U-Pb dating of metamorphic minerals: age of metamorphism and cooling history of Pan-African granulites and early Proterozoic eclogites in Tanzania

Precambrian Research 104 (2000) 123–146

U–Pb dating of metamorphic minerals: Pan-Africanmetamorphism and prolonged slow cooling of high pressure

granulites in Tanzania, East Africa

Andreas Moller a,b,*, Klaus Mezger b,1, Volker Schenk a

a Mineralogisch-Petrographisches Institut, Uni6ersitat Kiel, 24098 Kiel, Germanyb Max-Planck-Institut fur Chemie, Postfach 3060, 55020 Mainz, Germany

Received 1 November 1999; accepted 4 May 2000

Abstract

U–Pb monazite and zircon ages reveal that the high pressure granulites from eastern Tanzania were metamor-phosed during a Pan-African tectonothermal episode. These mineral ages range from 610 to 655 Ma and indicate thatpeak metamorphic conditions were diachronous in the different granulite domains. U–Pb titanite and rutile agesdefine integrated cooling rates of 2–5°C/Ma for all investigated granulite areas, and suggest a common process forthe post-metamorphic histories of the different granulite areas. Prolonged slow cooling-rates are consistent withnear-isobaric cooling in the deep crust after the metamorphic peak. The process responsible for crustal thickeningduring heating did not produce isostatic instability and fast erosion-driven or tectonic exhumation. The thermalhistory determined in this study is not consistent with the collision of East- and West-Gondwana as the cause ofgranulite facies metamorphism. Palaeomagnetic data have shown that this collision did not occur until 550 Ma, whenthe Pan-African granulites in Tanzania had already cooled below 500°C. The high pressure granulites of easternTanzania are thus interpreted as having attained their metamorphic peak prior to the final amalgamation ofGondwana, probably in an active continental margin setting. © 2000 Elsevier Science B.V. All rights reserved.

Keywords: Mozambique Belt; Pan-African orogeny; U–Pb geochronology; Monazite; Titanite; Rutile

www.elsevier.com/locate/precamres

1. Introduction

Metamorphic pressure–temperature–time (P–T– t) paths provide essential constraints for anymodels that relate metamorphism to tectonic pro-cesses. The different tectonic settings can be in-dicative of the plate-tectonic scenario that led tometamorphism and the formation of an orogenicbelt. In order to unravel the evolution of a com-

* Corresponding author. Present address: Institut fur Geo-wissenschaften, Johannes Gutenberg Universitat Mainz, Post-fach 3980, D-55099 Mainz, Germany. Tel.: +49-6131-3925584; fax: +49-6131-3924769.

E-mail address: [email protected] (A. Moller).1 Present address: Institut fur Mineralogie, Universitat Mun-

ster, Corrensstr. 24, 48149 Munster, Germany.

0301-9268/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.

PII: S 0301 -9268 (00 )00086 -3

A. Moller et al. / Precambrian Research 104 (2000) 123–146124

plex orogenic belt such as the Mozambique Belt(MB) of East Africa, the direct coupling ofgeochronologic data with petrologic information(Appel et al., 1998) and crustal residence ages(Moller et al., 1998) is of paramount importance.In this study, both the prograde and retrogradethermal histories are reconstructed using U–Pbages obtained on metamorphic minerals with dif-ferent closure temperatures including monazite,titanite and rutile. Because most of the mineralssampled in this study were extracted from gran-ulite facies metasediments they can be consideredto be most likely of metamorphic origin. Thisstudy also compares published U–Pb zircon agesand K–Ar, Ar–Ar, Rb–Sr on hornblende, biotiteand muscovite data from different granulite ter-

ranes in Tanzania for their consistency with newU–Pb ages. The scarcity of age data for thePan-African orogen of East Africa has led someauthors to use ages determined on different gran-ulite complexes for an integrated interpretation ofthe whole orogenic belt (Maboko et al., 1985,1989; Muhongo and Lenoir, 1994). However, itcan be shown that it is important to know the ageof metamorphism for each area separately forP–T– t path construction, because rock units jux-taposed today may have been at different crustallevels and experienced different P–T histories, butthe same tectonic and metamorphic processes.Samples from 17 locations (metapelitic gneisses,orthogneisses, marbles and calcsilicates) withinthe MB were chosen to cover the different partsof the respective granulite complexes in easternTanzania.

2. Geologic setting: granulite complexes in thePan-African Belt of Tanzania

One of the most influential contributions thatshaped the understanding of the African Precam-brian geology was the definition of the Mozam-bique Belt by Holmes (1951). He recognised thediscontinuity of geological structural trends be-tween the Tanzania craton and its eastern hinter-land and showed that these areas had to beyounger than the craton. Subsequently Shackleton(1967) proposed that the MB has a complex his-tory and suggested that the belt is composed ofArchaean basement and several younger metasedi-mentary sequences. The MB then served as one ofthe classical examples for rejuvenation (i.e. nonew crustal material added during orogenic cycle)of Archaean and Early Proterozoic basement(Watson, 1976). However, it was also proposedthat the MB is a product of late Precambrianplate collision following ocean closure (Burke etal., 1977 McWilliams, 1981).

Stern (1994) proposed the term ‘East Africanorogen’ for the areas covered by the older terms‘Arabian–Nubian shield’ and ‘MozambiqueBelt’’, because it is appropriate to view the wholearea as the product of one Neoproterozoic

Fig. 1. Simplified geological map of eastern Tanzania,modified from Coolen (1980). Important granulite domains inthe Mozainbique Belt are indicated by shaded areas. Newlyrecognised granulite occurrences in the Mozambique Belt afterAppel et al. (1998). The western limit of Pan-African meta-morphic influence on the Proterozoic Usagaran Belt is indi-cated by a dashed line after Gabert and Wendt (1974) andPriem et al. (1979).

A. Moller et al. / Precambrian Research 104 (2000) 123–146 125

Wilson-cycle (see inset of Fig. 1). The Arabian–Nubian shield contains large tracts of Pan-Africanjuvenile crust and abundant ophiolites and is in-terpreted by Stern (1994) as a collage of accretedterranes. In contrast, the MB with its high-gradegneisses resembles the deeply eroded root of anorogen formed by a single collision event betweenEast- and West-Gondwana. The MB experiencedfurther uplift during Phanerozoic rifting, some ofit associated with the development of the EastAfrican Rift. This interpretation supports themodel of Hoffman (1991), i.e. the MB was formedby fan-like closure of a previously existingMozambique ocean, with the hinge of the fansomewhere in South Africa. Since this fan neverfully closed, crustal shortening was most intensein the southern part of the belt. Stern (1994)argues further that the exposure of granulites atthe surface in Kenya and Tanzania is evidencethat crustal thickness of the orogen was greatestand collision most intense in these areas, becausetoday the granulites are found within the crust ofnormal thickness of approximately 35km (e.g.KRISP Working Party, 1995).

Within Tanzania, geochronological resultsshow that the Mozambique Belt of Holmes (1951)has to be subdivided into a Pan-African (lateProterozoic) domain to the east and an Usagaran(=Ubendian, Early Proterozoic) domain to thewest (Fig. 1). A tentative subdivision in southernTanzania was based on progressively older Rb–Srbiotite ages towards the west (Wendt et al., 1972;Priem et al., 1979) interpreted as the result of thedecreasing Pan-African thermal overprint on theEarly Proterozoic rocks (Gabert and Wendt,1974). U–Pb dating of metamorphic monaziteand titanite from eclogite-facies rocks places themain metamorphic event in the Usagaran domainat 2000 Ma (Moller et al., 1995). Appel et al.(1998) suggest that distinctive decompression tex-tures in the Usagaran Belt and cooling textures inthe Pan-African granulites can be used to distin-guish the two belts.

To distinguish the two metamorphic events weendorse the use of the terms ‘Pan-African Belt ofEast Africa’ or the ‘East-African Orogen’ pro-posed by Stern (1994) for the Pan-African gran-ulite facies gneisses of eastern Tanzania and use

the name ‘Usagaran Belt’ or ‘Ubendian–Usagaran Belt’ for the region where the mainmetamorphic event occurred at about 2 Ga (Fig.1). The term Pan-African is used in this study forthe time span from about 650 to 550 Ma, relevantto and encompassing metamorphic events in thecircum-Indic region related to the formation ofGondwana.

The Pan-African Belt in Tanzania consists ofArchaean to Proterozoic rocks (e.g. Moller et al.,1998) metamorphosed under granulite facies con-ditions (e.g. Bagnall, 1963; Sampson and Wright,1964; Coolen, 1980; Appel et al., 1998) during thePan-African orogeny (e.g. Coolen et al., 1982;Maboko et al., 1985, this study). Some of thegranulite complexes (Fig. 1) apparently form faultbounded mountain ranges, interpreted as tectonicklippen (e.g. Shackleton, 1986), namely the Pareand Usambara Mountains (Bagnall, 1963; Bagnallet al., 1963) and the Uluguru Mountains (Samp-son and Wright, 1964).

Previous petrologic and geochronological stud-ies have been carried out mainly on the Furuacomplex (Coolen, 1980; Coolen et al., 1982), theWami River complex (Maboko et al., 1985), andthe Uluguru Mountains (Muhongo, 1990;Maboko et al., 1989). The granulite complexeswithin the Mozambique Belt exhibit striking simi-larities in lithology, structure and grade of meta-morphism (Coolen, 1980; Appel et al., 1998).Petrologic studies reveal similar peak metamor-phic conditions of 810940°C and 9.5 to 11 kbarand a similar P–T path for an extensive areawithin the Pan-African Belt including the Pare,Usambara and Uluguru Mountains granulitecomplexes and some adjacent lowland areas (Ap-pel et al., 1998).

3. Analytical methods

Heavy minerals were separated using routineprocedures which involved steel jaw-crusher andsteel roller-mill, Wilfley table, Frantz magneticseparator and heavy liquids. Mineral fractionswere then hand-picked under a binocular micro-scope to avoid inclusions and obtain grain orfragment fractions of similar size, shape and


colour. Titanite was cleaned in pure alcohol in anultrasonic bath for about 15 min, washed in warmdistilled 3 N HCl for about 10 min to removesurface contamination, and twice rinsed in dis-tilled water. Monazite was washed in warm dis-tilled water only prior to dissolution. Rutile waswashed in warm 0.5 N HF for about half an hour,zircon in hot 6 N HCL for about 15 min.

Uranium and Pb concentrations were deter-mined by isotope dissolution with a 233U/205Pbmixed spike, added before dissolution to allowoptimum homogenisation with the sample. Ele-ment concentrations in weighed mineral fractionsare known to about 0.2%, calculated from analy-tical errors alone. All zircon-, monazite- and ru-tile-fractions were digested in 3 ml Savillex®

screw-top beakers in a Krogh-style or Parr®

Teflon® bomb within a screw top steel containerat 210°C. Monazite dissolved in 0.5 ml 7 N HNO3

and 0.5 ml 6.2 N HCl after 1–3 days. Rutiledissolved within a few days in a mixture of 0.5 mlconcentrated HF and five drops of 7 N HNO3.Titanite fractions were digested overnight in theoven in a mixture of 0.5 ml concentrated HF andten drops of 7 N HNO3 after boiling for 12–24 hon the hot-plate. The zircon fraction dissolved inconcentrated HF and ten drops of 7 N HNO3 inthe oven within 10 days. Dissolution was checkedoptically for each sample, under a microscopewhere necessary.

Uranium and Pb were separated with ion-ex-change Teflon® columns filled with about 0.5 mlof DOWEX AG 1X8® anion exchange resin (e.g.Krogh, 1973; Tilton, 1973). Pb chemistry for mon-azite, rutile, titanite, and feldspar employed theHBr–HCl method, whereas Pb from zircon wasseparated with HCl. Uranium was separated withthe HCl–HNO3 method. Five total proceduralblanks were determined between 44 and 123 pgwith an average of 80 pg. The Pb-isotope ratiosmeasured for the blank were: 206Pb/204Pb: 18.53;207Pb/204Pb: 15.69; 208Pb/204Pb: 35.90.

