ORIGINAL PAPER

K. Rankenburg Æ J. C. Lassiter Æ G. Brey

Origin of megacrysts in volcanic rocks of the Cameroon volcanicchain – constraints on magma genesis and crustal contamination

Received: 25 January 2002 / Accepted: 27 August 2003 / Published online: 7 February 2004� Springer-Verlag 2004

Abstract Lavas of the Biu and Jos Plateaus, NorthernCameroon Volcanic Line (CVL), contain abundantgenetically related megacrysts of clinopyroxene, garnetand subordinately plagioclase, ilmenite and amphibole.P, T-estimates of crystallization for the primitive groupof cpx and gnt megacrysts are 1.7–2.3 GPa and�1,400 �C. Because crustal thickness in these areas isonly �30 km (�0.9 GPa), megacrysts must have formedwithin the lithospheric mantle. Primitive Biu and Joslavas are isotopically heterogeneous in Sr-Nd isotopespace (87Sr/86Sr=0.70285–0.70360 and �Nd=7.5–4.6).Biu Plateau megacrysts overlap the range of Biu lavas inSr-Nd isotope composition, indicating that crustal con-tamination of Biu lavas was minor. Jos Plateau lavas areisotopically more enriched than their associated mega-crysts. Therefore an additional contamination of Joslavas due to assimilation of continental crust (�5%) orenriched shallow lithospheric mantle is indicated. Lavasof Biu and Jos Plateau do not reflect simple fraction-ation or equilibrium crystallization products, but insteadreflect mixing between primary melts and their frac-tionated derivatives.

Introduction

The Cameroon Volcanic Line (CVL) comprises agenetically related series of Cenozoic intraplate volca-

noes that extends for 1,600 km from the Island of An-nobon in the South Atlantic Ocean to the continentalinterior of West Africa (Fig. 1). There are 12 main vol-canic centers, which range in age from 30 Ma to present(Fitton and Dunlop 1985) and 17 anorogenic ringcomplexes in the continental sector with ages rangingfrom 66 to 30 Ma (Fitton 1987; Kraml et al. 2001).Magmatism of this line cannot be a simple expression ofthe African plate sliding over a fixed mantle hotspotbecause there has been no systematic change in the focusof magmatism for 30 Ma and Plio/Pleistocene volcanismyounger than 5 Ma is dispersed all along the line (e.g.Annobon, Bioko, Mt. Cameroon, Jos and Biu Plateau).Fitton (1980) showed that basaltic rocks in the oceanicand continental sectors are geochemically and isotopi-cally (Sr, Nd) similar and suggested that a line or zone ofhot asthenospheric mantle is upwelling underneath theCVL, forming the source magmas without any sub-stantial involvement of the lithosphere. However, thereis clear isotopic evidence for interaction with the conti-nental crust in some evolved magmas of the continentalsector. For example, Marzoli et al. (1999) pointed outthat there is isotopic evidence for the involvement ofcontinental crust in the most evolved silicic volcanoes ofOku, Bambouto and Ngaoundere, which have 87Sr/86Sras high as 0.705–0.714. It is unclear whether the heter-ogeneity observed in more primitive magmas is relatedprimarily to heterogeneity in the asthenospheric sourceregion of the lavas and represents mixing of a depletedand a more enriched end member, or if part of thisheterogeneity is due to melt interaction with continentallithospheric mantle, or melt interaction with continentalcrust. One way of constraining the role of crustal con-tamination in a suite of lavas is to compare the isotopiccomposition of the lavas with cogenetic phenocrystic orxenocrystic phases that precipitated in the mantle. Lavasfrom Biu and Jos Plateau contain abundant megacrystsof clinopyroxene (cpx) and garnet (gnt). In this paper,we examine petrologic and petrographic evidenceshowing that cpx and gnt megacrysts precipitated fromCVL magmas within the lithospheric mantle. We then

Editorial responsibility: I. Carmichael

K. Rankenburg (&) Æ G. BreyInst. fuer Mineralogie, Senckenberganlage 28,60054 Frankfurt/Main, GermanyE-mail: ranken@illite.u-strasbourg.fr

J. C. LassiterAbt. Geochemie, Max-Planck-Institut fuer Chemie,Postfach 3060, 55020 Mainz, Germany

Present address: K. RankenburgAbt. Geochemie, Max-Planck-Institut fuer Chemie,Postfach 3060, 55020 Mainz, Germany

Contrib Mineral Petrol (2004) 147: 129–144DOI 10.1007/s00410-003-0534-2

compare the isotopic and trace element compositions ofthe host lavas with those of the megacrysts to constrainthe role of crustal contamination and magma mixing inthe genesis of CVL magmas.

Geologic setting: The Benue Trough and the Biuand Jos Plateaus

The continental sector of the CVL is of y-shaped form(cf. Fig. 1). Whereas previous publications consider theBiu Plateau as the end of the NNW branch of the con-tinental sector of the CVL, the Jos Plateau is usually notassigned to CVL volcanism because of its offset to theline axis. However, as we will show in a later section, thetiming of the Jos Plateau volcanism is very similar toother CVL volcanic centers. The major and trace ele-ment chemistry of the Biu and Jos Plateau lavas aresimilar and also span a similar range in isotopic com-positions, overlapping the data of the CVL as a whole.Extensive geologic mapping might reveal a more dis-persed pattern of volcanism in the continental sector of

the CVL in which the Jos Plateau is not as isolated as toour present knowledge. For the remainder of this paperwe assume that the Jos Plateau volcanism is geneticallyrelated to the CVL.

The Mesozoic to early Cenozoic magmatism of theBenue Trough can be divided into two principal do-mains: In the Northern Benue, magmatism is charac-terized by transitional alkaline and tholeiitic basalts andsome dispersed peralkaline acidic rocks with ages rang-ing from 147–106 Ma (Maluski et al. 1995). TheSouthern Benue shows two additional episodes withalkaline intrusive rocks of 97–81 Ma age, and alkalineintrusions followed by tholeiitic volcanism of 68–49 Maage (Maluski et al. 1995). In the Northern Benue CVL-related magmatism restarted in Miocene times withintrusion of trachyte-phonolite plugs which have beendated by Grant et al. (1972) between 22–11 Ma andcontinued with the magmatic province of the Biu Pla-teau.

According to Turner (1978), the Biu Plateau mainlyoverlies granitic basement rocks, whereas to the W andN, basalts of the Biu Plateau have spread over Creta-ceous sediments, mainly arkosic Bima sandstone. TheBiu Plateau was constructed in three stages during twoperiods of volcanism: (1) An early fissure type eruptionand (2) formation of relatively large tephra ring volca-noes and building up of localized thick lava piles (up to250 m) in the southern part of the plateau. Lavas of thisplateau building stage range in composition from hy-normative basalt to basanite and K/Ar-ages range from5.35 to 0.84 Ma (Grant et al. 1972; Fitton and Dunlop1985). Extensive weathering and laterite formation sug-gests a break period after this episode. (3) Resumptionof igneous activity with formation of over 80 NNW-SSEaligned cinder cones with similar chemistry to earlierbasalts. A rough estimate of the age of the last magmaticperiod is <50,000 years based on diffusional constraintsof He in mantle xenoliths of the CVL (Barfod et al.1999) and >25,000 years based on pollen dating of maarsediments from the Biu Plateau (Salzmann 2000). Al-though the younger volcanoes of the Biu Plateau areprominent landscape features, individually they are onlysmall volcanic structures. Many of them are quite typicalcinder cones, examples include some of the mostprominent peaks: Pelamabelu and Ga Tila. Others arebroader and lower with wide flat crater floors corre-sponding more closely to tuff cones and rings (e.g. Mi-ringa), or cut into country rock below general groundlevel (maars). Recent volcanics also include one occur-rence of phonolite, which contains inclusions of spinel-peridotite and megacrysts of amphibole.

As with the Biu Plateau, volcanic activity of JosPlateau (located approximately 400 km to the NW ofthe central CVL axis; cf. Fig. 1) occurred during twoperiods and thus the basalts have been divided into anearlier and more recent group (McLeod et al. 1971).There are no isotopic age determinations available forthe older basalts, but Wright (1976) suggests a Paleoceneage, roughly synchronous with the Benue Trough fold-

Fig. 1 Geological map showing the eruption ages of the majorvolcanic centers of the Cameroon Volcanic Line and the Gulf ofGuinea (adapted from Fitton and Dunlop 1985, ages from Fittonand Dunlop 1985; Halliday et al. 1990; Lee et al. 1994; Ngounounoet al. 1997). The Jos volcanics are shifted �400 km to the NW ofthe line axis and are usually not included in CVL magmatism.However, no occurrence of continental Cenozoic volcanism hasbeen recorded west of the Jos Plateau. Sample locations areindicated as black triangle (Biu Plateau) and white square (JosPlateau)

130

ing and uplift. The more recent activity formed a groupof 22 cinder cones. Radiometric K-Ar ages (Grant et al.1972) suggest, unlike on the Biu Plateau, continuousvolcanism between 2.1 and 0.9 Ma. Since the phonolitesof both Biu and Jos Plateaus carry peridotitic xenoliths,Irving and Price (1981) inferred that extensive fraction-ation occurred at mantle depths. The crust-mantleboundary underneath the Biu/Jos Plateaus is con-strained from seismic data to depths of only 0.8/0.9(±0.2) GPa or 28/30(±6) km, respectively (Po-udjom-Djomani et al. 1995).

Sampling and analytical techniques

Samples (lavas and inclusions) described in this workwere collected during a field excursion in 1998. A sum-mary of the 30 sampling locations along with theirgeographic coordinates as determined by GPS is given inTable 1 (Biu Plateau) and Table 2 (Jos Plateau). Sam-ples were wrapped in paper and labeled plastic bags andshipped to the University of Frankfurt in steel boxes. Todate, only the volcanic rocks and the megacrysts havebeen extensively studied. However, two xenolith-derivedcpx have been studied in detail by EMP and ion probe toenlighten the significant differences between cpx mega-crysts and cpx derived from associated disaggregatedperidotites or pyroxenites.

