Journal of Petrology

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Miyano, T. Articles by Beukes, N. J. Search for Related Content Social Bookmarking          What's this?

Journal of Petrology | Volume 38 | Number 5 | Pages 651-676 | 1997
© Oxford University Press 1997

Mineralogy and Petrology of the Contact Metamorphosed Amphibole Asbestos-bearing Penge Iron Formation, Eastern Transvaal, South Africa

T. Miyano¹ and N. J. Beukes^2,*

¹ Institute of Geoscience, the University of Tsukuba Ibaraki 305, Japan
² Department of Geology, Rand Afrikaans University PO Box 524, Auckland park, Johannesburg 2006, South Africa

Received August 15, 1996; Revised typescript accepted December 4, 1996

	ABSTRACT

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
The mineralogy and petrology of the Penge Iron Formation of the Transvaal Supergroup, situated in the contact metamorphic aureole of the Bushveld Complex, was studied in detail to investigate the petrogenesis of crocidolite- and amosite-bearing rocks at Penge and Mafefe, in eastern Transvaal, South Africa. The rocks are laterally equivalent to the diagenetic to very low-grade metamorphic Asbesheuwels iron-formation succession in Griqualand West. Metamorphism at Penge apparently took place at 420–460°C and 2.6 ± 0.8 kbar as the result of the intrusion of the Bushveld Complex. The maximum metamorphic temperature at Mafefe was some 40–80°C lower than at Penge. Bulk-rock composition was a major factor in controlling the distribution of crocidolite and amosite in the succession. Riebeckite and crocidolite are absent from Al-rich rock units containing biotite with Si/Al>3, but are commonly associated with ferri-annite with Si/Al ratios of the order of three. In contrast, amosite is well developed in carbonaceous Al-rich biotite–grunerite hornfels. Couplets of amosite–crocidolite seams, developed in magnetite–grunerite banded iron-formation, may have formed from alternating mesobands of iron-silicate–riebeckite known from the diagenetic to very low-grade metamorphic Kuruman Iron Formation of the Asbesheuwels Subgroup. There is also evidence, especially in the case of one specific asbestos reef, that riebeckite–crocidolite have been replaced by grunerite–amosite with increasing grade of metamorphism from Mafefe towards Penge. In contrast, small amounts of grunerite appear to have been replaced by riebeckite during retrograde metamorphism.

KEY WORDS: asbestos; iron-formation; metamorphism; mineralogy; Transvaal Supergroup

	Introduction

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
The Penge Iron Formation of the Transvaal Supergroup is situated in the contact metamorphic aureole of the Bushveld Complex in eastern Transvaal, South Africa (Fig. 1). The iron-formation succession is unique in the sense that it contains deposits of both crocidolite (asbestiform riebeckite) and amosite (asbestiform grunerite), mined extensively up to the late 1980s (Beukes & Dreyer, 1986; Dreyer & Söhnge, 1992), in close spatial relationship. It is thus the only area known in the world where the relationship between the formation of crocidolite and amosite can be studied. In addition, some deep exploration diamond drill cores are available that can be used for establishing what influence the composition and stratigraphic setting of original iron-formation lithofacies had on variations in metamorphic mineral assemblages and chemical composition of certain metamorphic minerals in the succession.

View larger version (51K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 1. Simplified geological map showing location of boreholes MF2 and PA13, and the distribution of the Penge Iron Formation. Isotherms in the metamorphic aureole of the Bushveld Complex were postulated from Sharpe & Chadwick, (1982).

 
Metamorphic assemblages in iron-formations have been studied very extensively [see Klein, (1983) for a review] but integrated stratigraphic–metamorphic studies have received little attention. Similarly, the geology and mineralogy of the amphibole asbestos-bearing iron-formations in the Penge area have been studied by many workers (Peacock, 1928; Hall, 1930; Kirkman, 1930; Reinecke & McClure, 1934; Du Toit, 1945; Vermaas, 1952; Cilliers, 1964; Dreyer, 1982; Beukes & Dreyer, 1986; Dreyer & Söhnge, 1992) but they concentrated mainly on the economic potential of the rocks and detailed petrological data are not available. An exception is Miyano & Klein, (1983), who evaluated the phase relations of grunerite and riebeckite in iron-formation on the basis of available and predicted thermodynamic data. They showed that the phase relations are a function of oxygen fugacity [f(O₂)], sodium–hydrogen activity ratio [a(Na⁺)/a(H⁺), or sodium availability] and temperature at constant pressure. They also noted that crocidolite (or riebeckite) may readily be replaced by amosite (or grunerite) with increasing temperature at fixed a(Na⁺)/a(H⁺), or with decreasing a(Na⁺)/a(H⁺) at fixed temperature. This process may explain the presence of relict textures of crocidolite in amosite in iron-formation from the Malipsdrift (Miyano & Klein, 1983) and Penge areas (Miyano & Beukes, 1984a). However, no detailed mineralogical and petrographic data are available to evaluate these suggestions properly.

The purpose of the present study is therefore to document, in detail, the mineralogy and petrology of the contact metamorphosed Penge Iron Formation in the Penge and Mafefe areas (Fig. 1). Major aims are (1) to establish a petrogenetic model for spatially related crocidolite and amosite deposits, (2) to document the relationships between lithostratigraphic composition of the iron-formations and the character of the metamorphic mineral assemblages and (3) to calculate the temperature and pressure conditions in the iron-formation during intrusion of the Bushveld Complex.

	Geological Setting

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
The Pretoria and Chuniespoort Groups of the Transvaal Supergroup are exposed along the northeastern margin of the Bushveld Complex, which intruded Transvaal strata some 2052–2061 m.y. ago (Walraven et al., 1990) (Fig. 1). The Pretoria Group unconformably overlies the Chuniespoort Group, which consists of the Malmani Dolomite and the Penge Iron Formation (Fig. 1). The Penge Iron Formation varies considerably in thickness owing to folding and erosion that preceded deposition of the Pretoria Group (Schwellnus et al., 1962; Button, 1973). Cilliers, (1964) reported a thickness, excluding diabase sills, of 74 m for this succession at Kromellenboog and 178 m at Penge (borehole PA13) (Fig. 1). Further to the north, at Mafefe (borehole MF2), the succession is 670 m thick (Fig. 2) (Beukes, 1978). Samples for the present study were selected mainly from diamond drill cores PA13 at Penge and MF2 at Mafefe with a few samples coming from outcrops and underground mine stopes at Malipsdrift, Mafefe and Penge (Figs 1 and 2).

View larger version (31K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 2. Lithostratigraphic columns of the Penge Iron Formation at Penge (PA13) and Mafefe (MF2). Numbers on left of the PA13 column denote sample locations for bulk chemistry. (Note difference in scale between the two profiles.)

 
According to Beukes, (1978, 1983), the Penge Iron Formation is stratigraphically equivalent to the crocidolite-bearing Asbesheuwels Subgroup, comprising the Kuruman and Griquatown Iron Formations, in Griqualand West (refer to Fig. 2 for correlation). The Asbesheuwel Subgroup has an age of 2450 Ma (Trendall et al., 1990). The diagenetic to very low-grade mineral assemblages of the Asbesheuwels Subgroup are well known (Beukes, 1983; Miyano & Beukes, 1984b; Beukes & Klein, 1990) and could be used for comparison with the thermally metamorphosed Penge Iron Formation.

The iron-formation is only gently folded with fold intensity increasing northwards from Penge to Mafefe. Amosite and riebeckite asbestos deposits are preferentially developed in tensional stress areas along anticlinal and synclinal hinge lines of some of the folds (Beukes & Dreyer, 1986; Dreyer & Söhnge, 1992). Several NNE-trending faults displace the succession and dolerite dykes, intruded along these faults, may be of post-Karoo (Mesozoic) age (Schwellnus et al., 1962). Where these dykes intersect amosite seams, the fibres are often altered to hard silicified material (Cilliers, 1964; Day, 1978). Two older diabase sills intrude the Penge Iron Formation in both the Penge (PA13) and Mafefe (MF2) areas (Fig. 2), and transgress from the Penge Iron Formation to the Timeball Hill Formation of the Pretoria Group (Keep, 1962). These sills have been regarded as forerunners of the intrusion of the Bushveld Complex (Schwellnus et al., 1962). However, Miyano et al., (1987) have suggested that they postdate peak thermal metamorphism related to intrusion of the Bushveld Complex.

Limits of the contact metamorphic aureole of the Bushveld Complex have been mapped by Button, (1973) and Sharpe & Chadwick, (1982). The Penge–Mafefe area falls well within this metamorphic aureole (Fig. 1). The grade of contact metamorphism increases upwards in the Transvaal Supergroup towards the floor of the Bushveld Complex with the result that, towards the upper part of the Pretoria Group, argillaceous sediments have been altered to cordierite–sillimanite or andalusite hornfelses (Sharpe & Chadwick, 1982; Kaneko & Miyano, 1990). In the Penge area andalusite deposits are commonly developed within pelitic rocks of the Timeball Hill Formation of the Pretoria Group 900–1300 m above the Penge Iron Formation (Human, 1975; Hammerbeck, 1986; Human & Collins, 1986).

