MAGMATIC PROCESSES LEADING TO …MAGMATIC PROCESSES LEADING TO COMPOSITIONAL DIVERSITY IN IGNEOUS ROCKS: BOWEN (1928) REVISITED TIMOTHY L. GROVE† and STEPHANIE M. BROWN ABSTRACT

MAGMATIC PROCESSES LEADING TO COMPOSITIONAL DIVERSITYIN IGNEOUS ROCKS BOWEN (1928) REVISITED

TIMOTHY L GROVEdagger and STEPHANIE M BROWN

ABSTRACT Here we present a perspective on the evolution of thought on theorigin of compositional diversity in igneous rocks starting with the pioneer NormanLevi Bowen In pursing this question of diversity which was first clearly identified byDaly (1914) Bowen established the utility of experimentally determined phase equilib-ria as an aid to understanding geologic processes His work ultimately led him toattribute igneous rock diversity to a singular path of fractional crystallization Wesummarize the evolution of understanding acquired by petrologists during and afterBowenrsquos time Experimentalists beyond Bowen were crucial in furthering the under-standing of the origin of the diversity of igneous rocks by discovering that more thanone fractional crystallization path can occur in nature at a minimum differentiationcan either be dry (tholeiitic) or hydrous (calc-alkaline) We also reassess the fivealternative igneous processes that may give rise to compositional diversity that Bowenconsidered but found to be wanting These are magma mixing liquid immiscibilitySoret diffusion compositional gradients in liquids and contamination of magma byforeign material (assimilation) These processes play important roles in igneouspetrogenesis that is roles larger than Bowen envisioned yet fractional crystallizationremains fundamentally important

Keywords fractional crystallization magma mixing crustal assimilation liquidimmiscibility Soret diffusion experimental petrology phase equilibrium

introductionNorman Levi Bowen was the pioneer in establishing the field of experimental

igneous petrology and by doing so illustrated the value of laboratory investigations ofsilicate melt - crystal systems for understanding magmatic processes In the preface ofhis opus ldquoThe Evolution of the Igneous Rocksrdquo (Bowen 1928) he states ldquoWhile rocksthemselves remain the best aid to the discussion of their origin by fractional crystalliza-tion much light is thrown upon the problem by laboratory investigations of silicatemeltsrdquo Bowen was convinced that fractional crystallization was the most importantprocess that led to the compositional diversity exhibited by igneous rocks Fractionalcrystallization had initially gained traction after Charles Darwin flushed out thehypothesis in his 1844 book ldquoGeological Observations on the Volcanic Islandsrdquo but itwas still considered a subordinate igneous process (Daly 1914) Bowen stronglydisagreed further stating in his preface ldquoOn the question of the relative importance offractional crystallization as compared with other processes I can lay no claim to anopen mind But upon the relative importance of the various factors that may inducefractional crystallization there is much room for an open mindrdquo In chapter 1 ldquoTheProblem of the Diversity of Igneous Rocksrdquo Bowen offers up and systematically rejectsperhaps with unintended prescience other ldquorival processesrdquo that had been proposedalong with fractional crystallization magma mixing liquid immiscibility Soret diffu-sion compositional gradients in liquids and contamination of magma by foreignmaterial (assimilation)

In this review we will discuss what has been learned ldquofrom the rocks themselvesrdquoand ldquofrom the various factors that induce fractional crystallizationrdquo that derived fromBowenrsquos ground breaking experimental studies We will also discuss the evolution of

Department of Earth Atmospheric and Planetary Sciences 54-1220 Massachusetts Institute of Technol-ogy Cambridge Massachusetts 02139

dagger Corresponding author tlgrovemitedu

[American Journal of Science Vol 318 January 2018 P 1-28 DOI 10247501201802]

1

knowledge on the five aforementioned rival processes that Bowen brought up in chapter1 We will survey the evolution of the discovery of new types of compositional diversity andof the advances in experimental studies since publication of Bowenrsquos work It is curious towonder if Bowen had had the opportunity to study the full complement of igneous rocksknown today might he have reconsidered his perspective that those rival processes wereunimportant and secondary to fractional crystallization

Bowen and the Evolution of Experimental Igneous PetrologyBowen and his colleagues started experimental petrology from scratch only aided

by an absolute temperature scale the so-called Geophysical Lab Temperature scalewhich his immediate predecessors (Day and Allen 1904) had established Bowen usedthis temperature scale to construct the first petrologic phase diagrams His firstexperimental studies that established the phase equilibria of common rock formingminerals were performed at atmospheric pressure in air using Pt-wound resistancefurnaces Samples were placed in containers (usually a Pt tube) which were thensuspended on thin wires and heated to high temperatures Quenching was achieved bymelting the wire using an electric current which caused the sample to drop into a dishof water placed under the open furnace tube (this quenching technique is still widelyused today) Exploring phase space required for each experimental temperature ofinterest the preparation of many starting materials that were loaded into separate Ptcapsules and then simultaneously hung together Experimental run products wereexamined and identified with an optical microscope after placing crushed samples onglass slides in oils with known refractive indexes which allowed them to establish foreach bulk composition whether it was above its liquidus in a crystal liquid field orbelow the solidus

The first compositional space Bowen studied for his Ph D thesis was the binarysystem Nepheline ndash Anorthite (Bowen 1912 MIT) He then studied the binary solidsolution Albite ndash Anorthite (Bowen 1913) the system MgO ndash SiO2 in which heestablished the revolutionary peritectic melting behavior of Enstatite to Forsterite Liquid (Bowen and Andersen 1914) and the ternary system Forsterite ndash Diopside ndashSiO2 (Bowen 1914) In 1915 his colleague Olaf Andersen published the ternarysystem Forsterite ndash Anorthite ndash Silica (Andersen 1915) and Bowen published thephase diagram for Diopside ndash Albite ndash Anorthite (Bowen 1916) Bowen was able toapply these carefully established experimental melting relations to observed naturaligneous rock mineral associations and compositions to forge the concepts thatappeared in his 1928 book ldquoThe Evolution of the Igneous Rocksrdquo It seems a miracu-lous accomplishment for one person to take the information in this handful of simplesystem phase diagrams and create a fully developed theory for the workings ofmagmatic processes

Bowen developed a whole new laboratory-based sub discipline of the EarthSciences Experimental petrology continues to evolve but our debt to Bowen isenormous As new tools and techniques became available the field of experimentalpetrology moved on from simple systems to the study of chemically complex naturalsystems A major step forward was the application of the electron probe micro-analyzer(EPMA) to chemical characterizations of experimental run products (Green andRingwood 1967 Kushiro 1972) For the first time crystals and coexisting meltcompositions could be characterized in the experimental charge and a complete phasediagram could be built using only a single starting composition Another breakthroughwas the development of techniques for buffering oxygen fugacity in experiments atone atmosphere and at higher pressures (Nafziger and others 1971 Sato 1971Huebner 1971) A further advance was to include iron in 1-atmosphere experimentsThis challenge was first solved by several groups (for example Walker and others1977) during experimental investigations of the returned Apollo lunar samples They

2 TL Grove and SM BrownmdashMagmatic processes leading to

prevented iron loss by using metallic iron capsules held in evacuated silica glass tubesThis new approach allowed accurate measurements of FeO in melts and mineralsunder the extremely reducing conditions of lunar mare differentiation Soon afterApollo techniques for suspending droplets of natural magmas on FePt alloys weredeveloped (Walker and others 1979) and the calibration of the solution properties ofthese alloys allowed investigations of natural silicate systems to be undertaken over awide range of oxygen fugacities (Grove 1981)

The influence of dissolved volatiles was another important variable that Bowenexplored late in his career in his study of the origin of granites (Tuttle and Bowen1958) Technology developed by Tuttle (1948) allowed exploration of the effects ofdissolved magmatic H2O on phase relations up to 04 GPa and temperatures up to800 degC The groundbreaking work by Tuttle and Bowen (1958) on the melting of thesimplified granite system (Quartz ndash Albite ndash K-Feldspar) in the presence of waterdemonstrated the importance of this volatile component on phase relations Experi-ments with dissolved H2O within the higher temperature range relevant for crystalliza-tion of basalt magma became possible through the development of the internallyheated pressure vessels (Yoder 1950 1952) which were applied to measuring thesolubility of H2O in iron-free silicate melts (Burnham and Jahns 1962) The firstexperiments on the solubility of H2O in iron-bearing basalt melts were made byHamilton and others (1964) The next step of controlling oxygen fugacity in iron-bearing and H2O-bearing experiments was first done by Holloway and Burnham(1972) who used a Shaw membrane (Shaw 1967) (also they were able to report bothmineral and melt compositions analyzed by EPMA) Williams (1966 1968) developedan externally heated cold seal pressure vessel (the TZM) that can be used at tempera-tures up to 1200 degC and pressures up to 300 MPa The TZM began to be used forH2O-bearing experiments at controlled oxygen fugacity in iron bearing melts (Dixonand Rutherford 1979) Through these experimental developments it became possibleto examine the influence of oxygen fugacity dissolved magmatic volatiles andpressure on the crystallization of natural magmas

Since Bowenrsquos time high pressure ( 300 MPa) experimental techniques havesignificantly expanded our understanding of the mineralogy of the Earthrsquos mantle andthe diversity of magmas produced by mantle melting Either developed or applied toEarth Science problems after Bowenrsquos career ended these methods include thepiston-cylinder (Boyd and England 1960) the diamond anvil cell (DAC Merrill andBassett 1974) and the multi-anvil device (Kawai and Endo 1970) DAC and multi-anvil devices derived from the simple-squeezer ndash opposed anvil devices developed byPercy W Bridgman (Bridgman 1952) and eventually their development allowed forthe attainment of the high temperatures and ultra-high pressures necessary forsimulating lower mantle conditions Shallower pressures could be achieved using thepiston cylinder and multi-anvil devices which were used to investigate the significanteffects of pressure (Green and Ringwood 1967 OrsquoNeill 1981 Takahashi 1986Walter 1998) and variable H2O content (Gaetani and Grove 1998) on the composi-tions of mantle melts and the depth of critical phase transitions in the Earthrsquos uppermantle Ringwood and Major (1970) determined the stability relations of majorite andringwoodite and associated the appearance of these minerals with the 410 and 520 kmseismic discontinuities At pressures then too great to conduct experiments Ringwood(1962) hypothesized the upper ndash lower mantle transition zone seismic discontinuity at670 km was caused by a phase transition from a dominantly ringwoodite mantleassemblage to one containing bridgmanite (Ringwood called it MgSiO3 perovskite)and ferropericlase Continued evolution of the opposed anvil and DAC methodsallowed Ringwoodrsquos hypothesis to be directly tested as obtainable experimental pressuresincreased making the reproduction of the mineralogy throughout the Earthrsquos mantle

3compositional diversity in igneous rocks Bowen (1928) Revisited

possible Liu (1974 1975) first confirmed bridgmanite and ferropericlase as the stableassemblage below the 670 km discontinuity Improvements in DAC technology havemore recently led to the synthesis of a post-perovskite phase (Murikami and others 2004Oganov and Ono 2004) that may be a phase transition linked to the Drdquo discontinuity at thebase of the lower mantle (2900 km) right above the core ndash mantle boundary

ldquoWhile Rocks Themselves Remain the Best Aid rdquoBowen was limited in 1928 at the time of writing ldquoThe Evolution of the Igneous

Rocksrdquo by the amount of information available on the compositional variabilityexhibited by igneous rocks which only included major element data (and no traceelement or isotopic data) At the time of publication of Bowenrsquos book Henry SWashington had compiled basalts from the British Tertiary Province the Deccan Trapsin India Hawaii and various Pacific ocean islands and had published average composi-tions in his paper ldquoThe average chemical composition of igneous rocksrdquo (Clark andWashington 1922) Most of the 5519 analyses were silicandashrich samples By contrasttoday the PetDB database contains 1174126 bulk rock analyses and 655267 analysesof volcanic glasses as of October 1 2016 and includes a range of iron ndash rich to silica ndashrich samples Also during Bowenrsquos time the Skaergaard intrusion in east Greenlandhad not yet been discovered (Wager and Deer 1939) and basalts from mid-oceanridges had not yet been dredged from the ocean floor (Muir and others 1964)although Bowen understood that the oceanic crust was likely more mafic than thecontinental crust Plate tectonic theory had not been conceived and would not beuntil the 1960rsquos