All mineral separates were prepared following theprocedure of Zindler and Jagoutz (1988). Megacrystswere coarsely crushed, sieved and washed in 1 N HCland distilled water to improve surface quality. A subsetof handpicked grains were mounted in epoxy and stud-ied in detail by electron microprobe (Jeol JXA-8900 RL)at the University of Frankfurt. EMP-inspection alsorevealed the occurrence of melt inclusions (evolvedbenmoreitic compositions) in some ilmenite megacrysts.Fe3+-contents of cpx and gnt megacrysts were calcu-lated based upon charge balance. Highly evolved cpxcompositions yield Fe3+/Fetot of as much as 35%. Inorder to ensure accurate input data for thermobaro-metric equations, we have checked Fe3+-contents inde-pendently by Moessbauer spectroscopy on selected cpxand gnt megacrysts. The results are consistent withinerror.

Megacrysts were further analyzed for trace elements(REE, Ba, K, Sr, Y, Zr and Nb) by secondary ion massspectrometry (SIMS) on a recently upgraded CamecaIMS-3f at the MPI in Mainz. A detailed description ofthe ion probe conditions is given in Hellebrand et al.(2002). In summary, negative oxygen ions were used as aprimary source, using an accelerating potential of12.5 kV and 20 nA beam current. Positive secondaryions were extracted using an acceleration potential of4.5 kV with a 25 eV energy window, a high-energy offsetof )80 V, and fully open entrance and exit slits. The spotsize for these conditions was 15–20 lm. Each measure-ment consisted of six cycles, where in each cycle 16O,30Si, 39K, 88Sr, 89Y, 90Zr, 93Nb, 133Cs and all masses

between 134 and 180 were analyzed in this order. 30Si(3.1% isotopic abundance) was used as a reference mass,as determined by electron microprobe analysis. For eachcycle, time-corrected mass-to-30Si ratios were formedafter dead time and background corrections. A matrixscheme to correct for isobaric interferences was applied(Zinner and Crozaz 1986). Sensitivity factors weredetermined for each element on the well-studied stan-dard glasses KL2-G, ML3B-G, StHs6/80-G, BM90/21-G and ATHO-G, whereas the well-studied komatiiteglass GOR132 (Jochum et al. 2000) was used as anexternal standard. Overall absolute deviation from thestandard is <10% for Ba, K, Y, Zr and REE except Eu,Gd, Yb and <20% for Eu, Gd, Yb, Sr. Niobium, whichis present in the standard in the ppb range, revealed thelargest error of 60%. Because the HREE-concentrationsin some evolved cpx megacrysts reach the detection limitof the ion probe, results show considerably more scatteras compared to the standard glass. Moreover, detectionlimits are dependent on MREE concentrations due topeak overlapping and therefore are individual for eachsample (Zinner and Crozaz 1986).

Sr and Nd isotope analyses of basalts and megacrystswere carried out at the MPI in Mainz using a FinniganMAT 261 multicollector TIMS instrument in staticmode. Sieved and handpicked rock chips (�100 mg ofthe 0.75–1.5 mm fraction) were leached in hot 6 N HClfor 1 hour and washed ultrasonically in deionized waterbefore dissolution in HF-HNO3. Concentrates ofmegacrysts (150–250 mg) were handpicked for isotopeanalysis under a binocular microscope in dark andbright field. Grains were then leached twice in hot 2.5 NHCl for 20 minutes, then in cold 5% HF for 15 minutesin an ultrasonic bath, then rinsed with cold 2.5 N HCl toremove fluoride complexes and finally rinsed in deion-ized water. In a second microscopic reexamination allgrains with visible reaction rims were removed to insure100% optically pure separates. Grains were then dis-solved in Teflon beakers using HF-HNO3. Analyses ofleached and unleached separates revealed no significantdifferences (cf. Table 3).

After sample digestion, Sr and Nd were separated inconventional cation exchange columns following theprocedures outlined by White and Patchett (1984). Srwas loaded with TaF5 on single tungsten filaments; Ndwas measured as metal on double rhenium filaments.87Sr/86Sr and 143Nd/144Nd are normalized to86Sr/88Sr=0.1194 and 146Nd/144Nd=0.7219 respec-tively. Over the course of this study repeated analyses ofthe NBS987 standard provided a mean 87Sr/86Sr of0.71024±0.00007 (2r). Nd results were calibratedagainst the La Jolla standard which gave143Nd/144Nd=0.51186±0.00005 (2r). Because sampleswere run on two separate instruments, all data have beennormalized to reference values for standards:87Sr/86Sr=0.710250 for NBS987 and 143Nd/144Nd=0.511850 for the La Jolla standard. Blanks are <120 pgand <13 pg for Sr and Nd respectively, and are con-sidered negligible.

131

Table

1Volcanoes

oftheBiu-Plateaufrom

Nto

S

Sample

location

Petrography

Megacrysts

ZaguHill10�56’85N

12

�05’11E

Trachybasalt,somexenoliths

Rare

cpx

JiguHill10

�55’22N

12�00’55E

JiguMaar:basanite.

Jigu1:alk.basaltsomexenoliths

No

OhneNamen

(X)10�54’40N

12

�01’21E

Alk.basalt,noxenoliths

No

PelamabeluHill(Pelajung)10

�50’55N

12�06’15E

Trachybasalt,somexenoliths(�

1cm

)No

Koroko10

�50’22N

11

�56’70E

Phonolite,sm

allxenoliths

Amphibole,phlogopite

KraterSW

PelamabeluHill(Pelaalt)10

�50’14N

12�05’65E

Basaltic

trachyandesite,

strongly

weathered,manyxenoliths

Numerouscpx,plagioclase

Dutsin

Dam

Hill10�49’30N

12

�02’56E

Dam:basaniteDam

2:trachybasaltnumerousxenoliths

Garnet,cpx,plagioclase,ilmenite,

amphibole,

apatite,olivine

DutseBugorHill10�49’24N

12�05’98E

Alk.basalt,xenolithsoxidized

Cpx,plagioclase

NowaHill(K

rater)

10�48’87N

12�01’30E

Scoriaonly

Cpx,plagioclase,ilmenite,

zircon,corundum,spinel

Hugel

SEBugor10�48’54N

12

�06’64E

Alk.basalt,somexenoliths,cpxcumulates

BosHill10�47’34N

12�01’81E

Alk.basalt,cpx-ilm

cumulates

Hilia

Hill10�47’32N

12�05’22E

Hilia

1+

2:alk.basalt,xenolithsupto

30cm

,cpx-cumulates,somefelsic

Cpx,garnet,plagioclase,ilmenite,

spinel

Tamza

Hill10�46’42N

12�03’57E

Trachybasaltmanyxenoliths(upto

10cm

)Surrounded

bylaterite

containingcpx,garnet,ilmenite

GufkaHill10�44’18N

12�04’78E

Alk.basaltxenoliths,alsofelsic

Cpx,garnet,plagioclase

Dutsin

MiringaHill10�44’04N

12�07’30E

Basaltic

trachyandesitenumerousxenoliths(felsic,

maf.

cumulates*,peridotites*

upto

50cm

)Veryabundant:garnet,ilmenite*

*,cpx,plagioclase,

spinel,olivine,

phlogopite,

apatite

Dutsin

MaldauHill10�42’83N

12

�06’04E

Alk.basaltxenolithsupto

10cm

Plagioclase,ilmenite,

garnet,cpx

Ga-G

uldumburHill10�42’30N

12�07’41E

Alk.basalt,few

xenoliths

No

PelamabeluHill210�41’18N

12�07’85E

Alk.basaltnumerousxenolithsupto

10cm

No

WigaHill10

�40’79N

12�00’25E

Alk.basalt

Few

cpx,plagioclase

Gwaram

Hill10

�39’89N

12�07’96E

Basaltic

trachyandesitefew

xenoliths

Few

plagioclase,cpx,amphibole

Ga-Zumta

Hill10�38’96N

12

�06’84E

Trachybasaltsomexenoliths

Plagioclase

Ga-H

izshiHill10

�38’72N

12�08’10E

Trachybasalt,altered

peridotitesupto

40cm

,cpxcumulates

Plagioclase

upto

5cm

,cpx

Tum

10

�36’85N

12�06’62E

Tum:alk.basaltE’Tum:trachybasalt,sm

allxenoliths

No

KudangirHill10

�34’91N

12�02’30E

Alk.basalt,nosample,veryfew

xenoliths

No

Gumja

Hill10�33’45N

12

�06’15E

Alk.basaltxenolithsoxidized,manyfelsic

Cpx,plagioclase,spinel

GaTilaHill10

�32’68N

12�08’56E

Tila1:alk.basalt,maar.TilaStr.:alk.basalt,flow

Cpx,plagioclase

*Additionalanalysesofxenolith-derived

cpxgrains.

**Additionalanalysesofilmenite-hosted

meltinclusions

132

Results

Major element, trace element and isotope data of rep-resentative megacrysts from Biu and Jos Plateau arepresented in Tables 3, 4, and 5. Detailed chemicalanalyses of the lavas will be given in a separate paper(Rankenburg et al., in review) but lavas from Biu andJos Plateau are very similar in eruption ages, major andtrace element chemistry and span a range from basanitesto phonolites with mg# ranging from 68.5 to 17.5 (all Feas FeO). In terms of isotopic compositions the volcanicson the Biu and Jos Plateau span much of the range ofprimitive CVL lavas (Fig. 2). 87Sr/86Sr ratios in lavasfrom the Biu Plateau range from 0.70285–0.70350 and�Nd from 7.5–4.8. Lavas from the Jos Plateau have al-most identical major and trace element characteristics,but extend to more enriched isotopic compositions(87Sr/86Sr=0.70320–0.70360 and �Nd=5.1–4.6).

The megacryst-suite of Biu and Jos Plateau was de-scribed in detail by Wright (1970) and Frisch and Wright(1971) and comprises chemically homogeneous crystalsof cpx, gnt, plagioclase (plag) and ilmenite (ilm) withdiameters of up to several cm, while crystals of olivine(ol), amphibole (amph), spinel (sp), apatite (apa), zircon(zr) and blue corundum (cor) are extremely rare. Onlyilmenite, spinel, zircon, corundum and evolved compo-sitions of cpx occasionally show planar faces. The mostabundant phase is a subcalcic titaniferous aluminousaugite with poor cleavage, conchoidal fracture and aglassy appearance on freshly broken surfaces. In thinsection cpx megacrysts occasionally show signs ofdeformation and frequently channels of fluid inclusions.Some contain small drop-shaped polyphase assemblagesof Fe-Cu-Ni-sulfides, which are identical to those de-scribed in mantle xenoliths (e.g. Guo et al. 1999) and areinterpreted as exsolution from monosulfide solid solu-tion that was initially present as an immiscible sulfideliquid during growth of the megacrysts.

