Perfect Fractional Crystallization Research Articles

PETROMODELER (Petrological Modeler): a Microsoft® Excel© spreadsheet program for modelling melting, mixing, crystallization and assimilation processes in magmatic systems

PETROMODELER (Petrologic Modeler) is a Microsoft® Excel© spreadsheet program which numerically and graphically models magmatic processes including melting, crystallization, assimilation and mixing by using trace elements and isotopic ratios. Melting models include (a) batch, (b) dynamic (continuous) and (c) fractional melting for (1) instantaneous and (2) cumulate melts, (3) residual solid and (4) total residue. These models can also be used to treat modal and non-modal melting. Crystallization processes modelled in the program include: (1) perfect equilibrium and (2) perfect fractional crystallization (PEC and PFC), (3) equilibrium crystallization-imperfect fractional crystallization (EC-IFC), (4) zoned crystallization-imperfect fractional crystallization (ZC-IFC), and (5) combined assimilation and fractional crystallization (AFC). Mixing between two end-member compositions can also be modelled by the program. The main advantages of the program are that; (1) crystallization and mixing processes can be performed on a starting composition which may be chosen from; (a) any melting model result, or (b) any sample composition entered into the 'samples' tables, (2) the results of any model can be exported as a graphic file (GIF) and as tables, (3) changes in any parameters are simultaneously updated onto all diagrams and tables. PETROMODELER also calculates other useful parameters, such as normative mineralogy, Mg#, Eu/Eu*, \epsilonSr and \epsilonNd, \sigma(DM) (depleted mantle Nd model ages) of a given sample. Some classification diagrams for volcanic rocks are also included in the program. Conversion of element abundances on the basis of wt% and ppm can also be performed.

A trace-element geochemical model for imperfect fractional crystallization associated with the development of crystal zoning

Although many petrological studies of volcanic rocks have suggested that crystallization proceeds within magma bodies, highly compatible trace elements do not display the marked variations and extreme depletions predicted to result from perfect fractional crystallization. Imperfect crystal–liquid separation is a key process in explaining this paradox. The presence of suspended crystals greatly affects variations in highly compatible elements, and has been quantitatively modeled by assuming perfect equilibrium between the suspended crystals and the liquid (equilibrium crystallization and imperfect separation; ECIS); however, volcanic rocks generally contain zoned phenocrysts that reflect the absence of solid-state equilibration. The present study develops a mass-balance model for zoned crystallization and imperfect separation (ZCIS). The ZCIS process is more efficient than the conventional ECIS process in generating depleted compatible elements. These two end-member models are able to explain the compositional range of igneous rocks that experienced imperfect fractional crystallization under natural conditions. The predicted compositional regions in bivariate trace-element diagrams successfully account for the sizes and shapes of the regions defined by whole-rock and melt-inclusion data from the Bishop Tuff, CA, USA.

On the Skaergaard intrusion and forward modeling of its liquid line of descent: A reply to “Principles of applied experimental igneous petrology” by Morse, 2008, Lithos 105, pp. 395−399

Forward modeling based on an experimental investigation successfully duplicated main features of the gabbros in the Skaergaard layered series [Thy, P., Lesher, C.E., Nielsen, T.F.D., and Brooks, C.K. 2006. Experimental constraints on the Skaergaard liquid line of descent. Lithos 92, 154–180.]. The foundation for the modeling was equilibrium melting experiments that were controlled by temperature and oxygen fugacity at low-pressure conditions. The experimental techniques and methods were chosen to represent a reasonable approximation to the inferred emplacement and crystallization conditions of the Skaergaard intrusion. The dike rocks used as starting materials define a strong differentiation trend that represents the Skaergaard liquid line of descent. This suite of dikes allowed liquidus conditions to be defined for a range of composition and temperature. The initial redox conditions were chosen based on measured and calculated estimates for the gabbros of interest. The melting experiments defined liquidus and subliquidus conditions that were used to understand crystallization of the lower and middle zones of the Skaergaard layered series, and can be extrapolated to the upper zone gabbros assuming perfect fractional crystallization. The forward modeling reproduces the cryptic variation seen in the main gabbro minerals (olivine, augite, plagioclase) well into the upper zone and provides reasonable liquidus temperatures and compositions. It can be shown that based on the assumption of Fe–Ti oxide modes in the middle and upper zones, a range of oxygen fugacity trends can be obtained. We repeat our previous conclusion that iron depletion and strong reduction in oxygen fugacity in the upper zone are only feasible for very high Fe–Ti oxide modes that exceed the experimental evidence as well as the observations from the gabbros. A strong drop in oxygen fugacity in the upper zone requires a significant sink for Fe–Ti oxides that so far has not been identified. We thus reject Morse's [Morse, S.A., 2008. Principles of applied experimental igneous petrology: a comment on “Experimental Constraints on the Skaergaard liquid line of descent” by Thy, Lesher, Nielsen, and Brooks, 2006, Lithos 92: 154–180. Lithos 105, pp. 395−399.] contention that we violated in our original study established principles of applied experimental igneous petrology. Such principles dictate that experimental and forward models are carefully tested against field observations before petrologic processes can be verified.