Isotope ratios were measured on a FinniganMAT 261 mass-spectrometer in multi-collectorstatic mode on Faraday cups, using single Refilaments. A secondary electron multiplier (SEM)was used for measuring 204Pb when high ratiosmade it necessary, and for some U analyses in

dynamic mode. Pb was loaded with H3PO4 andsilica-gel (Cameron et al., 1969). The measured Pbisotopic ratios were corrected for fractionationwith a mass discrimination factor of 0. 1%/amu,based on 23 analyses of 50 ng of equal atomSRM-982, measured during this study in compari-son with the values recommended by Todt et al.(1996). Reproducibility of the 207Pb/206Pb ratio ofthe SRM-982 standard (average: 0.466512) was0.033%. Within-run reproducibility was muchhigher, with an average of 0.0021% at 2s confi-dence level. The measurements of 206Pb/204Pb ra-tios with 204Pb on the SEM were corrected with afactor of 1.0038, determined from five measure-ments of SRM-982. Most U was measured asoxide after loading with H3PO4 and silica-gel.Based on repeated analyses of 100 ng SRM-U500standard, a mass fractionation correction factor of0.01%/amu was applied to samples measured instatic mode and a correction factor of 0.3%/amuto SEM dynamic measurements. Reproducibilityfor the 235U/238U ratio of the standard (staticmode) was 0.29%, with an average within-runreproducibility of better than 0.04%.

For some samples, U was loaded with graphitedispersed in a water/alcohol-solution and mea-sured as U+ at temperatures between 1650 and1740°C. Reproducibility estimated from sevenU500 standards loaded with graphite was 0.28%for Faraday cup in static mode. Fractionation wascorrected with a factor of 0. 1%/amu. Mass frac-tionation was strongly time-dependent with thesegraphite loaded samples and care was taken toheat up all samples in the same manner and avoidacquisition times longer than approximately fiveblocks of 20 measurements each.

4. Closure temperature estimates

For a valid interpretation of mineral ages, infor-mation on the closure temperature (Tc) for parent/daughter systems is essential. The closuretemperature is defined as the temperature of thesystem at the time given by its apparent ages(Dodson, 1979). This Tc depends on grain size andshape as well as on geologic parameters like cool-ing rate and also on crystallographic parameters.


Table 1Summary of approximate closure temperaturesa

Mineral Grain size (mm)Tc (°C) References

U–Pb system100–300 Schenk (1980, 1990), Bingen and van Breemen (1998), Parrish and800 (peakMonazite

metamorphism) Whitehouse (1999), this study200–30 000Titanite Mezger et al. (1991), Gromet (1991), Cherniak (1993), Scott and St-Onge(630–730)

(1995), Zhang and Scharer (1996)200–500650 This study130–430 Mezger et al. (1989)380–420Rutile

K–Ar and Ar–Ar system160450–500 e.g. Harrison (1981)Hornblende

–Muscovite e.g. Hanson and Gast (1967)350–400–300 e.g. Harrison et al. (1985)Biotite

Microcline 125–250150–200 Harrison and McDougall (1982)

Rb–Sr systemMuscovite 450–500 – e.g. Harrison and McDougall (1982)

–350 e.g. Harrison and McDougall (1982)Biotite

a Tc values are chosen for selected minerals at different grain sizes for slow cooling rates of 1–10°C/Ma.

For some parent/daughter systems in someminerals experimental data is available (e.g. U–Pb in titanite, Cherniak (1993); K–Ar in horn-blende, Harrison (1981) U–Pb in monazite,Smith and Giletti (1997)). For other systems onlyempirical values are available and many of themmay need further refinement. A correlation ofexperimental results with well controlled naturalgeologic settings is still wanting for many miner-als. Table 1 summarises Tc for minerals relevantto this study and the choice preferred by theauthors, which is pivotal for the interpretation ofthe geochronological data and the cooling his-tory.

4.1. Monazite

It is generally accepted that the Tc of monaziteis at least 700°C for slowly cooled rocks. How-ever, there is ample evidence that the Tc may besignificantly higher as indicated by field datafrom the Hercynian crustal section of Calabria,Italy, (Schenk, 1980, 1990) or the Valhalla com-plex in British Columbia (Spear and Parrish,1996). A single grain U–Pb study by Bingen andvan Breemen (1998) in amphibolite to granulitefacies rocks shows that monazite growth ages

can be preserved through 850°C metamorphismunder dry conditions. A study by Parrish andWhitehouse (1999) also suggests higher Tc. Re-cent experiments on Pb diffusion rates in monaz-ite by Smith and Giletti (1997) suggest thatcircular or elongate monazite grains of 100 mmradius should have closure temperatures of only630–720°C in regions which cool at rates be-tween 1 and 10°C/Ma. The authors caution thatuncertainties in their closure temperature calcula-tions may be as high as 140°C. Comparison withthe examples from geochronological field studiesin granulites (see above) leads us to concludethat the calculations of Smith and Giletti (1997)underestimate the closure temperature of monaz-ite and are not accurate enough for application.

We conclude that monazite ages from thisstudy may be interpreted as growth ages andthus date the peak of the granulite facies meta-morphic event in eastern Tanzania which reachedtemperatures of 810940°C (Appel et al., 1998)in most areas. Estimates of a lower Tc may thenbe due either to growth of new monazite at tem-peratures below its Tc or, alternatively, to recrys-tallisation or Pb-loss induced by deformationand/or fluids rather than by diffusion.


4.2. Titanite

Experimental as well as empirical estimates forthe Tc of the U–Pb system in titanite are avail-able. Mezger et al. (1991) estimated a closuretemperature of 630°C for titanite crystals of 1 cmdiameter at a cooling rate of 2°C/Ma from fieldstudies in the Adirondack Mountains. In otherparts of the Grenville Orogen, titanite preservedtheir U–Pb ages although the surroundinggneisses were later migmatised, which indicatesthat Tc may be at least as high as 650°C (Mezgeret al., 1992) for larger grains. Experimental stud-ies of Cherniak (1993) yield a closure temperatureof approximately 630°C for a diffusion radius of500 mm at 2°C/Ma cooling rate. Cherniak (1993)thus concluded that effective diffusion radius maybe smaller than grain size.

Evidence for a higher closure temperature oftitanite in slowly cooled rocks was presented byScott and St-Onge (1995). Their combination ofthermobarometry and U–Pb dating suggests thatthe Tc of 100 mm–1 mm diameter titanite lies inthe range 660–700°C, higher than all previousestimates. This conclusion is now supported byother studies (e.g. Corfu (1996), Verts et al.(1996)) and by discordance patterns observed inrocks which experience brief thermal events,where discordant titanite data can be interpretedwith the episodic Pb-loss model (e.g. Tucker et al.(1986), Haggart et al. (1992)). Similar evidencewas presented by Gromet (1991), where titanitegrains of 500 to 2000 mm diameter showed strongdiscordance to an upper intercept and even titan-ite grains of 250 mm diameter showed some inher-itance although this rock experienced only about650°C during a metamorphic event, possibly re-lated to the brevity of the overprint. From inher-ited magmatic titanite in a syenite intrusion,Zhang and Scharer (1996) deduced a closure tem-perature for volume diffusion of Pb in excess of710°C. They suggest that titanite is always closedto Pb at its crystallisation temperature and thatthe closure temperature concept may be mislead-ing for metamorphic titanite, a contention notsupported by the data of this study. Importantfactors in all these studies are the time of titanitegrowth relative to the onset of cooling and the

duration of the metamorphic event in case thetitanite had formed previously. Both may limit theability to determine a closure temperature fromthese mineral ages for slowly cooled terranes.

Most titanite fractions analysed in this studyconsist of whole grains with a diameter of 200–500 mm. Based on the studies of Cherniak (1993),Gromet (1991), and Scott and St-Onge (1995) Tc

of 650°C for titanite from granulites-facies rockswith slow cooling rates has been used in this studyas a conservative estimate.

4.3. Rutile

An estimate for the closure temperature for Pbin rutile was given by Mezger et al. (1989), basedon comparison with K–Ar and 40Ar/39Ar ages ofhornblende and biotite. It was suggested that theclosure temperature is ca. 430930°C for slowcooling rates of 2–10°C/Ma. This value is consis-tent with U–Pb ages on rutile obtained byScharer et al. (1986). Most published U–Pb agesof rutile, are concordant or only slightly discor-dant. Inheritance of older age information is onlylikely when a metamorphic overprint does notexceed greenschist-facies conditions (Moller et al.,1995).

4.4. Mica and amphibole

For the K–Ar and Ar–Ar method, the state ofrecrystallisation of the minerals (Dallmeyer et al.,1990), and their chemical composition (Lee, 1993)appear to be critical in the control of Ar release.Excess Ar is a particular problem for biotite andhornblende. In general this excess Ar can be takenin from the growth environment or it can beinherited from an older event. However, it is alsopossible that minerals lose K, but not Ar, duringlow temperature alteration and this also results inan apparently old age. Since biotite, but not mus-covite, is generally much more prone to yield oldAr–Ar ages, this may indicate that chloritisationis the cause for the apparent excess ages. There-fore, only perfectly fresh biotite and hornblendecan be used for high precision K–Ar and Ar–Argeochronology in metamorphic rocks.


Table 1 shows that Ar–Ar ages from horn-blende are expected to be close to, but youngerthan U–Pb titanite ages from the same area andolder than U–Pb ages of rutile. Similarly, Ar–Arages of biotite are expected to be close to, butyounger than, U–Pb ages obtained on rutile. Thisis important in the discussion of inherited orexcess Ar, which appears to be very common inbiotite and hornblende of granulites from theMozambique Belt (data of Priem et al. (1979) andMaboko et al. (1989), discussed below).

5. Results

Sample locations are illustrated and geochrono-logical results summarised in Fig. 2. Major andtrace element analyses for most of the samples arereported by Appel (1996). Descriptions of sample

locations and mineral assemblages can be foundin Table 4.

Several reasons — geological and analytical —may explain discordant analyses (Pb-loss, over-growth core relations, incomplete dissolution).This study uses Pb-loss as the most likely interpre-tation and unless specifically stated the 207Pb/206Pb of discordant analyses is the most likelyminimum age for this mineral fraction. Reverselydiscordant results are quite common with monaz-ite and often indicate excess 206Pb (from short-lived 230Th) due to preferential incorporation ofTh over U during monazite growth (Scharer,1984), with monazite remaining below its closuretemperature. For such reversely discordant re-sults, this study uses the 207Pb/235U age as the bestestimate for the true age since this ratio is notaffected by excess 206Pb. Because concordant re-sults could also be the result of a combination of

Fig. 2. Simplified geological map with the sample locations and the U–Pb geochronological results on monazite, titanite, rutile andon one zircon fraction from the Pare and Usambara Mountains, Umba Steppe and Uluguru Moutains. Map compiled from quarterdegree sheets of the Tanzanian Geological Survey. For location within Tanzania see Fig. 1.


Fig. 3. Concordia diagram for monazite from the Pare andUsambara Mountains and the Umba Steppe. The resultsindicate an age difference of about 15 Ma. Two pairs ofdiscordant and near concordant monazite from the Pare andUsambara Mountains are shown (T115, A108). For bothlocations, monazite with the higher U content are more discor-dant than monazite with lower U content.

5.1. Monazite and zircon

Most of the monazite used in this study wasseparated from metasedimentary rocks. The mon-azite occur as pale to bright yellow spheres orellipsoids free of inclusions and range in diameterfrom 100 to 600 mm.

Two monazite fractions from the Pare Moun-tains (metapelites A16 and T115b) plot slightlyabove concordia (Fig. 3). Their 207Pb/235U ages of64192 Ma are interpreted as the true ages offormation. Another fraction from metapelite T115is discordant with a similar 207Pb/206Pb age of64092 Ma.