Solitary crystals of garnet are solid solutions of pyrope(61–73 wt%), almandine (14–23 wt%), grossular (5–10 wt%) and andradite. The crystals display the samehomogeneous nature as cpx; some of them have an iden-tical sulfide inclusion suite as cpx. Larger unfragmentedpieces of several centimeters diameter are well rounded bymagmatic resorption and show reaction rims of usually<100 lm, which extend into the crystal along cracks.

Abundant ilmenite megacrysts are solid solutions ofFeTiO3 (54–65 wt%), MgTiO3 (6–15 wt%) and Fe2O3

(20–40 wt%) and often show planar faces. Ilmenitecrystals form a bimodal distribution when plotted inMg-Fe-Ti triangular space. Ilmenite that formed asexsolution from cpx in our pyroxenite xenolith is char-acterized by high MgTiO3 contents (24–34 wt%). Thefew solitary ilmenite grains with similar high MgTiO3

contents are therefore interpreted as disaggregatedxenolithic material. Plagioclase megacrysts are clearcrystals of oligoclase to andesine with a mean of an25.They often are rounded by magmatic resorption and upto several centimeters in size. Plagioclase intergrownwith evolved cpx defines the most calcic compositions.

Figure 3 shows chondrite-normalized trace elementconcentrations of cpx. All of the megacrysts were foundto be chemically unzoned within error of major and traceelement analysis. The most primitive cpx compositions ofBiu and Jos Plateaus with mg# of 84–75 are indistin-guishable. Decreasing mg# of cpx is correlated withenrichment of the highly incompatible elements, butheavy rare earth elements (HREE) show progressivelyfractionated patterns with Tm, Yb and Lu below detec-tion limit of the ion probe. Figure 3 also includes traceelement patterns of a cpx from a spinel-lherzolite xeno-lith and a cpx from a pyroxenite xenolith. Both xenolithswere brought to the surface along with the subrecent BiuPlateau volcanics, which also host the megacrysts.

Figure 4 shows the results of Sr-Nd isotope analysisof Biu and Jos Plateau lavas and megacrysts. Biu Pla-teau lavas range in 87Sr/86Sr from 0.70286 to 0.70350and �Nd from 7.4 to 4.7. Biu Plateau cpx megacrystsoverlap host lava compositions, but span a narrowerrange from 0.70290 to 0.70328 and �Nd from 6.9 to 5.0,i.e. the most depleted and most enriched lava composi-tions are not sampled by megacryst compositions. JosPlateau cpx megacrysts overlap with those from BiuPlateau (87Sr/86Sr from 0.70316 to 0.70321 and �Nd from6.2 to 5.6) but are isotopically significantly less enrichedthan their host lavas (87Sr/86Sr from 0.70317 to 0.70359and �Nd from 5.2 to 4.6).

Discussion

Origin of megacrysts

Megacrysts—cogenetic or exotic to their host magma?

Different models for the genesis of megacrysts in alka-line basalts have been proposed and were summarized

Table 2 Volcanoes of the Jos-Plateaus from N to S

Sample location Petrography Megacrysts

Dai 9�24’20 N 9�10’63 E Alk. basalt Cpx, plagioclaseKerang 9�20’65 N 9�11’49 E Basanite, numerous xenoliths up to 25 cm, some felsic Cpx, gnt, plagioclase, ilmenite,

amphibole, olivineAmpang 9�19’07 N 9�11’99 E Trachybasalt, small xenoliths Cpx, plagioclasePidong 9�17’34 N 9�12’24 E Chain of five volcanoes. Pidong S: trachybasalt (xenolithic?).

Pidong M: basanite, maar alternating layers of granite and lapilliCpx, plagioclase

133

by e.g. Schulze (1987). Megacrysts could represent xe-nocrystic phases derived from disaggregated peridotite-or pyroxenite-xenoliths, or they could be magmaticphases from an earlier magmatic event unrelated to re-cent CVL volcanism, or they could be genetically relatedto their host lavas. In this section, we evaluate thesepossibilities for the megacrysts of the Biu and JosPlateau.

Lavas from the Biu and Jos Plateau contain abun-dant upper mantle xenoliths, raising the possibility thatassociated megacrysts represent disaggregated phases ofsuch peridotite and/or pyroxenite xenoliths. Figure 5shows a comparison of cpx megacryst compositions withthose predicted by the pMelts code (Ghiorso et al. 2002)for a mean primitive Biu Plateau basalt composition andfractional crystallization. Clearly, megacryst major ele-ment compositions and their systematic covariationswith decreasing mg# are consistent with crystallizationfrom an evolving magma. Whereas gnt megacrysts fromboth Biu and Jos Plateau are similar in major and traceelements, cpx from Jos Plateau comprise the mostprimitive compositions and have somewhat higher Na2Ofor a given mg#. Cpx derived from the lherzolite andpyroxenite xenoliths sampled on the Biu Plateau aremore calcic than megacryst cpx and have significantly

more depleted (lherzolite), or more enriched (pyroxenite)trace element patterns (Fig. 3) for a given mg# than cpxmegacrysts. Pyroxenite-derived gnt differs from mega-cryst gnt in its lower TiO2 and higher MnO content for agiven mg#.

Temperatures calculated from primitive cpx-gntmegacryst intergrowths at an assumed pressure of2 GPa, yield high temperatures of �1,430±37 �C(Krogh 1988). A change of ±0.5 GPa in the pressureestimate translates into a shift of temperature of +30�/)56 �C. The most evolved cpx-gnt megacryst inter-growth, which has considerable Fe3+/Fetot in cpx of�35%, yields a Fe3+-corrected temperature of only�1,100 �C. This is comparable to ambient lithospherictemperatures, as recorded by the spinel-lherzolite xeno-lith sampled on the Biu Plateau, which yields a tem-perature of �1,200 �C using the thermometer of Breyand Kohler (1990). The high (�1,430 �C) temperaturesrecorded in primitive cpx and gnt megacrysts and thecompositional differences between the megacrysts andxenoliths preclude formation of the megacrysts viasimple xenolith disaggregation. Instead, we propose thatthe megacrysts are magmatic phases.

Because magmatism on both the Biu and Jos Plateauoccurred in several time periods, we have to evaluate the

Table 3 Representative major and trace element analyses of cpx megacrysts

Sample Kerang Ampang Kerang Mir 48–16 Ker Ker** Dam 45 Dam+ Ampang Mir 21 Ker 2 Dam 60–4Comment Jos Jos Jos Biu Jos Biu Biu Jos Biu Jos Biu

SiO2 (wt%) 50.33 50.49 50.37 49.16 49.82 49.92 50.17 49.36 49.26 49.85 48.94TiO2 0.51 0.52 0.50 0.50 0.59 0.62 0.61 1.04 0.59 0.82 0.81Al2O3 7.31 7.40 7.32 7.56 7.78 7.99 7.98 8.49 8.83 8.36 9.03Cr2O3 0.21 0.20 0.20 0.14 0.08 0.06 0.08 0.01 0.04 0.01 0.03FeO 3.37 3.83 3.04 3.56 4.08 4.31 5.62 5.27 5.68 5.96 6.16Fe2O3

* 2.11 1.73 2.65 2.31 1.79 2.02 0.55 1.10 1.48 0.84 1.14MnO 0.12 0.12 0.13 0.14 0.14 0.15 0.15 0.13 0.16 0.14 0.16MgO 17.06 17.03 16.97 16.81 16.05 16.74 16.55 14.06 15.69 14.73 14.62CaO 16.58 16.14 16.38 16.36 16.69 16.07 15.81 17.83 15.32 16.68 16.34NiO 0.05 0.05 0.06 0.05 0.04 0.03 0.04 0.03 0.04 0.03 0.05Na2O 1.18 1.26 1.34 0.99 1.27 1.15 1.07 1.44 1.29 1.42 1.27Total 98.84 98.79 98.98 97.59 98.33 99.06 98.65 98.77 98.37 98.85 98.55Nb (ppm) 0.23 0.21 0.22 0.30 0.22 0.18 0.17 0.43 0.23 0.26 0.23La 1.43 1.07 1.08 1.17 1.10 1.18 1.24 1.95 1.05 1.51 0.87Ce 4.35 3.98 3.48 4.11 3.90 4.09 4.37 7.69 4.36 5.80 3.65Pr 0.97 0.80 0.75 0.80 0.79 0.76 0.78 1.55 0.81 1.07 0.82Sr 49.1 43.7 43.1 38.1 40.1 42.7 45.5 71.3 41.5 63.4 43.5Nd 4.45 3.77 3.96 4.34 4.03 4.22 4.80 7.91 4.58 6.42 4.74Zr 12.4 12.1 11.8 13.8 14.0 15.2 17.0 26.6 15.7 19.8 18.9Sm 1.61 1.39 1.38 1.16 1.59 1.42 1.49 2.64 1.84 2.13 1.80Eu 0.71 0.54 0.50 0.51 0.61 0.55 0.62 0.85 0.72 0.73 0.75Gd 1.49 1.74 1.53 1.81 2.02 1.62 2.54 2.72 1.53 2.07 2.15Tb 0.26 0.23 0.27 0.24 0.25 0.33 0.29 0.39 0.30 0.32 0.31Dy 1.54 1.68 1.64 1.39 1.64 1.61 1.94 2.26 1.91 1.93 2.39Ho 0.29 0.27 0.31 0.29 0.29 0.33 0.35 0.36 0.34 0.34 0.47Y 6.29 6.38 7.01 6.97 7.63 7.66 8.05 8.62 7.99 8.34 9.88Er 0.72 0.59 0.66 0.63 0.79 0.69 0.94 0.75 0.78 0.85 1.01Tm 0.11 0.11 0.13 0.09 0.09 0.11 0.12 0.06 0.12 0.07 0.16Yb 0.40 0.59 0.48 0.54 0.68 0.54 0.89 0.41 0.82 0.60 0.76Lu 0.03 0.03 0.03 0.07 0.08 0.04 0.08 0.02 0.05 0.11 0.1287Sr/86Sr n.m. n.m. n.m. n.m. 0.703161 0.703171 n.m. 0.703030 0.703207 0.703057 0.703187 0.703070143Nd/144Nd n.m. n.m. n.m. n.m. 0.512958 0.512973 n.m. 0.512938 0.512926 0.512950 0.512955 0.512990�Nd 6.2 6.5 5.9 5.6 6.1 6.2 6.9

n.m., not measured, b.d., below detection limit. *Fe2O3 from stoichiometry; **not leached