Liquid and magma viscosity in the anorthite‐forsterite‐diopside‐quartz system and implications for the viscosity‐temperature paths of cooling magmas

We measured the viscosity of 17 liquids in the systems anorthite‐forsterite‐quartz and anorthite‐diopside‐forsterite, representing analogs of dacitic, basaltic andesitic, and basaltic magmas. The three series lie in the anorthite liquidus fields and represent liquid lines of descent. The viscosity of evolving basaltic and basaltic andesite liquids changes little during cooling because the viscosity decrease associated with removal of polymerized anorthite component from the melt offsets the viscosity increase associated with cooling. A minimum liquidus viscosity in the basaltic system is encountered in the anorthite stability field. In contrast, dacitic liquids appear insensitive to the compositional changes during anorthite crystallization, and liquidus viscosity increases monotonically between anorthite and the anorthite‐forsterite‐quartz eutectic. Using the Einstein‐Roscoe equation to calculate the rheological effect of crystals, we show that magma viscosity always increases during crystallization for closed systems and that crystallinity is the dominant control on the viscosity of all three magma series. We introduce the concept of viscosity paths, which may approach one of four end‐members: (1) equilibrium crystallization (closed system) and (2) perfect fractional crystallization (open system) are both associated with changing residual liquid compositions; (3) supercooling without crystal growth maintains constant liquid composition, while (4) increasing crystallinity at constant liquid composition is only permissible for eutectic starting compositions. Because of the dependence of crystal‐liquid segregation rates on liquid viscosity and crystal buoyancy, feedback relations exist between evolving liquid composition, crystallinity, and magma viscosity. Crystal‐liquid segregation rates may increase or decrease during progressive cooling and crystallization of basaltic liquids.

Chemistry of the Shiant Isles Main Sill, NW Scotland, and Wider Implications for the Petrogenesis of Mafic Sills

Major and trace element data for the Tertiary, Shiant Isles Main Sill, NW Scotland, are used to discuss its complex internal differentiation. Vertical sections through the sill exhibit sharp breaks in chemistry that coincide with changes in texture, grain size and mineralogy. These breaks are paired, top and bottom, and correspond to the boundaries of intrusive units, confirming a four-phase multiple-intrusion model based on field relations, petrography, mineralogy and isotopes. Whole-rock chemistry is consistent with this model and necessitates only minor revisions to the intrusive and differentiation mechanisms previously proposed. The rocks contain strongly zoned minerals (e.g. olivine Fo70–5, clinopyroxene Mg# = 75–5, plagioclase An75–5) indicating almost perfect fractional crystallization, but whole-rock compositions do not show such extreme variations. Thus, while residual liquids became highly evolved in situ, they mainly became trapped within the crystal network and did not undergo wholesale inward migration. Some inward (mainly upward) concentration of residual liquids did occur to form a ‘sandwich horizon’, but the more volatile-rich, late-stage liquids that did not crystallize in situ appear to have migrated to higher levels in the sill to form pegmatitic horizons. Parental liquid compositions are modelled for the intrusive units and it is concluded that the original parent magma formed by partial melting of upper mantle that was more depleted in LREE than the sources of most Scottish Tertiary basaltic rocks. Incompatible trace elements in the picrodolerite–crinanite intrusive unit support isotope evidence that its parent magma was contaminated by crustal material. Attempts to reconcile the chemical characteristics of the sill with a recently proposed petrogenetic model based on a single intrusion of magma differentiated by novel, but controversial, processes fail comprehensively. It is predicted that the complex petrogenetic history of the Shiant Isles sill is not unusual and could become the model for other large (>50 m thick) sills.