Given that there were only a small number of rocks that had been analyzed at thetime Bowen focused on 335 subalkaline basalts for which compositions had beendetermined and that he thought of as potential parental magmas In the end Bowenrsquosskepticism of using rocks that might not represent liquids caused him to base hisdiscussion of the liquid line of descent followed during fractional crystallization ofnatural magmas to rock associations from subduction zones (such as Lassen Californiaand Katmai Alaska) because they were at the time the only rocks that clearlypreserved liquid compositions (Bowenrsquos glassy rocks) He also extensively discussedthe liquid versus cumulate compositional relationship between the lavas and intrusiverocks respectively of the British Tertiary province Ironically this particular rockseries subsequently was shown to represent a different differentiation trend (a tholei-itic trend) than the one Bowen discussed (a calc-alkaline trend) due to processes helargely dismissed modification by crustal assimilation and magma mixing (Sparks1988 Kerr and others 1999)

A critical assumption made by Daly (1914) and Bowen (1928) was the existence ofa single primary magma whose composition was effectively the same everywhere (thatis basaltic not granitic) This view persisted into the mid 1960rsquos (Engel and others1965) It was not until high pressure - high temperature experimental studies onprimary basalts of variable composition revealed the existence of a low pressurethermal divide which separated silica under-saturated (alkali) basalts from sub-alkalineand tholeiitic basalts (Yoder and Tilley 1962) The thermal divide causes liquids oneither side of the divide to follow distinct down temperature crystal fractionation pathsthat produce different residual liquid compositions This discovery conflicts withBowenrsquos hypothesis that both silica-saturated and silica under-saturated magma couldbe produced by means of fractional crystallization from essentially the same basalticparent although he acknowledged in his preface that this conclusion should beldquoregarded as resting on a less certain foundationrdquo because it lacked direct experimen-tal evidence He reasoned that silica under-saturated residual liquids could be pro-duced under certain circumstances either by minimal early olivine-fractionationsome later pyroxene fractionation and the incongruent melting of orthoclase or by


quartz fractionation in the presence of water Bowenrsquos rationale here exemplifies histhinking that the only ldquoflexibility in the course of crystallizationrdquo he would entertainwas in the extent of efficient fractionation (that is non-fractional crystallization couldoccur) which he did not consider to be a true alternate liquid line of descent

In 1967 Green and Ringwood addressed the question as to the origin of theprimary liquids on either side of the thermal divide by using high-pressure experimen-tal petrology and thereby demonstrated the existence of a range of primary mantle ndashderived melts whose compositions changed as a function of mantle melting depthThey found that at greater depths in the mantle silica-undersaturated alkali-olivinebasalts were generated while at shallower depths olivine tholeiites and silica - saturatedtholeiites were produced Therefore there could be many different mantle-derivedprimary magmas of differing chemical composition that might be expected to undergodifferent fractional crystallization processes Thus this established that silica-undersaturated basalts follow a distinct alkalic differentiation series on the nepheline-normative side of the thermal divide that was not recognized in Bowenrsquos time In otherinstances the wide range of compositional variability in mid-ocean ridge primarybasalts records mantle melting over a range of temperatures and depths (Klein andLangmuir 1987 Kinzler and Grove 1992) yet these diverse primary magmas followthe same low-pressure tholeiitic differentiation trends

It is prudent for us to keep in mind that the debate on the nature of primarymagmas during Bowenrsquos time centered on if they were maficbasaltic or felsicgraniticin composition rather than if basaltic primary melts could have significantly variablecompositions In truth there was not enough data for petrologists at the time torigorously consider this variable and Bowen himself defined a basaltic magma as anymagma that ldquoon rapid crystallization gives rise to a rock having intermediate plagio-clase and clinopyroxene as its principle constituentsrdquo So while Bowen was correct inthinking that most igneous rocks are indeed ultimately derived from basaltic primarymelts his assumption that primary basalts are globally uniform and that variations inprimary basalt composition would not lead to appreciable differences in the course offractional crystallization was clearly an oversimplification In 1928 the theory of platetectonics was decades away and so Bowen would not have had much appreciation forthe fact that mantle melting occurs by different processes in different locations (that isBowen thought that adiabatic decompression melting was the only important mantlemelting process) Accordingly he would also not have realized that his strict criteria fortrustworthy data of which there was not much of to begin with might inadvertentlylead to him to overlook critical igneous variations by only focusing on variations atsubduction zone or plateau basalts settings Importantly subduction zone primarymagmas are distinct from other primary magmas in that they contain significantamounts of dissolved H2O (Sisson and Grove 1993 Gaetani and Grove 1998) andtheir major element compositions can range from olivine tholeiite to quartz tholeiiteIt is differentiation of these H2O-bearing basaltic magmas at crustal levels that leads tothe development of the calc-alkaline trend While Bowen and others appreciated therole of H2O on phase equilibria in rocks with hydrous minerals [that is typically morefelsic rocks for example Tuttle and Bowen (1958)] this perspective did not extend tobasalts and it was not until Sisson and Grove (1993) discovered that mafic magmascould be hydrous without crystalizing hydrous minerals that an appreciation forhydrous basaltic parental magmas and their effect on differentiation trends wasestablished (Carmichael 2002)

fractional crystallization of basalt under anhydrous (dry) conditionsthe tholeiitic trend

The analyzed lava suites from Lassen California and Katmai Alaska showed atrend of iron-depletion and silica-enrichment that Bowen interpreted as the liquid line


of descent followed by a crystallizing basaltic melt His colleague Clarence NormanFenner disagreed arguing that the normal course of fractional crystallization shouldbe one of iron-enrichment (Fenner 1926 1931 1948) Fenner based his reasoning onplutonic rocks by separately analyzing the chemical compositions of the bulk rock andthe pyroxene contained within Fenner found that the pyroxene always had a higherMgFe than the whole rock and concluded that a differentiation trend that resultedfrom crystallization of plagioclase pyroxene should show iron-enrichment Furthersupport for iron-enrichment during fractional crystallization came from analysis oflayered rocks from the Skaergaard intrusion by Wager and Deer (1939) Wager andDeer concluded ldquothe trend of fractional crystallization of the Skaergaard magmaduring the early and middle stages supports Fennerrsquos view that during fractionalcrystallization of basalts there is absolute enrichment in ironrdquo Wager and Deer (1939)also pointed out that after 95 percent solidification and extreme fractionation thesilica contents of the residual magmas did not ldquorise beyond the limits of normal basicrocksrdquo They went on to state ldquothe normal calc-alkaline series of igneous rocks isfrequently considered to represent the result of crystal fractionation of basalt magmaFrom the evidence of the Skaergaard it appears that crystal fractionation of basalt leadsto ferrogabbro and not to intermediate rocks of the calc-alkaline seriesrdquo Both Fennerand Wager and Deer favored mixing of basalt and silicic magma to explain thecalc-alkaline series which we will discuss in the Magma Mixing and Assimilation sectionfound below

Suites of igneous rocks that exhibit iron-enrichment at constant SiO2 becameknown as the tholeiitic series (Daly 1952) Because Bowen only relied on liquids forevidence he was not convinced of the iron-enrichment differentiation trend found inthe plutonic Skaergaard environment The first true liquids measured that preservediron-enrichment tends were basaltic lavas from Iceland (Carmichael 1964) and theGalapagos (McBirney and Williams 1969) years after Bowenrsquos passing in 1956 Kuno(1965) analyzed the products of in-situ differentiation within single lavas flows (that issegregation veins containing from 14 to 18 wt FeO) that followed the iron-enrichment trend in lava flows from Hawaii Japan and California The most spectacu-larly preserved tholeiitic liquid line of descent is associated with the GalapagosSpreading Center (Byerly 1980 Perfit and Fornari 1983 Fornari and others 1983Juster and others 1989) Submarine pillow lavas erupted from the Galapagos Spread-ing Center are mantled by glassy chill margins that would have immediately gotten theattention of Bowen The maximum FeO content within this glass suite is 18 weightpercent FeO at 37 percent MgO (fig 1) Galapagos Spreading Center magmas rangefrom basalt to Fe-Ti rich basalt andesite and rhyodacite

Juster and others (1989) performed one-atmosphere experiments on a Galapagosparental composition over a range of oxygen fugacities from QFM (quartz-fayalite-magnetite buffer) to NNO2 (two log units above the Nickel ndash Nickel oxide buffer)The liquidus of the starting composition is saturated with olivine (oliv) and plagioclase(plag) After a small amount of oliv plag crystallization augite joins the crystallizingassemblage (fig 2) Olivine plagioclase and augite co-precipitate over a short tempera-ture interval before pigeonite joins the crystallizing assemblage at a peritectic reactionboundary At this reaction boundary olivine melt react to form plagioclase augite pigeonite Olivine disappears a Fe-Ti rich spinel phase saturates andcrystallization continues with the assemblage plag augite pigeonite spinel

The resulting saturation boundaries inferred from these experiments along withanalyses of the glassy chill margins of Galapagos Spreading Center lavas from 85 degW areshown in figure 1 As is evident in the variation diagrams crystallization leads to strongenrichment in FeO at nearly constant SiO2 and decreasing Al2O3 that is the result ofearly and modally abundant crystallization of plagioclase (Grove and Baker 1984) At


Fig 1 (A) Saturation boundaries determined in experiments on a Galapagos lava (Juster and others1989) are plotted along with glass compositions measured in mid-ocean ridge basalts from 15 segmentsalong the Galapagos Spreading Center (Gale and others 2013) in the pseudoternary projections Olivine ndashAugite ndash Quartz and Olivine Plagioclase ndash Quartz using oxygen units (see Tormey and others 1987) (B)Experimentally determined liquid lines of descent from Juster and others (1989) are plotted with glassanalyses from the same 15 segments along the Galapagos Spreading Center as in figure 1A (Gale and others2013) on MgO vs SiO2 MgO vs Al2O3 MgO vs TiO2 and MgO vs FeO variation diagrams


low-pressure and under anhydrous conditions plagioclase dominates the crystallizationassemblage (oliv plag 3070 by weight) Once augite joins as a crystallizing phaseplagioclase still dominates the precipitating assemblage (oliv plag augite 166024by weight) and iron-enrichment continues until FeO has increased to 16 weightpercent Iron-enrichment continues at the reaction boundary as iron-rich olivinedissolves and silicate phases with lower-iron contents (plagioclase augite pigeo-nite) crystallize This iron-enrichment continues until olivine reacts out andor Fe-Tirich spinel joins the crystallizing assemblage at which point SiO2 and Al2O3 increasewith decreasing MgO FeO and TiO2 until the residual melt evolves to a rhyodacitecomposition

The compositional variability followed by mid-ocean ridge basalts (MORBs)during low pressure anhydrous fractional crystallization as exemplified by the Galapa-gos Spreading Center lavas (fig 1) is now well established as the tholeiitic oriron-enrichment trend The first research on the compositions and petrographiccharacteristics of MORB did not appear until the work of Muir and others (1964) Thesubsequent experiments on MORB liquids showed that primitive magmas wouldcrystallize abundant plagioclase resulting in the inferred iron-enrichment trend The

