Canadian Mineralogist VoI. 17, pp.337-349 (1979)

THEASSOCIATION OF SOMENICKEL SULFIDE DEPOSITS WITH KOMATIITICVOTCANISM IN RHODESIA

D.A.C. WILIAMS* lohannesburg Consolidated Investment Company Limited, P.O. Box 976, Randfontein 1760, South Afrir.a

ABSTRAcT ba, d6crits ici en d6tail; nous pr6sentons aussi les faits saillants concernant cinq autres de ces gise- During the past ten years a number of small ments. Un relev6 r6gional et une r6-annotation des sulfide deposits have been discovered in cartes existantes couvrant 230 000 kmz du socle the Archean greenstone belts of the Rhodesian Rhod6sien montrent clairement que la min6ralisa- Midlands. These include the Damba deposits, de- tion est associ6e au volcanisme komatiitique, plus scribed here in some detail, and five other occur- r6pandu dans le groupe Bulawayen que ne l'indi- rences whose principal features are summarized. quent les cartes. Au moins quatre s6quenceskoma- A regional investigation and a reannotation of tiitiques possbdent la m6me stratigraphie simple et existing maps covering 230,000 km2 of the Rhode- r6gulitsre (komatiitg p6ridotitique- pyrox6- sian basement show that the deposits are clearly nitique-komatiite basaltique), dont les roches basal- assbciatedwith komatiitic volcanism and that koma- tiques constituent la majeure partie dans chacune tiitic volcanic rocks are more widespread in the des s6quences. Le magmatisme, dans le groupe Bulawayan Group than the maps suggest. At least Bulawayen, a 6t6 thol6itique et komatiitigue, tant6t four of the komatiite sequences have the same simultan6ment, tant6t alternativement. Tous les simple and regular stratigraphy of peridotitic koma- gisements se trouvent prbs du contact des groupes tiite - pyroxenitic komatiite - basaltic komatiite, Sebakwien et Bulawayen, et forment partie int6- with the basaltic varieties making up the major grale de la toute premiere unit6 volcanique du part of individual successions. Komatiitic and groupe Bulawayen. Les gisements Damba sont dans tholeiitic magmatism was both concomitant and une extrusion p6ridotitique; dans un cas, la min6ra- overlapping within the Bulawayan Group. All the lisation semble localis6e au-dessus d'un conduit nickel deposits are found at, or close to, the contact d'alimentation volcanique. Plusieurs des gisements between the Sebakw.ian and Bulawayan Groups and sont intimement recoupds par des rntrusions de are an integral part of the very first volcanic unit p6ridotite qui repr6senteraient des necks ou venues in the Bulawayan Group. The Damba deposits are volcaniques; ces gisements sont situ6s en .des within an extrusive peridotite unit and include one centres d'une activit6 volcanique ant6rieure et instance of mineralization overlying what may have post6rieure i la mise en place du minerai. Tant been a feeder conduit. Several of the other deposits par son emplacement que par sa forme, la s6quence occur in, or are closely associated with, cross- komatiitique du gisement Shangani indioue un cutting peridotite bodies that are interpreted as volcanisme de type fissural. volcanic necks or vents; the deposits seem to lie (Traduit par in sites that have acted as volcanic centres both la R6daction) prior to and following peridotite-ore emplacement. Both the detailed setting and configuration of the INTRoDUcTIoN komatiitic sequen@ at the Shangani deposit suggest that volcanism was of a fissure type. During the late 1960s and early 1970s a number of nickel sulfide depoeits were dis- SoMMAIRE covered in the Rhodesian Midlands (Fig. 1). Although by world standards tley are small Au cours de la dernidre d6cennie,plusieurs petits and of low grade, two of the discoveries are gisementsnickelifbres ont 6t6 d6couverts dans les currently in production at Shangani (Philpott ceintures de roches vertes arch6ennesen Rhod6sie 1975, Viljoen et al. L976) and Epoch. The centrale. Parmi ceux-ci figurent les gisementsDam- others include Damba, Hunters Road (Brand 1976) and Selukwe which are either subeconomic or await a more favorable economic climate for EPresent address: Exploration Department, The development. Several other minor or untested Broken Hi.ll Profrietary Co. Ltd., c.P.O. Box occurrences have also been located. L923, Pertb, Western Australia, 6001. As part of a broad-scale evaluation of the 337 338 THE CANADIAN MINERALOGIST

ll-l posr BAsEMENTRocKs [.-TIlvounor. MEraMoRPHrczoNEs l--l .RANrrEs "*r,"t"t, [l on..nsro*r ".rrs AAEA DIECUSSED

olo@@t6

FIo. l. Locality map, adapted from Bliss & Stidolph (1969). base-metalpotential of the Rhodesian basement GSNnRALGsor,ocv complex, a compilation study and reannotation of existing maps was implemented over some The Rhodesian Archean granite-greenstone 230,000 km' of this granite-greenstoneterrain. terrain provides some of the best information The basic data were derived from published and available anywhere on the relationships between unpublished maps of the Rhodesia Geological granite and greenstoneand between the various Survey at scales of between 1:50,000 and volcanic rock types. This is due to compre- 1:1,000,000.Along with company information, hensive, good-quality Survey mapping, conti- this was compiled at a scale of 1:250,000 and nuity of exposure and relatively low grades of augmented by four months field work. The meta.morphism. Bliss & Stidolph (1969) and Survey maps have a spread of publication dates Wilson (1973) have provided recent summaries from 1920 to 7975, resulting in major terrni- of the geology and possible development of the nological differences between the sheets. The Rhodesian craton which update Macgregot's field work was thus directed at resolving these (1951) classic account. The official Rhodesian differences and at examination of lithologies of geological map (1:1,000,'QQQ)has also been exploration interest with the aim of producing recently revised (7th edition, 1977), a unified geological map that reflects currenf Macgregor (1951) divided the greenstone knowledge. belt successionin Rhodesia into three "systems" Part of this work, coupled with a detailed (now called Groups): the Sebakwian Group at examinationof the Damba nickel deposits,has the base, overlain in turn by the Bulawayan and shown a rather striking associationof the nickel Shamvaian Groups. This subdivision is still re- deposits with komatiitic volcanism in the Mid- tained though the extent of the Sebakwian and lands area. Some of these results are presented Shamvaian Groups is less than as originally here; a more detailed discussionof the geologi- defined: cal and geochemical results is to be given else- Shamvaian Group: Poorly sorted sediments where (Williams in prep.). Note that the (grits, greywackes, arkoses and conglome- definitions of peridotitic, pyroxenitic and basal- rates with some felsic volcanic rocks. Not tic komatiite used here are those ot Arndt et aI. discussed further in this paper. (1977); see also discussion of this terrninology Bulawayan Group: Dominantly komatiitic, by Williams & Furnell (1979). Distinction be- basaltic and andesitic volcanic rocks with tween basaltic komatiite and tholeiite is based associated sedimentary units (arkose, grit, largely on texture, MgO content and field asso- phyllite, conglomerate and banded iron-for- ciation with spinifex-textured ultramafic koma- mation). tiites. Sebakwian Group: Mafic and felsic volcanic NICKEL SULFIDE DEPOSITS IN RHODESIA