Fractional crystallization of hydrous basaltic ?arc-type? magmas and the formation of amphibole-bearing gabbroic cumulates

The Rymmen and Eriksberg gabbros preserve the crystallization products of hydrous basaltic liquids, ranging from anorthosite and troctolite to highly fractionated oxide gabbro, diorite, and leucotonalite. The high initial water content of the liquids that filled the magma chambers (1.5–3 wt%) allowed amphibole to saturate as an interstitial mineral even in the earliest formed cumulates. The increasing water content of the system helped to produce a thick sequence of cumulates with anorthitic plagioclase (An87–92). The growth of interstitial amphibole caused much more interstitial liquid to be incorporated into the cumulates than might be expected for dryer basaltic systems, producing a very imperfect fractional crystallization process. Comparisons of models of the trace-element evolution of the two intrusions reveal that the overall enrichments of some elements are adequately predicted by a perfect fractional crystallization model but all agree better with an imperfect fractional crystallization model. This reflects the change in the bulk distribution coefficients of the solid assemblage upon amphibole saturation, with the retention of elements compatible in amphibole in the crystal pile and the loss of elements incompatible in amphibole from it. These results stress the importance of understanding the physical, as well as the chemical processes involved in fractional crystallization and in particular the role of postcumulus processes.

Minor Phases as Carriers of Trace Elements in Non-Modal Crystal–Liquid Separation Processes II: Illustrations and Bearing on Behaviour of REE, U, Th and the PGE in Igneous Processes

Minor phases which strongly concentrate selected trace elements, termed here ‘carrier-phases’, release relatively large amounts of those elements to the liquid phase when they are eliminated during partial melting and glean relatively large amounts of those elements when they first appear during progressive crystallization. It is characteristic of such relationships that concentrations of the selected trace elements in the bulk residues of partial melting will rise to a peak somewhat before the last of the carrier-phase is eliminated during progressive melting. In the liquids produced during equilibrium partial melting a corresponding peak in the concentration of the trace element occurs at the point where the carrier-phase is eliminated; the corresponding peak in trace element concentration in the liquids produced by accumulated perfect fractional melting is found somewhat above that point. These peaks become more sharply accentuated as the distribution coefficient of the trace element into the carrier-phase increases. The highest trace element concentration in a partial melt liquid product is found in the small drop of liquid produced during perfect fractional melting at the point where the carrier-phase is eliminated. Still higher concentrations may be found in the first cumulates containing the carrier-phase which precipitate during perfect fractional crystallization but the corresponding liquids do not contain exceptionally high concentrations. Under favourable conditions a large proportion of the available mass of a trace element in a magmatic system may be transferred from the solid to the liquid phases or vice versa with only a small change in the mass fraction of liquid in and energy content of the system. Within that range, separation of otherwise very similarly behaved trace elements becomes possible. Further complexities arise and the opportunities for separation increase when two carrier-phases compete with differing success for the same group of trace elements.