134

possibility that megacrysts could be cognate to an oldermagmatic event (e.g. the 147–106 Ma event in thenorthern Benue) unrelated to recent CVL volcanism.The megacrysts preserve magmatic textures and tem-peratures and lack compositional zoning or exsolutionlamellae. The pyroxenite xenolith that we sampled onthe Biu Plateau on the other hand is similar to thosestudied by Sautter and Harte (1988, 1990), Sen andJones (1988), Wilkinson and Stolz (1997), in that itshows (in contrast to any megacryst) exsolution lamellaeof garnet and ilmenite in matrix cpx, strong diffusionalgradients towards the lamellae, and a recrystallizedmetamorphic texture in parts of it. The temperatureestimate for the cpx-gnt exsolution interface in the BiuPlateau pyroxenite of �900 �C represents a freezingtemperature, because further cooling to �630 �C isindicated by cpx-sp thermometry. The history of thispyroxenite-xenolith is therefore interpreted as initialformation as cpx-sp-ilm accumulate from a basalticliquid and subsequent isobaric cooling from high initialtemperatures down to ambient lithospheric tempera-tures. Such cooling leads to decreasing subsolidus Al-and Ti-solubility in cpx (Velz 1999; Sepp and Kunzmann2001), and ultimately exsolution of an Al- and/or Ti-richphase such as gnt and ilm respectively. Because recrys-

tallization, diffusional gradients or exsolution are notobserved in cpx megacrysts, it is unlikely that themegacrysts were precipitated by an earlier magmaticevent and stored in cooler lithosphere before theyerupted to the surface along with the recent volcanism.

Although it is difficult to give an upper time limit forthe development of exsolution lamellae in cpx, similarHawaiian xenoliths with a maximum age of the litho-sphere of 90 Ma do show exsolution lamellae (Sen1988). More time constraints can be deduced from thelength of diffusion profiles of magmatic cpx-gnt mega-cryst intergrowths. Assuming a lithospheric temperatureof 1,200 �C (mean temperature recorded by CVL mantlexenoliths of Lee et al. 1996), the observed Fe-Mg con-centration gradient in cpx adjacent to gnt (Fig. 6) wouldhave been generated in only 5 years. Changing thelithospheric temperature by 100� translates into a timefactor of �10, i.e. at 1,000 �C the profile is generatedwithin 500 years. Thus the gnt-cpx assemblage cannothave been stored after precipitation at �1,400 �C forany significant amount of time in cooler lithosphere. Acomplete rehomogenisation of initially colder materialon the other hand seems unlikely based on the magmatictexture and lack of zonation in very slow diffusing spe-cies such as aluminum in cpx.

Dam++ Mir 60–1 Pela alt Mir+Grt Mir 15 Pid M Mir 45–7 Gumja Mir+Pl. Mir 45 Lherz. Pyrox.Biu Biu Biu Biu Biu Jos Biu Biu Biu Biu Biu Biu

49.05 48.38 48.57 48.75 48.38 48.09 47.19 47.36 44.71 46.08 50.88 48.200.74 0.73 0.95 0.75 1.03 1.46 1.33 1.68 2.16 1.73 0.04 1.229.18 9.52 9.56 9.58 10.02 8.98 10.71 9.24 12.28 10.83 5.53 7.070.02 0.02 0.01 0.02 0.01 0.01 0.01 0.01 0.01 b.d. 1.37 0.016.37 5.25 6.08 5.83 6.72 5.57 5.52 5.67 6.38 6.35 1.40 3.131.01 2.04 1.16 1.57 1.11 2.24 3.18 2.58 4.15 4.96 2.05 2.810.15 0.15 0.15 0.14 0.14 0.13 0.13 0.14 0.13 0.14 0.10 0.1114.72 14.29 13.96 13.98 13.03 12.74 11.84 13.25 8.67 8.97 17.46 13.0616.08 16.69 16.91 16.92 16.90 17.57 17.54 17.21 18.29 18.02 18.48 20.350.04 0.03 0.03 0.04 0.02 0.01 0.02 0.02 0.02 b.d. 0.06 0.011.28 1.35 1.33 1.38 1.52 1.71 1.81 1.44 2.17 2.40 0.97 1.3398.67 98.45 98.72 98.95 98.88 98.53 99.28 98.60 98.96 99.53 98.35 97.320.19 0.27 0.34 0.24 0.35 0.55 0.53 0.67 1.28 1.86 0.36 0.101.05 1.18 1.68 1.26 1.60 2.49 1.90 2.49 4.36 5.52 0.21 4.473.90 4.43 6.50 4.57 6.04 9.06 8.23 10.3 17.0 23.3 0.50 17.60.84 0.95 1.39 0.93 1.20 1.91 1.64 2.06 3.40 5.24 0.08 3.6144.6 47.8 60.8 50.1 61.1 90.7 75.3 84.5 121 133 9.87 1154.69 4.88 7.23 5.21 6.31 10.01 9.30 11.7 20.2 27.7 0.38 20.418.6 18.8 25.8 20.9 25.1 40.3 37.5 49.6 86.0 115 1.24 87.41.66 2.03 2.60 1.89 2.65 3.57 3.22 4.63 6.82 9.54 0.13 6.340.67 0.82 1.02 0.76 1.07 1.28 1.10 1.70 2.55 3.78 0.10 2.451.57 2.11 2.24 2.02 3.00 4.17 3.38 5.07 6.74 10.1 0.48 6.780.36 0.38 0.43 0.30 0.34 0.53 0.45 0.82 0.95 0.99 0.10 0.951.84 2.02 2.36 1.90 2.57 2.85 2.67 3.94 4.07 5.21 0.93 4.880.31 0.36 0.48 0.28 0.33 0.51 0.39 0.64 0.56 0.62 0.21 0.778.15 8.42 10.2 8.29 9.59 10.9 10.4 15.8 12.2 14.9 5.47 18.00.80 0.87 0.93 0.86 0.90 0.85 0.75 1.38 0.96 1.05 0.69 1.390.10 0.09 0.13 0.10 0.11 0.12 0.09 0.14 0.04 0.15 0.12 0.110.57 0.56 0.54 0.60 0.49 0.35 0.31 0.70 b.d. b.d. 0.81 0.380.06 0.09 0.01 0.07 0.15 b.d. b.d. b.d. b.d. b.d. 0.15 b.d.0.702900 0.703057 0.703276 0.703119 0.703002 0.703210 n.m. 0.703001 0.703058 n.m. n.m. n.m.0.512973 0.512950 0.512893 0.512961 0.512933 0.512925 n.m. 0.512975 0.512951 n.m. n.m. n.m.6.5 6.1 5.0 6.3 5.8 5.6 6.6 6.1

135

Table

5Representativeanalysesofplagioclase,ilmeniteandspinel

megacrysts

Sample

(plag)Mira

Mirb

Mirc

Hilia

2Gwaram

Mir+

cpx

Sample

(ilm

)Mir-1

Mir-2

Mir-3

Dam

Mir-5

Mir-6

Mir-7

Mir-8

Sample

(sp)

Hilia

Gumya

Mir

Comment

Biu

Biu

Biu

Biu

Biu

Biu

Comment

Biu

Biu

Biu

Biu

Biu

Biu

Biu

Biu

Comment

Biu

Biu

Biu

SiO

2(wt%

)59.58

59.51

60.19

59.56

61.84

57.86

Exs.

incpx

SiO

2(wt%

)0.06

0.06

0.07

TiO

20.04

0.03

0.04

0.04

0.02

0.05

SiO

2(wt%

)0.02

0.24

0.03

0.03

0.03

0.03

0.02

0.03

TiO

20.53

0.34

0.50

Al 2O

324.61

24.27

24.14

24.36

21.81

25.05

TiO

250.51

46.55

46.92

41.16

40.53

38.87

35.86

30.62

Al 2O

358.55

61.98

60.07

FeO

0.21

0.20

0.20

0.18

0.20

0.24

Al 2O

31.05

0.97

0.90

1.48

1.48

1.50

1.66

1.33

Cr 2O

30.05

1.74

0.09

MnO

0.01

0.01

0.00

0.01

0.01

0.01

Cr 2O

30.04

0.03

0.03

0.04

0.04

0.03

0.02

0.01

FeO

16.50

11.99

14.01

MgO

0.03

0.02

0.03

0.02

0.02

0.03

FeO

27.92

26.67

29.96

29.64

29.82

29.23

27.51

24.83

Fe 2O

3*

8.75

4.47

7.67

CaO

6.12

4.98

5.35

5.05

3.98

7.02

Fe 2O

3*

11.28

13.97

14.68

23.98

23.57

26.31

31.42

41.12

MnO

0.13

0.11

0.13

Na2O

7.48

8.22

7.95

8.21

8.47

7.14

MnO

0.32

0.28

0.30

0.17

0.16

0.15

0.12

0.15

MgO

16.55

19.51

18.28

K2O

0.91

1.02

0.97

1.04

1.20

0.73

MgO

9.54

8.17

6.64

4.03

3.57

3.09

2.51

1.43

NiO

0.19

0.33

0.16

Total

99.03

98.27

98.91

98.50

97.55

98.16

CaO

0.02

0.41

0.01

b.d.

b.d.

b.d.

b.d.

b.d.