1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18

Tormey and others (1987)POO82N2

Juster and others (1989)Oliv

Plag

Augite

Pig

Ilm

Tmt

Qtz

Fig 2 Experimentally determined liquid lines of descent at 1 atm for the primitive KANE fracture zonebasalt (All96-18) and the Galapagos basalt (POO82N2) The black diamonds indicate the temperature of anexperiment All 96-18 had a higher liquidus temperature (it is more primitive) and crystallizes olivineolivine plagioclase olivine plagioclase augite POO82N2 initially crystallizes olivine plagioclaseand soon reaches the peritectic reaction boundary olivine liquid plagioclase augite pigeoniteOnce all the olivine has been consumed ilmenite joins the crystallization assemblage followed bytitanomagnetite (tmt) and then quartz


pre-MORB petrologists who discovered iron-enrichment did so indirectly by calculat-ing the residual liquids that were generated by in situ fractional crystallization inSkaergaard-like magma chambers Figure 3 shows several estimates of the iron-enrichment of residual liquids inferred for the Sakergaard from Wager and Brown(1968) McBirney and Naslund (1990) Tegner (1997) Thy and others (2009) Toplisand Carroll (1995) and Hunter and Sparks (1987) Note in figure 3 that all of theseestimates of iron-enrichment cross into the experimentally determined field forsilicate liquid immiscibility (Charlier and others 2013) Also note that the tholeiitictrends all show Fe-enrichment during the bulk of crystallization At 90 to 95 percentcrystallization there is a rapid Fe-depletion and Si-enrichment trend that passesthrough the field of silicate liquid immiscibility We will return to this issue in theMagma Mixing and Assimilation section found below

Fractional Crystallization of Basalt under Hydrous (Wet) Conditions The Calc-Alkaline TrendThe compositional evolution of sub-alkaline lavas during fractional crystallization

considered by Bowen to be the true and only path of liquid evolution is now referred toas the calc-alkaline trend It is characterized by depletion in FeO with increasing SiO2Yet inescapable field evidence from the Skaergaard led petrologists to increasinglyrecognize that many if not most basalts followed a tholeiitic Fe-enrichment pathduring fractional crystallization and so they looked for alternative ways to produce thecalc-alkaline trend In a review Kennedy (1955) suggested that water dissolved in themelt might be important in promoting calc-alkaline differentiation and therebyapparently resolved the Bowen ndash Fenner controversy Kennedy thought that theaddition of H2O would cause oxidation of the magma and the early precipitation of aFe-oxide mineral could produce the silica enrichment ndash iron depletion trend Mean-while Osborn (1959) Presnall (1966) and Roeder and Osborn (1966) carried outexperiments on dry FeO-bearing basalt analog systems at 1-atm over a range ofoxidizing conditions and found that an iron-depletion trend would only occur under

McBirney and Naslund 1990

Jakobsen and others 2005iron-rich immiscible liquid

Tegner 1997

Thy and others 2009

Wager and Brown 1968

Toplis and Carroll 1995

Hunter and Sparks1987

40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for

silicate liquid immiscibility

Jakobsen and others 2005silica-rich immiscible liquid

Fig 3 SiO2 vs FeO variation diagram (from Charlier and others 2013) showing liquid lines of descentthat have been proposed for the Skaergaard intrusion East Greenland The two stars (Jakobsen and others2005) are compositions of immiscible melts measured in Skaergaard melt inclusions in apatite Gray shadedregion is the field of liquid immiscibility experimentally determined by Charlier and Grove (2012)


very very oxidizing conditions - conditions much more oxidizing than those found inthe Earthrsquos crust Although they produced a negative result (regarding Fe-depletiontrends in oxidized dry systems) they proved that there must be more than one type ofliquid line of descent

Sisson and Grove (1993a 1993b) performed the first experiments that examinedthe effects of H2O on basalt crystallization at 100 and 200 MPa and at oxygen fugacitiesrelevant for the crust (Nickel ndash Nickel oxide buffer or NNO) They found thatcrystallization in the presence of dissolved H2O at crustal pressures has three dramaticeffects (1) the liquidus temperature is lower (2) the order of phase appearancechanges and (3) the compositions of the crystallizing minerals change systematicallyUnder one-atmosphere anhydrous conditions at the QFM buffer a primitive high-magnesian andesite (Krawczynski and others 2012) crystallizes olivine as the liquidusphase at 1230 degC (fig 4) Plagioclase joins the crystallizing assemblage at 1200 degCfollowed by orthopyroxene at 1195 degC and then augite at 1180 degC Olivine reacts outand plagioclase augite spinel co-crystallize down to the lowest temperatureinvestigated (1110 degC) This crystallization results in an iron-enrichment trend Under200 MPa H2O-saturated conditions at the NNO olivine is the liquidus at 1160 degCaugite spinel join the crystallizing assemblage at 1080 degC and orthopyroxene (opx)appears at 1030 degC Olivine reacts with the liquid and plagioclase appears at 990 degCAmphibole joins the crystalizing assemblage at 910 degC For this composition the effect

900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc

Medard and Grove 2008

Krawczynski and others 2012

and Grove and others 2003Amph + Olv coexisting

olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870

olv=846cpx=852opx=870

cpx= 827opx= 811amph= 799


olv=903

olv=909

olv=862olv=

817olv=903

Fig 4 Pressure ndash Temperature diagram (after Krawczynski and others 2012) showing H2O saturatedphase appearance sequence for a primitive magnesian andesite from Mt Shasta California at conditions ofthe NNO buffer Magnesium number (Mg molar MgO(MgO FeO)) compositions of the Fe-Mgsilicates are shown for each experiment


of H2O is to suppress plagioclase appearance by 210 degC The early crystallization ofolivine augite and opx and the delay in plagioclase crystallization leads to irondepletion and silica enrichment (see fig 5 in Krawczynski and others 2012)

The effects of H2O on phase appearance relative to the dry iron-enrichmenttrend are critical to the characteristic early calc-alkaline silica-enrichment trendthat develops during fractional crystallization (Sisson and Grove 1993a 1993b)

Fig 5 Pseudoternary projections (see Tormey and others 1987) of the differences between saturationboundaries that control basalt fractional crystallization under anhydrous (dry) conditions at low pressure(1-atm black) at the QFM buffer and H2O-saturated (wet) conditions at 200 MPa (gray) at the NNO bufferLower left is Olivine - Clinopyroxene ndash Quartz the top projection is the upper half of Plagioclase ndash Olivine ndashClinopyroxene and right hand triangle is the upper half of the Olivine - Plagioclase ndash Quartz subprojectionUnder dry conditions a parental basalt liquid (black star) crystallizes olivine as a liquidus phase and saturatesearly on with plagioclase (d1 on the pseudoternaries) Plagioclase and olivine both crystallize and the liquidundergoes iron-enrichment at constant SiO2 contents until the liquid saturates with augite (d2) where ironenrichment continues (see fig 2) Under water-saturated conditions the same parental basaltic liquid (blackstar) also crystallizes olivine as a liquidus phase but for longer and does not quickly saturate with plagioclasewhich rapidly decreases FeO and MgO in the residual melt This happens because the effect of H2O is toshrink the plagioclase primary phase volume and expand the olivine and augite phase volumes Thefractionating melt next saturates with augite (w1) and then reaches plagioclase olivine augitesaturation (w2) and follows a path of SiO2 enrichment


Plagioclase (an iron-poor mineral) stability is significantly depressed causing it to nolonger be an early crystallizing phase when modest amounts of H2O (3ndash6 wt ) arepresent in the melt at upper crustal pressures Instead olivine and augite crystallize atthe liquidus and these feromagnesian minerals deplete the melt in FeO and MgOwhile enriching it in SiO2 When plagioclase appears at lower temperatures it is inlesser proportion to the Fe-Mg silicates and it is very anorthite-rich (that is CaO-richSiO2-poor and Na2O-poor) further promoting SiO2- and alkali- enrichment All silicatephases are affected by the presence of water in comparison to anhydrous conditionsplagioclase appearance is lowered by over 200 degC and olivine and augite appearance islowered by 100 to 150 degC (fig 4) However the addition of H2O does not have a similareffect on the appearance temperature of iron-rich oxide phases that is the appear-ance temperature remains the same in both anhydrous and hydrous melts In anhy-drous systems oxides become stable 150 to 200 degC below the liquidus which is thesame temperature offset the hydrous liquidus is depressed by compared to theanhydrous liquidus resulting in stable oxides near the liquidus Thus silica-freeiron-bearing oxides fractionate earlier further causing silica-enrichment and iron-depletion

The changes in the saturation boundaries caused by the addition of H2O areillustrated in figure 5 in pseudo-ternary mineral component projections The plagio-clase primary phase volume shrinks at high melt H2O contents Thus a basaltic meltthat has olivine and plagioclase as near-liquidus phases under anhydrous conditionsfinds itself in the olivine only primary phase volume under hydrous conditions Suchmelts crystallize olivine until they reach the olivine augite saturation boundary atwhich point they crystallize these two phases until multiple saturation with olivine augite plagioclase occurs Liquids residual to hydrous fractional crystallizationthereby become enriched in normative plagioclase and then in SiO2

Although Bowen did not know about the dramatic effects that H2O could havehad on phase equilibria one can imagine that he would have been fascinated bythe effects on mineral composition phase stability and appearance temperature Thevapor-saturated phase relations shown in figure 4 for a primitive basaltic andesite(Krawczynski and others 2012) show that increasing dissolved water contents of themelt by increasing pressure at vapor-saturation leads to further dramatic changes inthe mineral appearance temperature and mineral phase stability Increased H2Osolubility at 500 MPa (melt H2O of 8 wt ) continues to destabilize plagioclaseappearance (300 oC) and increase the stability field and change the composition ofamphibole At 800 MPa (melt H2O of 12 wt ) the olivine liquidus temperature isdepressed by 150 degC and augite and opx appear simultaneously 10 degC below theliquidus followed by a high-magnesian amphibole 60 degC below the liquidus Krawc-zynski and others (2012) show that the liquid lines of descent defined by crystallizationat 500 and 800 MPa under vapor-saturated conditions are difficult to distinguish from200 MPa crystallization paths primarily because the compositional effects of Mg-Fesilicate crystallization are similar (olivine ndash augite ndash opx) and thus the major elementmelt compositional paths are not diagnostic of crystallization pressure at vapor-saturation

Conversely the compositions of amphibole olivine opx and cpx are very sensitiveto crystallization pressure A prime example of this is the Mg variations of opx augiteand amphibole in the first appearing minerals in vapor-saturated experiments from200 to 800 MPa (fig 4) Amphibole Mg changes from 747 at 200 MPa to 799 at 500MPa and to 833 at 800 MPa Krawczynski and others (2012) used this variation inmineral composition (along with its variation with fO2) to calibrate an amphiboleMg geobarometer ndash hygrometer When this barometer-hygrometer is applied toamphiboles preserved in mixed andesites from Mt Shasta volcano in N California


they are found to have crystallized over a depth range from 276 to 948 MPa (fromshallow crust to the base of the crust beneath Mt Shasta) with the higher pressurehigh-Mg amphiboles having crystallized from melts that contained up to 14 to 15weight percent H2O Similar high-Mg amphiboles occur in lavas from Mt St HelensRedoubt El Reventador and Soufriere Hills (Ridolfi and others 2010) This suggeststhat meltingcrystallization processes in subduction zone magmas occur over a widerrange of magmatic water contents

bowenrsquos rival igneous processes

Magma Mixing and AssimilationBowen on p 1 of his textbook ldquofound [magma mixing] to fail so completelyrdquo that

it could not even be considered as an important igneous processes In the context ofthe time this remark is understandable because hypotheses were in play that explainedthe compositional variability in igneous rocks by mixing of two primary end-membermagmas basalt and rhyolite (for example Fenner 1937 1948) However there werealso some excellent discussions and descriptions of magma mixing and mingling inlavas from the Lassen region (Finch and Anderson 1930) Hakone volcano in Japan(Kuno 1936) and the silicic ash flow tuffs of the San Juan Mountains in Colorado(Larsen and others 1938a 1938b) where compositional and textural evidence convinc-ingly showed that magmas of contrasting composition had intermixed These authorsused Bowenrsquos work on the appearance sequence of minerals to argue for the mixing ofmafic and silicic melts Bowen seems to have paid no attention to these works and theigneous petrology community fell silent on this topic for nearly four decades Magmamixing was ldquorediscoveredrdquo in the 1970rsquos at subduction zone volcanoes (Anderson1976) in mid-ocean ridge basalts (Dungan and Rhodes 1978) and as a trigger forexplosive silicic eruptions (Sparks and others 1977)