Monazite from metapelite T137 and metagrani-toid T121 from the Usambara Mountains are alsoreversely discordant and yield 207Pb/235U ages of62192 and 62492 Ma, similar to the 207Pb/206Pbage of 62492 Ma obtained from concordantmonazite fraction A108b (Fig. 3). A second, dis-cordant fraction of monazite from metapeliteA108a has a slightly higher 207Pb/206Pb age of62992 Ma. Monazite in metapelite T137 hasbeen observed in thin section to occur mostly aslarge (\300 mm) grains attached to high-Ti gran-ulite facies biotite and hence as part of the highgrade assemblage. Backscatter electron (BSE)imaging indicates strong ‘patchy’ zoning, reflect-ing zonation in Th content (Fig. 4). U–Pb datingby laser-ICP-MS yielded a 206Pb/238U age of618915 Ma from four analyses on two grains(Moller and Jackson, unpublished data) support-ing the multi-grain-isotope dilution results of thisstudy. Two other analyses indicate some Pb loss,but no evidence of older growth phases within themonazite has been found.

A fraction of monazite grains from semipeliteA 144 in the Umba Steppe yields a 207Pb/206Pbage of 60992 Ma (Figs. 3 and 5), the discordance(1.5%) may be attributed to recent Pb-loss, possi-bly due to weathering.

Four fractions of monazite were analysed fromthree metapelite samples of the northern, north-eastern and eastern Uluguru Mountains (Fig. 6).Two monazite fractions from sample P1 show nodifferences in colour or grain size and differ onlyslightly in their Th and U contents. Fraction P1ais 1% discordant and has a 207Pb/206Pb age of

Fig. 4. Backscatter-electron image of monazite in metapeliteT137 showing pronounced patchy zoning probably related togrowth inhibition at low H2O activity under granulite faciesconditions.

reverse discordance and Pb-loss, their 207Pb/206Pbhas also to be taken as a minimum age. For agecalculations of some mineral fractions (rutile andsome titanite) very sensitive to corrections for Pbblank and initial common Pb, the 206Pb/238U agecan be regarded as the best age estimate.


66993 Ma, whereas fraction P1b is 0.6% re-versely discordant with a 207Pb/235U age of 65792 Ma. Sample P1b could have been affected bysome late disturbance, and the reverse discor-dance observed may only be the remainder of anoriginally much higher discordance. The age of

P1a may either indicate the presence of an inher-ited Pb component (detrital core) or reflects theage of monazite growth during prograde meta-morphism. We consider the latter case more likelyand the result of fraction P1a is, therefore, notused to calculate cooling rates for this location.However, this result may have some bearing onthe discussion of the age and duration of highgrade metamorphism in this part of the UluguruMountains.

Monazite from the graphite-rich metapelite P9has a high a 208Pb/206Pb ratio of 31.4 and plotssignificantly above concordia (2%) with a 207Pb/235U age of 64692 Ma. The monazite fractionanalysed from the eastern Uluguru Mountains(T28) is also slightly reversely discordant with a207Pb/235U age of 65392 Ma (Fig. 6). The resultsfrom three metapelite samples of the northern andeastern Uluguru Mountains thus span an agerange of about 11 m.y. between 64692 and65792 Ma (Table 2; samples P1b, P9, T28).

Monazite and zircon were separated from ameta-qtz-diorite (T46) originating from the north-western Uluguru Mountains. This meta-qtz-dior-ite shows evidence of all three deformationalphases observed in the surrounding granulite-fa-cies rocks (Appel et al., 1998). Its emplacement,therefore, must have been pre-peak-metamor-phism and pre-deformation. Two monazite frac-tions from the sample have different U–Th–Pbcontents but a very similar age. The smaller mon-azite grains. with a higher Th/U ratio are slightlyreversely discordant at a 207Pb/235U age of 62492Ma. The larger monazite size fraction has a lower208Pb/206Pb ratio and is slightly discordant at a207Pb/206Pb age of 62592 Ma. The zircon frac-tion analysed from this sample consists of 19long-prismatic, clear and euhedral grains (\ l40mm) without evidence of older cores when ob-served in transmitted light. The U–Pb result isslightly discordant (1%) with a 207Pb/206Pb age of62693 Ma, overlapping the age of the monazitefractions (Fig. 6).

5.2. Titanite

The geological map (see Fig. 2) shows very fewmarbles and calcsilicate rocks in the Pare and

Fig. 5. Concordia diagram for monazite, titanite and rutilefrom the Umba Steppe. Different fragments of the sametitanite grain from sample A141 yield similar 207Pb/206Pb agesclose to the 207Pb/206Pb age of 60992 Ma obtained formonazite from sample A144.

Fig. 6. Concordia diagram for monazite and a zircon fractionfrom the Uluguru Mountains. Note the age difference betweenmonazite and zircon from the granodiorite of NW UluguruMountains and the monazite from metapelites of the N and EUluguru Mountains.

A.

Moller

etal./

Precam

brianR

esearch104

(2000)123

–146

132Table 2U–Pb isotope dataa

Sample, mineral Ages (in Ma)Rock type Discordancef (%)Wt. (mg)b Pb (ppm) U (ppm) Isotopic ratios

206Pb/207Pbc 208Pb/206Pbd 207Pb/206Pbd 206Pb/238Ue 207Pb/235Ue 206Pb/238U 207Pb/235U 207Pb/206Pb

Pare Mountains21.50 0.082368 0.10473 0.88055 642920.35 64192389 63994 0.6MetapeliteA16, Mnz 185 597

0.8082 0.061064 0.09612T115a, Mnz 0.80872Metapelite 59292 60292 64092 −7.50.32 470 3050 5870029.87 0.063380 0.10495 0.88088 64392 641922110 63594790 1.3280T115b, Mnz Metapelite 0.15

610Calcsilicate 0.6612 0.083598 0.09663 0.80160 59592 59892 61093 −2.55.49 19.4 130A26, Tit167.4Metabasite 0.9988 0.146452 0.08937 0.74589 552910 56699 623916 −11.44.52 18.2 98T114, Tit

0.0181 0.061373 0.08635 0.69205 53492 534922140 53594A16, Rt −0.16.860.5513.31Metapelite0.2226 0.058517 0.08517 0.68126 52792 52892T115, Rt 53094Metapelite −0.610.89 2.9 31.1 9750

Usambara Mountains23.75 0.064764 0.10173 0.84275 625920.40 621921260 60692 3.0570MetapeliteT137, Mnz 2640

4210 28900 3.189 0.061022 0.10041 0.84068 61793 62093 62992 −2.00.22A108a, Mnz 1560Metapelite3020 22600 3.382 0.060666 0.10182 0.85021 62592 62591 62492 0.20.12A108b, Mnz 1180Metapelite

100.3 0.061912 0.10178 0.84821 62592 624924470 61992T121, Mnz 0.948042600.22Charnockite0.3119 0.178682 0.08191 0.64797 50892 50794T137, Rt 506917Metapelite 0.34.00 4.89 39.5 118.30.0446 0.070951 0.08384 0.66902 51992 52092830 52594T139, Rt −1.18.20.678.30Meta-qtz-di.

A108, Rt 0.0718Metapelite 0.059622 0.08526 0.68112 52792 52791 52793 0.012.78 1.14 13.7 4050

Umba Steppe1170 9730 5.949 0.061302 0.09740 0.80778 59992 60192 60992 −1.60.50A144, Mnz 690Grt-Bt gneiss

0.0181 0.061365 0.07358 0.61103 45892 4849212300 61292A141a, Tit −25.225401742.82Marble0.0502 0.061741 0.09333 0.77705 57592 58492A141b, Tit 61792Marble −6.82.38 111 1240 70600.0371 0.071373 0.08306 0.65759 51492 51392928 50897A144a, Rt 1.3362.93.70Grt-Bt gneiss

A144b, Rt 0.0025Grt-Bt gneiss 0.057805 0.08318 0.66102 51595 51594 51693 −0.215.52 2.4 32 16800

Uluguru Mountains0.0810 0.062140 0.10075 0.84223 61992T46, Zrng 62092Meta-qtz-di. 62693 −1.20.59 37.1 370 6100

18.52 0.064354 0.10180 0.84866 62592 624922140 62093Diorite 0.7T46a, Mnz sm 77013400.088300Diorite 11.88 0.061860 0.10106 0.84457 62192 62292 62592 −0.80.18 2460 2170T46b, Mnz la

1040 6110 7.995 0.063559 0.10828 0.92330 66392 66492 66993 −0.90.25P1a, Mnz 880Metapelite8.514 0.063433 0.10740 0.90936 65892 657925060 65492910 0.61020P1b, Mnz Metapelite 0.20

2190Metapelite 31.37 0.065035 0.10612 0.88956 65092 64692 63293 2.90.23 960 320P9, Mnz7370Metapelite 4.621 0.062217 0.10683 0.90205 65492 65392 64892 1.00.19 510 980T28, Mnz

0.9208 0.092502 0.09415 0.78521 58096 58897453 621922P8a, Tit −6.626043.64.71Calcsilicate280 553 0.7989 0.086545 0.10183 0.84817 625911 62498 61894 1.1P8b, Tit Calcsilicate 2.79 48.1

0.4862 0.081095 0.09893 0.82422 60895 61095681 61997Marble −1.7T25a, Tit 11515.85.78630Marble 0.0504 0.082931 0.09638 0.80358 59394 59993 62194 −4.44.76 10.8 112T25b, Tit831Marble 0.5163 0.077363 0.09989 0.83277 61497 61595 62093 −1.04.74 17 121T25c, Tit

0.0479 0.105575 0.09956 0.82986 61296 61495317.5 62096P88a, Tit −1.311612.24.43Calcsilicate0.4831 0.100277 0.09992P88b, Tit 0.83243Calcsilicate 61493 61592 62195 −1.13.83 19.5 136 359.10.0314 0.063062 0.08853 0.71904 54792 550912740 564922.1 −3.025.3P1, Rt Metapelite 10.27

1910Metapelite 0.2263 0.064848 0.08909 0.71589 55094 54893 54093 1.812.02 2.5 25.5P9, Rt25.3 2120 0.0976 0.062982 0.08065 0.63767 50095 50194 50595 −0.9T28, Rt 12.2Metapelite 2.0

a Ages and errors (2s) are calculated with Pbdat and ISOPLOT for Excel 2.0.6 software, after Ludwig (Ludwig, 1980, 1994); Zrn, zircon; Mnz, monazite; Tit, titanite; Rt, rutile; sm, small; la, large.b Most Mnz weights estimated from size, error about 10–20%, other samples weighed toB1% error.c Measured ratio.d Measured ratio, corrected for spike, 80 pg Pb blank, 0. 1% mass fractionation per a.m.u.e Corrected for spike, 80 pg Pb blank, 0. 1% mass fractionation per a.m.u. and common Pb composition determined from leached coexisting K-feldspar or plagioclase (Moller et al., 1998).f Discordance of result expressed as deviation of the 207Pb/206Pb age from the 206Pb/238U age = [207Pb/206Pb age/206Pb/238U age/100]−100.g Ninteen clear, pink, euhedral, prismatic grains, \140 mm mesh size, length to width ratio \6.

reviewer

Highlight

reviewer

Cross-Out

reviewer

Replacement Text

4

reviewer

Highlight


Fig. 7. Concordia diagrams for titanite.

isotope composition of A26 as no analysis ofcoexisting plagioclase was available. The possibleerror associated with the correction is small con-sidering the narrow range of common Pb compo-sition of feldspars from the Pare and UsambaraMountains (Moller et al., 1998). The result sug-gests that cooling through the closure temperatureof titanite in the South Pare Mountains occurredin the same age range as in the North PareMountains.