Total

101.31

100.54

100.99

Or(%

)5.2

5.8

5.5

5.9

6.9

4.2

NiO

0.08

0.04

0.02

0.03

0.02

0.01

0.00

0.01

Cr#

0.1

1.8

0.1

Ab

65.3

70.6

68.8

70.3

73.9

62.1

Na2O

0.02

0.02

0.02

0.01

0.02

0.02

0.04

0.01

An

29.5

23.6

25.6

23.9

19.2

33.7

Total

100.79

97.34

99.52

100.56

99.26

99.23

99.17

99.54

87Sr/

86Sr

0.703096

0.703044

0.703062

0.703036

0.702961

n.m

.Fe 2O

3(%

)0.10

0.13

0.14

0.23

0.23

0.25

0.31

0.40

143Nd/1

44Nd

0.512981

0.512940

0.512941

0.512955

0.512947

n.m

.FeT

iO3

0.56

0.56

0.62

0.62

0.64

0.63

0.59

0.54

� Nd

6.7

5.9

5.9

6.2

6.0

MgTiO

30.34

0.30

0.24

0.15

0.14

0.12

0.10

0.06

n.m

.,notmeasured.b.d.,below

detectionlimit.*Fe 2O

3from

stoichiometry

Table

4Representativeanalysesofgntmegacrysts

Sample

Mir+

grt

DamknolleKer

GA

Schliff1

Schliff2Schliff5Mir1

Mir2

Ker

GB

Mir13Mir21Mir23Mir37eMir37fMir37hMir37gMir37iDam

45–14

Pyroxenite

exsolution

Comment

Biu

Biu

Jos

Jos

Jos

Jos

Biu

Biu

Jos

Biu

Biu

Biu

Biu

Biu

Biu

Biu

Biu

Biu

Biu

SiO

2(w

t%)

40.96

39.73

41.30

41.18

41.12

40.74

41.72

41.23

41.65

40.73

41.38

41.19

41.37

41.27

40.89

40.93

40.24

40.34

40.25

TiO

20.33

0.36

0.41

0.44

0.48

0.44

0.40

0.40

0.38

0.47

0.36

0.42

0.37

0.30

0.38

0.37

0.51

0.40

0.14

Al 2O

323.40

22.56

23.57

23.40

23.27

22.91

23.65

23.39

23.64

23.30

23.76

23.63

23.55

23.55

23.51

23.43

22.85

22.78

22.96

Cr 2O

30.02

0.01

0.00

0.00

0.01

0.03

0.01

0.00

0.10

0.00

0.05

0.01

0.10

0.05

0.01

0.01

0.01

b.d

0.01

FeO

10.94

18.65

10.81

12.63

13.09

10.78

11.06

12.76

10.04

12.88

9.63

10.05

7.82

8.75

9.68

10.22

14.30

12.82

11.87

Fe 2O

3*

1.32

0.77

0.59

0.21

0.00

1.26

0.00

0.10

0.13

0.87

0.66

0.64

1.60

1.74

1.74

1.65

1.51

1.81

1.25

MnO

0.36

0.55

0.34

0.37

0.39

0.35

0.35

0.37

0.32

0.38

0.33

0.34

0.31

0.32

0.33

0.35

0.41

0.39

0.54

MgO

17.11

10.73

17.57

16.34

15.98

17.43

17.54

16.28

18.57

15.74

18.57

18.12

19.58

18.93

18.06

17.75

14.46

15.21

16.04

CaO

5.76

7.31

5.59

5.76

5.72

5.27

5.52

5.76

5.15

6.02

5.20

5.38

5.21

5.21

5.43

5.47

6.19

6.36

5.59

NiO

0.02

0.00

0.00

0.01

0.00

0.01

0.00

0.01

0.01

0.00

0.01

0.01

0.01

0.01

0.01

0.01

0.00

0.01

0.01

Na2O

0.03

0.06

0.03

0.03

0.04

0.04

0.03

0.03

0.03

0.03

0.02

0.03

0.03

0.03

0.02

0.02

0.05

0.04

0.03

Total

100.25

100.79

100.22

100.39

100.11

99.28

100.30100.34100.05

100.45

99.98

99.84

99.97

100.17

100.08

100.23

100.55

100.1898.69

143Nd/144Nd0.5129530.512960

0.5129480.512982

� Nd

6.1

6.3

6.0

6.7

b.d.,below

detectionlimit.*Fe 2O

3from

stoichiometry

136

In summary, we find that the magmatic textures,temperatures and lack of zoning or exsolution lamellaeindicate that the cpx and gnt megacrysts do not derivefrom an older period of volcanism unrelated to the recentCVL. Therefore, the megacrysts must be derived frommagmas related to the recent (�5 Ma to present) mag-matism on Biu and Jos Plateau. The sequence of mega-cryst crystallization is deduced on the basis ofmg#and thesystematics of mutual intergrowths as cpx fi cpx±gntfi cpx±gnt, amph fi cpx±gnt, amph, plag, apa, ilm.

Where did the megacrysts grow – pressure estimates

Unlike temperature, estimates of pressures of megacrystcrystallization are more difficult to constrain because of

Fig. 2 Collection of Sr-Nd isotopic compositions of the moreprimitive rocks (MgO >5 wt%) of the CVL. Data from Hallidayet al. (1988, 1990), Lee et al. (1994), Marzoli et al. (1999, 2000).Hatched fields: Biu and Jos Plateau lavas (Rankenburg et al., inreview). The continent-ocean boundary (c.o.b.), represented bylavas from Mt. Cameroon, Etinde and Bioko, ranges to moreundersaturated magmas such as nephelinites (Halliday et al. 1990)and is characterized by relatively unradiogenic Nd for a given Srisotopic composition

Fig. 3 Chondrite normalized trace element pattern of cpx mega-crysts measured by SIMS (solid lines). Evolved compositions havefractionated HREE with Tm, Yb and Lu below DL. There is nodetectable difference between the most primitive Biu and JosPlateau megacrysts which points to common source magma.Lherzolite-derived cpx is markedly different from megacryst cpxin its depleted pattern, pyroxenite-derived cpx shows higher traceelement concentrations for a given mg# and relative depletions inNb and Ti

Fig. 4 a) Sr-Nd isotope analyses of megacrysts. Biu Plateau cpxcompositions (dark stars) span almost the range of Biu Plateau lavacompositions (dark field). Arrows point towards possible contam-inants (DMM: depleted MORB mantle, CC: continental crust, EM:enriched mantle). Additional data for Biu plag megacrysts(triangles). Jos cpx (light gray stars) overlap with Biu cpx, but areisotopically significantly less enriched than Jos host lavas (whitefield). 2r-error bars from TIMS analyses as indicated in lower leftcorner

Fig. 5 Major element variations in cpx megacrysts (symbols) versuscpx compositions predicted by the pMelts code (small dots) at2.0 GPa assuming fractional crystallization (all Fe as FeO). Filledcircles: Biu Plateau, open circles: Jos Plateau, filled trianglesconnected by tie lines: intergrowths of megacrysts. Filled star markscpx from pyroxenite xenolith, open star cpx from sp-lherzolitexenolith. Although megacrysts as a whole define smooth trendsconsistent with fractional crystallization from an evolving magma,individual grains are chemically unzoned. The trends predicted formg# 85–70 are in good agreement with the variations observed inthe megacrysts as a whole. The kink in CaO of cpx megacrystsarises from onset of additional gnt fractionation (and not becauseof coexistence of opx), consistent with our data. However, thecalculation assumes constant pressure, which may not be the casein nature. This may also be the reason for the poor fit of the highlyevolved cpx megacrysts intergrown with plagioclase and apatite

137

the lack of a suitable barometer. However, results fromhigh-pressure experiments provide a matching sequenceof crystallization with consistent major element com-positions during fractionation of a basanitic/alkali-basaltic composition at upper mantle pressures. Therevised pMelts code of Ghiorso et al. (2002) can beconsidered a parameterization of such processes that isespecially well calibrated in the pressure range between 1and 3 GPa. Major element compositions of primitive,near-liquidus cpx and gnt calculated by the pMelts codeare consistent with those measured in the megacrysts.Figure 7 shows a simplified calculated phase diagram forone of the most primitive lavas erupted on the BiuPlateau, which is similar to the experimental olivinenephelinite data of Bultitude and Green (1971). Olivinepresent as small phenocrysts in the host lavas crystal-lized most probably during and after eruption to thesurface because of expansion of the olivine stability fieldat shallower pressure. This is consistent with calculatedolivine/whole-rock temperatures of 1,220 �C ()40/+20)(Putirka 1997). Equilibrium between olivine and whole-rock was assumed if calculated equilibrium olivinecompositions (with KD

ol/liq(Mg-Fe)=0.30±0.03, Roe-der and Emslie 1970) matched measured olivine phe-nocryst cores.

With increasing pressure phase relations indicate thatcpx replaces olivine as the first liquidus phase. Based onthe high Ni and Cr contents of most primitive cpxcompositions (Fig. 8a, b) and absence of olivine mega-crysts we conclude that crystal fractionation started inthe primary stability field of cpx, which requires pres-sures greater than �1.2 GPa. A fractionation path atpressures corresponding to that at the base of the crust(LP in Fig. 8) implies initial fractionation of olivine and

consequently depletion of Ni in the remaining liquid.The observed high concentration of Ni in most primitivecpx megacrysts is not consistent with a low-pressurefractionation history.

At higher pressures of �1.7-2.0 GPa cpx is joined bygarnet upon isobaric cooling. The observed depletion inheavy rare earth elements in evolved cpx (Fig. 8c) isconsistent with coprecipitation of garnet. If we combinethe phase diagram with the locus of possible P, T-pairsof one cpx-gnt intergrowth (cf. Fig. 7), we can furtherspecify the formation pressure of that intergrowth to begreater than �2.1 GPa. If we assume that most primitivecpx and basanitic melts found at the surface were inequilibrium at some point in their history, we can alsouse the cpx-liquid geothermometer of Putirka et al.(1996) and iteratively calculate a consistent P, T-estimateof 2.2 GPa and 1,380 �C from the most primitive cpxcompositions from Biu Plateau and a slightly higherestimate of 2.3 GPa and 1,390 �C from the most prim-itive Jos Plateau cpx megacrysts. This result is ratherinsensitive to the exact melt composition as long as it isprimitive basanite or alkali basalt.