Bowen discussed assimilation exhaustively and laid out the thermal energybalance that would be necessary for assimilation to occur He pointed out that mostmafic magmas are near or below their liquidi when they encounter crustal xenoliths orwallrocks in crustal magma chamber conduits and that the heat for assimilation mustcome from the latent heat of crystallization in the solidifying magma bodies them-selves Given this constraint Bowen did acknowledge that assimilation could indeedoccur for example he ended his chapter 10 on assimilation with the conclusionldquomagmas may incorporate considerable quantities of foreign inclusionsrdquo but arguedthat ldquoit is doubtful whether the presence of foreign matter is ever essential to theproduction of any particular type of differentiaterdquo By this Bowen meant that the resultsof assimilation would be to move the magma further along the same path it would havefollowed anyway because there was only one course of fractional crystallization Whatpetrologists discovered when examining the evidence preserved in erupted lavas wassomething quite different

At Medicine Lake a Cascade rear-arc volcano in N California Grove and others(1982 1988) Baker and others (1991) and Donnelly-Nolan and others (1990) carriedout field geochemical and experimental petrology studies of an observed calc-alkalinetrend preserved in a subset of the lavas erupted there It turns out that Medicine Lakein addition to erupting wet primitive basalts (Kinzler and others 2000) erupted dryprimitive high-alumina olivine tholeiites (HAOT) and experimental studies of thesemagmas under anhydrous low-pressure conditions showed that they crystallized alongiron-enrichment trends The liquidus phases olivine plagioclase crystallize untilabout 30 weight percent crystallization at which point augite also begins to crystallizeUltimately the liquid line of descent reaches a reaction boundary where olivine liquid react to form plagioclase augite pigeonite thereby continuing the trend ofiron-enrichment at low SiO2 contents However the 3000-year-old Burnt Lava andesitic


lava flow (033 km3) at Medicine Lake Volcano which contains both inclusions ofprimitive HAOT and melted granitic crust (Grove and others 1988 Donnelly-Nolanand others 2016) shows a calc-alkaline trend with between 568 to 58 weight percentSiO2 and not a tholeiitic trend It also contains a texturally and compositionallycomplex assemblage of minerals including Mg-rich and Fe-rich olivine (Fo88 andFo70) An-rich and An-poor plagioclase (An85 and An30) and Fe-rich augite (Mg 72) Grove and others (1988) used petrologic and geochemical evidence to model theprocesses that led to the formation of this mixed andesite They calculated a value of r(r mass assimilatedmass fractionated) of 135 which requires a large amount ofassimilation of granitic crust to produce the andesite of Burnt Lava In contrastestimates of the thermal energy budget for the shallow Burnt Lava magma reservoirpredict much lower r values near 025 (DePaolo 1981) Therefore the magma bodyitself could not supply enough heat to melt the wallrock alone requiring an opensystem assimilation process in which heat and mass transfer are decoupled Two viablerealistic processes using this constraint envisioned by Grove and others (1988) forforming the Andesite of Burnt Lava are illustrated in figure 6 as either intrusivegeometries of dike swarms (fig 6A) or in a magma chamber (fig 6B) In both casesdry primitive basaltic magma is emplaced at shallow crustal depths into subvolcanicgranitic country rock where it undergoes fractional crystallization to an iron-richtholeiitic magma The Mg of the crystallizing iron-rich augite preserved in the mixedandesite indicates that the FeO content of the iron-rich tholeiitic magma at that pointwould have reached 132 weight percent The latent heat released from crystallizationto this iron-rich differentiated liquid (63 wt crystallized) heats up and melts thesurrounding granitic wall rock These two subjacent magmas (the iron-rich tholeiiteand the melted granitic rock) then mix to form the hybrid Burnt Lava andesite whenanother batch of primitive HAOT is injected into the system Accordingly in this typeof model magma evolution is an open system characterized by fractionation assimila-tion and mixing with new inputs of recharged primitive basalt (FARM) in which heatand mass transfer are separated in time

Naturally this occurs because volcanic systems are episodic and magma chambersexperience repeated injections of primitive magmas Consequently the physicalprocesses of recharge and mixing set the mass fraction of the three-componentmixture resulting in a misleadingly large r value Models that assume a continuousprocess of assimilation occurring during fractional crystallization do not work theresimply is not enough heat

Another well-preserved example of the FARM process at Medicine Lake is foundin the lavas of the 10500-year-old Giant Crater lava field (44 km3 Donnelly-Nolan andothers 1990 Baker and others 1991) Mapping sampling petrological and geochemi-cal analyses reveal a process similar to that inferred for the Burnt Lava case Six eruptedunits preserve evidence for fractionation assimilation recharge and mixing Theearliest Group 1 lavas erupted are the most evolved (535 SiO2 and 53 MgO) andthe last erupted Group 6 HAOT lavas are the most primitive (48 SiO2 and 105 MgO) The most contaminated eruptive phase (Group 1) contains a texturally andcompositionally complex assemblage of minerals including Mg-rich and Fe-rich oliv-ines plus An-rich and An-poor plagioclase crystals Partly melted granitic xenoliths andinclusions of ferrobasalt containing olivine plagioclase and augite are also preservedThe latter phenocrysts record crystallization from an iron-rich tholeiitic liquid contain-ing 15 weight percent FeO The calc-alkaline compositional trend exhibited by the sixeruptive units is shown schematically in figure 7 along with the tholeiitic fractionationtrend that is required to produce the olivine plagioclase augite inclusions foundin the Group 1 lavas The open circles show a continuous assimilation and fractional


crystallization model with r 15 and the squares are models of a FARM process thatclosely matches the observed compositional variations

Major and trace element consequences of fractional crystallization can be exten-sively overprinted and modified in open magmatic systems Post-mixing crystallizationin hybrid magmas may move liquids along different saturation boundaries and thecrystallization paths may involve different mineral assemblages than would be ex-pected during closed-system fractional crystallization (Grove and others 1982) Crustalassimilation and mixing cause incompatible element abundances in magmas to reachconcentrations much greater than those expected from closed-system fractional crystal-lization alone For example the 10-fold increase in K2O in the Giant Crater lavas(fig 7) cannot otherwise be explained

1 1

22

33

A B

HAB Granite MeltedGranite

Ferrobasaltand crystals

Granitexenolith

Fig 6 Cartoon from Grove and others (1988) of two ways that FARM magma chamber processes couldtake place (A) A series of dikes of basalt intrude heat and melt the shallow crust as they crystallize (B) Amagma chamber where parental basalt crystallizes and heats and melts crust In both replenishment of themagma chamber by a fresh batch of parent magma triggers magma mixing


In summary both Burnt Lava and Giant Crater are examples of calc-alkalinetrends that exhibit silica enrichment and iron depletion but were produced byfractional crystallization along an iron-enrichment (tholeiitic) trend that haslargely been erased by mixing with melted granitic crust and new pulses ofprimitive HAOT The tholeiitic fractional crystallization trend is only partiallypreserved in the compositional variation of the last stage of magmatism at GiantCrater in the Group 6 lavas Importantly the distinctive calc-alkaline trend found atthese field sites is not a hydrous liquid line of descent but a multiple magmasmixing trend Therefore calc-alkaline trends can be produced by either hydrousfractional crystallization or by some combination of assimilation and magma

Fig 7 Compositional variation exhibited in the six eruptive units of the Giant Crater lava field on SiO2vs MgO TiO2 vs MgO K2O vs MgO and FeO vs MgO variation diagrams (see Baker and others 1991 andDonnelly-Nolan and others 1990) Group 1 is the most contaminated and forms when melted crust and ahighly differentiated ferrobasalt (black star fractionation trend shown in gray dashed line) are disturbed byreplenishment of the magma chamber by a fresh batch of primitive basalt Groups 2 and 3 form by a similarprocess Groups 4 and 5 form when mixing occurs between ferrobasalt a mixed magma left from a priormixing event and newly injected primitive basalt The last erupted unit Group 6 is the most primitive andthe compositional variation within that group is caused by fractional crystallization of olivine plagioclaseFerrobasalt composition calculated from compositions of olivine plagioclase augite-bearing magmaticinclusions found in Group 1 lavas Gray open circles are a continuous assimilation and fractional crystalliza-tion (AFC) calculation assuming R 15 and assimilation of granitic crust Squares are the FARM mixingmodels for each group (note that group 4 is divided into three subgroups)


mixing It turns out that Fennerrsquos mixing hypothesis which he largely based onfield relations to produce calc-alkaline trends (for example Fenner 1937 1948)had some basis in reality Bowen proposed that the compositional variability in thelavas and intrusive rocks at Mull were generated by calc-alkaline fractional crystalli-zation but subsequent studies (Sparks 1988 Kerr and others 1999) show thatassimilation and multiple magmas mixing also led to the observed compositionalvariations Thus contrary to Bowen magma mixing and assimilation do haveimportant consequences as magmatic differentiation processes

Compositional Gradients in LiquidsBowen also thought it unlikely that compositional gradients ldquoproduced by the

force of gravityrdquo could exist in magma chambers or be important influences inshaping compositional evolution However the Bishop Tuff a 07 million year old600 km3 eruption preserves a compositionally and thermally zoned magmachamber (Hildreth 1979 Hildreth and Wilson 2007) The compositional zoningformed when multiple batches of melt released from a deep crystal-rich mush zonedeeper in the magma chamber rose each to their own level of neutral buoyancy ina density stratified magma body Fractional crystallization was the dominantprocess that led to the observed compositional zonation (but see the Soret Effectsection below) Further evidence for compositional gradients in magma chamberscan also be found at Burnt Lava and Giant Crater (discussed in the Magma Mixingand Assimilation section) There dense ferrobasalt was trapped beneath less densemelted granite and when a parental mafic magma with intermediate density wasinjected into the magma chamber it rose through the ferrobasalt to the interfacebetween rhyolite and basalt triggering mixing The physical properties of theliquids (density viscosity) and the sequencing of intrusive processes govern thephysical processes leading to compositional zoning Huppert and Sparks (19801988) Campbell and Turner (1987) and McBirney and others (1985) outline theprocesses that control the development of compositional stratification when basaltmagma intrudes granitic crust

Liquid ImmiscibilityBowen devoted chapter 2 of ldquoThe Evolution of the Igneous Rocksrdquo to liquid

immiscibility beginning with the statement ldquoIn no case has any petrologist advocatingthis process been able to point out exactly how it is to be applied to any particular seriesof rocksrdquo Just before the publication of Bowenrsquos book Greig (1927a 1927b) showedexperimentally that liquid immiscibility occurred in the high - SiO2 portions of thetwo-component systems MgO SiO2 CaO SiO2 FeO SiO2 and Fe2O3 SiO2where known glassy rocks do not plot Bowen noted that no rocks had been found thatpreserved the requisite emulsion of globules of contrasting compositions expected tobe present in rocks if immiscibility were operating He also realized that immiscibleliquids would both crystallize a common mineral assemblage and have the sameequilibrium mineral compositions Based on this constraint he ended his discussionby pointing out that there is no evidence that basalt and rhyolite could be related byliquid immiscibility because these two compositions had been observed to crystallizemineral phases of extremely different composition (for example Na-rich plagioclasein rhyolite vs Ca-rich plagioclase in basalt) We now understand that contrastingmineral assemblages in basalt and rhyolite are produced by fractional crystallizationalong either the tholeiitic trend or the calc-alkaline trend which is in agreement withBowenrsquos conclusion