In the Umba Steppe interlayered calcsilicaterocks and marbles were found at the Umba river.A coarse-grained sample (A141) yielded anopaque to very dark brown titanite grain of morethan 0.5 cm diameter. Two fragments from thecore of the grain have variable U and Pb concen-trations and degree of discordance (Table 2, Fig.5) but similar 207Pb/206Pb ages of 61292 and61792 Ma. The high U content of the titanitefragments may have caused structural damageand Pb-loss. It can be concluded that the 207Pb/206Pb ages record the age of peak metamorphismand the effective closure temperature of this titan-ite grain is higher than the ca. 730°C calculatedfor grains with 0.5 cm diffusion radius by Cher-niak (1993) or alternatively that metamorphictemperatures did not exceed Tc after titanitegrowth.

Suitable titanite-bearing samples were onlyfound in the eastern part of the Uluguru Moun-tains (P8 and T25) close to the locations of themonazite and rutile samples. An additional sam-ple was taken from the southeast Uluguru Moun-tains (P88). To evaluate the reproducibility ofU–Pb ages of titanite from the eastern UluguruMountains, several fractions of grains wereanalysed for each sample. Their 207Pb/206Pb agesoverlap within error at 618–621 Ma. The variablydiscordant titanite fractions can be fitted on asingle regression line (Fig. 7b) despite being takenform localities about 60 km apart. The combinedintercept of titanite in the eastern Uluguru Moun-tains at 61992 Ma is interpreted as the age atwhich at least this part of the Uluguru Mountains(the crystalline limestone group of Sampson andWright (1964)) cooled through the closure temper-ature of titanite at ca. 650°C.

Usambara Mountains. Only two suitable samplesof titanite-bearing rocks could be collected (A26,T114) from the Pare Mountains. Sample A26from the North Pare Mountains is a calcsilicategneiss with the metamorphic assemblage Grt+Cpx+Hbl+Pl+Qtz+Cc+Scp+Tit9Kfs. Thetitanite is reddish-brown and does not exhibitcolour zonation. The 207Pb/206Pb age of 61093Ma (Fig. 7a) is interpreted as the minimum agefor closure at ca. 650°C. Titanite in metabasiteT114 from the South Pare Mountains is palebrown and only 150 to 200 mm in diameter. It hasa low 206Pb/204Pb ratio of 167.4 and yields animprecise 207Pb/206Pb age of 623916 Ma. Com-mon Pb correction was carried out with the Pb


5.3. Rutile

The rutile fractions used for U–Pb age determi-nations were obtained from metapelitic samplesthat also yielded monazite (A16, T115, T137,A108, A144, P1, P9, T28). Rutile in themetapelitic samples were mostly elongate grains(aspect ratio higher than 4) or fragments thereof.In all of these samples rutile is part of the high-pressure granulite-facies assemblage together withgarnet, sillimanite/kyanite, plagioclase, quartz9ilmenite. An exception is qtz-dioritic enderbiteT139 from the western Usambara Mountainswhich contains large, dark, short rutile grainswith an average diameter \250 mm. Rutile from

a single sample often spans a range of coloursfrom translucent reddish brown to almost opaqueand dark-brown to black. Care was taken to pickgrains of similar size and colour for each rutilefraction (4–16 mg) to avoid mixing of differentchemical compositions and possibly different dif-fusion behaviours.

Concordant rutile fractions yield 207Pb/206Pbages of 53594 and 53094 Ma for the North andSouth Pare Mountains, respectively (Fig. 8a). Re-sults of two rutile fractions from meta-qtz-dioriteT139 and metapelite A108 from the UsambaraMountains have indistinguishable 207Pb/206Pb agesof 52594 and 52793 Ma, respectively but sig-nificantly different 206Pb/238U ages of 5 1992 and52792 Ma. Rutile from sample T137 is stillyounger with a 206Pb/238U age of 50892 Ma, butit is slightly discordant and has a large uncer-tainty in 207Pb/206Pb (506917 Ma) due to its lowproportion of radiogenic Pb. Sample T139 wascollected just 10 km from metapelite T137 andbelongs to a suite of meta-qtz-diorites, some ofwhich cross-cut the foliation of the surroundinggranulites. Its rutile age is 11 Ma older than thatof rutile from sample T137 and could possibly beexplained by the larger diffusion radius of thestubby rutile grains from the meta-qtz-diorite.

Two rutile fractions from the Umba Steppewere picked from the same sample of semipeliticgneiss as the monazite (A144) and yield similar206Pb/238U ages of 51595 and 51492 Ma (Fig.5). The two rutile fractions from metapelite sam-ples P1 and P9 of the northern Uluguru Moun-tains are normally and reversely discordant,respectively. There is no reason to assume thatrutile could be affected by excess 206Pb since rutilegenerally has extremely low Th contents. Unlessother geological problems are responsible for thediscordance, the best estimate of the ‘true’ age isprobably the 206Pb/238U age (see discussionabove). The 206Pb/238U ages are indistinguishablewithin error at 55094 and 54792 Ma. They areinterpreted to date the time the rocks cooledbelow Tc. Rutile from metapelite T28 is concor-dant and has a 206Pb/238U age of 50095 Ma (Fig.8b). The slightly discordant ages at ca. 550 Mafrom the northern Uluguru Mountains are alsoabout 15 Ma older than the oldest rutile ageFig. 8. Concordia diagrams for rutile.


obtained in the Pare and Usambara Mountains, asimilar age difference as observed with themonazite.

6. Discussion

6.1. Discussion of pre6ious geochronologicalresults

The geochronologic results from this study canbe combined with published mineral ages andP–T estimates to derive quantitative P–T– tpaths for the different granulite segments of theMozambique Belt. Previously publishedgeochronological data for the granulites of easternTanzania are summarised in Table 3. They arerecalculated using modern decay constants whennecessary and re-interpreted (using the closuretemperatures summarised in Table 1). All ages arecombined to discuss and compare the coolinghistories of the different Pan-African granulitecomplexes in Tanzania.

Zircon ages are available from four of thegranulite complexes (Coolen et al., 1982; Mabokoet al., 1985; Muhongo and Lenoir, 1994). Theseages span a period of 70 Ma between 645 Ma and715 Ma and were originally interpreted as thetime of high grade metamorphism. Upper discor-dia intercepts at around 700 Ma (Maboko et al.(1985), Fig. 9a) are re-interpreted as intrusionages. However, the U–Pb results on the largemagnetic and size fractions yield mostly shortdiscordias intersecting concordia at a low angle(Maboko et al., 1985) and are, therefore,imprecise.

Studies which allow direct comparison of Rb–Sr with K–Ar data from the same terrane (An-driessen et al., 1985; Priem et al., 1979) or directcomparison of K–Ar and Ar–Ar data (Mabokoet al., 1989) reveal that many of the K–Ar andsome Ar–Ar ages of biotite and hornblende aretoo old and may be influenced by excess Ar (seeFig. 9b and c). Biotite Rb–Sr data and muscoviteK–Ar and Ar–Ar data, however, are consistentwith other geochronological results.

For the Wami River granulites correlation ofthe re-interpreted results yields a slow integrated

Fig. 9. Cooling paths reconstructed for Pan-African granulitecomplexes in the Mozambique Belt of Tanzania with pub-lished thermochronological data. (a): 1, results of U-Pb onzircon and K–Ar on biotite for the Wami River granulitecomplex (Maboko et al., 1985). (b): 2, Intrusion age of theanorthosite re-interpreted from U–Pb on zircon of Muhongoand Lenoir (1994); 3, U–Pb on zircon and monazite (thisstudy); 4, Ar–Ar and K–Ar on hornblende, biotite, muscoviteand K-feldspar for the NW-Uluguru Mountains and sur-rounding migmatite gneisses (Maboko et al., 1989); 5, Sm–Ndgamet-whole rock isochrons (Maboko and Nakamura, 1995).Tentative cooling path is indicated by dashed line. (c): 6,U–Pb on zircon (Coolen et al., 1982); 7, K–Ar on hornblende(Andriessen et al., 1985); 8, K–Ar on biotite and muscoviteand Rb–Sr on biotite and muscovite for the Furua granulitecomplex and surrounding migmatite gneisses (Priem et al.,1979). Error bars are given for the assumed uncertainties inthe closure temperatures of the minerals (925–930°C).Width of the symbols corresponds to approximately 5 Mawhich is larger than the analytical error for most of the U–Pbdata. Light areas indicate the maximum range of the inte-grated cooling paths (a) and (c) or an unlikely fast alternativecooling trajectory (b).


Tab

le3

Sum

mar

yof

publ

ishe

dge

ochr

onol

ogic

alda

tafo

rP

an-A

fric

angr

anul

ites

inE

-Tan

zani

aa

Are

aR

efer

ence

Roc

kty

peM

etho

dM

iner

alA

ge(M

a)R

emar

ks,

re-i

nter

pret

atio

n

Muh

ongo

and

Len

oir

u.i.,

met

amor

phic

6459

10Z

rnS

-Par

eM

ount

ains

U–P

bE

nder

bite

(199

4)50

39

20b

Cah

enan

dSn

ellin

gP

egm

atit

eK

–Ar

Bt

(196

6),

p.27

5909

25b

Exc

ess

Ar

K–A

rG

ranu

litic

gnei

ssB

t58

59

20b

Gra

nulit

icgn

eiss

Exc

ess

Ar

Cah

enan

dSn

ellin

gK

–Ar

Hbl

20km

Wof

N-P

are

(196

6),

p.27

Mou

ntai

nsC

ahen

and

Snel

ling

Hbl

–Scp

-Bt

gnei

ss20

kmN

ofU

sam

bara

K–A

rB

t48

89

20M

ount

ains

(196

6),

p.27

u.i.,

intr

usio

nsM

abok

oet

al.

(198

5)88

2–16

00,

701–

715

Zrn

U–P

bO

rtho

-gra

nulit

esW

ami

Riv

erO

rtho

-gra

nulit

esu.

i.,P

b-lo

ss,

met

amor

phic

Mab

oko

etal

.(1

985)

U-P

bZ

rn47

6–53

8,62

0–

642

Ort

ho-g

ranu

lites

Fiv

ere

sult

sM

abok

oet

al.

(198

5)R

b–Sr

Bt

458–

485

Mab

oko

etal

.(1

985)

719

8A

pF

issi

ontr

ack

Ort

ho-g

ranu

lites

25669

9,479

9l.i

.,l.i

.M

abok

oet

al.

(198

5)P

ara-

gnei

ss20

kmN

ofW

ami

Riv

erU

–Pb

Zrn

Muh

ongo

and

Len

oir

6959

4u.

i.,in

trus

ion

Ulu

guru

Mou

ntai

nsZ

rnA

nort

hosi

teU

–Pb

(199

4)63

39

7,61

89

16A

nort

hosi

te,

Tw

ore

sult

sM

abok

oan

dN

akam

ura

Sm–N

dG

rt-w

r(1

995)

Ort

ho-g

ranu

lite

5369

3,55

09

10,

Thr

eere

sult

sM

uhon

go(1

990)

Rb–

SrB

tO

rtho

-gra

nulit

es57

79

30(K

–Ar)

7029

14M

abok

oet

al.

(198

9)A

r–A

rH

bl62

89

3E

xces

sA

r(A

r–A

r)M

abok

oet

al.

(198

9)K

–Ar

Bt

9789

20,

15379

3356

09

11,

5299

10–,

–,�

580,

–,no

plat

eau

4879

2,49

59

2(K

–Ar)

4889

10,

5009

10M

abok

oet

al.

(198

9)A

r–A

rM

sM

abok

oet

al.

(198

9)(K

–Ar)

4349

9,45

29

9A

r–A

r42

29

2,45

09

2K

fsA

ppro

x.ag

esat

115,

80an

dO

rtho

-gra

nulit

eF

issi

ontr

ack

Nob

leet

al.

(199

7)A

p30

0,80

,30

20°C

Coo

len

etal

.(1

982)

Fur

uaco

mpl

exG

rt-2

Px

gran

ulit

eU

–Pb

Zrn

6529

10l.i

.,m

etam

orph

icP

riem

etal

.(1

979)

Bt

Fou

rtee

nR

esul

tsA

v.46

3(3

88–5

11)

Rb–

SrF

urua

com

plex

Ort

ho-g

ranu

lites

Tw

ore

sult

sP

riem

etal

.(1

979)

Rb–

SrM

s52

49

10,

5319

12S

urro

unds

414–

560

Fou

rtee

nre

sult

sP

riem

etal

.(1

979)

K–A

rB

tP

riem

etal

.(1

979)

K–A

rM

s48

29

14,

4839

14,

Thr

eere

sult

s48

79

15A

ndri

esse

net

al.