Although each of the above pressure estimates isdependant on certain model assumptions, the results areencouragingly mutually consistent. In each case, thepressure estimate for the primitive gnt and cpx mega-crysts is �1.7–2.3 GPa. Because the crust-mantleboundary is constrained from seismic data to depths ofonly 28–30(±6) km or 0.8–0.9(±0.2) GPa (Poudjom-

Fig. 6 Diffusion profile perpendicular to a cpx-gnt interface. 2r-Error bars (�0.8% rel.) derive from counting statistics of EPMA.The fitted error function is only dependant on time and the binaryMg-Fe interdiffusion coefficient D, which is itself a function oftemperature. The observed gradient can be generated by keepingthe assemblage at 1,200 �C in only 5 years, whereas a change of100� of the ambient temperature translates into a time factor of�10. The plateau concentration is homogeneous throughout thecrystal (�2 mm) and therefore was used to calculate the initialequilibration temperature of cpx-gnt. Since all relevant species havelower diffusivities in cpx than in gnt (Dimanov and Sautter 2000), itis sufficient to look at cpx profiles for it is the rate limiting exchangepartner

Fig. 7 Calculated phase diagram for a primitive Biu Plateau alkalibasalt using the pMelts code (Ghiorso et al. 2002). At pressuresabove �1.2 GPa cpx replaces ol as the first liquidus phase. Athigher pressures cpx is joined by garnet upon isobaric cooling.Dotted line marks locus of possible P, T-pairs for one cpx-gntintergrowth. A comparison of calculated and observed majorelement compositions of near liquidus cpx is given in Fig. 5. Phasesin parenthesis (left side) are calculated near-solidus fractionationassemblages at 1.0, 1.5 and 2.0 GPa respectively. The crust-mantleboundary (Biu: 30 km, Jos: 28 km) is based upon gravity data(Poudjom-Djomani et al. 1995). Note that for cpx and gnt to benear liquidus phases they have to fractionate in the mantle. Anadiabatic ascending path (arrow) will not saturate the magma incpx or gnt. Addition of 0.5 wt% water will lower the solidus by�40�K (Green 1973), whereas addition of CO2 will expand theliquidus phase volume of gnt against that of diopside and istherefore similar to increasing pressure (Adam 1988)

138

Djomani et al. 1995), this requires that the megacrystsgrew within the mantle, well below the crust/mantleboundary.

Even though the liquid chosen in Fig. 7 is not inequilibrium with mantle peridotite phases such as ol and

opx at high pressure, it can be derived from a primaryliquid such as the picritic basanite of Bultitude andGreen (1971) by initial fractionation of �17% olivine(Table 6). The origin of some rare large crystals ofolivine that were collected on the Biu Plateau, with highforsterite contents (fo91.6), strong internal deformationtextures and abundant melt and fluid inclusions, is as yetunclear, but these olivines may derive from very earlyfractions of CVL magmas more primitive than thoseerupted at the surface.

Because the majority of experimental data is obtainedeither from near liquidus conditions in basaltic systemsor near solidus conditions in peridotitic systems thereare few data published on evolved liquids which mightshed some light on the genesis of ilmenite, albite-richplagioclase, zircon and corundum megacrysts. Until aseries of experiments simulating fractional crystalliza-tion is available, the calculated near solidus fraction-ation assemblages in Fig. 7 are relatively uncertain.Although plagioclase is typically considered a low-pressure phase, there is one experiment by Rushmer(1993) on natural amphibolite which produced gar-net+cpx+ab-rich plagioclase coexisting with anandesitic melt at 1.8 GPa. The compositions of experi-mentally produced crystals (cpx, gnt, plag) in thatexperiment match those of the megacrysts closely, thuspermitting plagioclase precipitation within the mantle.One feature of such high-pressure plagioclase is lowanorthite content, consistent with the observed compo-sition of plagioclase megacrysts.

Further constraints on megacrysts growth

A liquid adiabatic ascending path with a typical dT/dzof �1�K km)1 (McKenzie and Bickle 1988) will notsaturate the magma in any phase (arrow in Fig. 7). Cpxand garnet cannot be derived by flow differentiation of arising magma unless significant heat is dissipated intothe wall rock. In this case we would expect a strongtemperature gradient and thus zoned crystals. However,the observed unzoned nature of the megacrysts and theirlarge size suggests derivation from a slow evolving orquasi-steady state magma chamber (O’Hara and Math-ews 1981). For equilibrium or steady state RTF (peri-

Table 6 Comparison of a primary composition after fractionationof 17% olivine with a mean composition of primitive Biu Plateaubasalts. Although the composition of Bultitude and Green (1971)was significantly poorer in Ti, the compositions agree fairly well

Pic. Bas. –17% ol Mean Biu prim.

mg# 63.8 64.1SiO2 45.2 45.6TiO2 1.5 3.1Al2O3 14.2 13.3FeO 10.5 11.2MgO 10.8 10.9CaO 11.4 9.9Na2O 3.3 3.3

Fig. 8a–c MgO vs. Cr (a), Ni (b) and Yb (c) of cpx megacrysts.Symbols as in Fig. 5. The equilibrium and fractional crystallizationmodels are calculated based on the phase proportions predicted bythe pMelts/Melts code for the same liquid as Fig. 7 at 2.0 GPa(HP) and 0.8 GPa (LP) respectively. Every point in the model curverepresents 5� cooling while numbers refer to the percentage of solidscrystallized. a An equilibrium type crystallization will not yield thelow concentrations of compatible elements (Cr and Mg) in theliquid and corresponding cpx. We also plotted the concentrationsin a steady state (RTF) magma chamber (O’Hara and Mathews1981), which for reasonable mass fractions of liquid crystallizing(x) and escaping (y) in each cycle also predicts too high Cr-concentrations. (D’s taken from Mysen 1978; Johnson 1998;McKenzie and O’Nions 1991; Nikogosian and Sobolev 1997). bLP fractionation starts out with olivine (�18 wt%) and quicklydepletes the liquids in Ni, which is inconsistent with high Niconcentrations in primitive cpx megacrysts. c The sharp bend in theHP model curve of Yb marks the onset of additional gnt-fractionation and therefore confirms the involvement of gnt inthe fractionating assemblage (filled triangles). A low pressurefractionation without the involvement of gnt will not depleteliquids and cpx in HREE and therefore is inconsistent with the data

139

odically replenished, periodically tapped, continuouslyfractionated) type crystallization we expect buffering ofcompatible elements such as Mg, Cr and Ni at relativelyhigh levels (e.g. Cr-concentrations >200 ppm in evolvedcpx, cf. Fig. 8a). However, what we observe is strongdepletion in these elements in paths that are consistentwith fractional crystallization.

Although pMelts modeling suggests that the observedspectrum of cpx compositions can be derived by up to65% fractional crystallization from a basaltic liquid, themegacrysts have to grow homogeneously directly fromthe melt because of extremely long diffusional equili-bration times. The diffusional relaxation of an initiallygrowth-zoned cpx of 5 cm in diameter for example re-quires for the slow diffusing species aluminum �220 Maat 1,350 �C (after Chakraborty and Ganguly’s 1991formulation t�a2/2D, and D(T) calculated with datafrom Jaoul et al. 1991). Although the effects of pressure,oxygen fugacity and H2O-activity on Al-diffusion is stilluncertain, the time scale is definitively longer than 5–10 Ma of recent volcanism on the Biu and Jos Plateaus.The evidence for fractional crystallization combinedwith the unzoned nature of the megacrysts thereforesuggests that megacrysts derived from relatively large(and therefore slow cooling) magma chambers thatformed within the lithospheric mantle during CVL vol-canism.

Isotopic constraints on crustal contaminationof Biu and Jos Plateau lavas

A full and more global discussion of the informationdrawn from comparison of the isotopic composition oflavas and megacrysts will be given in a separate paper(submitted to J. Pet). In brief, Sr-Nd isotope analyses(Fig. 4) indicate that the array of Biu and Jos Plateaulava compositions is consistent with mixing of a depletedasthenospheric component and a more enriched com-ponent such as continental crust, enriched lithosphericmantle or plume material. Because we have shown thatprimitive cpx and gnt megacrysts grew at mantle pres-sures, the isotopic composition of these megacrystsreflect the melt compositions before interaction withcrustal material could take place. Biu Plateau lavas andmegacrysts span a similar range in isotopic composition(Fig. 4), although most megacrysts are clustered at themore depleted end of the isotopic range defined by Biulavas. From this, we conclude that at least the isotopi-cally more depleted Biu Plateau melts did not signifi-cantly interact with enriched material such ascontinental crust or shallow enriched lithospheric mantleafter formation of the megacrysts. The observed isotopicheterogeneity in theses lavas therefore must be relatedeither to melt interaction with enriched lithosphericmantle before or during growth of the megacrysts, or toheterogeneity within the asthenospheric source region ofthe magmas. Because of the lack of correlation betweenindicators of fractionation such as mg# and isotopic

composition of the megacrysts we favor the secondpossibility.

In contrast, all Jos Plateau lavas have significantlymore enriched Sr and Nd isotope compositions thanassociated megacrysts. Because similar major elementcompositions, trace element patterns (Fig. 3) andoverlapping isotopic composition of cpx megacrystsfrom Biu and Jos Plateau (Fig. 4) strongly point tosimilar source magmas for both areas, the offset tomore enriched compositions of Jos Plateau lavas indi-cate an additional isotopic enrichment after formationof megacrysts. However, the gap in Sr-isotope com-positions between mean megacrysts (�0.7032) and la-vas (�0.7035) is rather small and can be accounted forby a relatively small addition of crustal material. Forexample, if we assume an assimilant similar to themean of literature data for basement rocks (23 granu-lites, gneisses, migmatites and granites analyzed byHalliday et al. (1988), Dickin et al. (1991) and Dadaet al. (1995) with [Sr]=351 ppm, [Nd]=66.7 ppm,87Sr/86Sr=0.7256 and �Nd=)16.2), addition of�5 wt% bulk crust is sufficient to explain the isotopicoffset between most depleted Biu and mean Jos Plateaulavas. This estimate was calculated with the conserva-tive assumptions of bulk rock assimilation and thatprimary magmas already had 800 ppm Sr, and there-fore represents an upper limit of contamination. Suchsmall degrees of crustal contamination will have neg-ligible effects on the major and trace element budgetsof the lavas. Given that the lavas of Biu and Jos Pla-teau on their part span almost the range of continentalCVL magmas with MgO >5 wt% as a whole (Fig. 4),we propose that crustal contamination of other primi-tive CVL magmas was relatively minor.