The Shiant Isles Main Sill: structure and mineral fractionation trends

AbstractThe Shiant Isles Main Sill, of Tertiary age, is a classic example of a composite, differentiated alkaline basic sill. The first unit to be intruded was a 2 m thick olivine teschenite which was emplaced with phenocrysts of olivine (mg > 83) [mg ≡ Mg#] and, perhaps, plagioclase. This was intruded by a 24 m thick picrite sill consisting of a mush of melt and suspended olivine phenocrysts (mg > 83) with a D-shaped modal profile. The 140 m thick picrodolerite-crinanite unit was formed by a magma carrying ∼ 10% olivine (mg > 83) as the main phenocryst phase, together with some calcic plagioclase phenocrysts, being emplaced into the top of the picrite unit before the host rock was completely solidified. The olivine phenocrysts settled towards the bottom to form the picrodolerites. In-situ differentiation processes occurred under conditions of almost perfect fractional crystallization, during which very strongly zoned ophitic crystals of olivine (fayalitic rims) and clinopyroxene (hedenbergitic rims), and zoned laths of plagioclase (anorthoclase rims), formed. The last unit consists of ∼ 2 m of granular olivine picrodolerite which was intruded into the upper crinanites, again before the host rock was fully solid.The mineral zoning patterns are interpreted using published cation diffusion coefficient data, and used to show that the picrite unit might have cooled to the blocking temperatures for Mg and Fe diffusion in < 5 years, and that even the relatively thick crinanite unit cooled very fast, so preserving the zoned Fe-Mg olivine and pyroxene compositions. The compositions of coexisting ilmenites and spinels define a redox trend which initially lies close to fayalite-magnetite-quartz buffer conditions, but later becomes more reducing and approaches magnetite-wustite buffer conditions. The final stages of development occurred during sub-solidus deuteric processes and involved formation of analcime and zeolites, as well as localized sulphide mineralization.

Differentiation of Ferro-Basaltic Magmas under Conditions Open and Closed to Oxygen: Implications for the Skaergaard Intrusion and Other Natural Systems

Because processes such as fractional crystallization and crystallization under conditions closed to oxygen are difficult to simulate in the laboratory there is a need for quantitative models of magma crystallization behaviour. Comparison of experimental data on an iron-rich basaltic composition with predictions of the MELTS free energy minimization algorithm shows that although liquidus temperatures and silicate mineral equilibria are predicted relatively well, the saturation of Fe-Ti oxides is not We have used the same experimental data to construct an alternative crystallization model based on known equilibrium phase relations, mineral-melt partitioning of major elements, and mass balance constraints. The model is used to explore the consequences of equilibrium and fractional crystallization in systems open and closed to oxygen. Liquid lines of descent for perfect equilibrium and perfect fractional crystallization are predicted to be very similar. In a system open to oxygen the model predicts that magnetite saturation leads to strongly decreasing iron and increasing silica contents of residual liquids, whereas systems closed to oxygen crystallize less abundant magnetite, leading to a less pronounced iron depletion in the liquid. Predicted bulk solid compositions and variations offo, with falling temperature agree well with those observed or inferred from the cumulates of the Skaergaard intrusion, but none of the predicted liquid lines of descent are consistent with the extreme iron enrichment proposed for this intrusion based on mass balance calculations. Compositional factors such as water and phosphorus are not thought to be the source of the discrepancy as the cumulates of the Basistoppen sill (which closely resemble those of Skaergaard) may be used to calculate a liquid line of descent in agreement with that predicted by the model for fractional crystallization closed to oxygen. A comparison of the predicted T-io, paths and liquid lines of descent with those inferred from natural systems suggests that volcanic centres such as Iceland and Hawaii evolve under conditions open to oxygen, whereas evidence from plutonic environments (e.g. Skaergaard and Kiglapait layered intrusions) suggests that they evolved under conditions more closed to oxygen. The compositional evolution of the melt phase in volcanic and plutonic systems may therefore be different, although the results of this study suggest that magnetite saturation will limit Fe enrichment in all environments to <20wt% FeO*, consistent with enrichments reported for volcanic glasses.

Open system crystallization and mixing in two-layer magma chambers

A modeling technique is presented for describing trace element behavior in tholeiitic magmas evolving by either perfect or imperfect fractional crystallization and magma mixing in periodically replenished two-layer reservoirs, in which eruption events occur at various times. The models are based on density versus mafic index calibrations, and can be used to examine two replenishment situations: (1) dense (picritic) liquid ponds at the base of the magma chamber or, (2) light liquid rises buoyantly through the resident magma and collects at the top of the reservoir. Input parameters are bulk solid/liquid distribution coefficients ( D) and enrichment/depletion factors ( e) i.e. the ratios between concentrations of trace elements in the erupted and incoming liquids, respectively. The results show graphically the relations between the weight fractions of incoming liquid ( W), the crystallized magma ( X 1 and X 2) and the erupted magma ( Y t ). These models can be used to distinguish imperfect fractional crystallization from Rayleigh's law-dependent fractionation. They are applied to mid-ocean ridge basalts (MORB) dredged from Galapagos Spreading Center (GSC) at 85 °W and to rocks from the Newark Island Layered Intrusion (NILI). The most likely mechanism to explain the GSC MORB variations is imperfect fractional crystallization. By contrast, the NILI variations appear to have been chiefly controlled by perfect fractional crystallization. The Y t versus W relations for these two situations are very different. Indications are that at Galapagos, the erupted magma amounts to less than 5 wt.% of the reservoir, whereas at NILI, half of the mass of the initial replenishing magma was erupted during a cycle.

