Myron G. Best. Igneous and metamorphic 2003 Blackwell Science

Подождите немного. Документ загружается.

Differentiation of Magmas

337

O + K

O MgO

THOLEIITIC TREND

Skaergaard

Palisades

Total Fe as FeO

CALC-ALKALINE TREND

Medicine Lake

(a)

12.22 Comparison of tholeiitic and calc-alkaline differentiation trends. (a) Data points are rocks from the Skaergaard intrusion (data from

McBirney, 1996; see also Table 12.6), the Palisades sill (data from Shirley, 1987), and the Medicine Lake volcanic ﬁeld (data from Grove

and Baker, 1984). (b) Palisades rocks are not plotted. The line with arrow through the Skaergaard rocks points in direction of more

evolved rocks, the two most Fe-rich lying off the diagram at the indicated coordinates.

bole that are absent generally from the cumulates, is

believed to have crystallized from locally more hydrous

and buoyant intercumulus melts rising diapirically

through the cumulus pile. The source of the water may

be the intercumulus melt itself or perhaps water ab-

sorbed from country rock.

The origin of widespread thick layers of anorthosite

in some layered intrusions, as much as 600 m in the

Stillwater, is unresolved but cannot be due to sec-

ondary processes.

Meter-thick layers (“reefs”) of pegmatitic pyroxen-

ite in the Bushveld and Stillwater are major world

repositories of platinum-group elements (PGEs), con-

sisting of Pt, Pd, Rh, Ir, Ru, and Os. The traditional

view has been that the PGE-bearing minerals were

concentrated as dense immiscible sulﬁde melts on the

intrusion ﬂoor. However, the intimate association of

the PGE minerals with pegmatite and hydrous miner-

als as well as exceptionally high concentrations of rare

earth elements (REEs) in pyroxenes are interpreted in

terms of redistribution of interstitial melt driven by

compaction of the cumulate layers (Mathez et al.,

1997).

12.4.3 Oceanic-Ridge Magma Chambers

After polybaric fractionation of olivine during their as-

cent from the mantle source, primary mid–ocean ridge

basalt (MORB) magmas experience further differenti-

ation at low P in crustal storage chambers beneath the

rift.

Intuitively, it seems unlikely that the crustal cham-

bers ﬁlled with magma would remain perfectly closed

systems throughout their entire history of solidiﬁca-

tion. In the actively spreading rift environment, fresh

(46.00, 308)

(49.09, 75)

Calc−alkaline

50 60 70

SiO

(wt.%)

FeO/MgO

(27.68, 308)

(28.67, 75)

0 10 20 30

Total Fe as FeO (wt.%)

Skaergaard

(b)

338 Igneous and Metamorphic Petrology

draughts of more primitive magma would be expected

to be episodically injected into the chamber of frac-

tionating magma from the underlying mantle source

(Figure 12.12). Three lines of evidence were found by

early investigators (e.g., Rhodes et al., 1979) indicating

chamber replenishment and mixing:

1. Disequilibrium phenocryst-melt compositions and

textures: Many larger olivines and especially pla-

gioclases have corroded, spongy cores (Fo

85–90

and

83–86

) that are surrounded by more evolved crys-

talline material.

2. Glass inclusions within the high-T cores have com-

positions similar to the most primitive MORB and

unlike the glass in the surrounding matrix.

3. Element variation trends follow those anticipated

by fractionation of major phenocrystic phases com-

bined with mixing of replenishing draughts of

primitive magma (Figures 12.23–12.25).

Thus, the limited compositional variation in most

MORBs reﬂects recurring replenishment and mixing of

relatively primitive mantle-derived magmas with frac-

tionating magma in rift chambers, moderating the ef-

fects of crystal-melt fractionation.