Contrasting crystallization/magma mixing historiesdeduced from megacrysts and lavas

In this section, we evaluate trace element budgets anddistributions between melts and fractionated phases inorder to constrain the evolution of the host magmas. Asalready shown in Fig. 8a, the megacrysts clearly followhigh-pressure fractionation trends. However, the hostlavas in Fig. 9 plot closer to straight binary mixing linesrather than to the constantly evolving liquids modeledby high- or low-pressure fractionation. We thereforeconclude that the major and trace element systematics ofthe lavas is mainly controlled by melt mixing betweenportions of primitive and evolved melts. A possibleevolved mixing end member might be the melt similar tothat found as inclusions in evolved ilmenite megacrysts(cf. open star in Fig. 9). Because most evolved cpxcompositions are intergrown with ilmenite of similarcomposition as the one containing the melt inclusion, weassume that the melt inclusion and evolved cpx are inequilibrium. For an estimation of Peq and Teq, we cal-culate 1.36 GPa and 1,160 �C using the model of Put-irka et al. (1996). This estimate is consistent with the

140

temperature calculated from the most evolved cpx/gntintergrowth which yields �1,100 �C.

Given these data, we then calculate a set of Dcpx/melt

(REE) for the most evolved cpx megacryst according toWood and Blundy’s (1997) model c, which calculatesDcpx/melt from cpx crystal major element composition,pressure and temperature. The resulting set of distribu-tion coefficients is associated with errors of )40/+67%(Wood and Blundy 1997). For the final inversion to meltconcentrations, however, we also have to consider theadditional errors introduced by the uncertainty of P andT, and errors from SIMS measurements of cpx mega-crysts.

For the primitive cpx composition, we minimized theeffect of cpx major element composition on trace ele-ment distribution by using the D-set of Johnson (1998),where the experimental P, T-conditions (2–3 GPa,1,310�–1,470 �C) match the forming conditions of themegacrysts and more important, synthesized cpx majorelement compositions match megacryst compositions.

Figure 10 shows chondrite-normalized (McDonoughand Sun 1995) rare earth element patterns of Biu and JosPlateau lavas along with two calculated trace elementpatterns of melts in equilibrium with most evolved andmost primitive cpx. The trace element pattern of melts inequilibrium with the most evolved cpx closely resemblesthe trace element pattern of the evolved, ilmenite-hostedmelt inclusions except that La and Ce in the calculatedpattern is lower. This similarity is consistent with ourassumption that the evolved cpx and ilmenite-trappedmelts are in equilibrium.

Although trace element concentrations of melts in-verted from primitive cpx compositions and meltsactually found at the surface overlap in Fig. 10, there isa slight discrepancy between the average of most prim-itive calculated and outcropping melts. However, if wetake into account the analytical errors, and the errorintroduced by the using a specific D-set, the differencebetween calculated primitive melts and melts found atthe surface might actually be non-existent. However, theD-set derived from Johnson’s experiment defines thelowest published values, and thus minimizes the dis-crepancy between inverted melts from megacrysts andmeasured melt compositions.

The influence of water/CO2 activity on mineral/meltpartitioning might in principle further decrease the dis-tribution coefficients in the cpx/melt system and therebybring our calculated melt compositions in line with meltcompositions observed at the surface. To our knowledgeno systematic investigations have examined mineral/meltpartitioning in the range where melts are not saturatedin water/CO2. However, such investigations are neededto differentiate between effects on the partitioningintroduced by changes in mineral chemistry and changesin water/CO2-activity. Assuming that the most signifi-cant factor on trace element partitioning is mineralmajor element chemistry, the D-set of Johnson still isour best choice because of the best compositionalagreement.

Fig. 9a,b Variation of Cr (a) and Yb (b) vs. MgO of basalts fromBiu and Jos Plateau. HP (2.0 GPa) fractionation model involvingcpx+gnt from pMelts, and LP (0.8 Gpa) fractionation modelinvolving ol+cpx+sp from Melts. Symbols for lavas and D’s asbefore, gray star marks phonolitic melt, white star marks meltinclusion in ilmenite megacryst. Liquids should quickly be depletedin elements compatible in cpx such as Cr. However, the liquid linesof descent predicted from the fractionation models are notobserved. Instead liquids more closely follow straight mixingtrajectories (arrows) from primitive melts to hypothetic evolvedmelts derived by fractional crystallization. The depletion in Yb ofevolved melts is consistent with a high-pressure fractionation pathinvolving gnt in the fractionating assemblage. The melt inclusion inilmenite is a possible mixing end member, whereas the phonoliteseems to have a separate fractionation history at low pressure

Fig. 10 Inversion of cpx trace element concentrations to meltconcentrations. Distribution coefficients for primitive cpx mega-crysts (stippled line) are taken from Johnson (1998), for evolved cpx(dotted line) the distribution coefficients were calculated accordingto Wood and Blundy (1997). Light gray field comprises lavas of Biuand Jos Plateau without the phonolite, solid line gives compositionof melt inclusions

141

To account for the evidence for magma mixingshown in Fig. 9 we propose a model in which magmasthat carry megacrysts are mixtures of a primitivemagma similar to that inverted from most primitive cpxmegacrysts with portions of an evolved magma, similarto that inverted from most evolved cpx megacrysts.Mixing of evolved and primitive magmas and entrain-ment of megacrysts may have occurred simultaneouslyas fresh batches of a deeply-sourced primitive magmaentered mantle magma chambers in the feeder system,triggering ascent and eruption. Because all megacrystsare precipitating from a uniform magma they will all lieon the same major element fractionation trends, con-sistent with the trends observed in Figs. 5 and 8.Mixing of primitive magmas with small quantities ofhighly evolved, trace element enriched magmas mayalso account for the elevated trace element concentra-tions observed in Biu and Jos lavas relative to calcu-lated melts in equilibrium with the primitive cpxmegacrysts.

Conclusions

We have shown based on chemical and diffusionalconstraints that at least cpx and gnt megacrysts aregenetically related to recent CVL lavas that broughtthem to the surface, though they may not be directlycognate to their host lava. Based upon this relationship,Sr-Nd isotope systematics suggests that crustal con-tamination was minor (<5%) in Biu and Jos Plateaulavas. The greater contamination of Jos lavas comparedto Biu Plateau lavas may also reflect increased assim-ilation of enriched shallow lithospheric mantle whichmight be thicker underneath Jos Plateau relative to theBiu Plateau due its greater distance to the Benue paleospreading center.

Lavas do not reflect simple fractionation or equilib-rium crystallization products, but instead reflect mixingof primitive and evolved batches of magma shortly be-fore eruption. An especially interesting result of suchmixing is that magmas with near ‘primary’ features (i.e.high mg#, Ni, Cr, containment of peridotitic xenoliths)can be significantly enriched in trace elements withoutsignificantly changing their major element compositions.Such mixing complicates efforts to infer the degree ofpartial melting or source enrichment by inversion of lavatrace element compositions.

Acknowledgements We are grateful to A. Hofmann and D.H.Green for their constructive criticism on this paper. D. Canil andJ. Blundy are thanked for their constructive comments on an earlierversion of this manuscript. We also thank J. Heliosch at theDepartment for Petrology and Geochemistry at the University ofFrankfurt for technical assistance. I. Raczek, E. Makiolczyk and E.Hellebrand are kindly acknowledged for their technical assistanceat the MPI in Mainz. Sincere thanks are given to G. Fitton forsending me some unpublished rock analyses of the Biu Plateau.This work was financially supported by Deutsche Forschungs-gemeinschaft project no. BR 1012/11–1.

References

Adam J (1988) Dry, hydrous, and CO2-bearing liquidus phase-relationships in the CMAS system at 28 kb, and their bearingon the origin of alkali basalts. J Geol 96:709–719

Barfod DN, Ballentine CJ, Halliday AN, Fitton JG (1999) Noblegases in the Cameroon line and the He, Ne, and Ar isotopiccompositions of high l (HIMU) mantle. J Geophys Res-Sol Ea104:29509–29527

Brey GP, Kohler T (1990) Geothermobarometry in 4-PhaseLherzolites II. New thermobarometers, and practical assess-ment of existing thermobarometers. J Petrol 31:1353–1378

Bultitude RJ, Green DH (1971) Experimental study of crystal-li-quid relationships at high pressures in olivine nephelinite andbasanite compositions. J Petrol 12:121–147

Chakraborty S, Ganguly J (1991) Compositional zoning and cationdiffusion in garnets. In: Ganguly J (ed) Advances in physicalgeochemistry, diffusion, atomic ordering, and mass transport:selected problems in geochemistry. Springer, Berlin HeidelbergNew York, pp 120–175

Dada SS, Briqueu L, Harms U, Lancelot JR, Matheis G (1995)Charnockitic and monzonitic pan-African series from North-Central Nigeria: trace-element and Nd, Sr, Pb isotope con-straints on their petrogenesis. Chem Geol 124:233–252

Dickin AP, Halliday AN, Bowden P (1991) A Pb, Sr and Nd iso-tope study of the basement and Mesozoic ring complexes of theJos Plateau, Nigeria. Chem Geol 94:23–32

Dimanov A, Sautter V (2000) ‘‘Average’’ interdiffusion of (Fe,Mn)-Mg in natural diopside. Eur J Mineral 12:749–760

Fitton JG (1980) The Benue Trough and Cameroon Line—amigrating rift system in West-Africa. Earth Planet Sc Lett51:132–138

Fitton JG (1987) The Cameroon Line, West-Africa: a comparisonbetween oceanic and continental alkaline volcanism. In: FittonJG, Upton BGJ (eds) Alkaline igneous rocks. Geol Soc SpecialPublication, pp 273–291

Fitton JG, Dunlop HM (1985) The Cameroon Line, West-Africa,and its bearing on the origin of oceanic and continental alkalibasalt. Earth Planet Sc Lett 72:23–38

Frisch T, Wright JB (1971) Chemical composition of high-pressuremegacrysts from Nigerian Cenozoic lavas. Neues Jahrbuch furMineralogie-Monatshefte 7:289–304

Ghiorso MS, Hirschmann MM, Reiners PW, Kress VC (2002) ThepMELTS: A revision of MELTS for improved calculation ofphase relations and major element partitioning related to partialmelting of the mantle to 3 GPa. GeochemGeophy Geosy 3:1030

Grant NK (1970) Geochronology of Precambrian basement rocksfrom Ibadan, Southwestern Nigeria. Earth Planet Sc Lett10:29–38

Grant NK, Rex DC, Freeth SJ (1972) Potassium-argon ages andstrontium isotope ratio measurements from volcanic-rocks inNortheastern Nigeria. Contrib Mineral Petr 35:277–292