002030.050 4

I -r'.I ^l \4 \r .{! i".-

BULAWAYO

Olg dlretln Frult Arenstona bctt GUrElO4 Nickat dapotit Tovn E Gr.lo sANo [f *^"oo. '(ALAHAR.ANo "owi ffi sxeuvereN ciouP (sEolMENT)

FELSTC tNfnUStVE uAFtc ,NfnustvE AHDESIIE LAUA/\UFF SEDIUENf THOLEII|'C EASAL7 XOUAT"'IC DASALT APO EilNTITE SETPENTTMTE(HOSILV'NIEUg'VEI

W|aesxwnN crou? crenrr:. rolllrrE oialr LOWII CWAXOA [-l

Ftc. 2. Simplified geology of part of the compilation area showing nickel deposit localities. Data sources are given in tle text. 340 THE CANADIAN MINERALOGIST

rocks, quartz-sericite schists, micaceous but it is usually characterized by quartzofeld- quartzites, phyllite and banded iron-forma- spathic sedimentary units, quartz-mica schists, tion with ultramafic intrusive rocks (of Bula- phyllites and quartzites, many of which are of wayan or later age) and some extrusive equi- volcaniclastic origin, together with banded iron- valents. formations, a wide variety of mafic schists and Bliss & Stidolph (7969, p. 309) noted that the some ultramafic lavas. The Sebakwian Group existence of the Sebakwian as a separate group is intruded by ubiquitous layered sills consisting is based on the recognition of an unconformity dominantly of , with minor meta- between it and the overlying Bulawayan Group pyroxenite and ,metagabbro (Macgregor 1937, together with a marked change in structure and Amm 1946, Tyndale-Biscoe 1949, Harrison metamorphic grade acrossthe contact. The type- 1969, Viljoen et al. 1976). area unconformity has been questioned by a In the western part of the Gwelo belt (35 number of workers (Harrison 1968, Bliss & km northwest of Gwelo: Fig. 2), Macgregor Stidolph 1969, Wilson L973) and, as a result, (1937) recognized a NE-SW axis of folding in many of the areas previously shown as being the Sebakwian Group that is not shown in the Sebakwian have been included within the Bula- overlying Bulawayan Group. However, in the wayan Group on the latest Rhodesia geological Shangani and Lonely Mine belts there is no map. The term Sebakwian will be used here evidence of any ,major time break despite the in the sense of Bliss & Stidolph (1969) but change from felsic to mafic volcanism. without the implication of an unconformity Bulawayan Group between it and the Bulawayan Group (Williarns in prep.). This contact marks an important The Bulawayan Group is generally regarded change in volcanism irrespective of whether a as consisting largely of basaltic and andesitic major time break is involved; to this author, volcanic rocks with locally abundant sedimentary the placing of Sebakwian-type sequencesinto units and banded iron formation and relatively the Bulawayan Group seemsunfortunate. minor acid volcanic rocks (Macgregor 1951, The geology of the area of interest is shown Bliss & Stidolph 1969). A number of authors, in Figure 2. This is only part of the geological including Macgregor (1928), Keep (1929) and compilation and has been greatly simplified for Phaup (1933), have describedultramafic rocks reproduction at this scale. The geology is with basaltic textures from within the Bulawayan largely a reinterpretation of the work of Mac- Group, but .few attempts were made to dis- gregor (1928, 1937), Keep (1929), Phaup tinguish these rocks on their maps. The main (L932, 1933), Ferguson (1934), Macgregor change to the existing published maps made in et al. (1937), Amm (1940, L946), Tyndale- compiling Figure 2 is the inclusion of these Biscoe (1,940, 1949), Worst (1956), Stowe rocks as a separate group of komatiitic basalt (1968) and Harrison (1969, 1970). The Shan- and peridotite. It is stressed that their distri gani granitic terrain forms the dominant struc- bution, as shown in Figure 2, is largely a result tural feature in the area and the greenstone of road traversing reconnaissanceand a reas- remnants dip outward from it at 40-80o. As sessmentof information in the various Survey many authors have observed (e.9., Macgregor bulletins. As a result the boundarips shown are 1951), the greenstone belts themselves are approximate and would need much further field dominantly synclinal with broadly north- work to be defined accurately. south-trending fold axes in the Lonely Mine, As can be seen in Figure 2, the Shangani, Filabusi, Belingwe and Gwelo belts are a major constituent of the Bulawayan Group. and east-westtrends in the Gwanda and Lower They also occur in the Sebakwian Group: a Gwanda belts. good example (shown as serpentinite in Fig. 2) occurs along the eastern edge of the Shangani Sebakwian Group belt (Viljoen et aI. 1.976). In the Bulawayan The oldest greenstone belt material in the Group itself the komatiites occur in remarkably area belongs to the Sebakwian Group and is consistent and simple sequenceslike that de- best developedin the Shangani,Lonely Mine, scribed by Viljoen et al. (L976). The sequence Gwelo and Selukwe belts where it rims the at Damba seemstypical of this volcanic asso- northern edge of the Shangani granitic terrain; ciation and will be describedin detail later. seeBliss & Stidolph (1969) and Wilson (1973) The komatiites give way to, and interfinger for discussions of pre-Sebakwian greenstone with, relatively monotonoussuccessions of pre- material. Generally the Sebakwian Group con- dominantly pillowed tholeiitic basalts,which in sists of both sedimentarv and volcanic rocks. places also rest on the underlying Sebakwian NICKEL SULFIDE DEPOSTTS IN RHODESIA