However, less common seaﬂoor basalts found lo-

cally along the Galapagos rift and other sites on the

East Paciﬁc Rise follow a tholeiitic fractionation trend

of extreme Fe-enrichment and limited enrichment in

silica and alkalies resembling the Skaergaard trend

(Figure 12.22). Apparently, in local more closed-system

chambers, 60–80% fractionation of olivine, plagio-

clase, and lesser clinopyroxene produces ferrobasalts

(13–16 wt.% total Fe as FeO) and less common ferro-

dacites (5–11 wt.% FeO, 64–71 wt.% SiO

12.5 ORIGIN OF THE CALC-ALKALINE

DIFFERENTIATION TREND

Several variation diagrams have been employed to dis-

tinguish between the tholeiitic differentiation trend of

Fe enrichment with limited felsic derivatives and the

calc-alkaline trend of limited Fe enrichment and abun-

dant felsic derivatives (Figures 2.17 and 12.22; see also

Miyashiro, 1974). The tholeiitic rock suite is typical of

differentiated basaltic intrusions, as just described, as

well as island arcs, whereas the calc-alkaline suite is

typical of continental margin arcs.

These contrasting rock suites and their associated

magmatic differentiation trends have been recognized,

and their origin debated, for many decades. Prior to

the mid-twentieth century, C. N. Fenner believed that

crystallizing basaltic magmas evolved along a trend of

Fe enrichment at nearly constant silica concentration,

12.23 Crystal-melt fractionation in MORB magmas. Analyzed glasses

from Bryan and Moore (1977) plot in shaded area. Filled

squares are analyzed olivine (Ol) phenocrysts and dashed line

is olivine control line. Accumulation or subtraction of olivine

produces most of the compositional variation in the glasses,

whereas limited fractionation of plagioclase (Pl) and pyroxene

(Px) from low-MgO magmas shifts the trend slightly toward

more FeO-rich compositions.

10 Kbar

1 atm

Mixing

Cpx + Ol

+ Pl fract.

Cpx + Pl

+ melt

Qtz

Opx +

Pl + melt

Ol + Pl + melt

Olivine

Fractionation

12.24 Compositional relations in seaﬂoor basalts plotted in terms

of normative Di, Ol, and Q. Plagioclase and/or spinel is stable

in all assemblages. Experimentally determined beginning-of-

melting invariant points in peridotite source rock at 20, 15, and

10 kbar, and 1 atm. Curved boundary line on 1 atm liquidus.

Shaded area encompasses nearly 2000 analyses of basalts that

are interpreted to have formed by polybaric fractionation of

olivine from primary melts generated at 10–20 kbar. Subse-

quent low-P fractionation of plagioclase, clinopyroxene, and

olivine from these primitive parent magmas within crustal

magma chambers yields evolved magmas extending down the

boundary line. Many of these derivative magmas have been

modiﬁed by mixing with more silica-rich magmas. Signiﬁcant

enrichment in Fe in derivative magmas cannot be represented

in this diagram. (Redrawn from Natland, 1991.)

Differentiation of Magmas

339

now designated the tholeiitic trend and exempliﬁed by

the extreme Fe enrichment of the Skaergaard intru-

sion. On the other hand, N. L. Bowen advocated a

trend of limited Fe enrichment and greater production

of felsic end products, exempliﬁed by calc-alkaline

rocks. It is now realized that there is a range of dif-

ferentiation processes that can operate on primitive

basaltic magmas, producing a continuous spectrum of

rock suites and trends whose end members are the

Skaergaard and the calc-alkaline trends. It should be

recalled that this continuum and its two end members

constitute the subalkaline suite (Section 2.4.5), which

accounts for about 90% of global magmatism (Fig-

ure 1.1); hence, contrasts in the evolution of tholeiitic

and calc-alkaline magmas merit special attention.

12.5.1 Tonga–Kermadec–New Zealand Arc

This continuous arc in the southwest Paciﬁc Ocean de-

veloped above the subducting Paciﬁc oceanic plate

during the late Cenozoic; the volcanic rocks in the arc

provide one of the clearest examples of contrasting

tholeiitic and calc-alkaline suites or trends (Figures

2.17 and 2.18). Tholeiitic rocks in the youthful Tonga-

Kermadec arc rest on oceanic crust, whereas calc-

alkaline rocks in the along-strike Taupo volcanic zone

of the North Island of New Zealand rest on continen-

tal crust (Figure 12.26). New Zealand obviously har-

bors a greater volume of silicic rocks than maﬁc-

intermediate composition rocks, whereas the reverse is

true of the oceanic Tonga-Kermadec part of the arc.