Ironically we additionally know now that liquid immiscibility does occur innatural magmatic systems but only in very FeO-rich liquids that form along theiron-enrichment (tholeiitic) differentiation trend that Bowen had dismissed The


experimental study of Roedder (1951) on the system Fayalite ndash Leucite ndash SiO2 led tothe discovery of a large field of liquid immiscibility in the center of this ternary systemThe first lunar samples from the Apollo 11 landing site were iron- and titanium ndashrichbasalts and these contained the abundant immiscible melt blebs in the mesostasis(Roedder and Weiblen 1970) The immiscible melts consisted of coexisting silica-richiron-poor and silica-poor iron-rich pairs and their discovery established the viability ofliquid immiscibility in natural systems More immiscible liquid blebs were discovered inthe groundmasses of Deccan Traps lavas (De 1974) in mid-ocean ridge basalts (Sato1978) and in a number of tholeiitic basalts (Philpotts 1979) Dixon and Rutherford(1979) experimentally produced liquid immiscibility in tholeiites and suggested thatplagiogranites in ophiolites and mid-ocean ridge settings might be produced by liquidimmiscibility This idea did not gain popularity because of the absence of a coexistingcomplimentary iron-rich low-silica liquid However evidence for silicate liquid immis-cibility has been identified in mafic layered intrusions including the Skaregaard(McBirney 1975 Jakobsen and others 2005 2011) Bushveld (VanTongeren andMathez 2012) Duluth gabbro (Ripley and others 1998) and Sept Iles (Namur andothers 2010 2012) where silica-rich and iron-rich melts can be observed to havesegregated at the mm- to meter scale and where melt inclusions in cumulus mineralspreserve coexisting immiscible liquids

Dry tholeiitic fractional crystallization experiments (Charlier and Grove 2012) atlow-pressure (1-atm) reproduced silicate liquid immiscibility in a broad range ofcompositions (44ndash56 wt SiO2 117ndash177 wt FeO and Mg between 29 and 36)They found that the solvus for liquid immiscibility in these natural tholeiitic systems laybelow 1020 degC as had been proposed by Philpotts (1979) and Philpotts (1982) Asshown in figure 8 the two-phase region is present at much lower temperatures thanhad been previously studied experimentally in 1-atm crystallization experiments ofmore primitive tholeiitic compositions (Grove and Bryan 1983 Juster and others1989 Toplis and Carroll 1995) The key to discovering the field of silicate liquidimmiscibility was to carry out crystallization experiments at low enough temperaturesso that the liquid line of descent intercepted the solvus Charlier and Grove (2012)found as had prior experimental studies of liquid immiscibility (Watson 1976 Visserand Koster van Groos 1979 Nasland 1983 Bogaerts and Schmidt 2006) thatincreasing K2O Na2O P2O5 and TiO2 contents in the melt also promotes thedevelopment of immiscible liquids whereas increasing CaO and Al2O3 contentspromotes the stabilization of a single liquid (fig 9)

With a more complete characterization of the extent of natural magma composi-tion space under which liquid immiscibility will occur under dry low-pressure condi-tions Charlier and others (2013) re-examined the major element compositions oftholeiitic basalts and their associated differentiates The composition gap that is acharacteristic of many tholeiitic provinces (fig 3) coincides with the experimentallydetermined immiscible liquid field of Charlier and Grove (2012) (fig 10) and permitsa role for large scale separation of silica-rich liquids in the late stages of tholeiiticdifferentiation The current perspective is that in volcanic tholeiitic systems theiron-rich silica-poor liquid is not present because it is too dense to erupt As we havenoted these Fe Ti P-rich plutonic products of immiscibility are found inplutonic environments Just as Bowen pointed out in Chapter 2 the immiscibilityprocess can be difficult to identify because both melts crystallize the same phases withthe same composition Both melts will evolve on the limbs of the solvus and exsolvecontinuously as the residual liquid changes in response to the removal of crystals Thisevolution by fractional crystallization can drive the bulk liquid out of the field ofimmiscibility back into the single-melt phase stability field So while it may be hard to


identify liquid immiscibility certainly occurs in nature and it plays a role in the lateststages of the tholeiitic differentiation trend

Soret EffectThe Soret effect is produced by chemical diffusion in a melt with a sustained

temperature gradient leading to the development of compositional zoning Bowendiscussed the Soret effect and concluded that it would not be an important process incausing compositional variation in magmas At the time the magnitude of the Soreteffect was not known and Bowen thought it would be small Bowen knew that heatdiffusion was much more rapid than diffusion of species in silicate melts He thoughtthat temperature gradients large enough for Soret diffusion to operate would be rarein magmatic systems and would equilibrate before Soret diffusion could have an effectThe magnitude of the Soret effect was first measured experimentally in a lunar melt(Walker and others 1981) and secondly in a terrestrial mid-ocean ridge basalt (Walkerand Delong 1982) In both compositions the effect of Soret diffusion was discovered tobe quite large The compositional effects were comparable to those created bycrystallization over a similar temperature interval and the diffusion of species that ledto the compositional gradients were the opposite of those created by crystal fraction-ation The Soret species appeared to be similar to network-forming and network-modifying components in silicate liquids Subject to a temperature gradient the lessdense network-forming components (SiO2 KAlO2 NaAlO2) diffused toward the hotend of the temperature gradient and the more dense network-modifying components(MgO FeO CaO) diffused toward the cold end of the thermal gradient (Lesher1986)

Fig 8 From Charlier and Grove (2012) Experimental results plotted as temperature vs NBOT ameasure of the degree of melt polymerization NBOT was calculated assuming T Si Al P Ti Starsindicate the composition and temperature of the four samples that were studied Black circles are theiron-rich immiscible melts and white circles are the silica-rich immiscible melts Experimental liquids thatfollow strong iron-enrichment trends but show no liquid immiscibility are shown from studies by Juster andothers (1989) as ldquoJampGrdquo and Toplis and Carroll (1995) as ldquoTampCrdquo


So while Soret diffusion could operate in temperature gradients in convectingmagma chambers Lesher and Walker (1991) concluded that Soret could not be amajor cause of chemical differentiation Hildreth (1979) initially called upon Soretdiffusion to produce the composition zoning observed in the Bishop Tuff but

Fig 9 From Charlier and Grove (2012) Immiscible melts from 1-atm experiments (A) Ternarydiagram with CaO and Al2O3 plotted at the base and SiO24 at the top (B) CaO and Al2O3 plotted at thebase and the sum of Na2O K2O P2O5 TiO2 at the top FeO in the iron-rich immiscible melts variesfrom 184 to 324 wt FeO Gray stars are the starting compositions used in their study Black circles are theiron-rich immiscible melts and white circles are the silica-rich immiscible melts Liquid lines of descent foriron-enrichment experiments of Juster and others (1989) and Toplis and Carroll (1995) are shown as JampGand TampC respectively


Fig 10 From Charlier and others (2013) Tholeiitic basalts and their differentiation products arecompared to experimental and naturally occurring immiscible melts in ternary composition space (A)Ternary with CaO and Al2O3 plotted at the base and SiO24 at the top (B) Ternary with CaO and Al2O3plotted at the base and the sum of Na2O K2O P2O5 TiO2 at the top Experimental melts are the onesshown in Figure 9 from Charlier and Grove (2012) and natural immiscible melts are from Charlier andothers (2013) and Phillpotts (1982)


withdrew that suggestion in light of Walker and Lesherrsquos experimental resultsHildreth and Wilson (2007) reinterpreted the zonation as a product of crystalfractionation Soret diffusion is more likely to occur in rare instances where veryhigh temperature melting processes have occurred leading to silicate melt diffu-sion rates faster than heat diffusion Two environments where this might happenare during terrestrial impact melting and the formation of tektites Delano andHanson (1996) infer temperatures for tektite reentry into the Earthrsquo atmosphereof 3000 degC and find compositional gradients that follow the expected Soretdiffusion trends When lightning strikes lead to the formation of fulguritestemperatures can exceed 10000 K (Paseck and others 2012) and in the rightcomposition target rock one might anticipate that a Soret compositional gradientcould develop While Soret diffusion may occur in nature it is only on a very smallscale

concluding remarksNorman Levi Bowen laid the groundwork for applying experimental petrology

to understanding the diversity of igneous rocks Bowen created the first phasediagrams using his experiments and established that fractional crystallization of aprimary magma would generate a wide range of residual liquid compositions and awider range of cumulate rock compositions sufficient to explain most igneousrocks During his time chemical analyses of liquids could only be made on bulkrocks forcing the limitation of comparing experiments to relatively sparse glassyigneous rocks Plutonic igneous rocks or even sparsely phyric quenched glassescould contain an unknown amount of crystal accumulation and accordingly wouldshow erroneous liquid lines of descent when graphed leading Bowen to notconsider any chemical evidence from them as trustworthy All of the rock composi-tions that matched his glassy criteria were from subduction zones and theyuniversally showed dramatic iron-depletion silica-enrichment trends He inte-grated that knowledge with the presence of abundant granitoids (that is extremelysilica-rich fractionates) and the fact that his experiments resulted in consistentrelative crystal appearances from basaltic parental magmas to conclude that therewas but one inevitable course of crystal fractionation Perhaps if the electronmicroprobe had been available to him so that he could have directly measuredmore liquid compositions he might have realized there could be more than oneliquid line of descent But hindsight is 2020 and regardless that Bowen did notget everything right his chemically and physically rigorous experimental approachto petrology has led us down a very fruitful path by providing invaluable tools forexpanding our understanding of magmatic systems

Since Bowenrsquos tremendous contribution to understanding the evolution of igne-ous rocks petrologists have realized that nature is in reality more complex than asingle course of fractional crystallization from effectively a single basaltic parentalmagma There are a variety of parental magma compositions which depending ontheir geologic context can undergo different and distinct liquid lines of descents suchas a dry tholeiitic trend a hydrous calc-alkaline trend and a low-pressure silica undersaturated alkaline trend Further significant igneous diversity can come from theother processes Bowen had considered but dismissed Field and geochemical evidencemake it clear that assimilation and magma mixing directly contribute to the variety ofigneous rocks Had Bowen realized that magmatic reservoirconduit systems aredynamic and episodically replenished he might have reframed his understanding ofmagma mixing and assimilation Comprehensive new experiments and field studieshave led to a better understanding of the role of liquid immiscibility Compositionalgradients in liquids are important in magma chambers undergoing replenishmentand Soret diffusion has its place when exceptionally high temperatures are rapidly


imposed such as during impacts or lightning strikes Even with all these new discover-ies over the last 80 years fractional crystallization has remained a critical igneousprocess - the repercussions of which are still fully being explored (for example Jagoutzand Klein this issue and references therein)

acknowledgments

Support for this work was provided through the National Science Foundationfrom grant EAR-1551321 The authors gratefully acknowledge the constructive reviewcomments of Mike Dungan Keith Putirka and one anonymous reviewer We alsothank the special editorial board members for the kind invitation to provide a paperfor this 200th Anniversary Volume of the American Journal of Sciences

REFERENCES

Andersen O 1915 The system Anorthite ndash Forsterite ndash Silica American Journal of Science Fourth Seriesv 39 p 407ndash454 httpsdoiorg102475ajss4-39232407

Anderson A T 1976 Magma Mixing - Petrological Process and Volcanological Tool Journal of Volcanol-ogy and Geothermal Research v 1 n 1 p 3ndash33 httpsdoiorg1010160377-0273(76)90016-0

Baker M B Grove T L Kinzler R J Donnelly-Nolan J M and Wandless G A 1991 Origin ofCompositional Zonation (High-Alumina Basalt to Basaltic Andesite) in the Giant Crater Lava-FieldMedicine Lake Volcano Northern California Journal of Geophysical Research-Solid Earth v 96n B13 p 21819ndash21842 httpsdoiorg10102991JB01945

Bogaerts M and Schmidt M W 2006 Experiments on silicate melt immiscibility in the system Fe2SiO4-KAlSi3O8-SiO2-CaO-MgO-TiO2-P2O5 and implications for natural magmas Contributions to Mineralogyand Petrology v 152 n 3 p 257ndash274 httpsdoiorg101007s00410-006-0111-6