	Lithofacies and Stratigraphy

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
The stratigraphic succession of the Penge Iron Formation intersected in core PA13 at Penge correlates only with the lower part of the formation exposed in core MF2 at Mafefe (Fig. 2). This is the result of much more erosion of the succession at Penge, relative to that at Mafefe, before deposition of the overlying Pretoria Group (Fig. 2). The major lithofacies of the succession, namely garnet-bearing black shale, biotite–grunerite hornfels and banded hornfels, gruneritic banded chert and magnetite–grunerite banded iron-formation (Fig. 2) are respectively equivalent to (1) carbonaceous shale, (2) stilpnomelane–siderite lutite and banded lutite, (3) siderite banded chert and (4) haematite–magnetite–siderite banded iron-formation lithofacies developed in the unmetamorphosed Kuruman Iron Formation (Beukes, 1978, 1980, 1984). It should be noted that the definition of lithofacies in the Penge Iron Formation is somewhat problematical. The grade of metamorphic alteration is such that a combination of metamorphic and sedimentary terms have to be applied. For example, the recrystallization of quartz grains in original chert bands is so slight that they remain microcrystalline and the units are best described as chert-bands. Similarly, some of the original black shale units in the succession have been so slightly affected by metamorphism that it would be misleading to describe them as either slate or hornfels.

The black shales in the Penge Iron Formation (Fig. 2) are composed mainly of biotite, chlorite and kerogen with abundant sulphides (mainly pyrrhotite) and Mn-rich garnet in certain zones. Biotite–grunerite hornfels is either black and carbonaceous or green in colour and locally highly magnetic. Banded hornfels (Fig. 2) contains prominent chert mesobands. Major mineable amosite reefs are developed in some of the carbonaceous biotite–grunerite hornfelses (Beukes & Dreyer, 1986).

Gruneritic banded chert (Fig. 2) consists of relatively thick (a few centimetres) grunerite microbanded chert mesobands alternating with thin (1–2 cm) microbanded grunerite mesobands. Ankerite and siderite or biotite microbands are locally preserved in some of the chert mesobands.

Magnetite–grunerite banded iron-formation forms several units, between 2 and 38 m thick, interbedded with biotite–grunerite hornfels (Fig. 2). The banded iron-formation is microbanded and two subfacies, typically with gradational contacts, are distinguished, namely grunerite-rich chert-mesobanded iron-formation and magnetite-rich grunerite chert-mesobanded iron-formation. In some cases, considerable amounts of biotite or ferri-annite are present in the banded iron-formation.

The various lithofacies display a rather systematic vertical stratigraphic distribution. This is best explained with reference to drill core MF2 at Mafefe, where one of the most complete profiles known of the Penge Iron Formation is preserved below the unconformity at the base of the Pretoria Group (Fig. 2). The bottom part of the Penge Iron Formation, in conformable contact with Malmani dolomite, is composed of gruneritic banded chert with thick interbeds of black carbonaceous shale (Fig. 2). It is important to note that the shale at Penge, in core PA13, contains tiny euhedral porphyroblasts of garnet. At Mafefe, in core MF2, garnets are absent from the shale. This succession is overlain by iron-formation with three prominent interbeds of green or black carbonaceous biotite–grunerite hornfels (Fig. 2). The iron-formation units are composed of stacked sedimentary cycles of biotite–grunerite hornfelsgruneritic banded chertmagnetite–grunerite banded iron-formation. The cycles, which are of the order of 1–10 m thick, correspond to the unmetamorphosed stilpnomelane lutitesiderite banded chertsiderite–magnetite±haematite banded iron-formation cycles of the Kuruman Iron Formation (Beukes, 1978, 1980, 1983, 1984).

This cyclical succession of iron-formation is overlain by three rather thick and monotonous units of magnetite–grunerite banded iron-formation containing considerable amounts of ferri-annite (core MF2, Fig. 2) . Correlative units in the Kuruman Iron Formation are composed of magnetite–siderite–greenalite banded iron-formation (Beukes, 1978, 1983). The magnetite–grunerite banded iron-formation units are separated by biotite–grunerite banded hornfelses (Fig. 2), which correspond to stilpnomelane–siderite banded felutites of the Kuruman Iron Formation (Beukes, 1978, 1980).

The upper one of the three thick magnetite–grunerite banded iron-formation units discussed above is capped by a thin riebeckitic biotite–grunerite hornfels overlain by a succession of iron-formation composed of stacked cycles of biotite–grunerite hornfelsgruneritic banded chertmagnetite–grunerite banded iron-formation with crocidolite (Fig. 2).

The top part of the Penge Iron Formation at Mafefe is mainly composed of magnetic biotite–grunerite banded hornfels and carbonaceous magnetic biotite–grunerite hornfels (Fig. 2). Microbanding is absent and some gruneritic chert grainstone units are present as well as a magnetic grunerite-bearing edgewise chert conglomerate. This upper part of the Penge Iron Formation correlates with the clastic-textured lutitic and granular siderite–greenalite–magnetite iron-formation succession of the Griquatown Iron Formation in Griqualand West (Fig. 2) (Beukes, 1978, 1983).

Amosite reefs are most commonly developed in magnetite–grunerite banded iron-formation (Fig. 2), but could also be present in carbonaceous biotite–grunerite hornfels (such as the reef denoted by D in PA13, Fig. 2). In contrast, crocidolite reefs are only developed in magnetite–grunerite banded iron-formation and restricted to the Mafefe area (core MF2, Fig. 2). Reefs denoted by A and B in core MF2 (Fig. 2) are composed essentially of crocidolite. The reef denoted by C in core MF2 (Fig. 2) represents an excellent example of a mixed fibre reef with alternating seams of crocidolite and amosite. It should also be noted that the reef denoted by B is composed of crocidolite at Mafefe and of amosite at Penge (Fig. 2). In the asbestos fields a single band of asbestos fibre (normally 0.5–30 cm thick) is referred to as a seam and a group of seams (commonly 1–3 m thick) as a reef (Cilliers, 1964).

	Mineral Abundances and Assemblages

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
Mineral assemblages and volumetric abundances of minerals in the Penge Iron Formation, excluding those in younger veinlets, were determined by point counting from core samples of PA13 and MF2 (Fig. 3a and b). Major minerals in the iron-formation (>5 vol. %) are quartz, grunerite, biotite (includes ferri-annite at Mafefe), carbonates, magnetite and riebeckite. Minor coexisting minerals (<5 vol. %) are shown as bars in Fig. 3. Of these, chlorite, tourmaline and garnet may constitute >5 vol. % in black shale and biotite–grunerite hornfels. Fayalite, clinopyroxene and hornblende are restricted to the metamorphic aureoles of diabase sills.

View larger version (62K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 3. (a) Mineral abundances and assemblages in core PA13. Mineral abbreviations were taken from Kretz, (1983) except for Fls for feldspar, Gn for greenalite and C for carbonaceous matter (kerogen). Biotite includes both biotite and ferri-annite. (Refer to Fig. 2 for legend of profile.) (b) Mineral abundances and assemblages in core MF2. (Refer to Fig. 2 for legend of profile.)

 
The carbonates include siderite, ankerite–dolomite and calcite. Siderite is more abundant at Mafefe (core MF2) than at Penge (core PA13), but scarce or absent in amosite- and/or crocidolite-rich layers. Calcite is more abundant, whereas siderite is absent in the metamorphic aureole of the diabase sills (Fig. 3a and b). Siderite and ankerite–dolomite are less abundant and calcite is more abundant in the Penge Iron Formation than in the diagenetic to very low-grade metamorphic Kuruman Iron Formation. Quartz, carbonates and magnetite appear to be more abundant at Mafefe than at Penge, whereas grunerite, chlorite, garnet and stilpnomelane are less abundant at Mafefe than at Penge (Fig. 3a and b). Grunerite tends to be more abundant and quartz and magnetite less abundant where amosite reefs occur. Riebeckite, which is intimately associated with ferri-annite (Si/Al>3), is abundant in MF2, but very scarce in PA13 at Penge (Fig. 3).

	Bulk Chemical Compositions

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
Major element chemical analysis (Si, Al, Ti, Fe, Mn, Mg, Ca, Na, K and P) was performed with X-ray fluorescence spectrometry (Phillips PW1404). The compositional data of the standard samples were taken from Ando et al., (1987). Wet chemical methods were used to determine CO₂, H₂O⁺, H₂O^– and total organic carbon contents. Fe²⁺/Fe³⁺ ratios were determined by the permanganate titration method. The analytical conditions were described by Tsunogae & Miyano, (1992). Results from samples in core PA13 are listed in Table 1 and sample positions are shown in Fig. 2.

View this table: [in this window] [in a new window]   Table 1 Bulk compositions of various rock types in core PA13 in the Penge Iron Formation

 
The chemical composition of the samples varies, depending on the relative abundances of constituent minerals in the succession. The SiO₂ content, a major component of quartz, grunerite and other silicates, ranges from 40 to 52 wt %. Fe₂O₃ is mainly present in magnetite, biotite (ferri-annite) and riebeckite, and ranges between 1 and 16 wt %. FeO is present in almost all minerals with the exception of quartz. Al₂O₃ occurs in Al-bearing minerals such as biotite, chlorite, stilpnomelane, ferri-annite and tourmaline. Although the average Al₂O₃ content is <1.0 wt % in grunerite banded chert, it ranges from 3 to 4 wt % in magnetite–grunerite banded iron-formation, from 4 to 8 wt % in biotite–grunerite hornfels, and exceeds 9 wt % in black shale. CaO and CO₂ contents vary proportionally to the abundance of calcite and ankerite–dolomite. The amounts of K₂O and Al₂O₃ are proportional to the abundance of biotite at Penge, but the Al₂O₃ content is not necessarily proportional to biotite at Mafefe because of the presence of Al-poor ferri-annite. Na₂O contents at Penge, where riebeckite is very scarce, is less than 1.0 wt %. K₂O is usually <1.0 wt % in grunerite banded chert; it varies between 2.0 and 4.0 wt % in magnetite–grunerite banded iron formation, and between 4.0 and 6.0 wt % in biotite–grunerite hornfels and black shale. Carbonaceous matter (kerogen) is a common constituent of both the black shale and the biotite–grunerite hornfels, with the shale containing up to 3.3 wt % organic carbon (Table 1).