Group and were apparently extruded concomi- tantly with the komatiites. Andesite lavas and widespreadagglomerates (Macgregor 1937, Mac- gxegor et al. 1937, Harrison 1970) commonly overlie either tholeiitic or komatiitic basalt in 0 lkm the western half of the map area, but in the vv easternhalf (Gwelo and Belingwe belts) thinner but other,wise similar andesites underlie the typical basaltic volcanic rocks of the Bulawayan Group. Sedimentaryunits within the Bulawayan Group are locally abundant (Belingwe, Gwelo, Filabusi and Bulawayo belts) and include banded iron-formations, phyllites, limestones (Bickle et aI. L975) as well as greywackes,grits and conglomerates (Macgregor 1937, Amm 1940).

Tns Nrcrnr, DnPosrrs

Although the nickel deposits so far located {? in the map area (Fig. 2) differ from one another, <4 they have associational characteristics that in- dicate a genetic link with the evolution of I Bulawayan Group volcanic rocks. The most L striking feature is the restriction of almost all the deposits to a stratigraphic interval of less 1 than 500 m, straddling the contact between the /t I Sebakwian and Bulawayan Groups. It seems that the mineralization is an intrinsic part of the first volcanic activity that initiated the Bulawayan Group. As will be shown in the following descriptions of some of the deposits, f'ffi, the associatedvolcanism is komatiitic. ,ffi: Damba The Damba nickel deposits provide a clear (*::.Y+++++ example of the relationship between mineraliza- ******! tion and volcanism; they will be dessribed in some detail although this is a preliminary report BULAWAYANGROUP based entirely on drilling information. Situated BASALTIC KOMATIITE 75 km north-northeast of Bulawayo in the MAFIC TUFF/AGGLOMERATE Lonely Mine belt (Fig. 2), the deposits are PYROXENITICKOMATIITE spread oyer a strike length of 7 km. The first PERIDOTITE of the deposits was discovered in 1968 below SEBAKWIANOROUP a showing of nickel-rich silicates and gaspeite, and subsequentexploration to 1975 had located OUARTZO-FELDSPATHICSEDIMENT five separate deposits. BANDED IRON FORMATION The regional geology has been outlined by INTRUSIVES Macgregor (1928) but surface'mapping in the LAYEREDCOMPLEX ( pcrldotlte/gobbro) area is hampered by poor exposure; the strati- GRANITE graphy in areas of that has been established a Nlckel mlnerolizqilon mineralization (Fig. 3) is based largely on Old gold mlnc drifling results. The sequence has a general x north-south strike and a regional westerly dip of 45-50o. The Sebakwian Group here consists Frc. 3. Damba surface geology. Map, in part' from of various amphibolites, hornblende and quartz- work by D.M. Anderson, MJ. and R.P. Viljoen. 342 THE CANADIAN MINERALOGIST mica schists,banded iron-formations and nearer to original rock type, the are the top, well-preserved,apparently volcanogenic, generally more common in thicker parts of the quartzofeldspathic sedimentary rocks and inter- peridotite and give way to talc-carbonate rocks layered pyritic carbonaceous shales. The se- in thinner zones. Primary textures are preserved quence is intruded by several large peridotite* in some of the less intensely altered rocks and pyroxenite-gabbro layered complexes and by a show completely serpentiuized subhedral to number of strikeaersistent multiple dolerite sills. euhedral olivine set in a matrix of subhedral to The Bulawayan Group appears to lie con- skeletal uralitized or serpentinizedclinopyroxene, formably on the quartzofeldspathic sedimentary minor chromite and a densely opaque-dusted rocks and minor cherts that terminate the groundmass of chlorite or serpentine. With in- Sebakwian Group. In the zone that includes tho creasedserpentinization the matrix is also com- nickel deposits (Figs. Z, 3), the Bulawayan pletely serpentinized. No systematic variation Group volcanism was komatiitic and the suc- in olivine grain-size could be detected; the cession can be divided into three units: Unit average size is 1.G-1.5 mm with some scattered 1 (at the base): ore-zone peridotite (G-200 m larger, lath-like crystals ranging from 3-10 mm thick); Unit 2: olivine-bearing pyroxenitic in length. kcrmatiite (0-300 m thick); Unit 3: basaltic Spinifex texture in a number of forms, is komatiite (up to 7 km thick). well developedin specific zonesin the peridotite. In most areas,unit 1, the ore-zone peridotite, Its rnain occurrenqe is at the top of the perido- rests directly on the underlying Sebakwian tite unit and it seems better developed in this Group, but in some parts a thin wedge (< 4O m position, where the peridotite is thicker (Fig. thick) of basaltic komatiite separates perido- 4). Internal spinifex units also occur and the tite from sedimentary units. Mapping and drill- maximum number intersected in any drillhole ing have shown that although the peridotite is four. The types represented include typical thickens and thins irregularly it is continuous platy varieties and examples with randomly over a strike of some 9 km. oriented, long (10 mm) skeletal crystals, witb The principal rock types that now make up both types showing widely varying grain sizes. the peridotite are serpentinites, carbonated ser- Most of the spinifex zones have arnygdales pentinites and talc--carbonate rocks. Although ranging in size from 3-10 mm, and filled with their distribution is mostlv erratic and unrelated chlorite, carbonate or sulfide.

SECTIONlNoEx (-l oaMBA-]J r, \\ Plan Foj$ion A.\ ":,:::::;."ii""ffi

BUIAWAYAN GROUP lNrRuslvEs ffi"en,oorrre ,_ Sl mrrc ruFrlrecLoMERArE [. ^I DOLERITE liESrrNERALrzafloN = evnorrnrc rcror,'. ffi SEBAKWIANGR()IJP liil QUAmz-FELDSnR rcRPHYRY mxr:r-rurunED peRrmnrE leo,4o ffi W;;*- 9 4o

Ftc. 4. Da^mbaand Silwane strike and dip sections. NICKEL SULFIDE DEPOSITS IN RHODESIA 343

Quite convincing flow-top breccias and pos- predominate. The deposits are in many ways sible ultramafic tuffs can be recognizedin several similar to the Otter Shoot at Kambalda (Keele drillholes, always associatedwith spinifex units. & Nickel 1974). The breccias seem to be more com'mon in the The distribution of mineralization in the flows with strongly amygdaloidal spinifex zones. Damba-Silwane area is shown in the strike and Flow-top breccias, widespread amygdaloidal dip sectionsin Figure 4. At Silwane, mineraliza- structures, repeated spinifex units and the tex- tion occurs as a well-defined concentration in tural and compositionalsimilarity between the the lower part of the peridotite and appears olivine-poor zones within the peridotite and confined to a basinJike depression in the olivine-rich basalt flows in the hanging wall footwall about 500 m in diameter. The mine- seem to indicate that the peridotite is in large ralization itself is a broadly discus-shapedbody part extrusive. up to 75 m thick at the centre but with thick- Unit 2 overlies the peridotite and is made up ness, grade and predictability falling off rapidly of olivine-bearingpyroxenitic and 'oglassy"koma- towards the margins. It is associated with a tiitic lavas. Individual flows, which are not pronounsed thickening of the host peridotite pillowed, are up to 8 m thick; thin tuff units and thus with serpentinites rather than talc- and cracked and broken flow tops can be seen carbonate rocks. The mineralization is typically marking the tops of many of the flows. Alfhough patchy, with interlayered mineralized and bar- now converted to tremolite, talc and chlorite, ren peridotite quite common, particularly on primary textures are generally well preserved the hanging-wall side of the orebody. Althougb and indicate that the basalts were olivine-phyric mineralization can usually be correlated from with a range in olivine content from l-3OVo. one drillhole to another, the zones in adjacent Drilling has intersected two areas of agglome- drillholes are commonly not at the same strati- rate in this unit; the agglomerate consists of graphic level within the peridotite. Below the irregular komatiitic basalt fragments, ranging centre part of the footwall depressiontwo drill- in size from 2-100 mm, set in a matrix of holes have intersected a 20-3O m thick, well- originally glassy welded tuff with shardlike rnineralized peridotite body some 20 m into the material and lapilli tuff. footwall sedimentary rocks (Fig. 4, section Unil 3 is made up of basaltic komatiite EF). This small peridotite body is interpreted (and some tholeiitic basalt) that is generally as a pipe-like feeder to the overlying peridotite. olivine-free and markedly lower in MgO contenl The principal 'mineralization at Damba (Fig. than the pyroxenitic basalts of Unit 2 (Williams 4, sections AB and GH) is located toward the in prep.), The basaltic komatiites are generally centre part of the peridotite. In dip section (GH pillowed and were predominantly clinopyroxene in Fig. 4) this central ,mineralization, typically and glass-rich basalts, ,many with appreciable of the bleb type, thins rapidly with depth, plagioclase. Some unpillowed examples show giving way to footwall disseminatedmineraliza- spectacular clinopyroxene-derived spinifex tex- tion which is confined to another, but much ture. Although it overlies the other two units smaller, depression in the footwall. where they are developed, basalt of Unit 3 off- The disseminated sulfide is a typical matrix- laps the underlying units along strike and even- type mineralization in that most of the sulfide tually rests directly on the Sebakwian Group. occurs in the interstices between the completely Similar basaltsare found. at the strike extremities serpentinized olivine crystals.' The small-scale of the peridotite, as wedgesbetween the perido- distribution of this type of mineralization is tite and the underlying Sebakwian Group. characteristically patchy and it most commonly Further along strike, basaltic komatiites of Unh occurs as irregular ovoid patches (2-5 cm in 3 interfinger with, and give way to typical dia,meter) of matrix sul.fide surrounded by bar- tholeiitic basalts that also rest directly on the ren peridotite. Only rarely does the sulfide com- Sebakwian Group. pletely fill the interstices between the olivines. Although five areas of mineralization have In the bleb- or'globule-style of rnineralization been located (from the south northwards, Damba individual blebs vary in diameter from 2-20 mm Sonth, Silwane, Damba, Flow and Fibre: Fig. and on average contain about SOVosulfide. In 3), Silwane and Da.mba are the only two in- form they range from small, welldefined vestigated in sufficient detail to outline ore spheres to larger oblate spheroids, becoming zones. A variety of types of mineralization can less sharply defined with increasing size. They be recognized, including several disseminated are associatedwith what appear to be chlorite- or types, blebs, schlieren and massive or vein sul- carbonate-filled amygdales of similar size and fide, but the disseminated and bleb varieties shape. In some examples the bottom half is 344 THE CANADIAN MINERALOGIST filled with sulfide and the top half with car- is a good correlation of the -magnetite bonate or chlorite. Similar structures in the type with a serpentinite host rock and of the Otter Shoot at Kambalda have been interpreted type regardlessof host rock. Sulfides as immiscible sulfide droplets (Keele & Nickel contacts between serpentinite and talc-carbonate 1974, Keele 1,975) or as amygdale fillings units, as well as between the respective ore (Willett 1975). The Damba examples require types, are generally abrupt. In contrast, the bleb further study but are here regarded as the or globule type of ore is almost invariably of the sulfur-rich phase associatedwith primary vesi- pentlandite type regardlessof host rock. Sulfides culation of the ultramafic liquid. in the footwall sediments appear to be of the Sul.fide veins up to 3 cm wide also occur pentlandite type. sporadically in the peridotite and are invariably A further generalization can be made in that rich in magnetite. Massive sulfides are seldom disseminated sulfide in talc-carbonate rocks is found in significant amounts but thin zones texturally more variable and showsclear evidence (up to 20 cm in thickness) do occur at the of at least local remobilization. It is also of footwall contact in some drillholes. It is quite lower and more erratic grade than disseminated possible that some of these contact zones are orq in an adjacent serpentinite. There is also thick veins rather than actual .massive sulfide a preferential association of sulfide and car- accumulations. Other types of mineralization bonate in the talc-carbonate rocks, and c,hlorite include a wide range of schlieren styles that rims on the sulfides are common. seem to result from local remobilization of Although talc-carbonate alteration seems to disseminated or bleb material during shearing have produced a more sulfur-rich (or nickel- associatedwith talc-carbonate alteration. poor) sulfide assemblagefrom the finely dis- Weak mineralization, as irregular blebs, seminated ore in the serpentinites, it does not patches or veins, also extends into the footwall seem to have had any effect on the sulfide sedimentary rocks in some places, particularly blebs. This may be a result of the widely differ- at Flow and to a lesser extent at Silwane. This ing surface areas of the sulfide asse.mblagesin mineralization does not penetrate further than the two types. Becauseof pervasiveserpentiniza- about 10 m from the peridotite and is most tion it has not been possibleto assessthe effects com.mon in the chert that generally terminates of the serpentinizationprocess itself on primary the sedimentary series. sulfide assemblage.The situation now is that the The principal sulfide minerals are pentlandite, sulfides found in blebs (be they amygdales or millerite and pyrite with widespread,minor chal- immiscible droplets) are markedly sulfur- copyrite, sporadic but locally abundant rich assemblages( pentlandite-pyrite-pyrrhotite) and pyrrhotite and traces of vallerite, gersdorf- compared with the spatially associated, but fite and chalcocite (Hendriks 1974). Ni:Cu usually stratigraphically lower, disseminated ratios averageabout l5:1, as is typical of nicke.l millerite-magnetite assernblage,which is relat- deposits associated with Archean ultramafic ively sulfur-deficient and of higher ni'ckel tenor. rocks (Naldrett 1973), A number of correla- Shangani tions can be made between differing sulfide assernblages,styles of mineralization and host- The Shangani nickel deposit (Philpott 1975, rock type although studies of the ore to date Viljoen et al. 1976) is situated about 35 km are preliminary. Two types of assemblageseem east-southeast of the Damba deposits in a to be developedthough there are exceptionsand separate greenstone belt, the geology of which gradationsbetween them (Hendriks 1974). The (Fig. 2), has been documented by Harrison first is a millerite-magnetite type with relatively (1969). minor amounts of pentlandite and pyrite and Viljoen et al. (L976) have recently described the second is a pentlandite-rich type with a the geology and mineralization of the deposit; high pyrite or pyrrhotite content and lessmagne- there is a marked similarity between the strati- tite. Violarite occurs in both types and is in graphic successionsin which the Damba and most casesan alteration product of either mille- Shangani deposits are situated. rite or pentlandite. Magnetite occurs in three' At Shangani, Viljoen et al. (1976) recognize forms: as magnetite rims on chromite that is the following stratigraphy in the mine area: quite separate from the sulfide association; as Esmyangene Formation (at the base): acid to primary magnetite forming an intrinsic part of intermediate tuffs terminated by a pyritic shale, the sulfide assemblageand as secondary magne- intruded by a large layered complex of peridotite, tite thaf is related to serpentinization. harzburgite, pyroxenite and gabbro; Makwe In the disseminatedand vein sulfide ores there Formation: peridotite and pyritic carbonaceous NTCKEL SI'LFIDB DEPOSITS IN RHODESIA 345 shale overlain by magnesium-rich metabasalts; appearsto be cut off by the large layered body EnsanguFormation: massiveand pillowed meta- that intrudes tlte Esmyangene Forrnation. Up- basalts. Viljoen et al. (1976) place the contact ward within the complex (Fig. 5, sections AB between the Sebakwian and Bulawayan Groups and CD) serpentinite is interlayered with am- at the Makwe-Ensangu Formation boundary; phibolite and some carbonaceousshales. horf,ever, as this contact is actually within a Viljoen et al. (1976) regard the complex as volcanic sequence,it seems.more reasonableto intrusive, emplaced into, or close to, an acid place it between the Esmyangene and Makwe volcanic vent. This author prefers to interpret Formations where mafic volcanism actually be- the complex itself as a volcanic neck that de- gan. The Bulawayan Group thus begins here veloped in the following way. It has a trench- with the peridotite unit ( = Unit 1 at Damba), or fissure-like form, with unlayered peridotite with associated pyritic shales and overlying in the stem and in the lower part of each lobe mapesium-rich basalts (- Unit 2 at Damba) (Fig. 5, section CD). This passesupward into of the Makwe Formation followed by the basalts a typical extrusive alternation of magnesiurn- and basaltic komatiites ( - Unit 3 at Damba) rich basalt, some of which has spinifex-style of the Ensangu Formation (Fig. 5). textures, and more peridotite, in places with The mineralized complex at Shangani (Viljoen sedimentary interlayers between units. Sulfide et al. 1976) lies within the EsmyangeneForma- liquid suspendedin the initial peridotite pulse tion and is a trenoh-like peridotite body (mush- was swept onto, or left behind on, the flatJying, room-shaped in cross section) that is at least flanking sections of the complex. The actual 1.7 km long (down dip) with the neck cutting site or sites of ultramafic volcanism would have across the regional stratigraphy (Fig: 5). The migrated along the trench with time. After the ore is predominantly disseminated and is a initial development of the mineralized complex, typical pyrhotite-pentlandite assemblageoccur- a pause in ultramafic volcanic activity, or its ring on the stratigraphically flatJying flanking migration along the fissure, allowed further ac- parts of the complex (Fig. 5, section CD). The cumulation of felsic tuffaceous material from downward extension of the stem of the complex ongoing but waning acid volcanism. Renewed

Frc. 5. Simplified plan and sectionsfor Shangani. Modified from Viljoen et al. (1976), 346 THE CANADIAN MINERALOGIST

ultramafic volcanism, yielding a peridotite with intrusive into the Bulawayan Group. Descrip finer grained suppended olivine crystals than tions of the general geology and petrology by previously, initiated the Bulawayan proper with Stowe (1968) suggestthat komatiitic volcanic the first part of -the Makwe Formation. This rocks are also developedin the area. peridotite (= -llnit 1 at Damba) probably issued from other sites along the fissure zone Hunters Road and was immediately followed first by the The Hunters Road nickel deposit (Brand pyroxenitic komatiite lavas ( = Unit 2 at Dam- 1976) is situated 80 km northeastof Shangani ba) that make up the rest of the Makwe Forma- (Fig. 2), on the eastlimb of the Gwelo syncline tion, and then by t*re,thick sequenceof basaltic in what appears to be a cross-cutting serpenti- komatiites and basalts of the Ensangu Forma- nite body. In this case the host serpentinite cuts tion (= Unit 3 at Damba). through a banded iron-formation lying at the Thus the mineralized complex can be regarded contact between acid to intermediate volcanic as the first tentative stage of Bulawayan Group rockS below, and pillowed tholeiitic basalts ultramafic volcanism in a zone of crustal weak- above (Macgregor 1937, Tyndale-Biscoe1949, ness where Viljoen et al. (1976) have recog- Harrison 1970), Brand (1976 and oral presen- gnized that felsic volcanism had been, or still tation, Salisbury) stated that disseminated sul- was, in progress. Viljoen et al. (1976) have fide mineralization occurs in a serpentinite in- also demonstrated that this zone acted as a truded along a tear fault and suggestedthat vent for the overlying volcanic rocks. Thus the both intrusive and extrusive ultramafic rocks mineralization itself is found on the horizontal may be represented.He also noted that layers or kink-structured sites flanking a vertical con- showing spinifex texture are developed in the duit; the interpretation outlined above shows body and that it contains large xenoliths of again the close link that exists in this region greenstone,banded iron-formation and acid vol- between nickel mineralization and the first canic rocks. volcanic event in the Bulawayan Group. At first glance this deposit would not appear to be the same Epoch at regional contact as the other deposits. Ifowever, the greenstone belt extend- The Epoch nickel mine is situated 85 km ing northwesterly from Gwelo is synclinal, with south of Shangani (Fig. 2). Although there a succession on the west limb of Sebakwian is no published account of the deposit, several Group schists overlain in turn by andesite and points of interest have been ascertained from Bulawayan Group mafic volcanic rocks (Fig. available maps (Ferguson 1934) and the pre- 2). On the east limb, where the nickel deposit sent investigation. Firstly, the deposit is also is located, a similar but far thicker suecession situated at or near the continuation of the of andesite is present and overlain by tholeiitic contact between the Sebakwian and Bulawavan basalts,but there is no evidenceof a Sebakwian- Groups. Secondly, like Shangani, it appears to type sequenceat the base. Structural informa- be associated with a cross-cutting peridotite tion in the Sebakwian Group on the west limb complex (Viljoen et al. 1976, p. 93) immedia- suggests that the andesites represent the first tely above an ultramafic-dominated ilavered volcanism of the Bulawayan Group in this area. sill intruded into the upper part of the Sebak- Therefore, the Hunters Road deposit does occur wian Group. Thirdly, the Bulawayan Group at a stratigraphic position marking the beginning Iavas that stratigraphically overlie the deposit of Bulawayan Group mafic volcanism as is the contain a significant komatiitic basalt com- case with the other deposits. In this case, how- ponent (Fig. 2). It is understood that the main ever, the mafic volcanism that follows is tholei- sulfides are millerite and pentlandite as found itic and not komatiitic. at Da.mba.Thus a cross-cutting peridotite at or Lower Gwelo near the base of Bulawayan Group komatiitic volcanism is a feature of another deposit in The west limb of the Gwelo syncline has the region. the Sebakwian Group at the base and it here consists of quartz-mica schists, quartzofeld- Selukwe spathic sedimentary units, banded iron-forma- Another nickel prospect occurs in the Selukwe tions and serpentinites (Macgregor 1937, Tyn- belt (Fig. 2) and again there is no published dale-Biscoe 1949), It is overlain in turn by a account of the occurrence. However, the de- thin but continuous andesite unit, a typical posit is also located near the contact between Damba-Shangani komatiite sequence, pillowed the Sebakwian Group and peridotites apparently tholeiitic bastlts, banded iron-formations and NICKEL SULFIDE DEPOSITS IN RHODESH 347 well-pr€served grits and conglomerates that make mineralization and volcanism throughout the up the synclinal core. region. Several deposits (Shangani, Epoch 1qd The komatiitic sequenceconsists of peridotitic Hunters Road) are in, or associated with' and pyroxenitic komatiites at the base, overlain cross-cuttingperidotite-complexes at the contact by basa,ltic and spinifex-textured "andesitiC' between the Sebakwiainand Bulawayan Groups. komatiite lavas (Williams in prep.). Dissem- At Shangani the mineralized complex has the inated nickel sulfide mineralization is present form of a fissure-filling with the trunk of the in the ultramafic unit at the base of the volcanic body cutting across the regional stratigraphy. pile but has yet to be fully evaluated. Again, This, together with the flanking position of the mineralization is at the same regional contact. mineralization and overall form of the complex, suggeststhat it is either a magma chamber or Noel an actual volcanic neck or vent. The latter alter- The Noel nickel arsenide deposit (Phaup native is preferred becauseof the interlayering 1933) is in the Lower Gwanda belt (Fig. 2) of peridotite, spinifex-textured amphibolite and has been known since 1928. The deposit (here regarded as komatiitic lava) and carbona- consists principally of chloanthite and niocolite ceous shale in thb upper part of the cornplex. associated with and silicate veinlets in The Hunters Road deposit is also in what a fault zone in an actinolitic serpentinite. Al- appears to be a cross-cutting complex and large though clearly hydrothermal, the deposit is also xenoliths-of country rock are recorded as occur- within a komatiite succession whish, in this ring within the serpentinite. Some of the mine- case, consists of a complex interlayering of ralization in this deposit is in the form of small basaltis-textured peridotites (Phaup 1933) and droplets like those described from Damba and basaltic and pyroxenitic komatiites. supports the view that the sulfides were carried as^an immiscible phase at the tirire of emplace- ment. The dif.ferent types of sulfide mineraliza' DtscussroN tion found at Damba suggestthat both sulfur- rich and sulfur-poor phases may have existed Many features of the Damba nickel deposits as droplets in close proximity to each other. point to a close relationship of ore with early Significantly, at least at Damba, the sulfur-rich peridotitic volcanism. The presence of amyg- assemblage is found' within markedly amyg- daloidal spinifex-textured units, probable flow- daloidal zones in the host peridotite. top breccias and ultramafic tuffs within the Using the concept that sulfides were trans' host peridotite all indicate an extrusive origin. ported is an immiscible phase within an olMne A possiblefeeder conduit (or vent) is recognized crystal - ultramafiq liquid mush, Naldrett under the Silwane mineralization and the pre- (i973) has speculated on some likely relation- it and country rock' His sence of an agglomerate zone 175 m'komatiites above ships botween ore, host within the overlying pyroxenitic diagrams are reproduced here in Figure 6 and (Unit 2) adds support to the idea of a vent in show possible reconstructions of thq settings for this locality. Shangani (Fig. 6, example L ot 2) and Silwane The mineralization occurs as blebs (at least (example3). some of which were immiscible droplets) and As ihere seems to be a close association be' as disseminated matrix sulfide that has a de- tween ore and volcanism, it is useful to examine finite appearance of being of a similar droplet the general relationships between the various style but dispersed by included olivine crystals' volcanic assemblages within the Bulawayan to produce the ovoid patches of disseminated Group and thus the regional setting of the de- sulfide. Mineralization occurs in a thick basin- posits. The present study has shown that koma- like depression or lava pool that overlies the tiitic volcanic roc'ks a.remore widespread in the possible feeder, which is itself concentrically Rhodesian greenstone belts than indicated on mineralized with higher grade ore in the centre existing,'maps.It has also shorrn that the simple passing outward into lower grade ore. Ttiese komatiite stratigraphy describedby Yilioen et al. factors all suggestthat the sulfide is an intrinsic (1976) at Shangani-anddescribed in detail here part of the host peridotite magma. Furthermore, for the Lonely Mine belt is"widespreadin the the emplacement of peridotite and ore was the region. Another example is developed on the first event of komatiitic volcanism within the western limb of the Grrelo belti Bickle et al. Bulawayan Group in this area. (1975) and.Nisbetet al. (1977) have described Features of many of the other deposits sug- a particular,ly extensive succession in the Be' gest that similar relationships exist between lingwe belt.,.Thesestratigraphic sequencesbegin 348 THE CANADIAN MINERALOGIST with a peridotite and pyroxenitic komatiite unit overlap and interfinger with apparently con- that ranges up to 1 km in thickness, followed comitantly extruded tholeiitic basalts. These are by a thick pile (up to 7 km) of basaltickoma- typically pillowed, weakly plagioclase-phyric tiite and some tholeiitic basalts. The basaltic basalts showing little of the skeletal crystalliza- komatiites commonly show clinopyroxenede- tion that is usually characteristic of the koma- rived spinifex texture or various textures caused tiite suite. However, the distinction between by skeletal clinopyroxene crystallization. More the tholeiites and low-MgO basaltic komatiites pyroxenitic or olivine-bearing examples are in- is often difficult and some uncertainty exists at tercalated with the basaltic varieties, but are thesecompositions. This overlap is also apparent not common. These komatiitic sequenceshave in their geochemistry (Williams in prep.) and widely varying strike extents ranging from 15- might suggestthat the two magma types result 80 km. from different levels of tapping of a common Along strike the komatiitic volcanic rocks magma stem. The rnore MgO-rich komatiitic rocks, which are the nickel sulfide hosts, would be expected to result from the deeper tapping. Higher up in the Bulawayan Group the inter- I. SUB-SURFACE MAOMARESERVOIR layering of volcanic types becomes more com- plex. In the Lonely Mine - Bulawayo belt a thick successionof andesite lavas overlies the tholeiitic and komatiitic volcanic suites and is, in turn, overlain by either more basaltic and pyroxenitic komatiites or by tholeiites. It is clear that in most of the Rhodesian greenstonebelts examined during this study, the simple koma- tiite to tholeiite evolution proposed for otler greenstonebelts in southern Africa (Viljoen & Viljoen 1969) does not apply. Rather there - 2. THICKEROLIVINE RICHPARTS OF seems to be random interlayering and over- FLOWSCLOSER TO FEEDER lapping of adjacent compositionally different volcanic suites throughout the succession.The clear evidence of contemporaneousor penecon- temporaneoustholeiitic and komatiitic volcanism, and in some casesacid volcanism (Lonely Mine belt), is particularly importpnt. Andesitic lavas and agglomeratescan occur either before koma- tiitic and tholeiitis activity (Belingwe, Gwelo, Hun,ters Road) or after (Lonely Mine 3. LAVA POOL OVERFLOWINGINTER - Bulawayo, Gwanda). MITTENTLYTO GIVE SERIESOF SPINIFE In su,rnrnary,six nickel deposits in the region CAPPEDFLOWS AT EDGES are found at or near the same regional contact; they seem to form an integral part of the very first event of komatiitic volcanism within the Bulawayan Group. Several of the deposits lie within cross-cutting peridotite complexes that can be interpreted as volcanic necks or vents lying in sites that aoted as volcanic centres both prior to, and follorring, peridotite-ore emplace- ment. The configuration of the Shangani mine- COUNTRYROCK ralized complex suggests that the ore-bearins PERIDOTITE(<50% OLIVINE) komatiitic volcanism was of a fissure tnre. PERIDOTITE(>50% OLIVINE) SULFIDE AcrNowlrpcEMENTS

Frc. 6. Speculationson likely The permission of JohannesburgConsolidated ' 'ore, relationshipsbetween host and country rock in nickel deposits of Investment Company Umited to publish these volcanic associationby Naldrett (1973). results is gratefully acknowledged. I arn indebted NICKEL SULFIDE DEPOSITS IN RHODESIA 349

(1929): geology of the Shabani to Company colleagues who ,have contributed KEnr, F.E. The Geol, Surv. Bull. L2, to this study, particularly R. Mason, D.E' Phil- mineral belt. S. Rhod. A.M. (1928): The geology of the pott, C. Willson, D.M. Anderson and L.P. Hen- Meccnrcon, country around the Lonely Mine, Bubi District' Viljoen who first driks and to R.P. and M.J. S. Rhod. Geol. Surv. Bull, tL. de- introduced me to the Damba and Shangani 0937): The geology of the country around posits. Staff of the Rhodesia Geological Survey Hunters Road. Gwelo District. S. Rhod. Geol' provided invaluable advice and comment during Surv. Bull. 8L. the regional investigation, and special thanks go (1951): Some milestones in the Precam- to N.M. Harrison and I.D.M. Robertson. I am brian of Southern Rhodesia. Geol. Soc. S. Atr. very grateful to I.D.M. Robertson' R. Mason, Proc. 54, 28-68. (1937): The P.G. Cochran and D.L. Garnett for criticizing FrncusoN, J.C. & Arvru, F.L. geology around the Mine, drafts of the manuscript. Special thanks go to of the country Queen's bubwayo District. S. Rhod. Geol. Sarv. BUIL SO' wife Gwynn for her geological help both my Neronrrr, AJ. (1973): Nickel sulfide deposits - his in the field and office, to Piet Britz for their classification and genesis, with special em- draughting and Elizabeth Vigus for typing the phasis on deposits of volcanic association. Can. manuscript. Inst. Mining Met. Trans. 76, 183-201. Nrsrrl E.G., BICKLE, M.J. & Menrnt, A. (1977)t ultramafic lavas of the Belingrrye RrrsRENces The mafic and greenstone belt, Rhodesia. J. Petrology 18' 521- (1940): The geology of the country 566. Arvrrra,F.L. geology the Antelope 'around S. Rhod. GeoL Surv. BuIl, 35. Pneup, A.E, (L932): Tbe of Bulawayo. Bull. 2L. (1946): The geologyof the Lower Gwelo gold belt. S. Rhod. Geol, Surv. (1933): geology of the Lower Gwanda gold belt. S. Rhod. Geol. Surv. BqIL 37, The BuIl. 24, Neronrm, A.J. & Pvrr, D.R. (1977): belt. S. Rhod. Geol. Surv. ARNDT,N.T. (1975): - a geochemical iron-rich tholeiitic lavas of Munro PHrl-porr, D.E. Shangani Komatiitic and sulphide deposit' northeast Ontario. l. Petrology L8, discovery of a nickel--copper Township, t974 (I. L. Elliot 319-369. 1r Geochemical Exploration eds.), Elsevier, Amsterdam. Brcrr-s, MJ, Mmrnt, A. & Nrsner, E.G. (1975): & W. K. Fletcher, Basaltic and peridotitic komatiites and stromato- Srown, C.W. (1968): The geology of the country lites above a basal unconformity in the Belingwe west atrd south of Selukwe. Rhod. Geol' Surv' greenstonebelt, Rhodesia.Earth Planet, Sci,Lett. Bull. 59. 27, 155-162. TYNDALE-BIwoE, R. (1940); The geology of the Br,rss,N.W. & Srrpor.pn,P.A. (1969): The Rhode- country around Gwanda. S. Rhod. Geol. Surv' sian basementcomplex: a review. Geol. Soc.S. Bull.36. Afr. Spec,Publ. 2,305-333. (1949): The geology of the country around BRAND,B.L. (1976): Hunters Road nickel. Mefal- Gwelo. S, Rhod, GeoI. Surv. BuIl,89. '76, Geol, Joc. ,S.Afr. Conf. Salisbury' logenesis VTLJoEN, M.J., BenNescoNl, A.' vAN CoLLER, N., 40 (Abstr.). Vn-loBN, R.P. (1976): The geol' geology coun- KrNrocr, E. & FnncusoN,J.C. (1934): The of the nickel deposit. Econ. GeoI' S. Rhod. ogy of the Shangani try around Filabusi, Insiza District. 71, 76-95. Geol. Surv. Bull. 27. of (1968): A re-assessmentof the & Vnrorry R.P. (1969): A reappraisal HARRrsoN,N.M. of shield areas stratigraphy of the Precambrian basement com- the granite--greenstone terrains based on the Barberton model. GeoI. 'loc' S' plex around Que Que, Gwelo District, Rhode- sia.Geol. ^Soc. ,S. Alr. 7L (Annex'), ll3'123. Afr. Spec. Publ. 2, U5'274, (1969): The geologyof the country around Wnr.nrr, G. C. (1975): Possible amygdaloidal Fort Rixon and Shangani.Rhod. Geol. Surv. sulfides in the Otter Shoot, Kambalda, Western Ball. 6L. Australia. Econ. GeoI. 70, 1L27, (L970)t The geologyof the country around Wrr-r-revrs, D.A.C. & FunNrr.r-, R.G. (1979): Re- Que Que. Rhod. GeoI. Surv. Bull. 67. assessment of part of the Barberton typs area' HENDRrcxs.L.P. (1974'lt Nature and modes of South Africa. Precambrian Res. (in press)' occurrenbeof thi ore-coDstituentsin sdmplesof WrlsoN, J.F. (1973): The Rhodesian Archaean Damba drill core. lohannesburg Consolidated craton - an essay in cratonic evolution. R. ,Soc. InvestmentInternal Rep. (unpubl.). Lond. Phil, Trans, A273,389-4L1. KsELs,R.A. (1975): Possibleamygdaloidal sulfides B.G. (1956): The geology of the country in.the Otter Shoot,Kambalda, Western Australia - Wonsl between Belingwe and West Nicholson. S. Rhod' reply.Econ. Geol.70, 1127'1129. a Sarv. BuIl. 43, & Nrcrrr,, EJJ'. (L974): The geologYof a primary millerite-bearing assemblageand super- gene alteration at the Otter Shoot, Kambalda' Received Augwt 1978; revised manuscript accepted Western Australia. Econ. Geol. 69, ll02'Llt7, December 1978.