Trace element geochemical effects of imperfect crystal-liquid separation

Abstract The strongly contrasted effects of melting or crystallization under perfect equilibrium conditions; of perfect fractional partial melting; and of perfect fractional crystallization upon the relative concentrations of the highly incompatible and highly compatible elements are well known and are surveyed briefly at the start. A major role for either of the two latter processes in the genesis of natural basaltic magmas and peridotitic residues is apparently excluded provided that the liquids or crystals are removed in infinitesimally small batches, the crystals and melts are perfectly separated and there is no flux of source material or residue in or out of the greater system during the process (either liquid or crystals must be removed from the immediate, or lesser, system if it is undergoing perfect fractional melting or perfect fractional crystallization). The consequences of small departures from (i) infinitesimal size of the batches of liquid or crystals removed, (ii) perfect separation of the crystal and liquid phases, or (iii) perfect closure of the greater system to fluxes of the source material and the residue are explored. These departures can very substantially reduce the differences between the trace element concentrations generated in the products and those expected in a perfect equilibrium process. These effects are illustrated here by means of 3D plots of the logarithm of relative concentration in the liquid phase as a function of the logarithm of the distribution coefficient and the logarithm of the mass fraction of liquid involved in each process. There are also significant complementary effects on the compositions of the residues or precipitates from these processes which are not illustrated in this paper, although the necessary equations and some worked examples are provided. With the values of parameters chosen in these examples, the fluxing of liquid or solid through a ‘box’ in which fractional crystallization or fractional melting is taking place proves to be by far the most important factor in modifying the results towards those of the equilibrium processes. Although much trace element behaviour in basalts can be quite successfully modelled using the perfect equilibrium melting relationships, it is unsafe to exclude on these grounds a major role for physical processes of fractional melting and fractional crystallization in the genesis of these magmas or to conclude that equilibrium partial melting is the principal physical process involved in their genesis.

Melt inclusions and crystal-liquid separation in rhyolitic magma of the Bishop Tuff

Melt inclusions in quartz phenocrysts from a single clast of pumice near the base of the plinian pumice fall of the Bishop Tuff were studied to test ideas concerning separation of melt and crystals in silicic magmas. Ten analyzed inclusions from the pumice clast are of high silica rhyolite composition with very low contents of the highly compatible elements Ba, Sr, and Eu, consistent with extensive fractionation. The concentrations of U, La, Ce, Mg, and Ca of these ten melt inclusions vary considerably as determined by ion microprobe. Petrologic considerations indicate that uranium is an incompatible element with a maximum bulk partition coefficient D of about 0.2 and that the evolution of the uranium content of the melt was controlled by crystallization of the magma. A minimum of 33 wt% perfect fractional crystallization is required to explain the observed range in uranium. However, only 17 wt% crystals occurred in the pumice clast. The greater calculated fraction of crystals requires significant separation of crystals and melt before the eruption of the plinian pumice fall in spite of the fact that crystal mixing (settling, etc.) did not occur in the Bishop magma.

Interpretation of open system petrogenetic processes: Phase equilibria constraints on magma evolution