The bulk compositions of three samples of diabase sills are shown in Table 1 (Nos 7, 11 and 12). Analyses 11 and 12 are similar in composition to that of magnesian basalt, but olivine is very rare or absent.

	Mineralogy and Mineral Chemistry

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
Electron microprobe analyses of minerals were performed on a Cameca CAMEBAX 355 equipped with three wavelength-dispersive spectrometers. The beam diameter was between 2 and 5 µm with 15 kV current, and acceleration voltage of 30 nA (on brass). Data reduction procedures were those of Henoc & Maurice, (1979). Standards used were albite for SiO₂ and Na₂O, corundum for Al₂O₃, synthetic MnTiO₃ for MnO and TiO₂, chromite for Cr₂O₃, haematite for Fe₂O₃, synthetic nickel oxides for NiO, periclase for MgO, wollastonite for CaO and orthoclase for K₂O.

Representative analyses of minerals are listed in Tables 2–5; extensive microprobe data tables are available from the authors on request. Fe²⁺/Fe³⁺ ratios in grunerite and hornblende were taken as a mean value of the minimum and maximum values based on crystal chemical constraints (Laird & Albee, 1981; Robinson et al., 1982). The same ratio in riebeckite was calculated on the bases of 13 cations, excluding Ca, Na and K. Fe₂O₃ in ferri-annite was computed assuming tetrahedral ferric iron (Miyano & Miyano, 1982).

View this table: [in this window] [in a new window]   Table 2 Representative microprobe analyses of carbonates

 

View this table: [in this window] [in a new window]   Table 5 Representative microprobe analyses of fayalite, clinopyroxene, feldspars, garnet and tourmaline

 
Quartz
Quartz with a mosaic texture and average grain size of 30–80 µm is ubiquitous throughout the succession and coexists with almost all the minerals. Asbestiform quartz is associated with amosite and occasionally also with crocidolite fibres in some asbestos seams.

Magnetite
Magnetite in the iron-formation is very pure and contains in total <0.5 wt % of TiO₂, SiO₂, Al₂O₃, MgO and NiO. Magnetite and carbonates appear to be less abundant and grunerite more abundant at Penge than at Mafefe if similar stratigraphic zones are compared.

In samples from core PA13 original magnetite crystals are often almost totally resolved into needles and prisms of grunerite (Fig. 4a), suggesting that much of the magnetite and most of the original iron-carbonates at Penge were transformed into grunerite according to the following reactions:

View larger version (106K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 4. (a) Grunerite prisms (white) cutting magnetite (black) in magnetite–grunerite–biotite hornfels (Sample PA13/54, 37 m). C, ankerite–dolomite. Grey interstitial material is biotite. (Note minute inclusions of kerogen in the prisms.) (b) Large prisms of grunerite (Gru) cutting riebeckite (Rk, grey to dark grey) (Sample MF2/51, 186 m). The grunerite also includes relict riebeckite, quartz (Q), ferri-annite (F) and minute particles of kerogen. Black euhedral grains are magnetite. (c) Hornblende (H, dark grey) coexisting with grunerite (needle), quartz (white) and biotite (B) in the contact metamorphic aureole of the lower diabase sill (Sample PA13/116A, 122 m). Biotite is much coarser and hornblende includes relict grunerite (white spots). (d) Grunerite prisms (white) rimmed by riebeckite (dark grey) in biotite–grunerite hornfels (Sample PA13/20, 17 m). Matrix consists mainly of biotite (dark grey), spotted albite (Ab), minor amounts of magnetite and pyrite (both black). (e) Altered crocidolite (dark grey) associated with fibres of quartz (white) and grunerite (light grey) (Sample MF2/84, 422 m). Black grains and band at bottom are composed of magnetite. Large riebeckite crystals (Rk, grey) cut the fibrous grunerite. (f) Chlorite (Chl, grey) and grunerite (white) prisms in biotite–grunerite hornfels (Sample PA13/93, 97 m). Matrix consists of biotite (dark grey) and fine grunerite needles. Small black grains are pyrite, magnetite and pyrrhotite. All photomicrographs were taken under polarized light and the bar scale represents 200 µm in length.

 

or

Reduction reactions, like the above, involving magnetite, are most likely to have taken place in the presence of kerogen (Fig. 4a). In contrast, magnetite does not display any evidence for reaction at Mafefe.

Carbonate minerals
Subhedral to euhedral ankerite–dolomite and siderite occur as fine to medium (<100 µm) grains. Calcite appears to be of metamorphic origin. It is usually coarser grained than the other carbonate minerals and occurs as anhedral to euhedral crystals; often interstitial with grains of quartz, magnetite and grunerite. Siderite microbands are commonly replaced by grunerite, suggesting the following reaction to have taken place:

Ankerite–dolomite and calcite also coexist with grunerite, indicating the following reaction also to have taken place:

MnO contents range from 1 to 4 wt % in siderite, from 0.5 to 10 wt % in ankerite–dolomite and from 1 to 5 wt % in calcite (Table 2). Calcites contain between 1 and 4 wt % FeO with very low MgO contents (<0.5 wt %). Magnesian ankerite–dolomite and magnesian siderite found in very low-grade metamorphic Kuruman iron-formation (Klein & Beukes, 1989) are scarce in the metamorphosed Penge Iron Formation.

Grunerite
Grunerite occurs as fibrous to acicular rosettes, or prismatic subhedral crystals (Fig. 4a), and coexists with almost all the other minerals present in the iron-formation. Grunerite grains in biotite hornfels and magnetite–grunerite banded iron-formation commonly contain minute inclusions of kerogen (Fig. 4a). Grunerite replaces riebeckite, ferri-annite and carbonates. Idioblastic prisms of grunerite are sometimes surrounded by blue riebeckite (Fig. 4b) and replace larger grains of magnetite, pyrrhotite and pyrite.

Amosite reefs consist of asbestiform grunerite, fibrous quartz, dispersed magnetite and minor amounts of ferri-annite or Al-poor biotite, pyrite, ankerite–dolomite and stilpnomelane. Amosite and grunerite are compositionally indistinguishable (Day, 1978).

The reaction of grunerite with carbonates and biotite produces hornblende (Al₂O₃>9 wt %) in the biotite-rich zones next to diabase sills (Fig. 4c). Grunerite in metamorphic aureoles of diabase sills coexists with quartz, clinopyroxene, magnetite, fayalite, ferri-annite and calcite, and commonly contains larger amounts of Al₂O₃, CaO and Na₂O than grunerite elsewhere in the succession.

The numbers of cations of Al³⁺(IV), Al³⁺(VI), Ca²⁺, Na⁺ and Fe³⁺ in the grunerite, based on 23 oxygens, are rather constant throughout the succession (Table 3 and Fig. 5a and b). However, in core PA13 these cation contents increase immediately above the lower diabase sill, and in the uppermost part of the core. In core MF2 these cation sums in grunerite increase in the uppermost 75 m of the core. The estimated Fe³⁺ content of grunerite in the Mafefe core (MF2) is in general higher than that in the Penge core (PA13) (Fig. 5). The suggestion that grunerite in the Penge Iron Formation contains Fe³⁺ should perhaps be further investigated by wet chemical analyses because Klein, (1964, 1968) came to the conclusion that only Fe²⁺ is present in the crystal structure of grunerite.

View this table: [in this window] [in a new window]   Table 3 Representative microprobe analyses of grunerite, riebeckite and hornblende in PA13 and MF2

 

View larger version (19K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 5. (a) Contents of Al(IV), Al(VI), Ca, Fe3+ and Na of grunerite in PA13. (b) Contents of Al(IV), Al(VI), Ca, Fe3+ and Na of grunerite in MF2.

 
Riebeckite
Riebeckite is abundant at Mafefe and Malipsdrift, but scarce at Penge. This may be in part due to a stratigraphic control because the lower part of the succession, preserved at Penge, is also poor in riebeckite at Mafefe (Fig. 3). Fibrous rosette-like, prismatic and asbestiform (crocidolite) varieties of riebeckite are known. As noted above, riebeckite is often clearly cut by grunerite (Fig. 4b). However, in some samples prismatic grunerite grains are rimmed by riebeckite (Fig. 4d) whereas in others large prismatic riebeckite crystals cut across asbestiform grunerite (Fig. 4e). Considering the wide stability range of riebeckite in iron-formation metamorphosed from low to high grade (e.g. Klein, 1983), some of the riebeckite might have formed during retrogressive metamorphism, depending upon the availability of sodium (Miyano & Klein, 1983; Miyano & Beukes, 1984a). Such retrograde riebeckite is common at Mafefe and not markedly different in composition from the earlier phase of riebeckite which is replaced by grunerite.