Bowen N L 1912 The binary system Na2Al2Si2O8 (Nephelite carnegieite) - CaAl2Si2O8 (Anorthite)American Journal of Science Fourth Series v 33 p 551ndash573 httpsdoiorg102475ajss4-33198551

ndashndashndashndashndashndash 1913 The melting phenomena of the plagioclase fieldspars American Journal of Science FourthSeries v 35 p 577ndash599 httpsdoiorg102475ajss4-35210577

ndashndashndashndashndashndash 1914 The ternary system Diopside - Forsterite - Silica American Journal of Science Fourth Seriesv 38 p 207ndash264 httpsdoiorg102475ajss4-38225207

ndashndashndashndashndashndash 1916 Das ternaumlre system Diopsid ndash Anorthit ndash Albit Zeitschrift fur Anorganische und AllgemeineChemie v 94 n 1 23ndash50 httpsdoiorg101002zaac19160940103

ndashndashndashndashndashndash 1928 The Evolution of the Igneous Rocks Princeton New Jersey Princeton University Press 332 pBowen N L and Andersen O 1914 The binary system MgO-SiO2 American Journal of Science Fourth

Series v 37 p 487ndash500 httpsdoiorg102475ajss4-37222487Boyd F R and England J L 1960 Apparatus for phase equilibrium studies at pressures up to 50 kilobars

and temperatures up to 1750 degC Journal of Geophysical Research v 65 n 2 p 741ndash748 httpsdoiorg101029JZ065i002p00741

Bridgman P W 1952 The resistance of 72 elements alloys and compounds to 100000 kgcm2Proceedings of the American Academy of Arts and Sciences v 81 n 4 p 167ndash1251 httpsdoiorg10230720023677

Burnham C W and Jahns R H 1962 A method for determining solubility of water in silicate meltsAmerican Journal of Science v 260 n 10 p 721ndash745 httpsdoiorg102475ajs26010721

Byerly G 1980 The nature of differentiation trends in some volcanic-rocks from the Galapagos spreadingcenter Journal of Geophysical Research-Solid Earth v 85 n B7 p 3797ndash3810 httpsdoiorg101029JB085iB07p03797

Campbell I H and Turner J S 1987 A laboratory investigation of assimilation at the top of a basalticmagma chamber Journal of Geology v 95 n 2 p 155ndash172 httpsdoiorg101086629117

Carmichael I S E 1964 The petrology of Thingmuli a tertiary volcano in eastern Iceland Journal ofPetrology v 5 n 3 p 435ndash460 httpsdoiorg101093petrology53435

ndashndashndashndashndashndash 2002 The andesite aqueduct Perspectives on the evolution of intermediate magmatism in west-central(105 ndash 99 degW) Mexico Contributions to Mineralogy and Petrology v 143 n 6 p 641ndash663 httpsdoiorg101007s00410-002-0370-9

Charlier B and Grove T L 2012 Experiments on liquid immiscibility along tholeiitic liquid lines ofdescent Contributions to Mineralogy and Petrology v 164 n 1 p 27ndash44 httpsdoiorg101007s00410-012-0723-y

Charlier B Namur O and Grove T L 2013 Compositional and kinetic controls on liquid immiscibilityin ferrobasalt-rhyolite volcanic and plutonic series Geochimica et Cosmochimica Acta v 113 p 79ndash93httpsdoiorg101016jgca201303017

Clarke F W and Washington H S 1922 The average chemical composition of igneous rocks Proceedingsof the National Academy of Sciences of the United States of America v 8 p 108ndash115 httpsdoiorg101073pnas85108

Daly R A 1914 Igneous rocks and their origin New York New York McGraw-Hill Book Company Inc563 p


ndashndashndashndashndashndash 1952 The Name ldquoTholeiiterdquo Geological Magazine v 89 p 69ndash70 httpsdoiorg101017S0016756800067339

Darwin C 1844 Geological Observations on the Volcanic Islands Visited During the Voyage of HMSBeagle Together with Some Brief Notices on the Geology of Australia and the Cape of Good HopeBeing the Second Part of the Geology of the Beagle Under the Command of Capt Fitzroy RN Duringthe Years 1832 to 1836 London England Smith Elder and Company 192 p

Day A L and Allen E T 1904 Temperature measurements to 1600 degC Physical Review Series 1 v 19p 177ndash185 httpsdoiorg101103PhysRevSeriesI19177

De A 1974 Silicate liquid immiscibility in deccan-traps and its petrogenetic significance Geological Societyof America Bulletin v 85 n 3 p 471ndash474 httpsdoiorg1011300016-7606(1974)85471SLIITD20CO2

Delano J W and Hanson B 1996 Liquid Immiscibility Cause of Compositional Heterogeneity inTektites Lunar and Planetary Science v 27 p 305ndash306

DePaolo D J 1981 Trace element and isotopic effects of combined wallrock assimilation and fractionalcrystallization Earth and Planetary Science Letters v 53 n 2 p 189ndash202 httpsdoiorg1010160012-821X(81)90153-9

Dixon S and Rutherford M J 1979 Plagiogranites as late-stage immiscible liquids in ophiolite andmid-ocean ridge suites An experimental-study Earth and Planetary Science Letters v 45 n 1 p 45ndash60httpsdoiorg1010160012-821X(79)90106-7

Donnelly-Nolan J M Champion D E Miller C D Grove T L and Trimble D A 1990 Post-11000-year volcanism at Medicine Lake Volcano Northern California cascade range Journal of GeophysicalResearch-Solid Earth v 95 n B12 p 19693ndash19704 httpsdoiorg101029JB095iB12p19693

Donnelly-Nolan J M Champion D E and Grove T L 2016 Late Holocene Volcanism at Medicine LakeVolcano Northern California Cascades United States Geological Survey Professional Paper 1822 59 phttpsdoiorg103133pp1822

Dungan M A and Rhodes J M 1978 Residual glasses and melt inclusions in basalts from DSDP legs 45and 46 Evidence for magma mixing Contributions to Mineralogy and Petrology v 67 n 4 p 417ndash431httpsdoiorg101007BF00383301

Engel A E J Engel C G and Havens R G 1965 Chemical charcteristics of the oceanic mantle and theupper mantle Geological Society of America Bulletin v 76 n 7 p 719ndash734 httpsdoiorg1011300016-7606(1965)76[719CCOOBA]20CO2

Fenner C N 1926 The Katmai magmatic province The Journal of Geology v 34 n 7 Part 2 p 673ndash772httpsdoiorg101086623350

ndashndashndashndashndashndash 1929 The crystallization of basalts American Journal of Science Series 5 v 18 n 105 p 225ndash253httpsdoiorg102475ajss5-18105225

ndashndashndashndashndashndash 1937 A view of magmatic differentiation The Journal of Geology v 45 n 2 p 158ndash168 httpsdoiorg101086624515

ndashndashndashndashndashndash 1948 Immiscibility of Igneous Magmas American Journal of Science v 246 n 8 p 465ndash502httpsdoiorg102475ajs2468465

Finch R H and Anderson C A 1930 The quartz basalt eruptions of Cinder Cone Lassen VolcanicNational Park California University of California Publications Bulletin of the Department of Geologi-cal Sciences v 19 p 245ndash273

Fornari D J Perfit M R Malahoff A and Embley R 1983 Geochemical Studies of Abyssal LavasRecovered by DSRV Alvin from eastern Galapagos Rift Inca Transform and Ecuador Rift 1 MajorElement Variations in Natural Glasses and Spacial Distribution of Lavas Journal of GeophysicalResearch-Solid Earth v 88 n B12 p 10519ndash10529 httpsdoiorg101029JB088iB12p10519

Gaetani G A and Grove T L 1998 The influence of water on melting of mantle peridotite Contributionsto Mineralogy and Petrology v 131 n 4 p 323ndash346 httpsdoiorg101007s004100050396

Gale A Dalton C A Langmuir C H Su Y and Schilling J-G 2013 The mean composition of oceanridge basalts Geochemistry Geophysics Geosystems v 14 n 3 p 489ndash515 httpsdoiorg1010292012GC004334

Green D H and Ringwood A E 1967 The genesis of basaltic magmas Contributions to Mineralogy andPetrology v 15 n 2 p 103ndash190 httpsdoiorg101007BF00372052

Greig J W 1927a Immiscibility in silicate melts American Journal of Science Fifth Series v 13 n 73p 1ndash44 httpsdoiorg102475ajss5-13731

ndashndashndashndashndashndash 1927b Immiscibility in silicate melts American Journal of Science Fifth Series v 13 n 74 p 133ndash154httpsdoiorg102475ajss5-1374133

Grove T L 1981 Use of FePt alloys to Eliminate the Iron Loss Problem in 1-Atmosphere Gas MixingExperiments Theoretical and Practical Considerations Contributions to Mineralogy and Petrologyv 78 n 3 p 298ndash304 httpsdoiorg101007BF00398924

Grove T L and Baker M B 1984 Phase equilibrium controls on the calc-alkaline vs tholeiiticdifferentiation trends Journal of Geophysical Research-Solid Earth and Planets v 89 p 3253ndash3274

Grove T L and Bryan W B 1983 Fractionation of pyroxene-phyric MORB at low pressure Anexperimental study Contributions to Mineralogy and Petrology v 84 n 4 p 293ndash309 httpsdoiorg101007BF01160283

Grove T L Gerlach D C and Sando T W 1982 Origin of Calc-Alkaline Series Lavas at Medicine LakeVolcano by Fractionation Assimilation and Mixing Contributions to Mineralogy and Petrology v 80n 2 p 160ndash182 httpsdoiorg101007BF00374893

Grove T L Kinzler R J Baker M B Donnelly-Nolan J M and Lesher C E 1988 Assimilation ofgranite by basaltic magma at Burnt Lava flow Medicine Lake volcano northern California Decouplingof heat and mass transfer Contributions to Mineralogy and Petrology v 99 n 3 p 320ndash343httpsdoiorg101007BF00375365


Grove T L Elkins-Tanton L T Parman S W Chatterjee N Muentener O and Gaetani G A 2003Fractional crystallization and mantle melting controls on calc-alkaline differentiation trends Contribu-tions to Mineralogy and Petrology v 145 n 5 p 515ndash533 httpsdoiorg101007s00410-003-0448-z

Grove T L Till C B and Krawczynski M J 2012 The Role of H2O in Subduction Zone Magmatism AnnualReview of Earth and Planetary Sciences n 40 p 413ndash439 httpsdoiorg101146annurev-earth-042711-105310

Hamilton D L Burnham C W and Osborn E F 1964 The Solubility of Water and Effects of OxygenFugacity and Water Content on Crystallization in Mafic Magmas Journal of Petrology v 5 n 2p 21ndash39 httpsdoiorg101093petrology5121

Hildreth W 1979 The Bishop Tuff Evidence for the origin of compositional zonation in silicic magmachambers Geological Society of America Special Paper 180 p 43ndash75 httpsdoiorg101130SPE180-p43

Hildreth W and Wilson C H 2007 Compositional Zonation of the Bishop Tuff Journal of Petrologyv 48 n 5 p 951ndash999 httpsdoiorg101093petrologyegm007

Holloway J R and Burnham C W 1972 Melting Relations of Basalt with Equilibrium Water Pressure Lessthan Total Pressure Journal of Petrology v 13 n 1 p 1ndash29 httpsdoiorg101093petrology1311

Huebner J S 1971 Buffering techniques for hydrostatic systems at elevated pressure in Ulmer G Ceditor Research Techniques for High Pressure and High Temperature New York Springer Verlagp 123ndash177 httpsdoiorg101007978-3-642-88097-1_5

Hunter R H and Sparks R S J 1987 The Differentiation of the Skaergaard Intrusion Contributions toMineralogy and Petrology v 95 n 4 p 451ndash461 httpsdoiorg101007BF00402205

Huppert H E and Sparks R S J 1980 The Fluid-Dynamics of a Basaltic Magma Chamber Replenishedby Influx of Hot Dense Ultrabasic Magma Contributions to Mineralogy and Petrology v 75 n 3p 279ndash289 httpsdoiorg101007BF01166768

ndashndashndashndashndashndash 1988 The Generation of Granitic Magmas by Intrusion of Basalt into Continental-Crust Journal ofPetrology v 29 n 3 p 599ndash624 httpsdoiorg101093petrology293599