We have used a computer model (TRACES) to simulate low pressure differentiation of natural basaltic magmas in an attempt to investigate the chemical dynamics of open system magmatic processes. Our results, in the form of simulated liquid lines of descent and the calculated equilibrium mineralogy, were determined for perfect fractional crystallization; fractionation paired with recharge and eruption (PRF); fractionation paired with assimilation (AFC); and fractionation paired with recharge, eruption, and assimilation (FEAR). These simulations were calculated in an attempt to assess the effects of combinations of petrogenetic processes on major and trace element evolution of natural systems and to test techniques that have been used to decipher the relative roles of these processes. If the results of PRF calculations are interpreted in terms of a mass balance based fractionation model (e.g., Bryan et al., 1969), it is possible to generate low residuals even if one assumes that fractional crystallization was the only active process. In effect, the chemical consequences of recharge are invisible to mass balance models. Pearce element ratio analyses, however, can effectively discern the effects of PRF versus simple fractionation. The fractionating mineral proportions, and therefore, bulk distribution coefficients (D) of a differentiating system are dependent on the recharge or assimilation rate. Comparison of the results of simulations assuming constant D with the results calculated by TRACES show that the steady state liquid concentrations of some elements can differ by a factor of 2 to 5. If the PRF simulation is periodic, with episodes of mixing separated by intervals of fractionation, parallel liquidus mineral control lines are produced. Most of these control lines do not project back to the parental composition. This must be an important consideration when attempting to calculate a potential parental magma for any natural suite where magma chamber recharge has occurred. Most basaltic magmas cannot evolve to high silica compositions without magnetite fractionation. Small amounts of rhyolite assimilation (assimilation/fractionation < 0.1), however, can drive evolving basalts to more silica rich compositions. If mass balance models are used to interpret these synthetic AFC data, low residuals are obtained if magnetite is added to the crystallizing assemblage. This approach works even for cases where magnetite was not a fractionating phase. Thus, the mass balance results are mathematically correct, but are geologically irrelevant.

Liquid collection in sidewall crystallization of magma: A comment on “liquid fractionation. Part I”, by A.R. McBirney, B.H. Baker and R.H. Nilson

The term liquid fractionation is inappropriate for a process driven by crystallization; liquid collection is proposed instead. The amount of evolved residual liquid produced by sidewall crystallization is not greater than that produced by any perfect fractional crystallization process and the fractionation products must be balanced somewhere by less evolved crystal deposits. There is no free lunch in petrology. Sidewall heating limits the local duration of liquid collection and the complementary crystal products should be sought in conduits. Flux ratios rather than diffusivities control the differences between thermal and chemical boundary layers.

The petrology and geochemistry of the Ocate volcanic field, north-central New Mexico

The Ocate volcanic field in north-central New Mexico is on the eastern flank of the Sangre de Cristo Mountains. It is a suite of Pliocene-Pleistocene basaltic and intermediate lavas which were erupted in five pulses over a period of ∼7 m.y. (8.3–0.8 m.y. B.P.) during rejuvenation of the Rio Grande Rift. Geochemical and petrographic criteria are used to define five major rock types: alkali olivine basalt (AOB), transitional olivine basalt (TOB), xenocrystic basaltic andesite (XBA), olivine andesites (OA), and dacite. The timing and eruptive volumes of these five lava types define a complex temporal sequence in which the transitional basalts and intermediate rocks are closely associated during the maximum in basaltic activity (4.5–2.0 m.y. B.P.). The alkali olivine basalts have higher incompatible and compatible trace-element abundances and higher normative nepheline (>2%) than do the TOB lavas. The two basalt types apparently were derived from physically discrete or heterogeneous source regions. The AOB lavas could be the products of smaller degrees of melting from similar sources, source heterogeneity, and/or melting at greater depth. The intermediate lavas were generated from basaltic parent magmas by varying degrees of fractionation, assimilation of crustal material, and mixing with silicic melts derived by crustal anatexis. The XBA lavas and dacites are characterized by disequilibrium phase assemblages which include olivine coexisting with biotite and quartz, melt inclusions of K- and Si-rich glass, and a bimodal distribution of plagioclase phenocryst compositions with modes at An60 and An25. The high concentration of Mg, Ni, and Cr in the intermediate lavas is inconsistent with simple fractionation of basaltic parent magmas. The linear trends defined on variation diagrams are consistent with mixing mafic and silicic end members. The lack of continuity between two hybrid types may be a reflection of contrasting subvolcanic environments in which the hybridization took place. The olivine andesites show little petrographic evidence for mixing. Chemical variations within this group define highly nonlinear trends more consistent with extensive crystal fractionation. Calculated liquid lines of descent from available mafic parent magmas (assuming perfect fractional crystallization) do not satisfactorily produce the observed intermediate compositions. If assimilation of a silicic crystal component is paired with crystal fractionation, such that the mass of crustal material is approximately equal to 50% of the fractionated solids, the major- and trace-element characteristics of the olivine andesites can be successfully modeled.