Riebeckite coexists with ferri-annite (Si/Al>3), quartz, grunerite, magnetite, carbonates and kerogen (Figs 3 and 6). A comparison of riebeckite from the Penge Iron Formation with that of the unmetamorphosed to very low-grade metamorphic Kuruman Iron Formation indicates that Mg-rich riebeckite, which is common in the Kuruman Iron Formation, is scarce to absent in the Penge Iron Formation (Table 3 and Fig. 6). Mg-rich siderite rarely coexists with riebeckite in both the Kuruman and Penge successions (Fig. 6a). Ankerite–dolomite coexisting with riebeckite in the Penge Iron Formation is somewhat more Fe rich than is the case in the Kuruman Iron Formation (Fig. 6b). The Al₂O₃ content of riebeckite from the Penge Iron Formation is greater than that of riebeckite from the Kuruman Iron Formation but the TiO₂ content is lower (Fig. 7).

View larger version (24K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 6. (a) Compositional variations of coexisting riebeckite, ferri-annite and siderite in the Penge (MF2) and Kuruman Iron Formations. (b) Compositional variations of coexisting riebeckite and ankerite–dolomite in the Penge (MF2) and Kuruman Iron Formations.

 

View larger version (14K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 7. Ti–Al contents of riebeckite in the Penge (MF2) and Kuruman Iron Formations (n is number of analyses).

 
Biotite and ferri-annite
Biotite with Si/Al=1.6–3.0 can compositionally be distinguished from ferri-annite with Si/Al>3 (Table 4 and Fig. 8). Biotite is further subdivided into biotite I and biotite II on the basis of the total amount of Al [Al(IV)+Al(VI)], as shown by bimodal Ti–Al distributions (especially for PA13) in Fig. 9. Total Al (based on 22 oxygens) ranges from 2.5 to 3.1 in biotite I, and it is between 2.0 and 2.3 in biotite II. In ferri-annite it is <2.0. Biotites I and II respectively replace stilpnomelane and ferri-annite.

View this table: [in this window] [in a new window]   Table 4 Representative microprobe analyses of biotite, ferri-annite ( F-ann), chlorite, stilpnomelane and greenalite (Gn)

 

View larger version (18K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 8. (a) The Si/Al ratios in ferri-annite and biotite from core PA13. (Note that the ratio is >3 in ferri-annite with tetrahedral Fe3+.) (b) The Si/Al ratios in ferri-annite and biotite from core MF2.

 

View larger version (19K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 9. Ti–Al contents of mica and stilpnomelane in the Penge (PA13 and MF2) and Kuruman Iron Formations. Total Al [Al(IV)+Al(VI)]<2 indicates presence of ferri-annite, 2.5–3.1 indicates biotite I, and 2.0–2.3 indicates biotite II.

 
Biotite coexists with grunerite, quartz, carbonates, magnetite, chlorite, tourmaline, garnet and trace amounts of pyrite and pyrrhotite. The assemblage biotite–grunerite–quartz ±carbonates±magnetite is ubiquitous throughout the Penge Iron Formation whenever biotite is present. The only exception is in the case of the black shale near the base of the succession, where biotite is present and grunerite absent. In the metamorphic aureole of diabase sills, biotite coexists with abundant calcite, hornblende, grunerite and small amounts of ankerite–dolomite and tourmaline (Fig. 4c).

Ferri-annite, a trioctahedral mica with tetrahedral Fe³⁺ (Table 4), is common in very low- to low-grade riebeckite-bearing iron-formations (e.g. Miyano & Miyano, 1982; Dyar & Burns, 1986; Miyano et al., 1990). In the Penge Iron Formation ferri-annite shows light reddish brown to pale yellow–green pleochroism in contrast to the green colours of biotite in the succession.

Ferri-annite is scarce in the lower part of the succession at Mafefe and its correlative units at Penge (Fig. 8). The abundance of ferri-annite increases upwards so that the middle part of the succession at Mafefe, above the stratigraphic level of core PA13, contains only ferri-annite. However, biotite is again abundant and ferri-annite scarce in the upper part of the succession at Mafefe. The distribution of riebeckite (Fig. 3) closely corresponds to that of ferri-annite (Fig. 8) at both Penge (PA13) and Mafefe (MF2). Biotite is scarce or absent from zones rich in riebeckite, ferri-annite and/or asbestiform riebeckite.

Ferri-annite in the Penge Iron Formation coexists with quartz, riebeckite, magnetite, carbonates (mostly ankerite–dolomite and calcite), grunerite and fayalite. The Si/Al ratio in ferri-annite ranges from 3 to 6 in core PA13 (Fig. 8a), but can be as high as 15 at Mafefe (Fig. 8b). If PA13 is correlated with MF2 it is apparent that ferri-annite at Penge is characterized by lower Si/Al ratios than at Mafefe for similar stratigraphic zones (see Fig. 8a and b). The Al(IV) and Ti contents of biotite and ferri-annite in MF2 (Fig. 10) are mirror images of the Si/Al ratios (Fig. 8b) and increase markedly towards the upper part of the succession.

View larger version (13K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 10. Contents of Al(IV), Al(VI) and Ti of biotite (+ferri-annite) in MF2.

 
Chlorite
Chlorite occurs as pale green to green 14 Å ripidolite. It commonly occurs as prisms, patches and/or aggregates of fine grains in the biotite-rich rocks of PA13, where it constitutes between 3 and 14 vol. %. The ripidolite coexists with biotite, grunerite, carbonates and/or minor amounts of retrograde stilpnomelane, pyrite and pyrrhotite (Fig. 4f). It is locally transgressed by grunerite needles. The Fe/(Fe+Mg) ratio of chlorite (Table 4) ranges from 0.65 to 0.85 in the Penge area and from 0.60 to 0.70 in the Kuruman succession. The ratio is similar to that of coexisting biotite. The Ti cation concentration is <0.02 (based on 14 oxygens) in most chlorites from both the Penge and Kuruman successions. However, in PA13, the Ti content in chlorite increases to 0.08 in black shale.

Chlorite is very scarce in MF2. It should also be noted that chlorite is scarce or absent in the zones where amosite, crocidolite and/or ferri-annite are developed, regardless of high alumina contents. Similarly, chlorite is very scarce or absent in low-grade Kuruman iron-formation lithofacies that contain abundant riebeckite and ferri-annite (Miyano & Beukes, 1984b).

Garnet
Porphyloblastic manganiferous garnet (MnO content 8–20 wt %, Table 5) is dispersed as medium-grained (0.1–1.0 mm) euhedral crystals in the black shale (Fig. 11a) near the base of the Penge Iron Formation at Penge in core PA13 (Fig. 3a). The crystals are zoned with Mg- and Fe-rich rims and Mn-rich cores (Table 5). The garnet consists of 35–63 mol % almandine, 20–49 mol % spessartine, 14–21 mol % grossular and 1–3 mol % pyrope. It normally coexists with biotite and quartz in a matrix consisting of variable amounts of grunerite, chlorite, calcite, ankerite–dolomite, tourmaline and potassium-feldspar.

View larger version (74K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 11. (a) Garnet porphyroblasts in black shale (Sample PA13/129, 199 m). Matrix consists of quartz (white), biotite (grey) and small amounts of feldspar, grunerite (needle), chlorite, pyrrhotite, pyrite, ankerite–dolomite and kerogen (black). (b) Stilpnomelane laths (dark grey) and grunerite prisms (white to grey) in biotite–grunerite hornfels (Sample PA13/98, 107 m). White spots in biotite matrix are calcite and albite. All photomicrographs were taken under polarized light and the bar scale represents 200 µm.

 
At Mafefe (core MF2), garnet (MnO 7–10 wt %) is limited to the carbonaceous biotite–grunerite hornfels in the uppermost part of the Penge Iron Formation (Fig. 3b). It is poikiloblastic, medium-grained and coexists with grunerite, biotite, tourmaline, magnetite and quartz. The zoning pattern is similar to that of PA13 and the crystals are composed of 65–71 mol % almandine, 16–21 mol % spessartine, 7–11 mol % grossular and 2–4 mol % pyrope.

Olivine
Fayalitic olivine (Fa_95–97, Table 5) occurs in or near magnetite layers in carbonate-bearing chert–magnetite bands in the contact aureole of the lower diabase sill at Mafefe (MF2). However, it is absent at Penge (PA13). The olivine is fine grained (10–50 µm) and euhedral in close proximity to contacts with the sill, but becomes medium grained and subhedral to euhedral in bands 2–3 m away from the contact with the sill. The olivine coexists with grunerite, quartz, ferri-annite, clinopyroxene, calcite and ankerite–dolomite. Some of the olivine is extensively altered into aggregates of fine-grained greenalite. The fact that grunerite and fayalite are commonly in contact suggests that the following reaction took place:

Clinopyroxene
Clinopyroxene occurs sporadically in the banded iron-formation in contact aureoles of the diabase sills. Chemically, it can be classified as a hedenbergitic pyroxene (Table 5). The clinopyroxene appears to have replaced ankerite–dolomite, as fine-grained aggregates of pyroxene are pseudomorphic after carbonate rhombohedra. The clinopyroxene coexists with grunerite and it is less magnesium rich than the clinopyroxene that occurs in the diabase sills (Table 5). Fe/(Fe+Mg) ratios in the pyroxene (0.58–0.72) are twice to three times that of clinopyroxene from the diabase sills.