Jagoutz O and Klein B 2018 On the importance of crystallization-differentiation for the generation ofSiO2-rich melts and the compositional build up of arc (and continental) crust American Journal ofScience v 318 n 1 httpsdoiord10247501201803

Jakobsen J K Veksler I V Tegner C and Brooks C K 2005 Immiscible iron- and silica-rich melts inbasalt petrogenesis documented in the Skaergaard intrusion Geology v 33 n 11 p 885ndash888httpsdoiorg101130G217241

ndashndashndashndashndashndash 2011 Crystallization of the Skaergaard Intrusion from an Emulsion of Immiscible Iron- and Silica-richLiquids Evidence from Melt Inclusions in Plagioclase Journal of Petrology v 52 n 2 p 345ndash373httpsdoiorg101093petrologyegq083

Juster T C Grove T L and Perfit M R 1989 Experimental constraints on the generation of Fe-Tibasalts andesites and rhyodacites at the Galapagos Spreading Center 85degW and 95degW Journal ofGeophysical Research-Solid Earth v 94 n B7 p 9251ndash9274 httpsdoiorg101029JB094iB07p09251

Kawai N and Endo S 1970 The generation of ultrahigh hydrostatic pressures by a split sphere apparatusReview of Scientific Instrumentation v 41 p 1178ndash1181 httpsdoiorg10106311684753

Kennedy G C 1955 Some Aspects of the Role of Water in Rock Melts Geological Society of AmericaSpecial Paper 62 p 489ndash504 httpsdxdoiorg101130SPE62-p489

Kerr A C Iturralde-Vinent M A Saunders A D Babbs T L and Tarney J 1999 A new Plate TectonicModel of the Caribbean Implications from a Geochemical reconnaissance of Cuban Mesozoic volcanicrocks Geological Society of America Bulletin v 111 n 11 p 1581ndash1599 httpsdoiorg1011300016-7606(1999)1111581ANPTMO23CO2

Kinzler R J and Grove T L 1992 Primary magmas of mid-ocean ridge basalts 1 Experiments andMethods 2 Applications Journal of Geophysical Research-Solid Earth v 97 n B5 p 6885ndash6926httpsdoiorg10102991JB02840

Kinzler R J Donnelly-Nolan J D and Grove T L 2000 Late Holocene hydrous mafic magmatism at thePaint Pot Crater and Callahan flows Medicine Lake Volcano N California and the influence of H2O inthe generation of silicic magmas Contributions to Mineralogy and Petrology v 138 n 1 p 1ndash16httpsdoiorg101007PL00007657

Klein E M and Langmuir C H 1987 Global correlations of ocean ridge basalt chemistry with axial depthand crustal thickness Journal of Geophysical Research-Solid Earth v 92 n B2 p 8089ndash81115httpsdoiorg101029JB092iB08p08089

Krawczynski M J Grove T L and Behrens H 2012 Amphibole stability in primitive arc magmas Effectsof temperature H2O content and oxygen fugacity Contributions to Mineralogy and Petrology v 164n 2 p 317ndash339 httpsdoiorg101007s00410-012-0740-x

Kuno H 1936 Chemical compositions of volcanic rocks from Izu and Hakone volcano VolcanologicalSociety of Japan Bulletin v 3 p 53ndash71

ndashndashndashndashndashndash 1965 Fractionation Trends of Basalt Magmas in Lava Flows Journal of Petrology v 6 n 2 p 302ndash321httpsdoiorg101093petrology62302

Kushiro I 1972 Determination of the liquidus relations in synthetic silicate systems with electronprobe analysis The system forsterite-diopside-silica at 1 atmosphere American Mineralogist v 57p 1260 ndash1271

Larsen E S Irving J Gonyer F A and Larsen E S III 1938a Petrologic results of a study of the mineralsfrom the Tertiary volcanic rocks of the San Juan region Colorado American Mineralogist v 23 n 7p 417ndash429

ndashndashndashndashndashndash 1938b Petrologic results of a study of the minerals from the Tertiary volcanic rocks of the San Juanregion Colorado American Mineralogist v 23 n 4 p 227ndash257


Lesher C E 1986 Effects of Silicate Liquid Composition on Mineral-Liquid Element Partitioning fromSoret Diffusion Studies Journal of Geophysical Research-Solid Earth v 91 n B6 p 6123ndash6141httpsdoiorg101029JB091iB06p06123

Lesher C E and Walker D 1991 Thermal Diffusion in Petrology in Ganguly J editor Diffusion AtomicOrdering and Mass Transport Advances in Physical Geochemistry v 8 p 397ndash451 httpsdoiorg101007978-1-4613-9019-0_12

Liu L 1974 Silicate perovskite from phase transitions of pyrope-garnet at high pressure and temperatureGeophysical Research Letters v 1 n 6 p 277ndash280 httpsdoiorg101029GL001i006p00277

ndashndashndashndashndashndash 1975 Post-oxide phases of olivine and pyroxene and mineralogy of the mantle Nature v 258p 510ndash512 httpsdoiorg101038258510a0

McBirney A R 1975 Differentiation of Skaergaard Intrusion Nature v 253 p 691ndash694 httpsdoiorg101038253691a0

McBirney A R and Naslund H R 1990 The Differentiation of the Skaergaard Intrusion - A DiscussionContributions to Mineralogy and Petrology v 104 n 2 p 235ndash240 httpsdoiorg101007BF00306446

McBirney A R and Williams H 1969 Geology and petrology of the Galapagos Islands Geological Soietyof America Memoirs v 118 p 1ndash197 httpsdoiorg101130MEM118-p1

McBirney A R Baker B H and Nilson R H 1985 Liquid Fractionation Part 1 Basic Principles andExperimental Simulations Journal of Volcanology and Geothermal Research v 24 n 1ndash2 p 1ndash24httpsdoiorg1010160377-0273(85)90026-5

Medard E and Grove T L 2008 The effect to H2O on the olivine liquidus of basaltic melts Experimentsand thermodynamic models Contributions to Mineralogy and Petrology v 155 n 4 p 417ndash432httpsdoiorg101007s00410-007-0250-4

Merrill L and Bassett W A 1974 Minature diamond anvil pressure cell for single crystal x-ray diffractionstudies Reviews of Scientific Instruments v 45 p 290 ndash 294 httpsdoiorg10106311686607

Muir I D Tilley C E and Scoon J H 1964 Basalts from the northern part of the rift zone of themid-Atlantic Ridge Journal of Petrology v 5 n 3 p 403ndash434 httpsdoiorg101093petrology53409

Murakami M Hirose K Kawamura K Sata N and Ohishi Y 2004 Post-perovskite phase transition inMgSiO3 Science v 304 n 5672 p 855ndash858 httpsdoiorg101126science1095932

Nafziger R H Ulmer G C and Woerman E 1971 Gaseous buffering for the control of oxygen fugacity atone atmosphere in Ulmer G C editor Research Techniques for High Pressure and High Tempera-ture New York Springer Verlag p 9ndash43

Namur O Charlier B Toplis M J Higgins M D Liegeois J P and Vander Auwera J 2010Crystallization Sequence and Magma Chamber Processes in the Ferrobasaltic Sept Iles LayeredIntrusion Canada Journal of Petrology v 51 n 6 p 1203ndash1236 httpsdoiorg101093petrologyegq016

Namur O Charlier B and Holness M B 2012 Dual origin of Fe-Ti-P gabbros by immiscibility andfractional crystallization of evolved tholeiitic basalts in the Sept Iles layered intrusion Lithos v 154p 100ndash114 httpsdoiorg101016jlithos201206034

Naslund H R 1983 The Effect of Oxygen Fugacity on Liquid Immiscibility in Iron-Bearing Silicate MeltsAmerican Journal of Science v 283 n 10 p 1034ndash1059 httpsdoiorg102475ajs283101034

Oganov A R and Ono S 2004 Theoretical and experimental evidence for a post-perovskite phase ofMgSiO3 in the Earthrsquos Drdquo layer Nature v 430 p 445ndash448 httpsdoiorg101038nature02701

OrsquoNeill H St C 1981 The transition between spinel lherzolite and garnet lherzolite and its use as ageobarometer Contributions to Mineralogy and Petrology v 77 n 2 p 185ndash194 httpsdoiorg101007BF00636522

Osborn E F 1959 Role of Oxygen Pressure in the Crystallization and Differentiation of Basaltic MagmaAmerican Journal of Science v 257 n 9 p 609ndash647 httpsdoiorg102475ajs2579609

Pasek M A Block K and Pasek V 2012 Fulgurite morphology A classification scheme and clues toformation Contributions to Mineralogy and Petrology v 164 n 3 p 477ndash492 httpsdoiorg101007s00410-012-0753-5

Perfit M R and Fornari D J 1983 Geochemical Studies of Abyssal Lavas Recovered by DSRV Alvin fromeastern Galapagos Rift Inca Transform and Ecuador Rift 2 Phase Chemistry and CrystallizationHistory Journal of Geophysical Research-Solid Earth v 88 n B12 p 10530ndash10550 httpsdoiorg101029JB088iB12p10530

Philpotts A R 1979 Silicate Liquid Immiscibility in Tholeiitic Basalts Journal of Petrology v 20 n 1p 99ndash118 httpsdoiorg101093petrology20199

ndashndashndashndashndashndash 1982 Compositions of immiscible liquids in volcanic rocks Contributions to Mineralogy and Petrol-ogy v 80 n 3 p 201ndash218 httpsdoiorg101007BF00371350

Presnall D C 1966 The Join Forsterite-Diopside-Iron Oxide and Its Bearing On Crystallization of Basalticand Ultramafic Magmas American Journal of Science v 264 n 10 p 753ndash809 httpsdoiorg102475ajs26410753

Ridolfi F Renzulli A and Puerini M 2010 Stability and chemical equilibrium of amphibole incalc-alkaline magmas An overview new thermobarometric formulations and application to subduction-related volcanoes Contributions to Mineralogy and Petrology v 160 n 1 p 45ndash66 httpsdoiorg101007s00410-009-0465-7

Ringwood A E 1962 Mineralogical constitution of the deep mantle Journal of Geophysical Research-SolidEarth v 62 n 10 p 4005ndash4010 httpsdoiorg101029JZ067i010p04005

Ringwood A E and Major A 1970 The system Mg2SiO4 ndash Fe2SiO4 at high pressures and temperaturesPhysics of the Earth and Planetary Interiors v 3 p 89 ndash108 httpsdoiorg1010160031-9201(70)90046-4


Ripley E M Severson M J and Hauck S A 1998 Evidence for sulfide and Fe-Ti-P-rich liquidimmiscibility in the Duluth Complex Minnesota Economic Geology v 93 n 7 p 1052ndash1062httpsdoiorg102113gsecongeo9371052

Roedder E 1951 Low Temperature Liquid Immiscibility in the System K2O-FeO-Al2O3-SiO2 AmericanMineralogist v 36 n 3ndash4 p 282ndash286

Roeder P L and Osborn E F 1966 Experimental Data for System MgO-FeO-Fe2O3-CaAl2Si2O8-SiO2 andtheir Petrologic Implications American Journal of Science v 264 n 6 p 428ndash480 httpsdoiorg102475ajs2646428

Roedder E and Weiblen P W 1970 Silicate Liquid Immiscibility in Lunar Magmas Evidenced byMelt Inclusions in Lunar Rocks Science v 167 n 3918 p 641ndash644 httpsdoiorg101126science1673918641

Sato H 1978 Segregation vesicles and immiscible liquid droplets in ocean-floor basalt of Hole 396BIPODDSDP Leg 46 in Dimitriev L Heitrtzler J Aguilar R Cambon P Dick H J B Dungan MErickson A Hodges F N Honnorez J Kirkpatrick R J Matthews D Ohnenstetter D PetersenN Sato H Schmincke H U and Kaneps volume authors Initial Repots of the deep Sea DrillingProject v 46 p 283ndash291 httpsdoiorg102973dsdpproc461181979