Two magmatic series with island arc affinities within the vourinos ophiolite

Within the Vourinos ophiolite evidence of two magmatic series has been preserved in cognate cumulates and in effusive and hypabyssal rocks, which constitute the earlier Krapa sequence and the younger Asprokambo sequence. The Asprokambo dyke basic magmas which are poor in incompatible elements and relatively Ni and Cr rich, bear some resemblance to very low Ti basalts (transitional to boninites) found in subduction related arcs or interarc basins. Krapa series magmas from sills, massive and pillow lavas are best equated with low-K tholeiites of island arc suites. Compositions of Al- and Ti- poor Cpx in lavas from both series are comparable to those in island arc basalts, the Asprokambo Cpx being richer in Ca and Cr than those from Krapa. The large volume of cumulates from the Krapa sequence displays the following crystallization order: Ol±Sp, Cpx, Pl±Opx, Mt. Periodic influx of fresh magma batches into the magma chamber occurred mainly during the formation of the lower cumulates (wehrlite, Ol-clinopyroxenite and melagabbro). The upper cumulates, gabbronorite and leucogabbronorite with minor Mt-bearing gabbronorite, crystallized in the upper levels of a magma chamber which became progressively smaller with time. In the Asprokambo sequence, Ol+Sp, Opx, Cpx, PI and Amph are the successively crystallizing phases. The ortho to heteradcumulates consist of websterite, Pl-websterite, gabbronorite, amphibole bearing leuconorite, diorite and granophyre. In cumulates, especially in the lower Krapa sequence, significant subsolidus reaction was probably induced by the persistence of high geothermal gradients linked to continuous magmatism. Petrological features indicate that the evolution of the Krapa series is more compatible with an intermediate fractional/equilibrium crystallization history in an initially open system, whereas nearly perfect fractional crystallization in closed system may have occurred in the small Asprokambo magma chambers. Chemical variations in the lavas of both series can be explained in terms of crystallization of the observed cumulates. Significantly, the Asprokambo intrusives have igneous Mg-hornblende and vanadium bearing, chromian, aluminous titaniferous magnetite, crystallization of which is responsible for the calcalkaline evolutionary trend of these rocks. Major and trace element modelling necessitates a two stage model for the petrogenesis of the Vourinos parental melts, involving high-degree remelting of previously depleted mantle sources favoured by the influx of subduction derived hydrous fluids. The primary magmas parental to the Krapa and Asprokambo series could have been derived respectively by 20 and 30% equilibrium partial fusion of variably depleted lherzolitic sources, leaving residua having a harzburgitic to dunitic composition.

Felsic mineral crystallization trends in differentiating alkaline basic magmas

Mineral compositional relationships have been studied in a variety of alkaline basic rocks from small, highlevel intrusions. These small intrusions must have cooled quickly as the phases are all zoned, especially those forming the matrix of some of the rocks which seem to have formed under conditions of almost perfect fractional crystallization.

The Fractional Crystallization of Plagioclase in the Hat Creek Basalt: Observations and Theory

Plagioclase in the Hat Creek high alumina olivine tholeiite decreases in anorthite (An) component with crystallization. Petrography reveals the composition of plagioclase that formed at various stages of crystallization represented by microporphyritic glassy tops of the flow and by segregation veins and their residual glasses. The plagioclase ranges from An79 in the cores of microphenocrysts to An9 rims on crystals in glass-bearing segregation veins as the fraction of crystallization increases from 0 to 0.9. The temperatures of crystallization corresponding to the various fractions of crystallization are estimated on the basis of olivine/liquid and magnetite/ilmenite geothermometry and linear interpolation. The temperatures ranged from 1237°C to 980°C. Supercooling was less than 20°C initially. Most of the plagioclase microphenocrysts have An consistent with the plagioclase geothermometer. Plagioclase in most of the segregation veins differs from that expected. The derived histogram of amount of plagioclase versus its composition reveals that the mode lies at the An-rich end of the spectrum with successively smaller amounts formed in more sodic intervals. The curve for composition versus crystallization extrapolates to pure albite at about 100% crystallized. Except for the anomalous segregation veins, the observations agree with a modified theory for perfect fractional crystallization of plagioclase in dry basaltic magma. The anomalous segregation veins apparently lost or gained residual liquid and some inherited microphenocrysts which were carried in during vein formation. The compositional spectrum of plagioclase in basaltic and gabbroic rocks can be used to substantiate closed system fractional crystallization.