Relatively primitive basalts in the two arc segments

have similar trace element concentrations and patterns,

suggesting a similar ancestry in the subarc mantle

wedge. However, the Taupo basalts have more ra-

diogenic Nd and Sr isotopic ratios that are consistent

with assimilation of an older sedimentary component

resembling metamorphosed Mesozoic sedimentary

rocks exposed to the east of the volcanic zone.

If the 16,000 km

of extruded rhyolite magma in the

Taupo volcanic zone originated wholly by fractionation

of parental basalt magma, about 10 times as much

basalt magma would have been required. This ques-

tionable volume plus the relative minor amounts of

contemporaneous intermediate andesite and dacite

make such an origin unlikely. On the other hand, the

high crustal heat ﬂow and intense geothermal activity

associated with the voluminous rhyolite volcanism

suggest a major heat source in the crust; this is specu-

lated to be underplated basalt magma from the man-

tle wedge. Compared to fractionation origin, rhyolite

magma generation by partial melting of older conti-

nental rocks requires a signiﬁcantly smaller mass of

basalt magma, approximately equal to that of the lower

crustal rocks being melted (Section 11.1.1). That the

erupted basaltic magmas in the Taupo zone appear to

show crustal contamination supports the hypothesized

model.

12.5.2 Factors Controlling Development of the

Calc-Alkaline Trend

Water is an essential component in arc magma genera-

tion, evolution, and typical explosive behavior. Water

promotes partial melting in the hydrated mantle wedge

overlying subducting wet oceanic crust. Subsequently,

these hydrous primitive magmas evolve along distinct

trends. High water fugacities stabilize amphiboles and

biotite and depolymerize the melt, expanding the sta-

bility ﬁeld of olivine while restricting plagioclase. En-

hanced olivine crystallization reduces Mg, Fe, Ni, and

Cr concentrations in residual melts. Restriction of pla-

gioclase precipitation to calcic compositions, An

70–90

enriches residual melts in Na and K. These factors are

some of those responsible for development of the felsic

calc-alkaline trend, rather than the Skaergaard trend of

increasing Fe.

In continental arcs, at least three possible factors,

not present in island arcs, might modify primitive mag-

mas derived from the mantle wedge, producing greater

proportions of felsic calc-alkaline rocks:

1. Sediment eroded off a high mountain terrain and

deposited in the adjacent ocean trench, as well as

tectonic slices carved off the edge of the overriding

plate, may be subducted with the oceanic crust.

This continental material may be partially melted,

Percent crystal fractionation

TiO

(wt. %)

Hypothetical

parent magma

Mixing line

Olivine + plagioclase

fractionation trend

Augite

fractionation

trend

12.25 Effects of crystal-melt fractionation and magma recharge and

mixing on the TiO

content of seaﬂoor basalts. A hypothetical

parent magma having 49.9 wt.% SiO

and 0.7 wt.% TiO

from which olivine, plagioclase, and augite of realistic compo-

sitions are fractionated produces residual melts along the

curved fractionation-trend lines. Intermittent mixing of such

residual melts with additional draughts of parental magma

would produce magmas whose compositions lie within the

shaded lens-shaped area bounded by the mixing line and the

curved fractionation trend lines. The shaded area represents

over 600 chemically analyzed seaﬂoor glasses. (Redrawn from

Rhodes et al., 1979.)

340 Igneous and Metamorphic Petrology

enhancing production of felsic magmas in the sub-

arc environment.

2. Subcontinental lithospheric mantle may have been

static (nonconvecting) and metasomatically en-

riched over long periods; ascending magmas scav-

enge these enrichments.