Hornblende
At both Penge and Mafefe hornblende (Table 3) is restricted to biotite-bearing layers next to diabase sills. Texturally, the hornblende crosscuts or surrounds earlier grunerite (Fig. 4c), carbonates and biotite. It is green to bluish green in colour. The Fe/(Fe+Mg) ratio in coexisting hornblende and grunerite appears to decrease towards the contact with sills (Fig. 12). The following reaction can be deduced from textural evidence:

View larger version (13K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 12. Compositional variation of coexisting grunerite and hornblende in PA13 and MF2. Distances from the lower sill contacts: A=0.9 m (PA13); B=3.7–5.2 m (PA13); C-1=13.1 m (MF2); C-2=44.5 m (1.1 m from the upper sill) (MF2).

 

Stilpnomelane
Stilpnomelane in cores PA13 and MF2 is typically brown to greenish brown in colour and has needle-like or ribbon-like form. The stilpnomelane retrogressively replaces biotite (Fig. 11b) and also occurs in veinlets. It is more abundant at Penge than at Mafefe (Fig. 3) and commonly coexists with biotite (but not with ferri-annite), quartz, grunerite and carbonates. The Fe/(Fe+Mg) ratio in stilpnomelane from PA13 varies from 0.8 to 0.9, a more restricted range than that seen in the low-grade metamorphic Kuruman iron-formation (0.50–0.85). The Fe/(Fe+Mg) ratio in MF2 varies between 0.65 and 0.85 (Table 4). The Ti content of stilpnomelane in samples from both Penge and Mafefe is <0.02 atoms based on 21.8125 oxygens (Miyano & Beukes, 1984b).

Greenalite
Pale green to brown greenalite aggregates represent a retrogressive metamorphic phase replacing fayalite, grunerite and clinopyroxene. Al₂O₃ is only present in trace amounts in the greenalite whereas the MnO and K₂O contents are respectively <1.5 wt % and <1.0 wt % . Calculations, balanced for structural formulae, indicate that the total number of cations is usually <10, suggesting the possible presence of ferric iron in the octahedral site (Table 4). The Fe/(Fe+Mg) ratio in greenalite ranges from 0.76 to 0.95.

Feldspar
Feldspar (Table 5) is restricted to biotite-rich mesobands. It is present in minor amounts and occurs as fine-grained (<40 µm) albite (An_0–10) and potassium-feldspar (Or_>95, Ab_<5) that locally coexist. The feldspars generally coexist with biotite, magnetite, quartz, grunerite and minor chlorite.

Tourmaline
Tourmaline occurs in biotite-bearing (Al-rich) mesobands in iron-formation next to diabase sills. It is also present in the diabase sills, and in iron-formation in the uppermost part of MF2. The tourmaline coexists with biotite, grunerite, calcite, hornblende and retrograde stilpnomelane. Larger tourmaline crystals enclose tiny grains of ankerite–dolomite, biotite and grunerite. Some crystals display zoning with a dark greenish brown core and light brown rim. This is related to the slight enrichment of iron in the rims as is illustrated by the corresponding Fe/(Fe+Mg) ratios (Table 5). Tourmaline from the sills is colourless and contains less FeO and more MgO [Fe/(Fe+Mg)=0.23–0.24] and Al₂O₃ than in the iron-formation.

Sulphide minerals
Sulphide minerals, including pyrrhotite, pyrite, sphalerite and chalcopyrite, constitute <1 vol. % of all the lithofacies. Pyrrhotite is most abundant, especially in the black shale. Pyrrhotite and pyrite are fine to medium grained and subhedral to euhedral in form. Larger grains are commonly transgressed by prismatic grunerite.

Kerogen
Kerogen (Fig. 4a) is common in iron-formation lithofacies throughout cores MF2 and PA13 (Fig. 3). However, it is absent in the contact aureoles of diabase sills. Kerogen is most abundant in the black shale of the lower part of the Penge Iron Formation where grunerite is scarce. Powder X-ray diffraction analysis indicates that the kerogen is non-graphitic and amorphous [also see Day, (1978)].

	Partition Coefficients

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
Fe–Mg partition coefficients (K_D values) for some coexisting mineral pairs in the Penge Iron Formation (PA13 and MF2) were calculated using total iron as Fe²⁺ for biotite, chlorite, stilpnomelane and tourmaline, and estimated Fe²⁺ for ferri-annite, grunerite and riebeckite (Fig. 13). The average Fe–Mg partition coefficient of grunerite–biotite is slightly higher at Penge (0.971; Fig. 13a) than at Mafefe (0.933; Fig. 13b). This partition coefficient displays a bimodal distribution at Penge, with samples between the two diabase sills showing a lower partition coefficient than those from directly above and below the sills (Fig. 13a). In the Mafefe core there are also two populations of grunerite–biotite Fe–Mg partition coefficients, with samples close to the contact with the sills displaying a lower coefficient (Fig. 13b).

View larger version (13K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 13. Partition coefficients of various coexisting mineral pairs in the Penge Iron Formation.

 
Other Fe–Mg partition coefficients from various mineral pairs are 1.545 for grunerite–riebeckite (Fig. 13c), 1.484 for ferri-annite-riebeckite (Fig. 13d), 0.934 for chlorite–biotite (Fig. 13e) and 1.813 for stilpnomelane–biotite (Fig. 13f). The partition coefficients for the last two mineral pairs appear to be similar at Penge and Mafefe. However, this is not the case for the mineral pairs grunerite–biotite and tourmaline–biotite.

With reference to Fe–Mg partition coefficients of tourmaline–biotite pairs (Fig. 13g), there appears to be a clear distinction between sample populations from Penge and Mafefe, with samples from Penge displaying a lower coefficient (0.523) than those from Mafefe (0.960). The mineral pair grunerite–riebeckite at Penge may display a lower Fe–Mg partition coefficient than at Mafefe (Fig. 13c). However, there are too few analyses available from Penge to completely substantiate this statement. The Fe–Mg partition coefficients of olivine and grunerite from metamorphic aureoles next to the diabase sills are variable and range between 3.93 and 4.43.

	Discussion

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
Metamorphic conditions
Pressure conditions during metamorphism of the Penge Iron Formation can only be estimated indirectly with reference to the overlying andalusite-bearing hornfels of the Timeball Hill Formation (Human, 1975) of the Pretoria Group. Metamorphic assemblages in the andalusite deposits consist of andalusite–garnet–staurolite–biotite, garnet–staurolite–biotite and andalusite–staurolite–biotite. The metamorphic conditions for these assemblages have been estimated at 500±50°C and 1–2 kbar by Human, (1975), and 535°C and 5.2 kbar by Sharpe & Chadwick, (1982). Kaneko & Miyano, (1990) recently estimated the P–T conditions of the andalusite zone on the farm Annesley 109 KT to be 540±30°C and 2.3±0.7 kbar. This was done on the basis of mineral assemblages from staurolite–andalusite–garnet–biotite hornfels, using the garnet–biotite geothermometer (Hodges & Spear, 1982; Ganguly & Saxena, 1984; Holdaway et al., 1988) and the garnet–plagioclase–andalusite–quartz geobarometer (Ganguly & Saxena, 1984; Koziol & Newton, 1988). The pressure for the Penge succession (PA13) at Penge is therefore at least 2.6±0.8 kbar, considering that it is situated 900–1300 m below the andalusite zone. For the purpose of calculating the temperature of metamorphism, a pressure of 2.5 kbar is thus assumed. The pressure conditions could not have been much different at Mafefe because of similarities between the successions known to have overlain the Penge Iron Formation at Penge and Mafefe (Button, 1973).

The biotite–garnet geothermometer (Hodges & Spear, 1982; Holdaway et al., 1988) applied to matrix biotite and manganiferous garnet rims from black shale near the base of the Penge Iron Formation at Penge, suggests metamorphic temperatures of 420–440°C at 2.5 kbar (Table 6). This temperature is supported by calcite–dolomite geothermometry (Table 6) applied to dolomite that occurs some 20–30 m below the Penge Iron Formation in the Malmani Subgroup. Calculated temperatures using this method range from 380°C to 420°C. These calculated temperatures (relative to that obtained from the andalusite deposits), suggest that the contact metamorphic geothermal gradient below the floor of the Bushveld Complex was of the order of 80°C±30°C/km.

View this table: [in this window] [in a new window]   Table 6 Metamorphic temperatures estimated from garnet–biotite and calcite–dolomite geothermometers in cores PA13 and MF2

 
Temperature conditions at Mafefe appear to have been slightly lower than at Penge if correlative stratigraphic units are considered. Biotite–garnet geothermometry applied to a sample 12 m from the top of the Penge Iron Formation at Mafefe gives a temperature of 420–460°C at 2.5 kbar (Table 6). This transforms to a temperature of 350–390°C in the shale at the base of the succession, if the Bushveld contact metamorphic geothermal gradient, estimated above, is applied to the Mafefe area. This temperature is 40–80°C lower than the temperature estimated from garnet–biotite assemblages in the equivalent shale at Penge, and may explain the absence of garnet from this shale at Mafefe.

The small difference in metamorphic temperatures at the two localities explains the rather similar Fe–Mg partition coefficients of chlorite–biotite, stilpnomelane–biotite and grunerite–biotite mineral pairs at Penge and Mafefe (Fig. 13). Similar partition coefficients for stilpnomelane–biotite mineral pairs suggest a similar cooling history in the two areas because in both localities stilpnomelane is a retrograde mineral.