(198

5)K

–Ar

Hbl

614–

665

Six

resu

lts

aN

ote:

biot

ite

and

horn

blen

deK

–Ar

data

ofA

ndri

esse

net

al.

(198

5)an

dP

riem

etal

.(1

979)

,an

dbi

otit

eR

b–Sr

(Mab

oko

etal

.,19

85)

are

too

num

erou

sto

list

indi

vidu

ally

and

only

the

age

span

ofal

lana

lyse

sis

give

n,re

fer

toth

eor

igin

alpa

pers

for

deta

ils;u

.i.is

the

uppe

rin

terc

ept

ofU

–Pb

disc

ordi

a,l.i

.is

the

low

erin

terc

ept.

bR

ecal

cula

ted

usin

gth

eva

lues

reco

mm

ende

dby

Stei

ger

and

Jage

r(1

977)

for

t/2

ofK

.


cooling rate between 2 and 4.5°C/Ma (Fig. 9a) forgranulites which have experienced similar P–Tconditions to the Pare, Usambara and UluguruMountains (Appel et al., 1998). Maboko et al.(1989) used K–Ar as well as the 40Ar–39Ar step-heating technique on hornblende, biotite, muscov-ite and K-feldspar from samples collected at theNW edge of the Uluguru Mountains, close to thenorthern edge of the anorthosite complex and tolocation T46 of this study (Fig. 2). Interpretationof the data of Maboko et al. (1989) is complicatedby the fact that only one of the samples in theirstudy (yielding the hornblende analysis and one ofthe biotite results) is from the granulite complexitself, whereas the other three samples (biotite,muscovite and K-feldspar) were collected fromfelsic gneisses and migmatites surrounding thegranulite complex. Growth of muscovite in theserocks may be retrograde (related to rehydrationreactions) in rocks which re-equilibrated to lowerT than the granulite complex itself. Therefore, themuscovite ages may have to be regarded as mini-mum ages. Despite their higher closure tempera-ture, K–Ar muscovite ages are without exceptionsignificantly younger than K–Ar biotite ages(Maboko et al., 1989). High K–Ar and Ar–Arages on hornblende and biotite from the granulitesample and biotite K–Ar and Ar–Ar results ofmigmatite samples are interpreted to reflect excessAr or were already interpreted in this way byMaboko et al. (1989). The hornblende Ar–Ar ageis identical to the zircon and monazite ages (T46,this study) as well as to two garnet Sm–Ndisochrons ages for granulites from the area(Maboko and Nakamura, 1995). Because instan-taneous cooling by about 350°C down to 450°C(shaded area in Fig. 9b) is considered very un-likely for these deep crustal (ca. 10 kb) granuliteswhich show no petrological evidence for decom-pression, this result is also interpreted to reflectexcess Ar. A tentative cooling path of about3°C/Ma is indicated by a dashed line (Fig. 9b).The garnet Sm–Nd isochron age of 618916 Maon texturally late, undeformed garnet coronas inthe anorthosite (Maboko and Nakamura, 1995) isimportant evidence in support of a single gran-ulite facies event, because it indicates that garnetgrowth during retrograde isobaric cooling oc-

curred just after or contemporaneously with zir-con and monazite growth in meta-qtz-diorite T46.

In the Furua complex, the combination of alower U–Pb intercept of zircon at 652910 Ma(Archaean upper intercept; Coolen et al. 1982)and Rb–Sr biotite and muscovite ages (Priem etal., 1979) yields a prolonged history of slow inte-grated cooling at a rate of about 2.5°C/Ma for thegranulite-facies rocks, and the surroundingmigmatitic gneisses (Fig. 9c). The results of K–Aron hornblende and biotite (Andriessen et al.,1985; Priem et al., 1979) are again interpreted asunreliable possibly because of the presence ofexcess Ar. The reconstruction of the cooling his-tory is thus based on Rb–Sr biotite and muscov-ite ages and K–Ar results on muscovite.

The geochronological data available prior tothis study and their interpretation did not providea conclusive picture of the precise age of Pan-African metamorphism in Tanzania, its spatialdistribution and the post-metamorphic coolinghistory. One of the problems appears to be theubiquitous presence of excess Ar in hornblendeand biotite.

6.2. Cooling histories of the granulites

Interpretation of the published U–Pb data onzircon for the Wami River complex, the Furuacomplex and data from the Pare Mountains(Muhongo and Lenoir, 1994), in combinationwith the new mineral data presented here, sug-gests that the granulite-facies event reached itspeak between 620 Ma and about 640–650 Ma.The range of ages from published U–Pb studieson zircon is consistent with the results of thisstudy. The data allow the tentative reconstructionof cooling histories for the granulites of the WamiRiver complex and the Furua Complex and possi-bly the NW Uluguru Mountains (Fig. 9), indicat-ing similar prolonged slow cooling with integratedcooling rates of 2–5°C/Ma over time intervals of140–240 Ma.

Different age domains can be distinguishedwithin the granulites of eastern Tanzania. Themonazite age of 640 Ma in the Pare Mountains isinterpreted as reflecting the time of peak meta-morphic conditions at about 800°C (Appel et al.,


1998), consistent with a U–Pb zircon upper inter-cept age of 645910 Ma on a granulite faciesgneiss from the Pare Mountains (Muhongo andLenoir, 1994). Monazite from the UsambaraMountains yields ages of 625–630 Ma which areabout 15 Ma younger than the time of peakmetamoiphism in the Pare Mountains. The mon-azite and titanite ages of 617–620 Ma from theUmba Steppe semipelites and marbles are inte-grated to date the peak of metamorphism, esti-mated at 750–800°C by Moller (1995) andsignificantly younger than the time of metamor-phism determined for the Pare and the UsambaraMountains.

Within the northern and eastern UluguruMountains monazite ages span an age range of atleast 11 m.y., possibly 23 m.y. It is unclearwhether the different ages reflect true differencesin the time of peak metamorphism or whetherthey can, in part, also be attributed to inheritanceor preservation of prograde growth, the latterbeing the preferred interpretation. The resultsmay be seen as the age envelope in which to placegranulite facies metamorphism in this part of theUluguru Mountains. More obvious diachronismof Pan-African metamorphism is indicated by the20 m.y. younger monazite and zircon ages onmeta-qtz-diorite sample T46 in the NW UluguruMountains than of monazite in the northern andeastern Uluguru Mountains (Fig. 10b). The mini-mum age difference is 20 m.y. Congruence of themonazite and zircon age of T46 (where the zirconfraction is interpreted as the product of crystalli-sation of partial melt during the earliest stages ofcooling from peak of metamorphism) is taken asstrong evidence for peak metamorphism at thistime (and a high Tc of monazite). The upperzircon intercept of 69594 Ma from theanorthosite (Muhongo and Lenoir, 1994) may beinterpreted as the age of crystallisation of theanorthosite body (Fig. 9b and Fig. 10b). It can beconcluded that there appear to be differences inthe timing of high grade metamorphism betweendifferent granulite complexes in the Pan-AfricanMozambique Belt of Tanzania on a relativelysmall scale of less than 100 km.

Using a Tc of 650°C for titanite results in aninitial cooling rate of about 5°C/Ma for the Pareand the Uluguru Mountains. Diachronism in thethermal history within the granulites of north-eastern Tanzania is supported by rutile (and somebiotite) cooling ages which show the same patternof regional age distribution as the monazite inmany areas (Fig. 2 and Fig. 10). Rutile ages areabout 100 Ma younger than monazite ages, indi-cating an integrated cooling rate of about 4°C/Ma. This supports the interpretation that the agedifference observed in the monazite results has ageological cause associated with Pan-Africanmetamorphism (Fig. 10a) and the subsequent un-roofing history. The Umba Steppe and the Usam-bara Mountains underwent a cooling history

Fig. 10. Cooling paths reconstructed for the different granulitecomplexes studied in the Pan-African part of the TanzanianMozambique Belt. (a) Pare and Usambara Mountains andUmba Steppe: 1, U–Pb age on zircon from Muhongo andLenoir (1994); 2, K–Ar ages of biotite from Cahen andSnelling (1966). (b) Uluguru Mountains; 3, Rb–Sr ages onbiotite from Muhongo (1990); 4, age of anorthosite re-inter-preted from U–Pb on zircon, (Muhongo and Lenoir, 1994); 5,Sm–Nd garnet-whole rock isochrons on anorthosite and or-thogneiss (Maboko and Nakamura, 1995). Dashed path forNW-Uluguru from Fig. 9.


Muhongo and Lenoir, 1994) and thus provide areliable estimate for the time of the Pan-Africanmetamorphism in the Mozambique Belt of Tanza-nia. The observed cooling paths are consistentwith the anti-clockwise isobaric cooling (ACW-IBC) P–T path deduced from petrological obser-vations and thermobarometry (Appel et al., 1998).

The cooling histories of the granulite terranes,are parallel but offset (Fig. 11). supporting thenotion that there are real age differences betweenmountain ranges (Pare Mountains and UsambaraMountains and Umba Steppe) and within a singlemountain range (Uluguru Mountains). These dif-ferences may be taken as evidence that these agedomains are separated by important but hiddenfaults, or may be explained by variations in thelocation of an external heat source. The progres-sive slowing of cooling rates is consistent with thehypothesis that a granulite-facies event is at leastin part driven by an external heat source, e.g. theintrusion of magmas into the lower crust (Anovitzand Chase, 1990; Oxburgh, 1990) and slow upliftrates.

7.2. Tectonic scenario for granulite metamorphismin eastern Tanzania

Previous geochronological results for the Tan-zanian granulites were interpreted to indicate fastuplift following tectonic crustal thickening afterplate-collision during the formation of Gondwana(e.g. Maboko et al., 1985, 1989; Muhongo andLenoir, 1994; Maboko and Nakamura, 1995). Acollision process would cause rapid decompres-sion after or during the thermal peak of metamor-phism, leaving behind decompression textures inthe rock record. Such rapid decompression isinvariably associated with fast cooling rates (Eng-land and Thompson, 1984; Bohlen, 1991) of \20°C/Ma as shown by data from the Alps andHimalayas (e.g. von Blanckenburg et al., 1989)and contrast with the slow integrated coolingrates in the Tanzanian granulites. It is concludedthat a continent collision scenario is not compat-ible with mineral texture (Coolen, 1980; Moller,1995; Appel et al., 1998), fluid inclusion (Hermsand Schenk, 1998) and geochronological evidence(this study).

Fig. 11. Compilation of the integrated cooling paths of gran-ulite complexes from the Pan-African Belt of NE Tanzania:Pare and Usambara Mountains and Umba Steppe (darkpaths) and different parts of the Uluguru Mountains (lightpaths). The range of cooling paths in the studied granuliteareas is indicated by the light shaded area. Paths constructedfrom previous geochronological data for the Wami River (1:Maboko et al., 1985) and Furua complex granulites (2: Coolenet al., 1982; Andriessen et al., 1985; Priem et al., 1979) shownfor comparison (black paths) fall on the same swath of coolingpaths.

similar but time-parallel to that of the PareMountains, with initial cooling rates of about5°C/Ma.