3. Contributions from the continental felsic crust it-

self are probably important. Where the continental

crust is thicker there is a tendency for a greater

proportion of K

O at a given SiO

content, greater

proportion of felsic rock types (dacite and rhyolite),

and higher proportion of K

O/Na

O in andesite

(Figure 12.27); these compositional parameters in-

dicate that the longer the path through which mag-

mas rise in their journey to the surface of the conti-

nental crust the more they take on felsic continental

characteristics. The ﬁrst factor listed is indepen-

dent of the path length and, thus, probably less sig-

niﬁcant. The “scatter” in the global data of Figure

12.27 likely reﬂects the compositional heterogene-

ity of the continental crust and other irregularities

in differentiation processes.

The greater the path length that ascending magmas

take through the continental crust the greater the op-

(a)

12.26 Tonga–Kermadec–New Zealand volcanic arc and arc rocks in the southwestern Paciﬁc. (a) Volcanic islands indicated by ﬁlled triangles.

Bathymetric contours are in fathoms (1 fathom  1.83 m). For a more detailed, larger-scale map of the North Island of New Zealand see

Figure 13.27. (b) Frequency distribution of silica in analyzed rocks from the Tonga-Kermadec island arc and from the North Island of

New Zealand. Tonga is chieﬂy basalt and andesite; sparse dacite is overrepresented. In New Zealand, rhyolite is estimated to be about

20 times more voluminous than andesite. (Redrawn from Ewart et al., 1977.)

50 55

70 75

SiO

(Wt.%)

NEW ZEALAND

Rhyolite

Basalt

Andesite

Dacite

TONGA − KERMADEC

Number of analyses

16,000

800

(b)

Differentiation of Magmas

341

portunity for assimilation and fractional crystallization

to produce felsic calc-alkaline magmas. Mantle-derived

basaltic magmas intruded in the lower crust can assim-

ilate felsic rock and mix with silicic partial melts gener-

ated by the heat from the crystallizing and usually frac-

tionating basalt magma. The combined effects of

simultaneous assimilation and fractional crystallization

(AFC) on evolving magma systems were ﬁrst formal-

ized by Taylor (1980) and further quantiﬁed by De

Paolo (1985) and Aitcheson and Forrest (1994). Their

equations express isotopic ratios and element concen-

trations of hybrid daughter magmas to be expected

from an assumed contaminant and parent magma

based on the following:

1. Estimated bulk distribution coefﬁcients

2. Ratio of the rates of assimilation to fractional crys-

tallization

3. Remaining melt fraction

4. Effects of magma recharge and eruption

The ﬁrst three parameters are commonly chosen to

provide the best ﬁt to the analytical data (Figure

12.28). This AFC model works best in cases in which

the contaminant and parent magma compositions dif-

fer signiﬁcantly.

The basalt-andesite-dacite-rhyolite suite in the

Medicine Lake volcano in the northern California

Cascade volcanic belt (Grove and Baker, 1984) is typ-

ical of many calc-alkaline suites. A general trend of in-

creasing

Sr/

Sr with increasing silica is consistent

with progressive assimilation of felsic crust in mantle-

derived magmas. Elevated incompatible element (e.g.,

Rb) concentrations in some evolved rocks indicate that

fractional crystallization alone cannot be responsible,

but combined fractionation and assimilation of felsic

rocks can (Figure 12.29a). Mixing of basaltic and rhy-

olitic magmas also seems required for some rock types

and is supported by mingled dacite and rhyolite lavas

and by disequilibrium phenocryst assemblages in

basaltic andesites and dacites; these assemblages in-

clude, in a single sample, Mg-rich olivine (Fo

), calcic

plagioclase (An

), reversely zoned orthopyroxene

with Fe-rich cores and Mg-rich rims, and reversely

zoned plagioclase with andesine cores and labradorite

0.5 1.0 1.5 2.0 2.5 3.0

% andesite + dacite + rhyolite

O/Na

O (wt. %) in andesite

Taupo, New Zealand

Tonga

O (wt. %) at 57.5 (wt. %) SiO

Crustal thickness (km)

12.27 Correlation between thickness of continental crust (30 km)

and compositional parameters. The graphed parameter is the

concentration of K

O at 57.5 wt.% SiO

from a best-ﬁt line

through a K

O versus SiO

wt.% variation diagram for a rock

suite; for example, values of K

O at 57.5 wt.% SiO

for the

New Zealand suite and, for comparison, the Tonga suite are

taken from Figure 2.18. Bars in diagram indicate variable

crustal thickness for the K

O value. Two other parameters not

graphed also show a positive correlation with crustal thick-

ness; these are the K

O/Na

O in andesite (57–63 wt.% SiO

)

and the proportion of andesite  dacite  rhyolite to basalt.