Trace element distribution patterns and their relationship to the crystallization of granitic melts

The distribution of Ba, Rb and Sr during crystallization of a granitic melt is examined in a number of theoretical models. The modes of crystallization considered are perfect fractional crystallization, perfect equilibrium crystallization, and an intermediate mode, incremental equilibrium crystallization. The effect of the degree of separation of cumulus minerals from melt during crystallization is also considered. Perfect fractional and incremental equilibrium crystallization (with small increments) are broadly similar, but differ in the final stages of crystallization in that the latter mode defines a finite trace element composition for the last solid. The effect of intercumulus melt in both modes of crystallization imparts a ‘liquid’ character to the solids, and suppresses the degree of enrichment of Rb and depletion of Ba and Sr in late solids and melts. Examination of trace element data for the Acid Phase of the Bushveld Igneous Complex in the light of these models suggests that these granites represent a suite of cumulate rocks, containing relatively large amounts of intercumulus melt.

Crystallization of Alkali Feldspar and Quartz in the Haplogranite System NaAlSi3O8-KAlSi3O8-SiO2-H2O at 4 Kb

Results of experimental studies in the system NaAlSi 3 O 8 -KAlSi 3 O 8 -SiO 2 -H 2 O, at 4 kb and from 650° to 1000°C, have been used to generate composition paths for liquid and crystal fractions as functions of temperature and bulk composition. In both the experimental work and the analyses of crystallization, attention has been devoted mainly to silicate liquid that is saturated with quartz and an alkali feldspar but unsaturated with respect to an aqueous vapor phase. Such liquid can be represented on a compound T-X surface that slopes, with decreasing H 2 O content in the system, toward SiO 2 for NaAlSi 3 O 8 -rich compositions and away from SiO 2 for KAlSi 3 O 8 -rich compositions. Compositions of the crystal fraction that separates from the liquid under equilibrium conditions in the temperature range 700° to 655°C have been calculated for an array of bulk compositions. Equilibrium crystallization history is characterized by numerous shifts and some reversals along paths that describe the composition of the crystal fraction. The various paths demonstrate that intensive parameters such as temperature, pressure, and H 2 O fugacity cannot be uniquely determined from the composition of a specific fraction. Paths also have been outlined, for a variety of bulk compositions, from the maximum temperature for separation of quartz ± alkali feldspar to the solidus temperature of 655° C under conditions of perfect fractional crystallization in which successive precipitates are immediately isolated from the silicate liquid. These crystal fraction paths are marked by discontinuities that indicate abrupt, discrete changes in composition, and with falling temperature, they tend to focus toward a compositional area centering about 20 NaAlSi 3 O 8 , 45 KAlSi 3 O 8 , 35 SiO 2 . The melting of crystalline assemblages that represent the haplogranite system can be considered as the inverse of equilibrium crystallization, fractional crystallization, or some combination of these idealized processes. Thus the composition path for an equilibrium crystal fraction can also serve for the tracing of compositional changes in a residue formed during equilibrium melting. A residue of given composition can be produced at constant elevated pressure under conditions representing many different combinations of temperature and bulk composition. The experimentally determined complexities of crystallization are reasonable indications of what can occur under natural plutonic conditions. They are potentially most useful, however, for the testing of genetic models based upon detailed studies of rocks in a geologic and petrographic context. They also lead to the conclusions that (1) the bulk composition of a granitic plutonite is not by itself sufficient for estimating conditions under which the rock was formed and (2) experimental data obtained for the haplogranite system under conditions that include the presence of an aqueous fluid phase ( P H 2 O ≈ P fluid = P tot ) are highly restricted in their pertinence to the crystallization histories of most granitic plutonites.