Fe–Mg partitioning between biotite and chlorite from Penge and Mafefe displays a strong linear relationship, suggesting that the partition coefficient is insensitive to the relatively small difference in metamorphic temperature (40–80°C) between the two areas (Fig. 13e). However, the coefficients for grunerite–biotite pairs appear to decrease with increasing temperature because grunerite–biotite pairs in the metamorphic aureole of the diabase sills display lower Fe–Mg partition coefficients (Fig. 13b).

The marked difference in Fe–Mg partition coefficients for biotite–tourmaline pairs from Penge and Mafefe (Fig. 13g) suggests that this coefficient is sensitive to temperature differences of 40–80°C. According to Henry & Guidotti, (1985), the partition coefficient for tourmaline and biotite in staurolite-grade metapelites of NW Maine is 0.504, which is smaller than the value at Mafefe and Penge (Fig. 13f). This suggests that partitioning decreases with increasing metamorphic grade; in agreement with the Penge–Mafefe situation. Inclusions of ankerite–dolomite, biotite and grunerite in tourmaline suggest that tourmaline formed during late-stage metamorphism, most probably at the time of intrusion of the diabase sills. Metamorphic temperatures in the contact aureole of the diabase sills, which overprints the metamorphic assemblage related to intrusion of the Bushveld Complex, are estimated to be of the order of 570–610°C on the basis of the empirical calibration [ln K_D=2328.2/T(K) – 1.27; Miyano & Klein, 1986] of the Fe–Mg partitioning between coexisting fayalite and grunerite.

As noted above, kerogen is common in the Penge Iron Formation at both Mafefe and Penge, and may have played an important role in the reduction of ferric iron to ferrous iron in minerals such as ferri-annite, magnetite and riebeckite during metamorphism. In the presence of kerogen these minerals appear to be replaced by grunerite and/or biotite, resulting in assemblages depleted in ferri-annite and riebeckite at Penge. However, at Mafefe maximum metamorphic temperature appears to have been some 40–80°C lower than at Penge. Kerogen may therefore have been less effective as a reduction agent at Mafefe, with the result that some rock units rich in ferri-annite and riebeckite are preserved there but not at Penge.

Phase relations
The phase relations in the iron-formation could be depicted in the model system Fe–K–Fe–Mg–Al–Si–C–O–H. However, because mineralogical data for K-bearing phases such as ferri-annite and biotite as a function of Si–Al are limited, we only discuss the phase relations in the systems Fe–Mg–Si–C–O–H and Na–Fe–Mg–Si–O–H.

Phase relations in the system Fe–Mg–Si–C–O–H may be examined quantitatively on the basis of the thermodynamic datasets of Miyano & Klein, (1986), Holland & Powell, (1990) and Saxena et al., (1993), and fluid (H₂O–CO₂) data of Kerrick & Jacobs, (1981), using activity–composition relations of Miyano & Klein, (1986, 1989) for solid phases. Topology of the phase relations in Fig. 14 is consistent with that of type A (Ol-bearing and Opx-free) of Miyano & Klein, (1986), but fayalitic olivine does not occur in the iron-formation at the metamorphic conditions of cores PA13 and MF2, except along contacts with diabase sills. The metamorphic fluid composition during intrusion of the Bushveld Complex is thus determined from the coexisting assemblage grunerite and siderite at the calculated pressure and temperature [reaction (21) in Fig. 14]. Although the occurrence of this assemblage is scarce, X(CO₂) values are 0.38–0.63 at 435–455°C, some 22 m from the top of the Penge Iron Formation in core PA13, and 0.20–0.68 at 420–460°C, some 17 m from the top of the formation in core MF2, where neither riebeckite nor ferri-annite is present. In contrast, X(CO₂) values range from 0.08 to 0.49 at 400–440°C between 158 and 323 m from the top of core MF2, where both riebeckite and ferri-annite occur.

View larger version (14K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 14. T–X(CO2) diagram at 2.5 kbar, showing the upper stability fields of grunerite plus quartz and siderite plus quartz. The set of continuous curves refers to end-member reactions indicated in Table 7. The dashed curve shows maximum displacement of the curve for reaction (21) owing to impurities in grunerite and siderite. Neither minnesotaite nor fayalitic olivine is stable within the estimated T–P range of the Penge and Mafefe assemblages. Temperature ranges from which estimates of X(CO2) were made (see text) are indicated on lower right side of diagram. They are as follows: A=435–455°C at 22 m from top of Penge Formation in PA13; B=420–460°C at 17 m from top of MF2; C=400–440°C at 158–323 m from top of MF2.

 
Phases in the system Na–Fe–Mg–Si–O–H are represented by haematite, magnetite, quartz, grunerite, riebeckite and acmite. Reactions predicted from the petrographic study of the iron-formation are listed in Table 7. No internally consistent thermodynamic datasets which include both riebeckite and acmite are available. The phase relations (Fig. 15) are therefore constructed on the basis of thermodynamic properties given by Miyano & Klein, (1983) for riebeckite and acmite, and of Helgeson et al., (1981) for aqueous species and other solid phases involved. Site occupancy data of Ulbrich & Waldbaum, (1976) were used for the activity–composition relation in riebeckite. Water activity [a(H₂O)] was assumed to be unity because the concentration of dissolved ionic species in the metamorphic fluid is unknown. Thus, the phase relations in Fig. 15 are only approximate. However, the upper stability limit of riebeckite in the presence of quartz at the HM (haematite–magnetite) buffer [reaction (19), Table 7] is the same as that of the experimental data of Ernst, (1962) (Fig. 15).

View this table: [in this window] [in a new window]   Table 7 Possible reactions used in the phase analyses in the two model systems Fe–Mg–Si–O–C–H and Na–Fe–Mg–Si–O–H

 

View larger version (26K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 15. (a) Log[a(Na+)/a(H+)]–T diagram at 2.5 kbar, showing stability fields of grunerite plus quartz and water, and riebeckite plus quartz and water. It shows that grunerite coexisting with riebeckite is stable below log[a(Na+)/a(H+)]=5.0±0.1 at the calculated temperature. Location of reaction (19) at 2.5 kbar is consistent with the experimental bracket of Ernst, (1962). Dotted curve denotes a shift caused by impurities in grunerite and riebeckite in the Penge and Mafefe assemblages. (b) Log[f(O2)]–T diagram at 2.5 kbar and a(H2O)=1.0, showing stability fields of grunerite plus quartz and water, and riebeckite plus quartz and water. Stippled area illustrated at log[a(Na+)/a(H+)]=6.0. (Note other shifts with values 5.0 and 5.5.) The stippled area is enlarged with increasing log[a(Na+)/a(H+)] values. Dashed curves refer to shifts caused by impurities in grunerite and riebeckite in the Penge and Mafefe assemblages.

 
According to the phase diagrams, grunerite is unlikely to coexist with acmite as well as haematite below 600°C. Figure 15a shows that with increasing temperature grunerite is stable at higher log[a(Na⁺)/a(H⁺)] values. At constant log[a(Na⁺)/a(H⁺)] riebeckite would be transformed to grunerite with increasing temperature and vice versa, consistent with observations in the Penge Iron Formation. If log[a(Na⁺)/a(H⁺)]=5.0±0.2 is taken as the upper stability limit of grunerite (Fig. 15a), the oxygen fugacity level, log[f(O₂)], ranged from about –20 to –30 for riebeckite and grunerite to coexist (Fig. 15b).

Genesis of crocidolite and amosite
Crocidolite at Mafefe and amosite at Penge and Mafefe are known to have formed in dilatational bands in the tensional stress areas of open folds (Dreyer, 1982; Beukes & Dreyer, 1986; Dreyer & Söhnge, 1992). However, the vertical distribution of riebeckite and associated crocidolite appears to be controlled by stratigraphic setting, i.e. bulk-rock composition of the original lithofacies, and by the grade of metamorphism. Riebeckite is absent from Al-rich rock units containing biotite with Si/Al<3, as is a common feature in the upper and lower parts of the succession at Mafefe. Similarly, riebeckite is virtually absent from the Penge core, which stratigraphically corresponds to the lower part of the Mafefe core (Fig. 2). In contrast, riebeckite is abundantly present in Al-poor rock units in association with ferri-annite and biotite with Si/Al ratios of the order of three at Mafefe.

It has been suggested that amosite of the Penge Iron Formation replaced crocidolite that formed at lower grade of metamorphism (Miyano & Klein, 1983; Miyano & Beukes, 1984b). This suggestion could also explain the absence or scarcity of crocidolite at Penge and its presence at lower-grade metamorphic conditions at Mafefe. It most probably applies to the asbestos reef denoted as B in Fig. 2. This reef is composed of riebeckite at Mafefe and of amosite at Penge (Fig. 2). The transformation reaction of riebeckite into grunerite, which occurs with increasing temperature and/or lower sodium activity and oxygen fugacity, can be written as

The reverse reaction also occurs, as is demonstrated by the transformation of grunerite to riebeckite, most probably during retrogressive metamorphism and/or increased sodium activity (Fig. 15).

However, it is also clear that the distribution of crocidolite and amosite in the succession is controlled by initial rock composition. This is best illustrated by the fact that grunerite–biotite hornfels may contain abundant amosite but never any crocidolite (Fig. 2). A feasible explanation would be original differences in the bulk chemical composition of various bands in the iron-formation and the development of either grunerite or riebeckite from these bands during folding and metamorphism. Couplets of iron-silicates (greenalite, ferri-annite and/or stilpnomelane) and riebeckite are known from the Kuruman Iron Formation (Van Wyk, 1987), and may be responsible for the formation of Mafefe-type amosite–crocidolite couplets as a result of metamorphism and deformation. The amosite deposits, with fibres longer than that known from any crocidolite deposit (Du Toit, 1945), are developed in Al-rich biotite (biotite I) hornfels where crocidolite is absent.