7. Conclusions

7.1. The P–T– t e6olution of Pan-African highpressure granulite terranes in Eastern Tanzania

The ages of the high pressure granulite peak-metamorphism in eastern Tanzania can be con-strained to an interval from 655 to 615 Ma andvaries significantly in the five different areas (thePare Mountains, the Usambara Mountains, theUmba Steppe, the NW Uluguru Mountains andthe N and E Uluguru Mountains). The ages ob-tained from monazite and zircon in this study arein the lower range of previously published U–Pbzircon ages for the Wami River Complex(Maboko et al., 1985) or coincide closely withzircon ages obtained for the Pare Mountains andthe Furua Complex (Coolen et al., 1982;


Slow cooling and all other evidence can be bestreconciled with a scenario in which Pan-Africangranulite facies metamorphism in eastern Tanza-nia was caused by underplating and intrusion ofmagmas into the crust, causing heating and burial(e.g. Oxburgh, 1990). Cessation of magmatic ac-tivity initiated cooling at rates of about 5°C/Mawhich slowed progressively. Post-orogenic col-lapse was either slow or evidence for it could notbe detected by integration of the presentgeochronologic data. An active continental mar-gin setting is a likely tectonic setting consistentwith the chemistry of the meta-igneous granulites(Appel, 1996).

The combined geochronologic evidence fromthe granulite terranes provides strong evidencethat the final collision of East- and West-Gond-wana has not directly caused granulite-faciesmetamorphism in the Pan-African domains ofTanzania. Crustal shortening and uplift of thegranulites due to continent collision is youngerthan the peak-metamorphic ages determined inthis study. The available data suggest that thegranulites of eastern Tanzania had already cooledto 450°C or below at about 550 Ma. This inter-pretation requires refinement of recent models forthe geodynamic evolution of the MozambiqueBelt of East Africa and the circum-Indic Pan-African orogen (e.g. Stern, 1994) on the whole.

7.3. Age distribution of Pan-African high gradeterranes in central Gondwana: implications forplate tectonic scenarios?

It is well accepted that the Pan-African Belt inCentral Gondwana is the site of a suture from thecollision of East- and West-Gondwana (e.g. Burkeet al., 1977). However, there are currently twodifferent plate-tectonic interpretations for the for-mation of granulites in the Pan-African Belt ofTanzania (and East Africa on the whole).

The first model proposes the collision of East-and West-Gondwana causing high-grade meta-morphism in East Africa before 600 Ma and issimilar to the collision and ocean-closure modelof Stern (1994). The age of metamorphism incentral Gondwana has been used repeatedly toconstrain the age of collision, based on the as-

sumption that metamorphism was caused directlyby the collision event. Kroner (1993), for example,concluded that the granulite-facies metamorphicevents in eastern and north-eastern Africa couldbe explained by ocean closure, terrane accretionand continental collision of East- and West-Gondwana during the time-span between 600 and800 Ma, consistent with the scenario of theSWEAT-hypothesis discussed by Dalziel (1991)and Moores (1991).

The second model, put forward by Meert et al.(1995) based on palaeomagnetic evidence andAr–Ar dating of Neoproterozoic lavas from Tan-zania, argues for polyphase accretion of terranesto form the Gondwana supercontinent, becausethey observed no common polar-wander-path forEast- and West-Gondwana before 550 Ma. Theyproposed specifically that the formation of Gond-wana took place in two distinct orogenic events.The Pan-African orogen is considered to representthe earlier collision of India, Madagascar, SriLanka, part of Antarctica and the Kalahari cra-ton with the Congo craton and the Arabian–Nu-bian Shield between 800 and 650 Ma, duringclosure of the northern part of the Mozambiqueocean. The second event occurred further to thesouth at 550 Ma and resulted in collision andamalgamation of the earlier ‘Pan-African’ orogenwith the rest of Antarctica and Australia to formGondwana at 550 Ma.

Neither of the two variations on the collisionmodel for the Pan-African Belt can be fully sup-ported by the present study as it leads to theconclusion that high-grade metamorphism priorto 600 Ma did not require a concurrent continentcollision. The collision of East- and West-Gond-wana did not cause the formation of granulites ineastern Tanzania and, therefore, the age of high-grade metamorphism cannot be used directly todate the collision event, an approach that hasbeen used in all previous interpretations.

The palaeogeographic reconstruction in Fig. 12summarises available geochronological data inter-preted as ages of metamorphism in the fragmentsof the central Gondwana collision zone. Peakgranulite-facies metamorphism is estimated tohave occurred after 600 Ma in the Buur area ofSomalia (Lenoir et al., 1994), southern and central


Madagascar (Andriamarofahatra et al., 1990;Guerrot et al., 1993; Nedelec et al., 1994; Paquetteet al., 1994; Tucker et al., 1999), South India(Buhl et al., 1983; Choudhary et al., 1992; Un-nikrishnan-Warrier et al., 1995) and also in theLutzow-Holm Complex of Antarctica (Shiraishi etal., 1992; 1994). Granulite-facies metamorphismin Sri Lanka appears to have a complex historywith evidence for a metamorphic event at 610 Ma,but magmatic activity lasting until 550 Ma (Bauret al., 1991; Holzl et al., 1994). Except for the

results from Somalia, these ages are consistentwith the second of two collision phases in themodel of Meert et al. (1995). Only complete P–T– t histories of these terranes will help to deter-mine whether collision was indeed the cause ofmetamorphism.

Pre-600 Ma granulites occur in eastern Tanza-nia (this study, Coolen et al. (1982), Maboko etal. (1985), Muhongo and Lenoir (1994)), inter-preted by Appel et al. (1998) to follow ACW-IBC-paths. Other ACW-IBC-granulites occur inAntarctica (Shiraishi and Kagami, 1992). Pre-600Ma granulites are also exposed in southernEthiopia (Teklay et al., 1998) and central Kenya(Key et al., 1989), but more data are needed inthese two areas to substantiate the existing resultsas ages of metamorphism. There are differentgroups of ages for the Pan-African metamorphicevent in different areas of the circum-Indic region,but no simple East–West pattern as proposed bythe two-phase collision model of Meert et al.(1995) is evident (Fig. 12). More P–T– t informa-tion on other granulite terranes is needed to estab-lish whether the apparent diachronism of pre- andpost-600 Ma metamorphic ages can be explainedbest by diachronism in the collision of East- andWest-Gondwana similar to the model of Meert etal. (1995), or whether it is the result of complexcontinental margin processes as diverse as thosefound in modern orogens.

Acknowledgements

This paper is a contribution to IGCP 348. A.M.thanks the colleagues and staff of the Max-Planck-Institut fur Chemie in Mainz for theirkind assistance and helpful discussions. We like tothank P. Appel for his close collaboration in thisproject and numerous fruitful discussions. Weacknowledge the detailed reviews of F. Corfu, A.Lanzirotti and an anonymous reviewer whichhelped to improve the manuscript significantly.Research permits from the Tanzanian Commis-sion for Science and Technology and supportfrom the University of Dar-es-Salaam (foremostS. Muhongo) and the Geological Survey of Tan-zania are also gratefully acknowledged. Research

Fig. 12. Palaeogeographic map of part of central Gondwana(modified after Windley et al. (1994), adapted from Kriegsman(1995)) with estimates for the age of Pan-African granulite-fa-cies metamorphism. Geochronological data from (1) Holzl etal. (1994): U–Pb Zrn+Mnz, (2) Baur et al. (1991): U–Pb Zrn(SHRIMP), (3) Shiraishi et al. (1992): U–Pb Zrn (SHRIMP),(4) Shiraishi et al. (1994): U–Pb Zrn (SHRIMP), (5) Choud-hary et al. (1992): Grt–Sm–Nd isochron, (6) Buhl et al.(1983): U–Pb Mnz+Zrn, (7) Unnikrishnan-Warrier et al.(1995): Grt–Sm–Nd isochron, (8) Lenoir et al. (1994): U–PbZrn, (9) Key et al. (1989): Sm–Nd+Rb–Sr isochron, (10) thisstudy: U–Pb Mnz+Zrn, (11) Muhongo and Lenoir (1994):U–Pb Zrn, (12) Maboko et al. (1985): U–Pb Zrn, (13) Coolenet al. (1982): U–Pb Zrn, (14) Paquette et al. (1994): U–Pb+Pb–Pb, Zrn, (15) Andriamarofahatra et al. (1990): U–PbZrn+Mnz, (16) Tucker et al. (1999): U–Pb Zrn, (17) Guerrotet al. (1993): U–Pb Zrn, (18) Nedelec et al. (1994): Pb–PbZrn, (19) Teklay et al. (1998): Pb–Pb Zrn, U–Pb Zrn, (20)Theunissen et al. (1992): U–Pb Zrn, (21) Shiraishi andKagami (1992): Grt–Sm–Nd isochron, (22) eclogite occur-rence (Nicollet, 1989).


Tab

le4

Min

eral

asse

mbl

ages

and

sam

ple

loca

tion

desc

ript

ions

Kfs

Bt

Qtz

Ms

Ky

Sil

Rt

Gr

Spl

Ore

Sam

ple

Sec.

Ky

Are

aSe

c.Si

lIn

cl.

inG

rtO

ther

Loc

atio

nR

ock-

type

Grt

P1

Met

apel

ites

Grt

P1

Kfs

Bt

Qtz

A01

6E

aste

rnsi

deof

N-P

are

Mou

ntai

ns,

1st

smal

lSi

lR

tSp

lO

reSi

lN

-Par

eM

etap

elit

eM

ouni

ains

brid

geso

uth

ofB

utu-

Riv

erM

etap

elit

eG

rtP

1K

fsB

tQ

tzSi

lR

tG

rO

reSi

lS-

Par

eF

oot

ofw

este

rnsl

opes

,ro

adou

tcro

pat

road

T11

5to

seco

ndar

ysc

hool

inH

edar

uM

ount

ains

Usa

mba

raK

fsB

tQ

tzK

ySi

lR

tT

137

Met

apel

ite

Spl

Ore

Sil

Roa

dou

tcro

p,M

ombo

-Lus

hoto

Roa

d,1.

5P

1G

rtkm

NE

ofM

ombo

toll

stat

ion

Mou

ntai

nsS-

Usa

mba

raM

etap

elit

eG

rtP

1K

fsB

tQ

tzK

ySi

lR

tG

rO

reK

yA

108

Roa

dcut

+bl

ocks

onsl

ope

belo

wro

ad,

21.8

Mou

ntai

nskm

Sof

Mah

anga

Mla

i,he

adin

gto

Kor

ogw

eG

rtP

1K

fsB

tQ

tzsM

sR

tG

rt-B

tgn

eiss

A14

4O

reO

utcr

opS

side

Um

bari

ver,

alon

gU

mba

Step

peU

mba

-Mw

akije

mbe

trac

k,11

.2km

Eof

Um

baru

bym

ine

Grt

P1

Bt

Qtz

Sil

Rt

P00

1R

iver

outc

rop,

100

mbe

low

Mor

ning

side

,ca

.sK

yN

Ulu

guru

Met

apel

ite

10km

sout

hof

Mor

ogor

oto

wn

cent

erM

ount

ains

Met

apel

ite

Grt

P1

Kfs

Bt

Qtz

Ky

Sil

Cre

ekou

tcro

p,1

kmN

NE

alon

gtr

ack

from

Rt

P00

9G

rsS

ilK

yN

EU

lugu

ruG

ulio

niD

ispe

nsar

yat

Mor

ogor

o-M

atom

boM

ount

ains

road

,T

opo

shee

t18

3-4

Grt

P1

Kfs

Bt

Qtz

Ms

Ky

Rt

Gr

Met

apel

ite

TO

28Sc

pSE

Ulu

guru

Roa

dou

tcro

p,8.

