geol. Surv. S. Afr., 28, 560 pp. Accepted for publication by the Society on 25.5.1974