(Data from Leeman, 1983; Hildreth and Moorbath, 1988.)

40 80 120 160

200

300

400

500

52.3

52.4

52.7

51.7

57.2

59.6

68.7

AFC

Fractional

crystallization

Rb (ppm)

12.28 K/Rb versus Rb in the Volcanes Planchon-Peteroa-Azufre at

3515S latitude in the central Andes. Numbers alongside solid

circles are weight percentage silica in rock samples that range

from basalt to dacite. For reference only are approximate

trends to be expected (1) if a parent magma, represented by

the rock that has 51.7 SiO

, were to evolve wholly by fractional

crystallization of olivine, pyroxene, and plagioclase or (2) if

fractional crystallization acted in concert with assimilation of

granite (AFC). Note that fractional crystallization alone is in-

adequate to account for these diverse evolved rocks because of

the similarity in partition coefﬁcients of K and Rb. In contrast,

assimilation of granite country rocks in concert with fractional

crystallization is a more viable diversiﬁcation process. (Re-

drawn from Hildreth and Moorbath, 1988.)

46 50 54 58 62 66 70 74

SiO

(wt. %)

120

160

200

240

280

120

160

200

Ni (ppm)

Rb (ppm)

(a)

Rhyolite

Dacite

Basaltic-andesite

Andesite

High-Al basalt

Mixing

AFC

Fractional crystallization

AFC

r = 0.11

r = 0.4

(b)

Parent magma

AFC

Mixing

12.29 Evolution of Quaternary lavas at Medicine Lake, northern California, by fractionation, assimilation, and mixing. Fractionation of

parental high-Al basalt magma modeled by extraction of olivine  plagioclase  clinopyroxene is shown by line with tick marks at in-

tervals of 0.1 in melt fraction to F  0.5. The fractionation model accounts for the observed variations in the high-Al basalts but not other

magmas that must have originated by mixing of the evolved high-Al basalt magmas with rhyolite magma together with combined assim-

ilation and fractional crystallization (AFC). The AFC model assumes the ratio (r) of the rate of assimilation of feldspathic sandstone and

rhyolite to the rate of fractional crystallization was 0.11 (long dashes) or 0.4 (short dashes). Fractionation involving magnetite and horn-

blende could not reproduce the whole suite of lavas. (Redrawn from Grove et al., 1982.)

Differentiation of Magmas

343

46 50 54 58 62 66 70 74 78

SiO

(wt. %)

46 50 54 58 62 66 70 74 78

SiO

(wt. %)

O (wt. %)

Cpx

Hbl

Crystal-melt

fractionation

Assimilation

and mixing

Partial melts

Granite

High-K

Shoshonitic

Low-K

Medium-K

(a)

(b)

Gabbro Diorite Tonalite Granodiorite

12.30 K

O-SiO

relations in calc-alkaline plutonic arc rocks. (a) Composition of rocks in the Sierra Nevada batholith, California (ﬁlled circles;

from data referenced in Bateman, 1992) and in the Coastal batholith of Peru. (Shaded area from Atherton and Sanderson, 1987.) (b) On

the left, trends of residual melts resulting from fractionation of indicated minerals in parental basalt and andesite magmas (open circles).

On the right, compositions of experimentally generated, dehydration partial melts of pelitic source rocks (open symbols),

rock (ﬁlled circles), and feldspathic sandstone (pluses). (Data from Conrad et al., 1988; Skjerlie and Johnson, 1993;

Patiño Douce and Harris, 1998; Pickering and Johnson, 1998.) Calc-alkaline batholithic rocks can be created from fractionated basaltic

and andesitic parental arc magmas that are mixed with partial melts of continental source rocks.