It is likely that the amosite and biotite I developed from a stilpnomelane-rich precursor (Beukes & Dreyer, 1986) according to a reaction such as

The iron-rich end-member formulae of stilpnomelane and chlorite were respectively taken from Miyano & Beukes (1984b) and Klein & Fink, (1976). The above reaction has been quantitatively discussed by Miyano & Klein, (1989) and proceeds to the right-hand side with increasing temperature.

Some amosite may also have developed from ferri-annite, a mineral commonly associated with riebeckite and crocidolite in the unmetamorphosed Kuruman Iron Formation. A possible reaction is as follows:

It is interesting to note that this reaction would tend to produce biotite II with a total Al content lower than that of biotite I derived from stilpnomelane. Derivation of biotite from either stilpnomelane or ferri-annite during metamorphism may thus explain the bimodal distribution of Al–Ti cations in the biotite assemblages of Mafefe and Penge (Fig. 11).

Trendall & Blockley, (1970) demonstrated that crocidolite in the Hamersley Group of Western Australia grew as dilatational bands between pre-existing microbands or mesobands in the iron-formation. The same applies to the crocidolite and amosite deposits of the Kuruman and Penge iron-formations, implying mass transfer of components towards low-stress areas of folds and into the dilatational asbestos seams (Beukes, 1978, 1986). It is thus important to note that the development of crocidolite and amosite seams is not envisaged as volume for volume transformations of earlier mesobands or microbands into asbestiform bands. Rather, it is suggested that primary bands of different composition seeded growth of either crocidolite or amosite in early stages of development of dilatational bands.

	Conclusion

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
The Penge Iron Formation, situated in the contact metamorphic aureole of the Bushveld Complex, represents a metamorphosed equivalent of the diagenetic to very low-grade metamorphic Asbesheuwels Subgroup, comprising the Kuruman and Griquatown Iron Formations, in Griqualand West. Vertical variations in the chemical composition of metamorphic minerals in the iron-formation are mainly controlled by bulk composition of original lithofacies. Al and Ti contents of biotite, ferri-annite and grunerite appear to increase sympathetically with increased amounts of siliciclastic material (shale) intermixed with the iron-formation. In contrast, Fe³⁺/Fe²⁺ ratios in these minerals appear to be related to the original oxidation state of iron in specific lithofacies and to the presence or absence of kerogen.

Garnet–biotite geothermometry indicates that the temperature of metamorphism near the base of the iron-formation succession at Penge was of the order of 420–460°C at 2.5 kbar. This is some 40–80°C higher than temperatures estimated from laterally correlatable rock units in the Mafefe area some 30–35 km northwest of Penge. Fe–Mg partition coefficients of chlorite–biotite and grunerite–biotite mineral pairs appear to be insensitive and those of biotite–tourmaline mineral pairs sensitive to relatively small differences of 40–80°C in metamorphic temperatures.

Bulk-rock composition was a major factor in controlling the distribution of crocidolite and amosite asbestos seams in the Penge Iron Formation. However, there is also evidence that riebeckite–crocidolite have been replaced by grunerite–amosite with increasing grade of metamorphism from Mafefe towards Penge. In contrast, small amounts of grunerite appear to have been replaced by riebeckite during retrograde metamorphism.

	Acknowledgements

 
Funding for this project came from the CSIR, Rand Afrikaans University and the Jim and Gladys Taylor Educational Trust in South Africa, and the MESC and the University of Tsukuba in Japan. We are grateful to Trevor Tregoning for his guidance during underground visits at the Penge Asbestos Mine, and to Elsa Maritz for typing the manuscript.

---

^* Corresponding author.

	References

TOP ABSTRACT Introduction Geological Setting Lithofacies and Stratigraphy Mineral Abundances and... Bulk Chemical Compositions Mineralogy and Mineral Chemistry Partition Coefficients Discussion Conclusion References

 
Ando A., Mita N., Terashima S. Values for fifteen GSJ rock reference samples. Geostandards Newsletter (1987) 11:159–166.[Web of Science]

Anovitz L. C., Essene E. J. Phase equilibria in the system CaCO₃–MgCO₃–FeCO₃. Journal of Petrology (1987) 28:389–414.[Abstract/Free Full Text]

Beukes N. J. Die karbonaatgesteentes en ysterformasies van die Ghaap-Groep van die Transvaal-Supergroep in Noord-Kaapland. Johannesburg: Rand Afrikaans University. 580. Ph.D. Thesis.

Beukes N. J. Lithofacies and stratigraphy of the Kuruman and Griquatown Iron-formations, northern Cape Province, South Africa. Transactions of the Geological Society of South Africa (1980) 83:69–86.

Beukes N. J. Palaeoenvironmental setting of iron-formations in the depositional basin of the Transvaal Supergroup, South Africa. In: Iron-Formation: Facts and Problems—Trendall A. F., Morris R. C., eds. (1983) Amsterdam: Elsevier. 131–209.

Beukes N. J. Sedimentology of the Kuruman and Griquatown Iron-formations, Transvaal Supergroup, Griqualand West, South Africa. Precambrian Research (1984) 24:47–84.[Web of Science]

Beukes N. J. The Transvaal sequence in Griqualand West. In: Mineral Deposits of Southern Africa—Anhaeusser C. R., Maske S., eds. (1986) Johannesburg: Geological Society of South Africa. 819–828.

Beukes N. J., Dreyer C. J. B. Amosite asbestos deposits of the Penge Area. In: Mineral Deposits of Southern Africa—Anhaeusser C. R., Maske S., eds. (1986) Johannesburg: Geological Society of South Africa. 901–910.

Beukes N. J., Klein C. Geochemistry and sedimentology of a facies transition-from microbanded to granular iron-formation-in the early Proterozoic Transvaal Supergroup, South Africa. Precambrian Research (1990) 47:99–139.[Web of Science]

Bickle M. J., Powell R. Calcite–dolomite geothermometry for iron-bearing carbonates: the Glockner area of the Tauern window, Austria. Contributions to Mineralogy and Petrology (1977) 59:281–292.[Web of Science]

Button A. A regional study of the stratigraphy and development of the Transvaal basin in the Eastern and Northeastern Transvaal. (1973) Johannesburg: University of Witwatersrand. 352. Ph.D. Thesis.

Cilliers J. J. le R. Amosite at the Penge asbestos mine. In: The Geology of some Ore Deposits of Southern Africa—Haughton S. H., ed. (1964) Johannesburg: Geological Society of South Africa. 579–591.

Day P. J. Mineralogical characterization of amosite asbestos. (1978) Johannesburg: Rand Afrikaans University. 246. M.Sc. Thesis.

Dreyer C. J. B. Amphibole asbestos in South Africa. (1982) Johannesburg: Rand Afrikaans University. 301. Ph.D. Thesis.

Dreyer C. J. B., Söhnge A. P. G. The Crocidolite and Amosite Deposits of the Republics of South Africa and Bophuthatswana (1992) Pretoria: Geological Survey of South Africa. 126. Handbook 12.

Du Toit A. L. The origin of amphibole asbestos deposits of South Africa. Transactions of the Geological Society of South Africa (1945) 58:161–206.

Dyar M. D., Burns R. G. Mossbauer spectral study of ferruginous one-layer trioctahedral micas. American Mineralogist (1986) 71:955–965.[Abstract]

Ernst W. G. Synthesis, stability relations, and occurrence of riebeckite and riebeckite–arfvedsonite solid solutions. Journal of Geology (1962) 70:689–736.[Web of Science]

Ganguly J., Saxena S. K. Mixing properties of aluminosilicates garnets: constraints from natural and experimental data, and application to geothermobarometry. American Mineralogist (1984) 69:88–97.[Abstract]

Hall A. L. Asbestos in the Union of South Africa (1930) Pretoria: Geological Survey of South Africa. 324. Memoir, 12.

Hammerbeck E. C. I. Andalusite in the metamorphic aureole of the Bushveld Complex. In: Mineral Deposits of Southern Africa—Anhaeusser C. R., Maske S., eds. (1986) Johannesburg: Geological Society of South Africa. 993–1001.

Helgeson H. C., Kirkman D. H., Flowers G. C. Theoretical predictions of the thermodynamic behavior of aqueous electrolytes at high pressures and temperatures: IV. Calculations of activity coefficients, osmotic coefficients, and apparent molal and standard and relative partial molal properties to 600°C and 5 Kb. American Journal of Science (1981) 281:1249–1256.[Abstract/Free Full Text]

Henoc J., Maurice F. Practical quantitative analysis of bulk specimens. In: Microanalysis and Scanning Electron Microscopy—Maurice F., Meny L., Tixier R., eds. (1979) Orsay, France: Les Editions de Physique. 281–317.