Sato M 1971 Electrochemical measurements and control of oxygen fugacty and other gaseous fugacitieswith solid electrolyte sensors in Ulmer G C editor Research Techniques for High Pressure and HighTemperature New York Springer Verlag p 43ndash99 httpsdoiorg101007978-3-642-88097-1_3

Shaw H R 1967 Hydrogen osmosis in hydrothermal experiments in Abelson P H editor Researches inGeochemistry volume 2 New York John Wiley and Sons p 521ndash541

Sisson T W and Grove T L 1993a Experimental Investigations of the Role of H2O in Calc-AlkalineDifferentiation and Subduction Zone Magmatism Contributions to Mineralogy and Petrology v 113n 2 p 143ndash166 httpsdoiorg101007BF00283225

ndashndashndashndashndashndash 1993b Temperatures and H2O Contents of Low-MgO High-Alumina Basalts Contributions toMineralogy and Petrology v 113 n 2 p 167ndash184 httpsdoiorg101007BF00283226

Sparks R S J 1988 Petrology and Geochemistry of the Loch Ba Ring-Dyke Mull (NW Scotland) AnExample of the Extreme Differentiation of Tholeiitic Magmas Contributions to Mineralogy andPetrology v 100 n 4 p 446ndash461 httpsdoiorg101007BF00371374

Sparks S R J Sigurdsson H and Wilson L 1977 Magma Mixing A Mechanism for Triggering AcidExplosive Eruptions Nature v 267 p 315ndash318 httpsdoiorg101038267315a0

Takahashi E 1986 Melting of a dry peridotite KLB-1 up to 14 GPa Implications on the origin of peridotiticupper mantle Journal of Geophysical Research-Solid Earth v 91 n B9 p 9367ndash9382 httpsdoiorg101029JB091iB09p09367

Tegner C 1997 Iron in Plagioclase as a Monitor of the Differentiation of the Skaergaard IntrusionContributions to Mineralogy and Petrology v 128 n 1 p 45ndash51 httpsdoiorg101007s004100050292

Thy P Lesher C E and Tegner C 2009 The Skaergaard liquid line of descent revisited Contributions toMineralogy and Petrology v 157 p 735ndash747 httpsdoiorg101007s00410-008-0361-6

Toplis M J and Carroll M R 1995 An Experimental-Study of the Influence of Oxygen Fugacity on Fe-TiOxide Stability Phase-Relations and Mineral-Melt Equilibria in Ferro-Basaltic Systems Journal ofPetrology v 36 n 5 p 1137ndash1170 httpsdoiorg101093petrology3651137

Tormey D R Grove T L and Bryan W B 1987 Experimental petrology of normal MORB near the KaneFracture Zone 22degndash25degN mid-Atlantic ridge Contributions to Mineralogy and Petrology v 96 n 2p 121ndash139 httpsdoiorg101007BF00375227

Tuttle O F 1948 A New Hydrothermal Quenching Apparatus American Journal of Science v 246 n 10p 628ndash635 httpsdoiorg102475ajs24610628

Tuttle O F and Bowen N L 1958 Origin of Granite in the Light of Experimental Studies in the systemNaAlSi3O8-SiO2-H2O Geological Society of America Memoir 74 153 p httpsdxdoiorg101130MEM74

VanTongeren J A and Mathez E A 2012 Large-scale liquid immiscibility at the top of the BushveldComplex South Africa Geology v 40 n 6 p 491ndash494 httpsdoiorg101130G329801

Visser W and Koster van Groos A F 1979 Effects of P2O5 and TiO2 on Liquid-Liquid Equilibria in theSystem K2O-FeO-Al2O3-SiO2 American Journal of Science v 279 n 8 p 970ndash988 httpsdoiorg102475ajs2798970

Wager L R and Deer W A 1939 The petrology of the Skaergaard intrusion Kangerdlugssuaq EastGreenland Meddelelser om Groslashnland v 105 p 1ndash352

Wager L R and Brown G M 1968 Layered Igneous Rocks Edinburgh Scotland Oliver and Boyd 588 pWalker D and Delong S E 1982 Soret Separation of mid-Ocean Ridge Basalt Magma Contributions to

Mineralogy and Petrology v 79 n 3 p 231ndash240 httpsdoiorg101007BF00371514Walker D Longhi J Lasaga A C Stolper E M Grove T L and Hays J F 1977 Slowly cooled

microgabbros 15555 and 15065 in Lunar Science Conference 8th Houston Texas March 14ndash18 1977Proceedings v 2 New York Pergamon p 1521ndash1547

Walker D Shibata T and Delong S E 1979 Abyssal Tholeiites from the Oceanographer Fracture-ZoneII Phase-Equilibria and Mixing Contributions to Mineralogy and Petrology v 70 n 2 p 111ndash125httpsdoiorg101007BF00374440

Walker D Lesher C E and Hays J F 1981 Soret separation of lunar liquid Proceedings of the Lunarand Planetary Science Conference 12B p 991ndash999

Walter M J 1998 Melting of garnet peridotite and the origin of komatiite and depleted lithosphereJournal of Petrology v 39 n 1 p 29ndash60 httpsdoiorg101093petroj39129

Watson E B 1976 Two-Liquid Partition Coefficients Experimental Data and Geochemical Implications


Contributions to Mineralogy and Petrology v 56 n 1 p 119ndash134 httpsdoiorg101007BF00375424

Williams D W 1966 Externally Heated Cold-Seal Pressure Vessels For Use To 1200 degC at 1000 barsMineralogical Magazine and Journal of the Mineralogical Society v 35 p 1003ndash1012 httpsdoiorg101180minmag196603527514

ndashndashndashndashndashndash 1968 Improved Cold Seal Pressure Vessels to Operate to 1100 degC at 3 kilobars American Mineralogistv 53 p 1765ndash1769

Yoder H S Jr 1950 High-low Quartz inversion up to 10000 bars Eos Transactions of the AmericanGeophysical Union v 31 n 6 p 827ndash835 httpsdoiorg101029TR031i006p00827

ndashndashndashndashndashndash 1952 Change of Melting Point of Diopside with Pressure The Journal of Geology v 60 n 4p 364ndash374 httpsdoiorg101086625984

Yoder H S Jr and Tilley C E 1962 Origin of Basalt Magmas An Experimental Study of Natural andSynthetic Rock Systems Journal of Petrology v 3 n 3 p 342ndash532 httpsdoiorg101093petrology33342

28 TL Grove and SM Brown

knowledge on the five aforementioned rival processes that Bowen brought up in chapter1 We will survey the evolution of the discovery of new types of compositional diversity andof the advances in experimental studies since publication of Bowenrsquos work It is curious towonder if Bowen had had the opportunity to study the full complement of igneous rocksknown today might he have reconsidered his perspective that those rival processes wereunimportant and secondary to fractional crystallization

Bowen and the Evolution of Experimental Igneous PetrologyBowen and his colleagues started experimental petrology from scratch only aided

by an absolute temperature scale the so-called Geophysical Lab Temperature scalewhich his immediate predecessors (Day and Allen 1904) had established Bowen usedthis temperature scale to construct the first petrologic phase diagrams His firstexperimental studies that established the phase equilibria of common rock formingminerals were performed at atmospheric pressure in air using Pt-wound resistancefurnaces Samples were placed in containers (usually a Pt tube) which were thensuspended on thin wires and heated to high temperatures Quenching was achieved bymelting the wire using an electric current which caused the sample to drop into a dishof water placed under the open furnace tube (this quenching technique is still widelyused today) Exploring phase space required for each experimental temperature ofinterest the preparation of many starting materials that were loaded into separate Ptcapsules and then simultaneously hung together Experimental run products wereexamined and identified with an optical microscope after placing crushed samples onglass slides in oils with known refractive indexes which allowed them to establish foreach bulk composition whether it was above its liquidus in a crystal liquid field orbelow the solidus

The first compositional space Bowen studied for his Ph D thesis was the binarysystem Nepheline ndash Anorthite (Bowen 1912 MIT) He then studied the binary solidsolution Albite ndash Anorthite (Bowen 1913) the system MgO ndash SiO2 in which heestablished the revolutionary peritectic melting behavior of Enstatite to Forsterite Liquid (Bowen and Andersen 1914) and the ternary system Forsterite ndash Diopside ndashSiO2 (Bowen 1914) In 1915 his colleague Olaf Andersen published the ternarysystem Forsterite ndash Anorthite ndash Silica (Andersen 1915) and Bowen published thephase diagram for Diopside ndash Albite ndash Anorthite (Bowen 1916) Bowen was able toapply these carefully established experimental melting relations to observed naturaligneous rock mineral associations and compositions to forge the concepts thatappeared in his 1928 book ldquoThe Evolution of the Igneous Rocksrdquo It seems a miracu-lous accomplishment for one person to take the information in this handful of simplesystem phase diagrams and create a fully developed theory for the workings ofmagmatic processes

Bowen developed a whole new laboratory-based sub discipline of the EarthSciences Experimental petrology continues to evolve but our debt to Bowen isenormous As new tools and techniques became available the field of experimentalpetrology moved on from simple systems to the study of chemically complex naturalsystems A major step forward was the application of the electron probe micro-analyzer(EPMA) to chemical characterizations of experimental run products (Green andRingwood 1967 Kushiro 1972) For the first time crystals and coexisting meltcompositions could be characterized in the experimental charge and a complete phasediagram could be built using only a single starting composition Another breakthroughwas the development of techniques for buffering oxygen fugacity in experiments atone atmosphere and at higher pressures (Nafziger and others 1971 Sato 1971Huebner 1971) A further advance was to include iron in 1-atmosphere experimentsThis challenge was first solved by several groups (for example Walker and others1977) during experimental investigations of the returned Apollo lunar samples They



























1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18



Plag

Augite

Pig

Ilm

Tmt

Qtz








Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES







































































































































































1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18



Plag

Augite

Pig

Ilm

Tmt

Qtz








Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES



































































































































































1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18



Plag

Augite

Pig

Ilm

Tmt

Qtz








Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES





























































































































































1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18



Plag

Augite

Pig

Ilm

Tmt

Qtz








Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES























































































































































1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18



Plag

Augite

Pig

Ilm

Tmt

Qtz








Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES


















































































































































1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18



Plag

Augite

Pig

Ilm

Tmt

Qtz








Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES
















































































































































1240

1200

1160

1120

1080

1040

Tem

pera

ture

(˚C

)AII96-18



Plag

Augite

Pig

Ilm

Tmt

Qtz








Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES



















































































































































Tegner 1997

Thy and others 2009




40 45 50 55 60 65 70 75 800

5

10

15

20

25

30

35

SiO2 (wt)

OeF

tot)

tw( field for







900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES
















































































































































900 1000 1100 1200

100

300

500

700

Temperature (degC)

Pre

ssur

e (M

Pa) olv in

olv out

op

x in

plag in

nix

pc




olv= 868cpx= 877opx= 853

olv= 833cpx= 848opx= 839amph= 817

olv= 871cpx= 892

olv= 854cpx= 870




olv=903

olv=909

olv=862olv=

817olv=903

























1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES




































































































































































1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES
































































































































































1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES



























































































































































1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES




















































































































































1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES
















































































































































1 1

22

33

A B



Granitexenolith

































acknowledgments


REFERENCES












































































































































































acknowledgments


REFERENCES









































































































































































acknowledgments


REFERENCES


































































































































































acknowledgments


REFERENCES






























































































































































acknowledgments


REFERENCES

























































































































































acknowledgments


REFERENCES






















































































































































acknowledgments


REFERENCES




















































































































































acknowledgments


REFERENCES















































































































































acknowledgments


REFERENCES











































































































































































































































































































































































































































































Documents

MAGMATIC PROCESSES LEADING TO …MAGMATIC PROCESSES LEADING TO COMPOSITIONAL DIVERSITY IN IGNEOUS ROCKS: BOWEN (1928) REVISITED TIMOTHY L. GROVE† and STEPHANIE M. BROWN ABSTRACT