2km

.S

ofM

tom

bozi

,M

ount

ains

Shep

pard

spa

ss

Bt

Qtz

Cpx

Opx

Hbl

P1

Rt

Scp

Tit

Ore

Oth

erO

rtho

gnei

sses

,ca

lcsi

licat

egn

eiss

esan

dm

arbl

esK

fsG

rt

Grt

P1

Kfs

Qtz

Cpx

Hbl

Cal

csili

cate

Roa

dcut

onol

dK

wa

Kih

indi

-Usa

ngi

road

,Sc

pA

026

Tit

Ore

SEN

-Par

eM

ount

ains

800

mfr

omK

wa

Kih

indi

villa

geM

etab

asit

eG

rtP

1Q

tzC

pxT

itT

114

Wes

tern

slop

e,cr

eek

outc

rop,

Nof

Saw

eni

S-P

are

Mou

ntai

nsO

utcr

opS

side

Um

bari

ver,

alon

gQ

tzC

pxH

blU

mba

Step

peSc

pT

itO

reC

cA

141

Cal

csili

cate

Um

ba-M

wak

ijem

betr

ack,

Sof

Ger

evi

hills

P1

Kfs

Bt

Qtz

Hbl

Cha

rnoc

kite

Qua

rry

atcr

ossr

oads

inL

ukos

i,ro

adto

T12

1O

reU

sam

bara

Mou

ntai

nsSh

ume

Met

a-qt

z-di

orit

e‘A

nort

hosi

te’

onQ

DS,

old

Mom

bo-L

usho

toP

1K

fsB

tQ

tzC

pxO

pxH

blU

sam

bara

Rt

T13

9O

reM

ount

ains

road

,10

0m

Eof

first

pass

from

sisa

lpl

anta

tion

inM

ombo

T02

5R

oad

outc

rop

Mat

ombo

-Mvu

haro

ad,

2.3

kmT

itC

cSE

Ulu

guru

Mar

ble

Mou

ntai

nsso

uth

ofM

torn

bozi

villa

geC

alcs

ilica

teG

rtP

1B

tC

pxP

008

Hbl

Cre

ekou

tcro

p,1

kmN

NE

ofN

EU

lugu

ruT

itM

orog

oro-

Mat

ombo

road

,at

Mad

abal

aM

ount

ains

scho

ol,

Top

osh

eet

183/

4K

fsQ

tzC

pxS

Ulu

guru

Hbl

P08

8Sc

pT

itC

reek

outc

rop,

100

mE

ofro

adto

Lub

asaz

i,C

alcs

ilica

teca

.15

kmN

from

Mvu

ha-D

utur

niro

ad,

Mou

ntai

nsT

opo

shee

t20

1/4

WU

lugu

ruG

rtP

1B

tQ

tzR

iver

outc

rop

Wof

road

Mor

ogor

o-M

geta

,H

blM

eta-

qtz-

dior

ite

T04

6Z

oiM

ount

ains

5.6

kmN

ofM

geta


was financially supported by the DeutscheForschungsgerneinschaft (DFG) through grantsSche 265-2/5 and Sche 265-6/1.

Appendix

See Table 4.

References

Andriamarofahatra, J., de la Boisse, H., Nicollet, C., 1990.Datation U–Pb sur monazites et zircons du dernierepisode tectono-metamorphique granulitique majeur dansle Sud-Est de Madagascar. C.R. Acad. Sci. Paris 310 (2),1643–1648.

Andriessen, P.A.M., Coolen, J.J.M.M.M., Hebeda, E.H.,1985. K–Ar hornblende dating of late panafrican meta-morphism in the Furua granulite complex of southernTanzania. Precambrian Res. 30, 351–360.

Anovitz, L.M., Chase, G.M., 1990. Implications of post-thrusting extension and underplating for P–T– t paths ingranulite terranes: a Grenville example. Geology 18, 466–469.

Appel, P., 1996. Hochdruckgranulite und Eklogite im Mozam-bique Belt von Tansania: eine petrologische und geo-chemische Studie. Dissertation, Christian-Albrechts-Universitat Kiel, Germany.

Appel, P., Moller, A., Schenk, V., 1998. High-pressure gran-ulite facies metamorphism in the Pan-African Belt of east-ern Tanzania: P–T– t evidence against granulite formationby continent collision. J. Metamorph. Geol. 16, 491–509.

Bagnall, P.S., 1963. The geology of the North Pare mountains.Rec. Geol. Surv. Tangan. 10, 7–16.

Bagnall, P.S., Dundas, D.L., Hartley, E.W., 1963. The geologyof the Mnazi and Lushoto areas: quarter degree sheets 90and 109, and part of 73 E. Rec. Geol. Surv. Tangan. 11,15–17.

Baur, N., Kroner, A., Liew, T.C., Todt, W., Williams, I.S.,Hofmann, A.W., 1991. U–Pb isotopic systematics of zir-cons from prograde and retrograde transition zones inhigh-grade orthogneisses. Sri Lanka J. Geol. 99, 527–545.

Bingen, B., van Breemen, O., 1998. U–Pb monazite ages inamphibolite- to granulite-facies orthogneiss reflect hydrousmineral breakdown reactions: Sveoconorwegian Provinceof SW Norway. Contrib. Mineral. Petrol. 132, 336–353.

von Blanckenburg, F., Villa, I.M., Baur, H., Morteani, G.,Steiger, R.H., 1989. Time calibration of a PT path fromthe western Tauern Window, eastern Alps: the problem ofclosure temperatures. Contrib. Mineral. Petrol. 101, 1–11.

Bohlen, S.R., 1991. On the formation of granulites. J. Meta-morph. Geol. 9, 223–229.

Buhl, D., Grauert, B., Raith, M., 1983. U–Pb zircon dating ofArchean rocks from the South Indian Craton: results from

the amphibolite to granulite facies transition zone at Kab-bal Quarry, southern Karnataka. Fortschr. Mineral. 61,43–45.

Burke, K., Dewey, J.F., Kidd, W.S.F., 1977. World distribu-tion of sutures — the sites of former oceans. Tec-tonophysics 40, 69–99.

Cahen, L., Snelling, N.J., 1966. The Geochronology of Equa-torial Africa. North-Holland, Amsterdam, pp. 1–195.

Cameron, A.E., Smith, D.H., Walker, R.L., 1969. Mass spec-trometry of nanogram-size samples of lead. Nature 41,525–526.

Cherniak, D.J., 1993. Lead diffusion in titanite and prelimi-nary results on the effects of radiation damage on Pbtransport. Chem. Geol. 110, 177–194.

Choudhary, A.K., Harris, N.B.W., Van Calsteren, P.W.C.,Hawkesworth, C.J., 1992. Pan-African charnockite forma-tion in Kerala, South India. Geol. Mag. 129, 257–264.

Coolen, J.J.M.M.M., 1980. Chemical petrology of the Furuagranulite complex, southern Tanzania. GUA Pap. Geol.Ser. 1 (13), 1–258.

Coolen, J.J.M.M.M., Priern, H.N.A., Verdurmen, E.A.T.,Verschure, R.H., 1982. Possible zircon U–Pb evidence forPan-African granulite-facies metamorphism in the Mozam-bique Belt of southern Tanzania. Precambrian Res. 17,31–40.

Corfu, F., 1996. Multistage zircon and titanite growth andinheritance in an Archean gneiss complex, Winnipeg RiverSubprovince, Ontario. Earth Planet. Sci. Lett. 141, 175–186.

Dallmeyer, R.D., Hermes, O.D., Ibarguchi, G.I., 1990. 40Ar/39Ar mineral ages from the Sictuate metagranite, RhodeIsland: implications for the chronology of late Paleozoicterrane accretion and resultant tectonothermal activity inNew England. J. Metamorph. Geol. 8, 145–157.

Dalziel, I.W.D., 1991. Pacific margins of Laurentia and EastAntarctica-Australia as a conjugate rift pair: evidence andimplications for an Eocambrian supercontinent. Geology19, 598–601.

Dodson, M.H., 1979. Theory of cooling ages. In: Jager, E.,Hunziker, J.C. (Eds.), Lectures in Isotope Geology.Springer, New York, pp. 194–202.

England, P.C., Thompson, A.B., 1984. Pressure–temperature–time paths of regional metamorphism I. Heat transferduring the evolution of thickened continental crust. J.Petrol. 25, 894–928.

Gabert, G., Wendt, I., 1974. Datierung von granitischenGesteinen im Dodoman- und Usagaran System und in derNdembera-Serie (Tanzania). Geol. Jahrbuch B11, 3–55.

Gromet, P.L., 1991. Direct dating of deformational fabrics. In:Heaman, L., Ludden, J.N. (Eds.), Applications of Radio-genic Isotope Systems to Problems in Geology. Mineralog-ical Association of Canada, Toronto, pp. 167–189.

Guerrot, C., Cocherie, A., Ohnenstetter, M., 1993. Origin andevolution of the West Andriamena Pan-African mafic-ul-tramafic complexes in Madagascar as shown by U–Pb, Ndisotopes and trace element constraints. Terra Abstr. EUG7, 387.


Haggart, M.J., Jamieson, R.A., Reynolds, P.H., Krogh, T.E.,Beaumont, C., Culshaw, N.G., 1992. Last gasp of theGrenville orogeny: thermochronology of the GrenvilleFront Tectonic Zone near Killarney, Ontario. J. Meta-morph. Geol. 101, 575–589.

Hanson, G.N., Gast, P.W., 1967. Kinetic studies in contactmetamorphic zones. Geochim. Cosmochim. Acta 31,1119–1153.

Harrison, T.M., 1981. Diffusion of 40Ar in hornblende. Contr.Mineral Petrol. 78, 324–331.

Harrison, T.M., McDougall, I., 1982. The thermal significanceof potassium feldspar K–Ar ages inferred from 40Ar/39Arspectrum results. Geochim. Cosmochim. Acta 46, 1811–1820.

Harrison, T.M., Duncan, I., McDougall, I., 1985. Diffusion of40Ar in biotite: temperature, pressure and compositionaleffects. Geochim. Cosmochim. Acta 49, 2461–2468.

Herms, P., Schenk, V., 1998. Fluid inclusions in high-pressuregranulites of the Pan-African belt in Tanzania (UluguruMts): a record of prograde to retrograde fluid evolution.Contrib. Mineral. Petrol. 130, 199–212.

Hoffman, P.F., 1991. Did the breakout of Laurentia turnGondwanaland inside-out? Science 252, 1409–1412.

Holmes, A., 1951. The sequence of Pre-Cambrian orogenicbelts in south and central Africa. In: Sandford, K.S.,Blondel, F. (Eds.), 18th International Geological Congress.Association of African Geological Surveys, London, pp.254–269.

Holzl, S., Hofmann, A.W., Todt, W., Kohler, H., 1994. U–Pbgeochronology of the Sri Lankan basement. PrecambrianRes. 66, 123–149.

Key, R.M., Charsley, T.J., Hackman, B.D., Wilkinson, A.F.,Rundle, C.C., 1989. Superimposed Upper Proterozoic col-lision-controlled orogenies in the Mozambique OrogenicBelt of Kenya. Precambrian Res. 44, 197–225.

Kriegsman, L.M., 1995. The Pan-African event in EastAntarctica: a view from Sri Lanka and the MozambiqueBelt. Precambrian Res. 75, 263–277.

KRISP Working Party, 1995. Group takes a fresh look at theLithosphere underneath southern Kenya. EOS Trans. Am.Geophys. Union 76, 73–82.

Krogh, T.E., 1973. A low-contamination method for hy-drothermal decomposition of zircon and extraction of Uand Pb for isotopic age determinations. Geochim. Cos-mochim. Acta 37, 485494.

Kroner, A., 1993. The Pan African Belt of Northeastern andEastern Africa, Madagascar, Southern India, Sri Lankaand East Antarctica: terrane amalgamation during forma-tion of the Gondwana supercontinent. In: Thorweihe, U.,Schandelmeier, H. (Eds.), Geoscientific Research in NEAfrica. Balkerna, Rotterdam, pp. 3–9.

Lee, J.K.W., 1993. The argon release mechanisms of horn-blende in vacuo. Chem. Geol. 106, 133–170.

Lenoir, J.-L., Kuster, D., Liegeois, J.P., Utke, A., Haider, A.,Matheis, G., 1994. Origin and regional significance of latePrecambrian and early Palaeozoic granitoids in the Pan-African belt of Somalia. Geol. Rundsch. 83, 624–641.

Ludwig, K.R., 1980. Calculation of uncertainties of U–Pbisotope data. Earth Planet. Sci. Lett. 46, 212–220.