Henry D. J., Guidotti C. V. Tourmaline as a petrogenetic indicator mineral: an example from the staurolite-grade metapelites of NW Maine. American Mineralogist (1985) 70:1–15.[Abstract]

Hodges K. V., Spear F. S. Geothermometry, geobarometry and the Al₂SiO₅ triple point at Mt. Moosilauke, New Hampshire. American Mineralogist (1982) 67:1118–1134.[Abstract]

Holdaway M. J., Dutrow B. L., Hinton R. W. Devonian and Carboniferous metamorphism in west–central Maine: the muscovite–almandine geobarometer and the staurolite problem revisited. American Mineralogist (1988) 73:20–47.[Abstract]

Holland T. J. B., Powell R. An enlarged and updated internally consistent thermodynamic dataset with uncertainties and correlations: the system K₂O–Na₂O–CaO–MgO–MnO–FeO–Fe₂O₃–Al₂O₃–SiO₂–TiO₂–C–H₂–O₂. Journal of Metamorphic Geology (1990) 8:89–124.[Web of Science]

Human D. R. The geology and metamorphic petrology of part of the basal argillaceous zone, Daspoort Stage, Pretoria Series on the farm Havercroft, North Eastern Transvaal. Petros (1975) 6:25–43.

Human D. R., Collins L. A. The Havercroft–Streatham andalusite deposit, Eastern Transvaal. In: Mineral Deposits of Southern Africa—Anhaeusser C. R., Maske S., eds. (1986) Johannesburg: Geological Society of South Africa. 1005–1008.

Kaneko Y., Miyano T. Contact metamorphism by the Bushveld Complex in the northeastern Transvaal, South Africa (in Japanese, with English abstract). Journal of Mineralogy, Petrology, and Economic Geology (1990) 85:66–81.

Keep F. E. Amphibole asbestos in the Union of South Africa. In: Transactions of 7th Commonwealth Mining and Metallurgy Congress 1 (1962) Johannesburg: 1961, South African Institute of Mining and Metallurgy. 93–120.

Kerrick D. M., Jacobs G. K. A modified Redlich–Kwong equation for H₂O, CO₂, H₂O–CO₂ mixtures at elevated pressures and temperatures. American Journal of Science (1981) 281:735–767.[Abstract/Free Full Text]

Kirkman H. L. Some notes on crocidolite and amosite occurrences in the Union. Transactions of the Geological Society of South Africa (1930) 33:13–18.

Klein C. Cummingtonite–grunerite series: a chemical, optical and X-ray study. American Mineralogist (1964) 49:963–982.[Web of Science]

Klein C. Coexisting amphiboles. Journal of Petrology (1968) 9:281–330.[Abstract/Free Full Text]

Klein C. Diagenesis and metamorphism of Precambrian banded iron-formations. In: Iron-formations: Facts and Problems—Trendall A. F., Morris R. C., eds. (1983) Amsterdam: Elsevier. 417–469.

Klein C., Beukes N. J. Geochemistry and sedimentology of a facies transition from limestone to iron-formation deposition in the Early Proterozoic Transvaal Supergroup, South Africa. Economic Geology (1989) 84:1733–1774.[Abstract/Free Full Text]

Klein C., Fink R. P. Petrology of the Sokoman Iron Formation in the Howells River area, at the western edge of the Labrador Trough. Economic Geology (1976) 71:453–488.[Abstract/Free Full Text]

Koziol A. M., Newton R. C. Redetermination of the anorthosite breakdown reaction and improvement of the plagioclase–garnet–Al₂SiO₅–quartz geobarometer. American Mineralogist (1988) 73:216–223.[Abstract]

Kretz P. Symbols for rock-forming minerals. American Mineralogist (1983) 68:277–279.[Abstract]

Laird J., Albee A. L. High-pressure metamorphism in mafic schist from northern Vermont. American Journal of Science (1981) 281:97–126.[Abstract/Free Full Text]

Miyano T., Beukes N. J. Mineralogy and petrology of the amphibole asbestos-bearing rocks of the Penge Iron Formation, Transvaal Supergroup. In: Abstracts, Geocongress 84, Geological Society of South Africa, Potchefstroom (1984a) Potchefstroom: Geological Society of South Africa. 109–113.

Miyano T., Beukes N. J. Phase relations of stilpnomelane, ferri-annite, and riebeckite in very low-grade metamorphosed iron-formations. Transactions of the Geological Society of South Africa (1984b) 87:111–124.

Miyano T., Klein C. Evaluation of the stability relations of amphibole asbestos in metamorphosed iron-formations. Mining Geology in Japan (1983) 33:213–222.

Miyano T., Klein C. Fluid behavior and phase relations in the system Fe–Mg–Si–C–O–H: application to high grade metamorphism of iron-formations. American Journal of Science (1986) 286:540–575.[Abstract/Free Full Text]

Miyano T., Klein C. Phase equilibria in the system K₂O–FeO–MgO–Al₂O₃–SiO₂–H₂O–CO₂ and the stability limit of stilpnomelane in metamorphosed Precambrian iron-formations. Contributions to Mineralogy and Petrology (1989) 102:478–491.[Web of Science]

Miyano T., Miyano S. Ferri-annite from the Dales Gorge Member iron-formation, Wittenoom area, Western Australia. American Mineralogist (1982) 67:1179–1195.[Abstract]

Miyano T., Beukes N. J., van Reenen D. D. Metamorphic evidence for early post-Bushveld sills in the Penge Iron Formation, Transvaal Sequence, Eastern Transvaal. South African Journal of Geology (1987) 90:37–43.[Abstract]

Miyano T., Miyano S., Beukes N. J. Behavior of alkali ions in low-grade metamorphosed Precambrian iron-formations. 15th International Mineralogical Association Meeting, Beijing, China, Abstract Volume (1990) 881–882.

Peacock M. A. The nature and origin of the amphibole asbestos of South Africa. American Mineralogist (1928) 13:241–285.[Web of Science]

Powell R., Condliffe D. M., Condliffe E. Calcite–dolomite geothermometry in the system CaCO₃–MgCO₃–FeCO₃: an experimental study. Journal of Metamorphic Geology (1984) 2:33–42.[Web of Science]

Reinecke L., McClure L. Variations in the quality of amosite asbestos at Penge, Transvaal. Transactions of the Geological Society of South Africa (1934) 36:29–39.

Rice J. M. Progressive metamorphism of impure dolomitic limestone in the Marysville aureole, Montana. American Journal of Science (1977) 277:1–24.[Abstract/Free Full Text]

Robinson P., Spear F. S., Schumacher J. C., Laird J., Klein C., Evans B. W., Doolan L. B. Phase relations of metamorphic amphiboles: natural occurrence and theory. In: Amphiboles: Petrology and Experimental Phase Relations. Mineralogical Society of America, Reviews in Mineralogy—Veblen D. R., Ribbe P. H., eds. (1982) 9B:1–227.

Saxena S. K., Chatterjee N., Fei Y., Shen G. Thermodynamic Data on Oxides and Silicates: an Assessed Data Set Based on Thermochemistry and High Pressure and Phase Equilibrium (1993) New York: Springer-Verlag. 428.

Schwellnus J. S. I., Engelbrecht L. N. J., Coertze F. J., Russell H. D., Malherbe S. J., van Rooyen D. P., Cooke R. The Geology of the Olifants River Area, Transvaal (1962) Pretoria: Geological Survey of South Africa. 85. [Explanation of Sheets 2429B (Chuniespoort) and 2430A (Wolkberg).].

Sharpe M. R., Chadwick B. Structures in Transvaal Sequence rocks within and adjacent to the eastern Bushveld Complex. Transactions of the Geological Society of South Africa (1982) 85:29–41.

Tsunogae T., Miyano T. Major element characteristics of high-grade quartzofeldspathic gneisses in the Northern Marginal Zone of the Limpopo Belt, Zimbabwe. Annual Report, Institute of Geoscience, University of Tsukuba (1992) 18:102–105.

Trendall A. F., Blockley J. G. The iron-formations of the Hamersley Group, Western Australia, with special reference to the associated crocidolite. Bulletin of the Geological Survey of Western Australia (1970) 119:353.

Trendall A. F., Compston W., Williams I. S., Armstrong R. A., Arndt N. T., McNaughton N. J., Nelson D. R., Barley M. E., Beukes N. J., de Laeter J. R., Retief E. A., Torne A. M. Precise zircon U–Pb chronological comparison of the volcano-sedimentary sequences of the Kaapvaal and Pilbara Cratons between about 3.1 and 2.4 Ga. In: Abstracts Volume, Third International Archaean Symposium, Perth, 1990 (1990) Perth: Geoconferences Ltd. 81–83.

Ulbrich H. H., Waldbaum D. R. Structural and other contributions to the third-law entropies of silicates. Geochimica et Cosmochimica Acta (1976) 40:1–24.[Web of Science]

Van Wyk C. J. Die mineralogie en geochemie van sedimentêre siklusse in die Kuruman- en Griquatown-Ysterformasies van die Transvaal Supergroep in Griekwaland-Wes. (1987) Rand Afrikaans University: Johannesburg. 230. M.Sc. Thesis.

Vermaas F. H. S. The amphibole asbestos of South Africa. Transactions of the Geological Society of South Africa (1952) 55:199–229.

Walraven F., Armstrong R. A., Kruger F. J. A chrono-stratigraphic framework for the north–central Kaapvaal craton, the Bushveld Complex and the Vredefort structure. Tectonophysics (1990) 171:23–48.[Web of Science]

CiteULike    Connotea    Del.icio.us    What's this?

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Miyano, T. Articles by Beukes, N. J. Search for Related Content Social Bookmarking          What's this?

Online ISSN 1460-2415 - Print ISSN 0022-3530

Copyright © 2010 Oxford University Press

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics