The differentiation of the Skaergaard intrusion

Conclusions We find no support for the claim that the Skaergaard magma followed the trend of common tholeiitic volcanic magmas, such as those of Iceland and the Scottish Tertiary. The end product of differentiation was not a large mass of rhyolite but an iron-rich, silica-poor liquid not unlike that deduced by Wager in 1960.The proposal that a large mass of rhyolitic liquid occupied the upper levels of the intrusion finds no support in the field. The thick series of ferrogabbos, which became richer in iron and poorer in silica until they reached a field of immiscibility cannot be reconciled with crystallization of a large mass of felsic magma. Mass-balance calculations that indicate otherwise are invalid, because they fail to take into account large volumes of rocks that differ in composition from those assumed in the calculations.While ignoring the existence of major units of the intrusion, Hunter and Sparks propose that lavas in Scotland and Iceland are more relevant to the liquid compositions than rocks that are intimately associated with the intrusion. Their argument that the Skaergaard Intrusion followed a trend of silica enrichment that is universal to tholeiitic magmas is based on an incomplete knowledge of the rocks and faulty calculations of mass-balance relations.We agree that much remains to be learned about the Skaergaard Intrusion and the basic mechanisms of magmatic differentiation. In this case, however, we are ready to hang our case on well-established field relations and a mass of laboratory data for what must be the most intensely studied body of rock on Earth.     

The Chilled Marginal Gabbro and Other Contact Rocks of the Skaergaard Intrusion

Rocks in the outer selvage of the Skaergaard intrusion havea range of textures and compositions, and among these are materialsrepresenting quenched Skaergaard magma. Pristine chilled marginalgabbro (CMG), however, is not ubiquitous at the intrusive contact,because many of the "contact" rocks have been hydrothermallyor metasomatically altered, contaminated with gneiss or olivinexenocrysts, while others contain accumulated minerals. Materialrepresenting quenched magma appears to be restricted to contactrocks that are texturally and mineralogically similar to diabase,and free of accumulated minerals. Where it exists, the CMG isfound within one to three meters of the exposed intrusive contactexcept at the roof of the intrusion where its thickness is greater.CMG was distinguished from the diverse group of contact rocksby petrographic and geochemical screening of over 80 specimens.Samples of CMG from the eastern and western margins and fromthe roof of the intrusion have relatively uniform compositionsimilar to that of ferrobasalt, and are noticeably richer iniron (mg-number=0?51-0?54), TiO₂ K₂O, and P₂O₅ than other unmodifiedcontact rocks. CMG's also have trace element compositions distinctfrom most other rocks in the outer Marginal Border Series (MBS).They have incompatible element contents up to 3–6 timesgreater than in LZa-type cumulates, negligible Eu anomalies,and Ni and Cr contents and Ni/Cr ratios that are among the lowestof rocks in the outer MBS. The results of melting experiments corroborate selection ofthis material as CMG. The composition of glasses obtained frompartial melting experiments of LZa-type cumulates are essentiallyidentical to those of the CMG. The 1-atm. liquidus phase relationsfor one of the CMG samples (KT-39) is largely consistent withthe sequence and composition of cumulus minerals observed withdistance inward through the MBS and upward through the LayeredSeries. Solidification of magma at the outer margin of the intrusionis interpreted to have involved locally efficient quench crystallizationfollowed by initial primocryst growth in an undercooled transitionzone a short distance inward that finally extended into regionsof near equilibrium crystallization. The similarity in composition between samples of chilled marginalgabbro from the exposed roof and sides of the intrusion, andthose of reconstituted trapped liquid from early cumulates inthe outer MBS suggests that a single magma, similar in compositionto ferrobasalt, was parental to the Skaergaard intrusion. Thisinterpretation corroborates geophysical evidence of a significantlysmaller mass for the intrusion than that estimated by Wager,and provides a basis for revision of models of its chemicalevolution. Samples chosen by Wager as chilled marginal gabbrobelong spatially, texturally, and compositionally to the groupof LZa-type cumulates in the MBS, and should no longer be regardedas chilled marginal gabbro.     

Petrology of the Upper Border Series of the Skaergaard Intrusion

The Upper Border Series of the Skaergaard intrusion consistsof a 960 m thick sequence of rocks that crystallized againstthe roof of the magma chamber. The texture and composition ofthe unit vary systematically from top to bottom as a resultof changes that occurred in the magma during the solidificationof the intrusion. The order of crystallization of primocrystminerals in the Upper Border Series was: olivine; + plagioclase;+ apatite; + ilmenite; + magnetite; + Ca-rich pyroxene;—olivine;+ olivine; + ferrobustamite. The major silicate phases varyfrom high-temperature compositions to low-temperature compositionswith increasing distance from the upper contact. Post-crystallizationre-equilibration has affected the compositions of the oxideminerals and to a lesser extent the compositions of olivineand Ca-rich pyroxene. The Upper Border Series differentiationsequence differs from the Layered Series sequence, in that:(1) apatite appears much earlier; (2) magnetite precipitatedbefore Ca-rich pyroxene rather than after it; (3) orthopyroxeneis much less common; (4) the plagioclase is systematically poorerin K₂O; and (5) the rocks are systematically richer in K₂O andSiO₂. The upper part of the Skaergaard magma appears to havebeen enriched in H₂O, K₂O, SiO₂, and P₂O₅ relative to the partthat was parental to the Layered Series.     

Pyroxenes of the Bushveld Intrusion, South Africa

New analyses are presented, for major, minor, and trace elements,of eleven Ca-rich pyroxenes, four bronzites, and two invertedpigeonites from the Bushveld layered basic intrusion. The twenty-threeanalyses now available are believed to represent the entireBushveld fractionation sequence. The Ca-rich pyroxene trendis from Ca_45.4Mg_49.6Fe_5.0 to Ca_42.7Mg_0.6Fe_56.8, the ferrohedenbergitesshowing no evidence of inversion from ferriferous ß-wollastonites.The Ca-poor pyroxene trend is from bronzite (Ca_2.8Mg_85.0Fe_12.2)through pigeonites to ferropigeonites (approximately Ca₃Mg₂₇Fe₅₅).All the pigeonitic pyroxenes have inverted to orthopyroxene. The compositional trends are remarkably similar to those ofthe Skaergaard pyroxene series, but the Bushveld sequence isthe most complete known for a single fractionated intrusion.The compositional and other variations of the pyroxenes, consideredtogether with those of the coexisting olivines and feldspars,leave little doubt that the Bushveld rocks originated by crystalaccumulation from a slowly cooled and fractionated intrusionof tholeiitic basalt magma. The slight but significant differences between the Bushveldand Skaergaard pyroxene trend characteristics can be explainedin terms of a displacement, in one intrusion as compared withthe other, of the liquidus and solidus surfaces relative tothe solvus and inversion surfaces in the system Wo—En—Fs.This may be due to minor differences in the initial magma compositionsof the two intrusions. Differences in the Mg/Fe ratios of Bushveldand Skaergaard coexisting pyroxene pairs are believed to bedue, at least in part, to the greater depth of the Bushveldmagma chamber. The Bushveld trends are briefly discussed in the light of recentexperimental studies on compositions within the Di-Hed-En-Fspyroxene quadrilate     

Parental magma of the Skaergaard intrusion: constraints from melt inclusions in primitive troctolite blocks and FG-1 dykes

Troctolite blocks with compositions akin to the Hidden Zone are exposed in a tholeiitic dyke cutting across the Skaergaard intrusion, East Greenland. Plagioclase in these blocks contains finely crystallised melt inclusions that we have homogenised to constrain the parental magma to 47.4–49.0 wt.% SiO₂, 13.4–14.9 wt.% Al₂O₃ and 10.7–14.1 wt.% FeO^T. These compositions are lower in FeO^T and higher in SiO₂ than previous estimates and have distinct La/Sm_N and Dy/Yb_N ratios that link them to the lowermost volcanic succession (Milne Land Formation) of the regional East Greenland flood basalt province. New major- and trace element compositions for the FG-1 dyke swarm, previously taken to represent Skaergaard magmas, overlap with the entire range of the regional flood basalt succession and do not form a coherent suite of Skaergaard like melts. These dykes are therefore re-interpreted as feeder dykes throughout the main phase of flood basalt volcanism.     

Electron-probe studies of the earlier pyroxenes and olivines from the Skaergaard intrusion,East Greenland

Pyroxenes and olivines from the earlier stages of fractionation of the Skaergaard intrusion (Wager and Brown, 1968; Brown, 1957) have been studied using the electron microprobe. The subsolidus trend for both Ca-rich and Ca-poor pyroxenes has been established, from the Mg-rich portion of the quadrilateral to the Hed-Fs join, together with the orientations of the tie-lines joining coexisting pyroxenes. For the Mg-rich Ca-poor pyroxenes, Brown's (1957) solidus trend has been modified slightly. From a study of a previously undescribed drill core, reversals in the cryptic layering have been found in the Lower Zone. The reversals are attributed to existence within the convecting magma chamber of local temperature differences. The Skaergaard magma temperatures are postulated to have passed out of the orthopyroxene stability field into the pigeonite stability field at EnFs ratios of 7228, for Ca-free calculated compositions, and specimen 1849, a perpendicular-feldspar rock, is interpreted as straddling the orthopyroxene-pigeonite transition interval. The cessation of crystallisation of Ca-poor pyroxene and the increase in Wo content of the Ca-rich pyroxene trend have been reexamined, and Muir's (1954) peritectic reaction (pigeonite+liquid=augite) has been confirmed. The composition at which Ca-poor pyroxene starts reacting with the liquid is postulated as Wo₁₀ En_36.7Fs_{53 3}. It is suggested that the cessation of crystallisation of Ca-poor pyroxene is sensitive to the amount of plagioclase crystallising from the liquid.A complete series of accurate olivine compositions for the whole Skaergaard sequence is presented for the first time, including the compositions of the Middle Zone olivine reaction rims.     

Major and trace element composition of Skaergaard plagioclase; geochemical evidence for changes in magma dynamics during the final stage of crystallization of the Skaergaard intrusion

Plagioclase separates from the Layered Series (LS), Upper Border Series (UBS), and Marginal Border Series (MBS) of the Skaergaard intrusion were analyzed to examine major and trace element variations. In general, plagioclase from the LS, UBS, and MBS show similar trends in major elements vs. crystallization: SiO₂, Na₂O, and K₂O progressively increase, and CaO and MgO progressively decrease with fractionation. No abrupt changes in the trends of major components of Skaergaard plagioclase during the differentiation of the intrusion are observed. Trace elements in plagioclase reflect changes in the Skaergaard magma and changes in plagioclase distribution coefficients with differentiation. Sr, Ga, and probably Ba are included elements in Skaergaard plagioclase, but were excluded from the other cumulus phases, and as a result systematically increased in the magma and plagioclase during differentiation. Be, Cs, Hf, Rb, Ta, U, and Zr, and the transition metals Co, Cr, Cu, Ni, Sc, V, and Zn were excluded elements in Skaergaard plagioclase, and remained low in plagioclase during differentiation. Changes in the abundances of these elements in plagioclase during differentiation reflect changes in their abundance in the magma. With the exception of the lower zone, which is enriched in the light rare earth elements, rare earth elements in LS plagioclase, in general, increase with differentiation of the Skaergaard intrusion, but decrease dramatically at the UZa/UZb boundary where abundant apatite first appears. Rare earth elements in UBS plagioclase followed a similar trend to LS plagioclase, except during the initial and final stages of differentiation. UBS plagioclase is much more enriched in rare earth elements during the final 20% of crystallization, except for Eu, which is similar in plagioclase from the two series. The observed trends suggest that the floor and roof sequences became isolated from each other and that the floor sequence may have been more reducing and the roof sequence more oxidizing during the final 20% of crystallization. As the Skaergaard magma ceased convection, or convected as isolated cells, during the final stages of differentiation, volatile elements may have accumulated in the UBS magma, resulting in an increase in ƒO₂, and a decrease in Eu/Sm in UBS plagioclase. The observed trends of rare earth elements in plagioclase from the LS and UBS fit well with theoretical calculations that assume closed-system crystallization, and would be difficult to reconcile with any model requiring significant discharge of magma from the chamber during the final 20% of crystallization. The enrichment of light rare earth elements in plagioclase, suggests that the lower part of the intrusion re-equilibrated with a late, light rare earth element-rich fluid or melt. The recharge model proposed by earlier workers to explain anomalous Sr and Nd isotopes appears unlikely in light of the two to fourfold enrichment of light rare earth elements in these samples. Received: 1 October 1999 / Accepted: 14 May 2000     

Partial melting below Tubuai (Austral Islands, French Polynesia)

Process identification diagrams based on trace element data show that mafic lavas from Tubuai, including alkali basalts, basanites, analcitites and nephelinites, result from different degrees of partial melting of an isotopically homogeneous mantle source. Our fractionation-corrected data are consistent with a batch melting model or a dynamic melting model involving a threshold value for melt separation close to 1% and degrees of melting ranging from 5–8% (alkali basalts) to 1.5–3% (nephelinites). The relative source concentration pattern, calculated using an inverse numerical method, shows an enrichment in highly incompatible elements. We propose that the Tubuai lava suite was derived from a two-stage partial melting process. Melting first affected the plume material located within the transition zone between garnet and spinel domains, producing alkali basalts and basanites. Then, the melting zone migrated upwards to the base of the overlying spinel-bearing lithospheric mantle, producing highly silica-undersaturated lavas. The lower lithosphere had previously been enriched by intrusion of pyroxenite veins representing plume-derived melts which percolated away from the main magma conduits. Received: 11 June 1996 / Accepted: 8 January 1997     

Plagioclase in the Skaergaard intrusion. Part 1: Core and rim compositions in the layered series

The anorthite content of plagioclase grains (X_An) in 12 rocks from the layered series of the Skaergaard intrusion has been studied by electron microprobe (typically ∼30 core and ∼70 rim analyses per thin section). Mean core compositions vary continuously from An₆₆ at the base of the layered series (LZa) to An_32–30 at the top. On the other hand, crystal rims are of approximately constant composition (An_{50 ± 1}) from the LZa to the lower Middle Zone (MZ). Above the MZ, core and rim compositions generally overlap. Profiles across individual plagioclase grains from the lower zone show that most crystals have an external zone buffered at X_An ∼50 ± 1. The simplest explanation for these features is that during postcumulus crystallization in the lower zone, interstitial liquids passed through a density maximum. This interpretation is consistent with proposed liquid lines of descent that predict silica enrichment of the liquid associated with the appearance of cumulus magnetite.     

Early stages in the differentiation of the Skaergaard magma as revealed by a closely related suite of dike rocks

Certain dikes in the Kangerdlugssuaq area of east Greenland correspond rather closely with the presumed bulk composition of the nearby Skaergaard intrusion and therefore allow a better characterization of the chemistry of the Skaergaard initial magma than has so far been possible. They indicate that Wager's chilled margin composition needs only minor revision, except in the case of Al₂O₃. Other dikes belonging to the same suite show a range of compositions extending up to the differentiation stage represented by the lower part of the Skaergaard middle zone and these allow an investigation, using computer least squares and Rayleigh fractionation modelling, of the early part of the Skaergaard differentiation trend which is otherwise unexposed. We conclude that the only important cumulus phases at this stage were olivine and plagioclase but clmopyroxene grew by heteradcumulus growth and effectively eliminated all intercumulus liquid. The liquids described here are similar to many ocean-floor basalts.     

Magma plumbing systems in large igneous provinces: Inferences from cyclical variations in Palaeogene East Greenland basalts

We present compositional data on a 1,250-m-thick sequence of sparsely porphyritic lavas that comprise the Geikie Plateau Formation, part of the ~55-Ma break-up-related flood basalts in East Greenland. Major element compositions are relatively restricted (6.3–7.6 wt% MgO; 2.2–2.4 wt% TiO₂), with two excursions to more evolved compositions (2.4–3.4 wt% TiO₂) that are similar to the inferred parental magma of the nearby Skaergaard Intrusion. Major and trace element calculations show that fractional crystallisation is the principal control on magma compositions, and the cyclical sequential variations imply regular magma chamber replenishment events. Isotopic data indicate minor crustal assimilation, but with different contaminants for the main group (amphibolitic gneiss) and evolved cycles (granulitic gneiss). Rifting episodes may have allowed more primitive magmas to ascend to shallow crustal levels and subsequently fractionate to more evolved compositions in a separate chamber, which was perhaps similar to the source of the Skaergaard Intrusion.Electronic Supplementary Material Supplementary material is available in the online version of this article at Editorial responsibility: I. Parsons     

Intraplutonic Quench Zones in the Kap Edvard Holm Layered Gabbro Complex, East Greenland

The Kap Edvard Holm Layered Gabbro Complex is a large layeredgabbro intrusion (>300 km²) situated on the opposite sideof the Kangerdlugssuaq fjord from the Skaergaard Intrusion.It was emplaced in a continental margin ophiolite setting duringearly Tertiary rifting of the North Atlantic. Gabbroic cumulates, covering a total stratigraphic thicknessof >5 km, have a typical four-phase tholeiitic cumulus mineralogy:plagioclase, clinopyroxene, olivine, and Fe–Ti oxides.The cryptic variation is restricted (plagioclase An_81–51,olivine Fo_85–66, clinopyroxene Wo_43–41 En_46–37Fs_20–11) and there are several reversals in mineral chemistry.Crystallization took place in a low-pressure, continuously fractionatingmagma chamber system which was periodically replenished andtapped. Fine-grained (0•2–0•4 mm) equigranular, thin(0•5–3 m), laterally continuous basaltic zones occurwithin an {small tilde}1000 m thick layered sequence in theTaco Point area. Twelve such zones define the bases of individualmacrorhythmic units with an average thickness of {small tilde}80m. The fine-grained basaltic zones grade upwards, over a fewmetres, into medium-grained (>1 mm) poikilitic, olivine gabbrowith smallscale modal layering. Each fine-grained basaltic zoneis interpreted as an intraplutonic quench zone in which magmachilled against the underlying layered gabbros during influxalong the chamber floor. Supercooling by {small tilde}50C isbelieved to have caused nucleation of plagioclase, olivine,and clinopyroxene in the quench zone. The nucleation rate isbelieved to have been enhanced as the result of in situ crystallizationin a continuously flowing magma. The transition to the overlyingpoikilitic olivine gabbro reflects a decreasing degree of supercooling. Compositional variation in the Taco Point sequence is typicalfor an open magma chamber system: olivine (Fo_{77–68 5})and plagioclase cores (An_80–72) show a zig-zag crypticvariation pattern with no overall systematic trend. Olivinehas the most primitive compositions in the quench zones andmore evolved compositions in the olivine gabbro; plagioclasecores show the opposite trend. Although plagioclase cores arebelieved to retain their original compositions, olivines re-equilibratedby reaction with trapped liquid. Some plagioclase cores containrelatively sodic patches which retain quench compositions. Whole-rock compositions of nine different quench zones varyover a range from 10 to 18% MgO although the mg-number remainsconstant at {small tilde}0•78. The average composition(47•7% SiO₂, 13•3%MgO, 1•57% Na₂O+K₂O) is takenas a best estimate of the parental magma composition, and isequivalent to a high-magnesian olivine tholeiite. The compositionalvariation of the quench zones is believed to reflect burstsof nucleation and growth of olivine and plagioclase during quenching. Magma emplacement is believed to have taken place by separatetranquil influxes which flowed along the interface between alargely consolidated cumulus pile and the residual magma. Theresident magma was elevated with little or no mixing. At certainlevels in the layered sequence the magma drained back into thefeeder system; such a mechanism is referred to as a surge-typemagma chamber system.     

Petrology of the Proterozoic Potato River Layered Intrusion, Northern Wisconsin, USA

The Potato River intrusion is a Keweenawan (1100 Ma) mafic plutonemplaced in Keweenawan volcanics and earlier Proterozoic metasedimentaryrocks along the southeastern flank of the Lake Superior syncline.It comprises the following lithostratigraphic zones: a thinto absent Border zone of altered olivine gabbro; a Lower zoneof olivine gabbro; a Picritic zone of picrite and troctolite;a Middle zone of olivine gabbro and leucogabbro; an Upper zoneof quartz leucogabbro and ferrogabbro; and a Roof zone of granophyricand granitic rocks. Fractional crystallization is evident fromcompositional changes in the rocks and cumulus minerals withstratigraphic height. Elements concentrated in the cumulus mineralsolivine and plagioclase (Mg, Fe²⁺, Al, Ca, Ni, Co, Cr, Sr) decreasewith height; elements concentrated in the trapped liquid (Na,K, La, Y, Zr, Nb, Rb, Ba) increase with height; and other elements(Ti, Fe³⁺, P, Ga, V, Sc, Cu, Zn) show complicated behavior relatedto the appearance of additional cumulus phases such as clinopyroxene,Fe-Ti oxides, and apatite. Lower zone rocks contain some sulfide,probably from sulfur derived from the country rock, and theUpper zone has sulfides probably precipitated from an immisciblesulfide liquid. The sulfide-bearing rocks have similaritiesto those of other intrusions, such as Bushveld, Stillwater,and Skaergaard. The picritic and troctolitic rocks of the Picritic zone indicatethat the intrusion was open to additional injections of maficmagma. Roof zone granophyric rocks are residual liquids intrudedalong the upper margin of the intrusion during regional tilting,but Roof zone granitic rocks are probably melted country rock.An attempt is made to estimate by reverse stratigraphic summationthe compositional path of the magma that solidified above thePicritic zone. The first compositions are highly aluminous,which suggests that the upper part of the intrusion has beenenriched in plagioclase by convection-aided crystal sorting.A complementary unit of mafic rocks is not exposed, but it couldbe present down dip. Some of the later compositions are similarto typical Keweenawan high-Al tholeiites. The magma did notundergo extreme iron enrichment, probably because of oxygenfugacity buffering.     

Kiglapait Mineralogy II: Fe-Ti Oxide Minerals and the Activities of Oxygen and Silica

Cumulus titanomagnetite and subordinate ilmenite first appearin the Upper Zone of the Kiglapait intrusion. They arrive graduallyand then reach abnormal abundances before falling to a sustainedcotectic mode near seven volume per-cent. The most Ti-rich titanomagnetites(to Usp ₆₆) are preserved in ore bands (layers) which solidifiedby adcumulus growth leading to the complete expulsion of interstitialsilicate liquid. Analyses from three of these ore bands, appliedto the solution model of Lindsley (1977), form a single lineararray in f_o2 versus 1/T ?K, with log f_o2 = (–28,283 ?89)/T + 11.03 ? 0.25. This array implies log f_o2 = –9.65at 1094 ?C, the model temperature of the Main Ore Band, consistentwith primary mineral compositions Ilm₈₉, Usp₇₉ and a weightmode of 18 per cent ilmenite. Silicate rocks yield another linear array, i.e. log f_o2 = (–37,910? 102)/T + 22.57 ? 0.61. This array is ascribed to closure ofsubsolidus reactions from initial compositions near Usp_79–80Ilm₉₀. The center of gravity of the data falling on this arraysuggests a primary mode of about 50 per cent ilmenite for thesilicate rocks, implying somewhat more reducing conditions ofcrystallization than for the ore bands. The modal overproductionrepresented by the ore bands is attributed to super-saturationin oxygen, which is demonstrated by the Al-depleted compositionsof titanomagnetite in ore bands, by direct evidence for elevatedf_o2 at the top of the Main Ore Band, and by abnormally magnesiansilicate mineral compositions in and near the ore bands. The primary titanomagnetite composition for average rocks isestimated at Usp₈₀ for the base, and Usp₇₃ for the top of theUpper Zone, from rock and mineral chemistry and observed textures.The idealized magma path for the Upper Zone runs from (T andlog f_o2) 1154 ?C, –9.0 to 960 ?C, –12.2. The orebands lie above this path and are interpreted as lying on themetastable extension of the Lower Zone path, which originatesat 1250 ?C, –8.1, on the WM buffer at 4 kbar total pressure. Silica activity is estimated from mineral compositions nearthe ore bands as applied to the FMQ equilibrium, and mappedfor the Lower Zone by an adjustment downward from the En-Fo-Silequilibrium, with resultant values near 0.55 relative to quartz= 1.0. The logarithmic oxygen/silica activity ratio (OSAR) coincideswith that of the Skaergaard intrusion in the Lower Zones. TheSkaergaard OSAR is offset downward from the Kiglapait trendduring MZ time, and remains below it at the end of crystallization.The more highly silicated Skaergaard magma was initially moreoxidized than the Kiglapait magma, but this relation was reversedafter the loss of olivine in the Skaergaard intrusion, as couldhave been predicted from theory and the mineralogy of the twointrusions.     

Geochemical Variation in the Greenstones of S.W. England

The characteristic rocks of the Upper Palaeozoic greenstonesof S.W. England are intrusive dolerites and extrusive basicpillow lavas with minor intermediate volcanics and ultrabasics(picrites). Pyroclastics are represented by keratophyric andbasic tuffs. The intrusive greenstones show varying degrees of alteration(spilitization) from a primary ilmenite-plagioclase-clinopyroxene?olivineassemblage to a hydrous low-grade spilite (or meta-dolerite)assemblage composed of variable proportions of albite, chlorite,epidote, calcite, and amphibole. Based on the distribution of elements little affected by secondaryprocesses (Ti, P, Y, Nb, and Zn), the intrusive greenstonescontain representatives of both the alkali olivine basalt andtholeiitic basalt magma series. Magmatic differentiation isgenerally minimal with the Devonian alkali basalt greenstonesbeing principally basaltic, while some of the Carboniferousalkali basalt greenstones tend towards mugearitic compositions.No intrusive acid differentiates have yet been reported. Apart from differences of magma type and minor differentiation,low-grade alteration or ‘spilitization’ has alsogoverned the geochemical variation seen in the greenstones.Spilitization caused (a) local redistribution of principallyCa (forming epidote-rich and calcite-rich patches) and Mg (formingchlorite-rich patches), together with their respective coherenttrace elements, and (b) the variable, but often limited, lossof Ca, Sr, K, Rb, and Ba from many bodies, together with a gainin Na and H₂O. Progressive hydration, however, caused a decreasein the oxidation ratio—a feature found to be common inmany spilitic suites and mainly governed by the relative distributionof chlorite versus epidote.     

Xenoliths of gneisses and the conformable,clot-like granophyres in the Marginal Border Group,Skaergaard intrusion,East Greenland

Results of this research and the earlier work of Wager and his colleagues indicate that contamination influenced the overall character of the Marginal Border Group of the Skaergaard intrusion. Precambrian gneisses were a major source of contamination and are identifiable as xenoliths in the Marginal Border. A variety of other xenoliths occur with a wide compositional range.The range of K _D values for partitioning of Fe and Mg in coexisting pyroxene pairs in xenoliths from various parts of the Marginal Border, and in hornfelses adjacent to it, is consistent with temperatures that rose steeply inward. Temperatures estimated on the basis of compositions of coexisting pyroxenes are also consistent with dehydration that exceeded the stability of amphibole (850°C). The texturally compatible association of some granophyres with gneissic xenoliths suggests that both formed during melting. These observations suggest that there was a similar range of temperature for formation of xenoliths and granophyres.The xenoliths of gneisses have bulk rock compositional features which indicate that lithophilic constituents were removed, causing an increase in basic constituents upon assimilation of the gneissic precursors. If we assume that granophyres and xenoliths of gneisses represent a consanguineous set formed by a fusion process from gneiss, enrichment and depletion factors for excluded trace elements are complementary and generally less than 10. Allowable enrichment in granophyres by crystal fractionation using the average for a large number of chilled marginal gabbro analyses as an initial composition, is also less than 10 for the same elements. The calculated low factors for enrichment of lithophile elements in granophyres by both mechanisms favor the hypothesis of partial melting of gneiss.     

Calculation of the composition of fractionated solid as deduced from chemical profiles in tholeiitic lava

The composition of solid precipitated from cooling magma is calculated from chemical differentiation profiles and an incompatible element monitor of fractionation. In a period of cooling, magma evolves to a derivative liquid and a solid; liquid compositions are obtained from the chemical profiles, and the fractionation interval via the incompatible element and the Rayleigh distillation equation. The composition of the solid, the only unknown, can be calculated for both equilibrium and continuous fractionation models. A sequence of low-K tholeiitic laves (basalt to rhyolite) from an Archean greenstone belt in northwestern Quebec are used to describe the procedure. Chemical trends in the lava exhibit strong iron enrichment followed by depletion and evolution to rhyolite. Trends in the calculated solid are amplified relative to lava, and bulk distribution coefficients, normative minerals, mg-number, and other parameters also show significant divergence. Oxidation state and amount of “trapped liquid” have important effects on solid composition. Solids have notably more basic chemical compositions than coexisting liquids, and this produces a “lag” in differentiation indices between the two. The high values of bulk distribution coefficients calculated for Fe, Ti and P (D ∼ 3 to 4+) in late stages of fractionation make it possible to produce the extremely Fe–Ti–P-rich solid fractions in layered igneous complexes from quartz-normative liquids. Received: 28 June 1999 / Accepted: 13 December 1999     

A discussion of Hunter and Sparks (Contrib Mineral Petrol 95:451–461)

Summary Silica-poor, iron-rich liquids like those estimated for Skaergaard by Wager and Brown are preserved in Labrador as chilled pillows. Iron-rich rocks in the Kiglapait Intrusion have likewise been produced by liquids poor in silica. There is no reason to expect such liquids in volcanoes, so their absence as lavas means nothing about their possible plutonic occurrence. Successful tests of the siliceous liquid hypothesis would include showing that the proposed rhyolitic liquids can produce the observed olivines in the range Fo 10–40. They would require showing that the observed low-f _{O 2} oxide mineral assemblages can occur at the high silica contents proposed, and that the silicate mineral assemblages and compositions can represent high-silica liquids. I doubt very much that these critical tests can be met.     

Feldspar geothermometry of the Hell Canyon Pluton,Boulder Batholith,Montana

The bulk compositions of the groundmass alkali feldspar from the Hell Canyon Pluton is 0.146mole% albite. The composition of the outermost zone of the oscillatory zoned plagioclase is 0.686 mole% albite, whereas the most calcic cores have a composition of 0.43 mole% albite. The structural state of the alkali feldspar is near orthoclase. Both composition of coexisting feldspars and structural state of the alkali feldspar are nearly constant throughout the pluton.Exsolved albite in the alkali feldspar have a composition of 0.965 mole% albite and the orthoclase host has a composition of 0.032 mole%. Singe crystal X-ray studies indicate that the albite intergrowths are coherent with the host.Equilibrium temperatures derived from the coexisting feldspar average 554 ° C; about 150 ° C, too low for the minimum solidus temperatures for reasonable emplacement pressures (2 kb). If this minimum solidus temperature is assumed, then the alkali feldspar has lost about 0.15 mole% albite. This loss was most likely caused by hydrothermal solutions associated with the crystallizing magma and equilibrated at about 550 ° C. However, based on the coherent albite intergrowths and the orthoclase structure state it can be inferred that the system was relatively free of volatiles below 500 ° C. Final equilibirium between orthoclase host and albite intergrowths occurred at about 311 ° C.     

Variation in the geochemistry of mantle sources for tholeiitic and calc-alkaline mafic magmas,Western Sunda volcanic arc,Indonesia

Quaternary lavas of the normal island-arc basalt—andesite—dacite association in the islands of Java and Bali range from those belonging to tholeiitic series over Benioff-zone depths of ～ 150 km to high-K calc-alkaline series over Benioff-zone depths of 250 km. More abundant and diverse calc-alkaline lavas are found over intermediate Benioff-zone depths. On average, basaltic lavas become slightly more alkaline (largely due to increased K contents) with increasing depth to the Benioff zone. Levels of incompatible minor and trace elements (K, Rb, Cs, Ba, Nb, U, Th, light REE) show a corresponding increase of almost an order of magnitude.Low average Mg-numbers (～ 0.52) and Ni and Cr abundances (15–25 and 35–60 ppm, respectively) of basaltic lavas suggest that few lavas representing primary mantle-derived magma compositions are present. Calculated primary basaltic magma compositions for most tholeiitic and calc-alkaline volcanic centres are olivine tholeiites with 15–30% ol. The single high-K calc-alkaline centre considered yielded transitional alkali olivine basalt—basanite primary magma compositions. These calculated magma compositions suggest that the percentage of mantle melting decreases with increasing depth to the Benioff zone (from >25 to <10%), while the corresponding depth of magma separation increases from ～ 30 to 60 km.Calculation of REE patterns for basaltic magmas on the basis of peridotitic mantle sources with spinel lherzolite, amphibole lherzolite or garnet lherzolite mineralogy, and model REE levels of twice chondritic abundances, indicates that change in the conditions of magma genesis alone cannot explain the observed change in light-REE abundances of basaltic lavas with increasing depth to the Benioff zone. Complementary calculations of the REE levels of mantle sources required to yield the average tholeiitic, calc-alkaline and high-K calc-alkaline basaltic magma indicate that light-REE abundances must increase from 2–3 to 7–8 times chondrites with increasing depth to the Benioff zone. The percentages of mantle melting favoured on REE evidence are lower than those indicated by major-element considerations.The observed variation in incompatible element geochemistry of mantle magma sources is thought to be related directly or indirectly to dehydration and partial-melting processes affecting subducted oceanic crust. The possible nature of this relationship is discussed.