Ludwig, K.R., 1994. Isoplot–a plotting and regression pro-gram for radiogenic-isotope data, Version 2.75. US Geo-logical Survey Open File Report, pp. 91-445.

Maboko, M.A.H., Nakamura, E., 1995. Sm–Nd garnet agesfrom the Uluguru complex of eastern Tanzania: furtherevidence for post-metamorphic slow cooling in theMozambique belt. Precambrian Res. 74, 195–202.

Maboko, M.A.H., Boelrijk, N.A.I.M., Priem, H.N.A., Ver-durinen, E.A.T., 1985. Zircon U–Pb and biotite Rb–Srdating of the Wami river granulites, eastern granulites,Tanzania: evidence for approximately 715 Ma old gran-ulite-facies metamorphism and final Pan-African coolingapproximately 475 Ma ago. Precambrian Res. 30, 361–378.

Maboko, M.A.H., McDougall, L, Zeitler, P.K., 1989. Datinglate Pan-African cooling in the Uluguru granulite complexof Eastern Tanzania using the 40Ar–39Ar technique. J. Afr.Earth Sci. 9, 159–167.

McWilliams, M.O., 1981. Palaeomagnetism and Precambriantectonic evolution of Gondwana. In: Kroner, A. (Ed.),Precambrian Plate Tectonics. Developments in Precam-brian Geology. Elsevier, Amsterdam, pp. 649–687.

Meert, L.G., van der Voo, R., Aytib, S., 1995. Paleomagneticinvestigation of the Neoproterozoic Gagwe lavas andMbozi complex, Tanzania and the assembly of Gondwana.Precambrian Res. 74, 225–244.

Mezger, K., Hanson, G.N., Bohlen, S.R., 1989. High-precisionU–Pb ages of metamorphic rutile: application to the cool-ing history of high-grade terranes. Earth Planet. Sci. Lett.96, 106118.

Mezger, K., Rawnsley, C.M., Bohlen, S.R., Hanson, G.N.,1991. U–Pb garnet, sphene, monazite, and rutile ages:implications for the duration of high-grade metamorphismand cooling histories, Adirondack Mts. New York J. Geol.99, 415–428.

Mezger, K., van der Pluijm, B.A., Essene, E.J., Halliday,A.N., 1992. The Carthage–Colton mylonite zone (Adiron-dack Mts.): the site of a cryptic suture? J. Geol. 100,630–638.

Moller, A., 1995. Pan-African granulites and Early Protero-zoic eclogites in the Precambrian basement of easternTanzania: P–T– t history and crustal evolution of thecomplex Mozambique belt. Dissertation, Christian-Al-brechts-Universitat Kiel, Germany.

Moller, A., Appel, P., Mezger, K., Schenk, V., 1995. Evidencefor a 2 Ga subduction zone: eclogites in the Usagaran Beltof Tanzania. Geology 23, 1067–1070.

Moller, A., Mezger, K., Schenk, V., 1998. Crustal age domainsand the evolution of the continental crust in the Mozam-bique Belt of Tanzania: combined Sm–Nd, Rb–Sr, andPb–Pb isotopic evidence. J. Petrol. 39, 749–783.

Moores, E.M., 1991. Southwest US–East Antarctic (SWEAT)connection: a hypothesis. Geology 19, 425–428.

Muhongo, S., 1990. Evolution of the Proterozoic granulitecomplexes in eastern Tanzania and implication for the


geodynamic evolution of the Mozambique Belt of EastAfrica. PhD Thesis, Technische Universitat Berlin,Germany.

Muhongo, S.M., Lenoir, J.-L., 1994. Pan-African granulite-fa-cies metamorphism in the Mozambique Belt of Tanzania:U–Pb zircon geochronology. J. Geol. Soc. Lond. 151,343–347.

Nedelec, A., Paquette, J.-L., Bouchez, P., Olivier, P., Ralison,B., 1994. Stratoid granites of Madagascar: structure andposition in the Panafrican orogeny. Geodin. Acta (Paris) 7,48–56.

Nicollet, C, 1989. L’Eclogite de Faratsiho (Madagascar): un casexceptionnel de metamorphisme de haute-P-Base-T auProterozoique superieur. Precambrian Res. 45, 343–352.

Noble, W.P., Foster, D.A., Gleadow, A.J.W., 1997. The post-Pan-African thermal and extensional history of the crys-talline basement rocks in eastern Tanzania. Tectonophysics275, 331–350.

Oxburgh, E.R., 1990. Some thermal aspects of granulite history.In: Vielzeuf, D., Vidal, P. (Eds.), Granulites and CrustalEvolution. Kluwer, Dordrecht, pp. 569–580.

Paquette, J.-L., Nedelec, A., Moine, B., Rakotondrazafy, M.,1994. U–Pb, single zircon Pb-evaporation, and Sm–Ndisotopic study of a granulite domain in SE Madagascar. J.Geol. 102, 523–538.

Parrish, R., Whitehouse, M., 1999. Constraints on the diffusiv-ity of Pb in monazite, its closure temperature, and itsU–Th–Pb systematics in metamorphic terrains, from aTIMS and SIMS study. Terra Abstr. EUG 10, 711.

Priem, H.N.A., Boelrijk, N.A.I.M., Hebeda, E.H., et al., 1979.Isotopic age determinations on granitic and gneissic rocksfrom the Ubendian–Usagaran system in southern Tanza-nia. Precambrian Res. 9, 227–239.

Sampson, D.N., Wright, A.E., 1964. The geology of theUluguru Mountains. Bull. Geol. Surv. Tanzan. 37, 1–69.

Scharer, U., 1984. The effect of initial 230Th disequilibrium. onyoung U–Pb ages: the Makalu case, Himalaya. EarthPlanet. Sci. Lett. 67, 191–204.

Scharer, U., Krogh, T.E., Gower, C.R, 1986. Age and evolutionof the Grenville Province in eastern Labrador form U–Pbsystematics in accessory minerals. Contrib. Mineral. Petrol.94, 438–451.

Schenk, V., 1980. U–Pb and Rb–Sr radiometric dates and theircorrelation with metamorphic events in the granulite-faciesbasement of the Serre, southern Calabria (Italy). Contrib.Mineral. Petrol. 73, 23–38.

Schenk, V., 1990. The exposed crustal cross section of southernCalabria, Italy: structure and evolution of a segment ofHercynian crust. In: Salisbury, M.H., Fountain, D.M.(Eds.), Exposed Cross-Sections of the Continental Crust.Kluwer, Dordrecht, pp. 21–42.

Scott, D.J., St-Onge, M.R., 1995. Constraints on Pb closuretemperature in titanite based on rocks from the Ungavaorogen, Canada: Implications for U–Pb geochronology andP–T– t path determinations. Geology 23, 1123–1126.

Shackleton, R.M., 1967. Complex history of the MozambiqueBelt. Annu. Rep. Sci. Results Univ. Leeds 11, 12–13.

Shackleton, R.M., 1986. Precambrian collision tectonics inAfrica. In: Coward, M.P., Ries, A.C. (Eds.), CollisionTectonics. Blackwell, Oxford, pp. 329–349.

Shiraishi, K., Kagami, H., 1992. Sm–Nd and Rb–Sr ages ofmetamorphic rocks from the Sør Rondane Mountains, EastAntarctica. In: Yoshida, Y. (Ed.), Recent Progress inAntarctic Earth Science. Terra Scientific, Tokyo, pp. 29–35.

Shiraishi, K., Hiroi, Y., Ellis, D.J., Fanning, C.M., Motoyoshi,Y., Nakai, Y., 1992. The first report of a cambrian orogenicbelt in East Antarctica—an ion microprobe study of theLutzow Holm complex. In: Yoshida, Y. (Ed.), RecentProgress in Antarctic Earth Science. Terra Scientific, Tokyo,pp. 67–73.

Shiraishi, K., Ellis, D.J., Hiroi, Y., Fanning, C.M., Motoyoshi,Y., Nakai, Y., 1994. Cambrian orogenic belt in EastAntarctica and Sri Lanka: implications for Gondwanaassembly. J. Geol. 102, 47–65.

Smith, H.A., Giletti, B.J., 1997. Lead diffusion in monazite.Geochim. Cosmochim. Acta 61, 1047–1055.

Spear, F.S., Parrish, R.R., 1996. Petrology and cooling rates ofthe Valhalla Complex, British Columbia, Canada. J. Petrol.37, 733–765.

Steiger, R.H., Jager, E., 1977. Subcommision on geochronology:convention on the use of decay constants in geo- andcosmochronology. Earth Planet. Sci. Lett. 36, 359–362.

Stern, R.L, 1994. Arc assembly and continental collision in theNeoproterozoic East African Orogen: implications for theconsolidation of Gondwanaland. Annu. Rev. Earth Planet.Sci. 22, 319–351.

Teklay, M., Kroner, A., Mezger, K., Oberhansli, R., 1998.Geochemistry, Pb–Pb single zircon ages and Nd–Sr isotopecomposition of precambrian rocks from southern and east-ern Ethiopia: implications for crustal evolution in EastAfrica. J. Afr. Earth Sci. 26, 207–222.

Theunissen, K., Lenoir, J.-L., Liegeois, J.-P., Delvaux, D.,Mruma, A.H., 1992. Empreinte panafricaine majeure dansla chaıne ubendienne de Tanzanie sud-occidentale:geochronolgie U–Pb sur zircon et contexte structural. C.R.Acad. Sci. Paris 314 (2), 1355–1362.

Tilton, G.R., 1973. Isotopic lead ages of chondritic meteorites.Earth Planet. Sci. Lett. 19, 321–329.

Todt, W., Cliff, R.A., Hanser, A., Hofmann, A.W., 1996.Evaluation of a 202Pb–205Pb double spike for high-precisionlead isotope analysis. In: Basu, A., Hart, S.R. (Eds.), EarthProcesses: Reading the Isotopic Code. American Geophys-ical Union, Washington, DC, pp. 429–437.

Tucker, R.D., Raheim, A., Krogh, T.E., 1986. Uranium–leadzircon and titanite ages from the northern portion of theWestern Gneiss Region, south-central Norway. EarthPlanet. Sci. Lett. 81, 203–211.

Tucker, R.D., Handke, M.J., Hamilton, M.A., Ashwal, L.D.,LeGrange, M., Rambeloson, R.A., 1999. U–Pb geochronol-ogy and isotope geochemistry of the Archean and Protero-zoic rocks of north-central Madagascar. J. Geol. 107,135–153.

Unnikrishnan-Warrier, C., Santosh, M., Yoshida, M., 1995.First report of Pan-African Sm–Nd and Rb–Sr mineral


isochron ages from regional charnockites of southern India.Geol. Mag. 132, 253–260.

Verts, L.A., Chamberlain, K.R., Frost, C.D., 1996. U–Pbsphene dating of metamorphism: the importance of sphenegrowth in the contact aureole of the Red Mountain pluton,Laramie Mountains, Wyoming. Contrib. Mineral. Petrol.125, 186–199.

Watson, J., 1976. Vertical movements in Proterozoic structuralprovinces. Philos. Trans. R. Soc. Lond. A 280, 620–640.

Wendt, I., Besang, C., Harre, W., Kreuzer, H., Lenz, H., Muller,

P., 1972. Age determinations of granitic intrusions andmetamorphic events in the Early Precambrian of Tanzania.24th Int. Geol. Congr. 1, 295–314.

Windley, B.F., Razafiniparany, A., Razakarnanana, T., Acker-mand, D., 1994. Tectonic framework of the Precambrian ofMadagascar and its Gondwana connections: a review andreappraisal. Geol. Rundsch. 83, 642–659.

Zhang, L.-S., Scharer, U., 1996. Inherited Pb componentsin magmatic titanite and their consequence for the interpre-tation of U–Pb ages. Earth Planet. Sci. Lett. 138, 57–65.

.

U-Pb dating of metamorphic minerals: age of metamorphism and cooling history of Pan-African granulites and early Proterozoic eclogites in Tanzania

Documents