Pl  Hbl  Qtz

rims. Elevated compatible element (e.g., Ni) concen-

trations at a particular silica content in basaltic an-

desites and andesites were apparently produced by

mixing of high-Al basalt and rhyolite magmas (Figure

12.29b).

The remarkably similar I-type granitic rocks of the

Sierra Nevada batholith in California and the Coastal

batholith of Peru are plotted in K

O-Si

O space in Fig-

ure 12.30a. These two plutonic calc-alkaline suites can

be compared to the K

O-Si

O plot in Figure 12.30b,

which shows trends of residual melts resulting from

fractionation of the indicated minerals from parental

basalt and andesite magmas. Parent basalt magma

evolves toward andesite chieﬂy by fractionation of

olivine. Fractionation of clinopyroxene, Fe-Ti oxides,

and hornblende from andesite magmas propels evolved

melts along the calc-alkaline trend exhibited by the

batholithic rocks in Figure 12.30a. Fractional crystal-

lization of water-rich primitive magmas generated in

the subarc mantle wedge causes residual melts to fol-

low the calc-alkaline trend (Grove et al., 1982), in

part by stabilizing hornblende, which is widely be-

lieved to be a major player in fractionating calc-alkaline

magmas. Sidewall crystallization and accompanying

convective melt fractionation (Figure 8.22) are prob-

ably important phenomena in production of calc-

alkaline andesite-dacite-rhyolite magmas. Thick conti-

nental crust plays an important role in arresting or at

least slowing maﬁc magmas during ascent, allowing

them to fractionate.

344 Igneous and Metamorphic Petrology

0.702 0.704 0.706 0.708 0.710

0.5120

0.5124

0.5128

0.5132

MORB

Island

BASALTS

Continental arc

0.51264

0.7052

arc

143

144

Sr/

12.31 Nd and Sr isotope ratios in arc basalts compared to MORB.

Basalts from western Paciﬁc island arcs are slightly more ra-

diogenic than MORB because of the slab-derived aqueous

ﬂuid component, whereas continental arc basalts (Japan,

Philippines, New Zealand, Ecuador, Central America, Lesser

Antilles) range to much more radiogenic ratios because of in-

teraction with highly radiogenic old continental crust in which

143

Nd/

144

Nd can be 0.510 and

Sr/

Sr 0.900. (Redrawn from

Tatsumi and Eggins, 1995.)

Thick continental crust also promotes assimilation

of country rock into primitive magmas. If open magma

systems are contaminated by surface-derived ﬂuids

and oxidized volcanic rocks, the relatively higher f

hybrid magmas can precipitate magnetite early in their

evolution, depleting residual melts in Fe and enhancing

a calc-alkaline trend of silica and alkali enrichment.

A high f

is consistent with the lack of a negative

Eu anomaly in many continental arc rocks. On the

other hand, under f

, reducing conditions, generally

in magma systems open only to replenishment by prim-

itive nonarc basalt magma, such as in MORB magma

chambers, the Fe-enriched tholeiitic trend develops.

Also plotted in Figure 12.30b are experimental

melts generated from pelitic (clay-rich progenitors)

and maﬁc source rocks by dehydration partial melt-

ing (Section 11.6.1). These partial melts are silicic,

mostly about 75 wt.% SiO

, but range widely in K

concentration, reﬂecting corresponding variations in

the source rock. Mixing of these partial melts with

primitive as well as fractionated magmas is obviously

an important process in production of calc-alkaline

magmas.

Magmas developed in thick continental crust com-

monly have relatively radiogenic Sr and Nd isotope

ratios created by contamination with ancient

Sr-

enriched rock (Figure 12.31). However, the absence

of an elevated

Sr/

Sr ratio cannot be taken as evi-

dence for no crustal contamination because in some

arcs, such as in the southern Andes, the crust is not

very old and has similar isotopic composition to that

of the primitive magmas. In this instance, contamina-

tion can only be discerned with trace element data or



O values if hydrothermally altered rocks were as-

similated.