Green DH (1973) Conditions of melting of basanite magma fromgarnet peridotite. Earth Planet Sc Lett 17:456–465

Guo JF, Griffin WL, O’Reilly SY (1999) Geochemistry and originof sulphide minerals in mantle xenoliths: Qilin, SoutheasternChina. J Petrol 40:1125-1149

Halliday AN, Dickin AP, Fallick AE, Fitton JG (1988) Mantledynamics—a Nd, Sr, Pb and O isotopic study of the CameroonLine volcanic chain. J Petrol 29:181–211

Halliday AN,Davidson JP, Holden P, Dewolf C, Lee DC, Fitton JG(1990) Trace-element fractionation in plumes and the origin ofHIMUmantle beneath the Cameroon Line. Nature 347:523–528

Hellebrand E, Snow JE, Hoppe P, Hofmann AW (2002) Garnet-field melting and late-stage refertilization in ‘residual’ abyssalperidotites from the Central Indian Ridge. J Petrol 43:2305–2338

Irving AJ, Price RC (1981) Geochemistry and evolution of lherz-olite-bearing phonolitic lavas from Nigeria, Australia, East-Germany and New-Zealand. Geochim Cosmochim Ac 45:1309–1320

142

Jaoul O, Sautter V, Abel F (1991) Nuclear microanalysis: A pow-erful tool for measuring low atomic diffusivity with mineralogicapplications. In Ganguly J (ed) Advances in physical geo-chemistry, diffusion, atomic ordering, and mass transport:selected problems in geochemistry. Springer, Berlin HeidelbergNew York, pp 198–220

Jochum KP, Dingwell DB, Rocholl A, Stoll B, Hofmann AW,Becker S, Besmehn A, Bessette D, Dietze HJ, Dulski P, Erz-inger J, Hellebrand E, Hoppe P, Horn I, Janssens K, JennerGA, Klein M, McDonough WF, Maetz M, Mezger K, MunkerC, Nikogosian IK, Pickhardt C, Raczek I, Rhede D, SeufertHM, Simakin SG, Sobolev AV, Spettel B, Straub S, Vincze L,Wallianos A, Weckwerth G, Weyer S, Wolf D, Zimmer M(2000) The preparation and preliminary characterisation ofeight geological MPI-DING reference glasses for in-sitemicroanalysis. Geostandards Newsletter-The Journal of Geo-standards and Geoanalysis 24:87–133

Johnson KTM (1998) Experimental determination of partitioncoefficients for rare Earth and high-field-strength elements be-tween clinopyroxene, garnet, and basaltic melt at high pres-sures. Contrib Mineral Petr 133:60–68

Kraml M, Kamdem JB, Keller J, Henjes-Kunst F (2001) Timeconstraints for Cameroon Line magmatism from single crystallaser 40Ar/39Ar dating of the Hossere Ngo anorogenic plutoniccomplex. EUG XI, Strasbourg, Journal of Conference Ab-stracts 4:596

Krogh EJ (1988) The garnet-clinopyroxene Fe-Mg geothermome-ter: a reinterpretation of existing experimental data. ContribMineral Petr 99:44–48

Lee DC, Halliday AN, Fitton JG, Poli G (1994) Isotopic variationswith distance and time in the volcanic islands of the CameroonLine—evidence for a mantle plume origin. Earth Planet Sc Lett123:119–138

Lee DC, Halliday AN, Davies GR, Essene EJ, Fitton JG, TemdjimR (1996) Melt enrichment of shallow depleted mantle: A de-tailed petrological, trace element and isotopic study of mantle-derived xenoliths and megacrysts from the Cameroon line.J Petrol 37:415–441

Maluski H, Coulon C, Popoff M, Baudin P (1995) Ar-40/Ar-39chronology, petrology and geodynamic setting of Mesozoic toEarly Cenozoic magmatism from the Benue Trough, Nigeria.J Geol Soc 152:311–326

Marzoli A, Renne PR, Piccirillo EM, Francesca C, Bellieni G,Melfi AJ, Nyobe JB, N’ni J (1999) Silicic magmas from thecontinental Cameroon Volcanic Line (Oku, Bambouto andNgaoundere): 40Ar/39Ar dates, petrology, Sr-Nd-O isotopesand their petrogenetic significance. Contrib Mineral Petr135:133–150

Marzoli A, Piccirillo EM, Renne PR, Bellieni G, Iacumin M,Nyobe JB, Tongwa AT (2000) The Cameroon Volcanic Linerevisited: Petrogenesis of continental basaltic magmas fromlithospheric and asthenospheric mantle sources. J Petrol 41:87–109

McDonough WF, Sun SS (1995) The Composition of the Earth.Chem Geol 120:223–253

McKenzie D, Bickle MJ (1988) The volume and composition ofmelt generated by extension of the lithosphere. J Petrol 29:625–679

McKenzie D, O’Nions RK (1991) Partial melt distributions frominversion of rare-Earth element concentrations. J Petrol32:1021–1091

McLeod WN, Turner DC, Wright EP (1971) The geology of theJos Plateau. Bulletin of the Geological Survey of Nigeria 32:17–23

Mysen BO (1978) Experimental determination of nickel partition-coefficients between liquid, pargasite, and garnet peridotiteminerals and concentration limits of behavior according toHenrys Law at high-pressure and temperature. Am J Sci278:217–243

Ngounouno I, Deruelle B, Demaiffe D, Montigny R (1997) Newdata on the Cenozoic volcanism of the Garoua valley (UpperBenue trough, Northern Cameroon). Comptes Rendus De LAcademie Des Sciences Serie Ii Fascicule a- Sciences De LaTerre Et Des Planetes 325:87–94

Nikogosian IK, Sobolev AV (1997) Ion-microprobe analysis ofmelt inclusions in olivine: Experience in estimating the melt-olivine partitioning coefficients of elements. Geokhimiya:149–157

O’Hara MJ, Mathews RE (1981) Geochemical evolution in anadvancing, periodically replenished, periodically tapped, con-tinuously fractionated magma chamber. J Geol Soc 138:237–277

Poudjom-Djomani YH, Nnange JM, Diament M, Ebinger CJ,Fairhead JD (1995) Effective elastic thickness and crustalthickness variations in West Central-Africa inferred fromgravity data. J Geophys Res-Sol Ea 100:22047–22070

Putirka K (1997) Magma transport at Hawaii: Inferences based onigneous thermobarometry. Geology 25:9–72

Putirka K, Johnson M, Kinzler R, Longhi J, Walker D (1996)Thermobarometry of mafic igneous rocks based on clinopy-roxene-liquid equilibria, 0–30 kbar. Contrib Mineral Petr123:92–108

Roeder PL, Emslie RF (1970) Olivine-liquid equilibrium. ContribMineral Petr 29:275–289

Rushmer T (1993) Experimental high-pressure granulites: Someapplications to mafic xenoliths and Archean terranes. Geology21(5):411–414

Salzmann U (2000) Are modern savannas degraded forests? - AHolocene pollen record from the Sudanian vegetation zone ofNE Nigeria. Veg Hist Archaeobot 9:1–15

Sautter V, Harte B (1988) Diffusion gradients in an eclogitexenolith from the Roberts Victor kimberlite pipe - 1. Mecha-nism and evolution of garnet exsolution in Al2O3-rich clino-pyroxene. J Petrol 29:1325–1352

Sautter V, Harte B (1990) Diffusion gradients in an eclogitexenolith from the Roberts Victor kimberlite pipe - 2. Kineticsand implications for petrogenesis. Contrib Mineral Petrol105:637–649

Schulze DJ (1987) Megacrysts from alkaline volcanic rocks. In:Nixon PH (ed) Mantle xenoliths. Wiley, New York, pp 433–452

Sen G (1988) Petrogenesis of spinel iherzolite and pyroxenite suitexenoliths from the Koolau Shield, Oahu, Hawaii—Implicationsfor petrology of the post-eruptive lithosphere beneath Oahu.Contrib Mineral Petrol 100:61–91

Sen G, Jones RE (1988) Exsolved silicate and oxide phases fromClinopyroxenes in a single Hawaiian xenolith—Implications foroxidation-state of the Hawaiian upper mantle. Geology 16:69–72

Sepp B, Kunzmann T (2001) The stability of clinopyroxene in thesystem CaO-MgO-SiO2-TiO2 (CMST). Am Mineral 86:265–270

Turner DC (1978) Volcanoes of the Biu basalt plateau, North-eastern Nigeria. J Min Geol 15:49–63

Van Breemen O, Pidgeon RT, Bowden P (1977) Age and isotopicstudies of some Pan-African granites from North-CentralNigeria. Precambrian Res 4:307–319

Velz S (1999) Die Verteilung von Spurenelementen zwischen Kli-nopyroxen, Granat, Orthopyroxen, Olivin und silikatischenSchmelzen im Druckbereich 1.5 GPa bis 3.0 GPa und beiTemperaturen von 1,150 �C bis 1,380 �C. Ph.D Thesis, Uni-versity of Koln, 172 pp (unpublished)

White WM, Patchett J (1984) Hf-Nd-Sr isotopes and incompatibleelement abundances in island arcs—Implications for magmaorigins and crust-mantle evolution. Earth Planet Sc Lett67:167–185

Wilkinson JFG, Stolz AJ (1997) Subcalcic clinopyroxenites andassociated ultramafic xenoliths in alkali basalt near Glen Innes,Northeastern New South Wales, Australia. Contrib MineralPetr 127:272–290

143

Wood BJ, Blundy JD (1997) A predictive model for rare Earthelement partitioning between clinopyroxene and anhydroussilicate melt. Contrib Mineral Petr 129:166–181

Wright JB (1970) High pressure phases in Nigerian Cenozoic lavas:Distribution and tectonic setting. B Volcanol 34:833–847

Wright JB (1976) Volcanic rocks in Nigeria. In: Kogbe CA (ed)Geology of Nigeria. Elizabethan Publishing, Lagos, pp 125–173

Zindler A, Jagoutz E (1988) Mantle cryptology. Geochim Cos-mochim Ac 52:319–333

Zinner E, Crozaz G (1986) A method for the quantitative mea-surement of rare-Earth elements in the ion microprobe. Int JMass Spectrom 69:17–38

144