SUMMARY

A wide spectrum of magma-generating conditions in

compositionally variable sources compounded with a

wide range of differentiation overprints account for the

wide spectrum of magmatic rock compositions found

on Earth (e.g. Figure 2.4).

Despite their treatment in separate chapters, magma

generation and differentiation in natural systems may

be a continuum of processes, acting sequentially or

possibly, to some degree, simultaneously. For example,

it is reasonable to believe that, as basaltic melts seg-

regate and begin to ascend buoyantly out of their

mantle source, some precipitation of olivine might im-

mediately occur during initial decompression and cool-

ing. Or, as felsic magma is generated in the lower con-

tinental crust, it may simultaneously mix with basaltic

magma that is providing heat for partial melting; as-

similation of wall rock may also concurrently occur.

To understand the origin of magmas, the petrologist

must sort out signatures related to primary generation

versus secondary differentiation in suites of comag-

matic rocks, using their mineralogical, elemental, and

isotopic compositions as well as their fabric and ﬁeld

relations.

Few magmas are likely to rise very far from their

source without overprinting modiﬁcations in composi-

tion. Overprinting processes include closed-system dif-

ferentiation—chieﬂy crystal-melt fractionation—and

open-system mixing of magmas and assimilation of

country rock. Crystal-melt fractionation by crystal set-

tling, ﬁlter pressing, and convective melt fractionation

is important in the differentiation of tholeiitic magmas

emplaced in continental and oceanic crust, producing

a trend of Fe enrichment with limited felsic differenti-

ates. These basaltic intrusions may be open to replen-

ishment by more draughts of primitive magma that mix

with the crystallizing magma. In MORB chambers this

replenishment and mixing apparently occur so fre-

quently that modiﬁcation of magma composition is

limited; consequently, most MORB constituting the

ﬂoor of the oceans is relatively uniform tholeiitic basalt.

Mantle-derived basaltic magmas can underplate or

stall within the lower continental crust, cool, fraction-

ally crystallize, and give off latent heat to allow assimi-

lation of country rock (AFC). Mixing of maﬁc and more

silicic magmas also occurs. Altogether, these processes

produce calc-alkaline rock suites typical of continental

margin magmatic arcs. The fractional crystallization ac-

Differentiation of Magmas

345

mas with continental crust a viable means of

creating highly potassic magmas that form the

shoshonite series? Discuss.

12.11 What are the different ways that andesitic

magmas can be produced in arcs? What trace

element and isotopic attributes would distin-

guish among these?

12.12 Rocks originating by mingling of two magmas

and by physical separation of immiscible melts

may be superﬁcially similar. What rock attrib-

utes would distinguish between these origins?

PROBLEMS

12.1 Evaluate the validity of hornblende fractiona-

tion from a parental andesite magma in produc-

ing rhyolite and dacite daughter magmas. Use

compositions from Table 2.2 and Appendix A

on two-element variation diagrams. Critically

discuss your results.

12.2 In Figure 12.29a, what was the approximate

proportion of rhyolite magma that mixed with

andesite magma that contains 60 wt.% SiO

create dacite that contains 69.5 wt.% SiO

12.3 Write a stoichiometrically balanced reaction for

assimilation of Al-rich minerals and rocks by

basalt magma to yield anorthite and enstatite

components in the hybrid contaminated magma

at the expense of diopside in the basalt. Use

SiO

to represent the aluminous material.

What bearing might this assimilation have on

the origin of some leuconorite?

12.4 Write balanced mineralogical reactions showing

how: (a) An initially subalkaline, silica-saturated

magma that contains normative enstatite and or-

thoclase could become silica-undersaturated with

normative diopside and leucite by assimilating

limestone. (b) Silica from the magma can metaso-

matize wall rock and xenolithic limestone and

dolomite, producing calcium silicate minerals

such as wollastonite and diopside. Discuss how

effective these reactions are in modifying the de-

gree of silica saturation in the original magma.

12.5 Examine the consequences of assimilation of

crystalline material into magma using the binary

system NaAlSi

-CaAl

as a model. Con-

sider a model magma that contains crystals of

plagioclase An

in equilibrium with melt into

which crystals of cool An

and of An

are in-

troduced. Compare and contrast the subsequent

evolution of the two magmas as they crystallize.

12.6 Le Roex et al. (1990) show that the basanite-

phonotephrite-tephriphonolite-phonolite se-

quence on Tristan da Cunha (Figure 2.16) was

created by crystal-melt fractionation. This

companying assimilation is particularly effective in pro-

ducing the calc-alkaline trend if it occurs under ele-

vated water and oxygen fugacities. Thicker continental

crusts allow more opportunity for AFC and mixing, so

that in thickest crusts, as in the central Andes of South

America, little, if any, primitive mantle-derived basaltic

magma is extruded and most shallow intrusions and

extrusions are of rhyolite, dacite, and andesite.

CRITICAL THINKING QUESTIONS

12.1 Essentially monomineralic layers of pyroxen-

ite, dunite, or anorthosite in large differenti-

ated basaltic intrusions can only form by crys-

tal accumulation because no magmas of

equivalent composition are known to occur.

What do you think is the evidence for no such

magmas? (Hint: Consider the volcanic record.)

12.2 How do monomineralic layers (see preceding

question) originate? Even the least “porous”

cumulate still has a signiﬁcant proportion of

interstitial basaltic melt entrapped between the

cumulus crystals.

12.3 Why are Sr and Nd isotopic ratios not used to

evaluate details of crystal-melt fractionation in

magma systems?

12.4 Explain the origin of the seven textural types

of olivine grains in the 1959 summit eruption

of Kilauea, Hawaii (Special Interest Box 12.1).

Which are mantle-derived, cumulates, or pre-

cipitates of ascending magma?

12.5 Mixed magmas erupted from volcanoes can

have an essentially homogeneous melt, but

crystals are commonly markedly inhomoge-

neous and display disequilibrium textures and

locally complex zoning. Why?

12.6 Why is assimilation probably insigniﬁcant in

the evolution of MORB magmas crystallizing

in oceanic crustal magma chambers?

12.7 Critically evaluate the origin of pegmatites, es-

pecially their distinctive fabric, in terms of the

Jahns-Burnham model involving volatile-over-

saturated magma systems versus the London

model involving kinetic factors in an under-

cooled magma system.

12.8 Contrast primary, primitive, and parent

magma—their differences and ways each can

be discerned.

12.9 (a) Does the immiscible Fe-rich glass in

mesostases of some tholeiitic basalts (column

1, Table 12.3) resemble any common silicate

rock? (b) In what ways does it differ from

other rocks of comparable silica content?

12.10 Examine Figures 12.30, 2.18, and 13.24. Is

contamination of mantle-derived basaltic mag-

Table 12.7. Concentrations of SiO

, Rb, and Sc and Distribution Coefﬁcients for Rb and Sc in Tristan da Cunha Lavas

ASANITE

HONOTEPHRITE

EPHRIPHONOLITE

HONOLITE

Cpx TiMag Pl Amph Alk feld

Rb 0.01 0.01 0.01 0.5 0.6

Sc 4 20 0.01 4 0.01

SiO

concentrations rounded to nearest whole number weight percentage; Rb and Sc, parts per million; A, aphyric; P, porphyritic; SP, sparsely porphyritic.

(From Le Roex et al., 1990.)

SPAAAASPSPASPAP A

SiO

44 45 44 46 46 46 44 46 44 46 50 48

Rb 73 64 62 75 73 78 74 71 65 76 112 90

Sc 18.1 17.1 22 14.5 14.7 13.9 17 12.7 14.6 10.3 4.4 10.1

AP PPA P P A A SPSPP

48 51 50 48 48 55 57 56 55 55 61 58 59

95 119 88 87 129 151 153 139 128 176 151 165

9.3 6.2 5.4 6.5 7.8 3.1 2.7 2.6 2.5 3.4 1.3 1.6 1.3