Petrology, geochemistry and genesis of newly discovered Mesoproterozoic highly magnesian, calcite-rich kimberlites from Siddanpalli, Eastern Dharwar Craton, Southern India: products

Page 1

ORIGINAL PAPER

Petrology, geochemistry and genesis of newly discoveredMesoproterozoic highly magnesian, calcite-rich kimberlitesfrom Siddanpalli, Eastern Dharwar Craton, Southern India:products of subduction-related magmatic sources?

N. V. Chalapathi Rao & A. Dongre & G. Kamde &

Rajesh K. Srivastava & M. Sridhar & F. V. Kaminsky

Received: 26 December 2008 /Accepted: 2 October 2009 /Published online: 27 October 2009# Springer-Verlag 2009

Abstract The Siddanpalli kimberlites constitute a newlydiscovered cluster (SKC) of Mesoproterozoic (1090 Ma)dykes occurring in the granite-greenstone terrain of theGadwal area in the Eastern Dharwar Craton (EDC),Southern India. They belong to coherent facies and containserpentinized olivines (two generations), phlogopite, spinel,perovskite, ilmenite, apatite, carbonate and garnet xeno-crysts. A peculiar feature of these kimberlites is theabundance of carbonate and limestone xenoliths of theeroded platformal Proterozoic (Purana) sedimentary cover

of Kurnool/Bhima age. Chemically, the Siddanpalli dykesare the most magnesium-rich (up to 35 wt.% MgO) andsilica-undersaturated (SiO2<35 wt.%) of all kimberlitesdescribed so far from the Eastern Dharwar Craton. The La/Yb ratio in the Siddanpalli kimberlites (64–105) isconsiderably lower than that in the other EDC kimberlites(108–145), primarily owing to their much higher HREEabundances. Since there is no evidence of any crustalcontamination by granitic rocks we infer this to be aspecific character of the magmatic source. A comparison ofthe REE geochemistry of the Siddanpalli kimberlites withpetrogenetic models for southern African kimberlitessuggests that they display involvement of a wide range inthe degree of melting in their genesis. The differentgeochemical signatures of the SKC compared to the otherknown kimberlites in the EDC can be explained by acombination of factors involving: (i) higher degrees ofpartial melting; (ii) relatively shallower depths of deriva-tion; (iii) possible involvement of subducted component intheir mantle source region; and (iv) previous extraction ofboninitic magmas from their geological domain.

Introduction

The Siddanpalli kimberlite cluster (SKC; Sridhar et al.2004) comprises three kimberlites (designated as SK-1,SK-2, and SK-3) that were emplaced in an area of 2.5 km2

within the granitoid rocks of the Gadwal granite-greenstoneterrane of the Eastern Dharwar Craton, southern India(Fig. 1). The SKC is considered to be a part of the Raichurkimberlite field (RKF) which is sandwiched between thewell known kimberlites fields of Narayanpet (NKF) in thenorth and Wajrakarur (WKF) in the south. The kimberlites

Editorial handling: K. R. Moore

N. V. Chalapathi Rao (*) : R. K. SrivastavaDepartment of Geology, Banaras Hindu University,Varanasi 221005 Uttar Pradesh, Indiae-mail: [email protected]

A. DongreDepartment of Geology, Institute of Science,Aurangabad, India

G. KamdeDepartment of Geology, RSTM (Nagpur) University,Nagpur, India

M. SridharGeological Survey of India,Eastern Region, Kolkata, India

F. V. KaminskyKM Diamond Exploration Ltd.,West Vancouver, BC, Canada

Present Address:N. V. Chalapathi RaoMineral Resources, Technical University of Clausthal,Adolph Roemer Straße,Clausthal-Zellerfeld 38678, Germany

Miner Petrol (2010) 98:313–328DOI 10.1007/s00710-009-0085-y

Page 2

of Undraldoddi, Mettimalkapur and Maliabad near Raichur(Shivanna et al. 2002) and those at Chagapuram (Ravi andSatyanarayana 2007) also occur in the RKF and theirdetailed petrography and geochemistry are not published.The SKC are yet to be processed for diamonds, but some ofthe kimberlites from the RKF have proven diamondiferous(Lynn 2005).

Mesoproterozoic kimberlites of 1.1 Ga are known to occurin largest numbers (~100) in the Eastern Dharwar Craton(EDC) of southern India. One of the kimberlites in the SKC(SK-1) yielded a 1090 Ma age (Kumar et al. 2007), whichsuggests that this cluster could constitute a part of thewidespread Mesoproterozoic kimberlite magmatism thatoccurs in the Southern Indian Shield. The SKC provides anopportunity to investigate the nature and evolution of thesub-continental or sub-lithospheric mantle in a new geolog-ical domain of the Indian Plate. In addition, the study ofthese rocks provides impetus to refine the existing modelsfor the emplacement and evolution of the Eastern DharwarCraton kimberlites, and in evaluating their genesis.

The Siddanpalli kimberlite occurrences are locatedwithin the Krishna River Basin, which has produced someof the famous Indian diamonds, such as Koh-i-Noor, Pitt

and Orloff. The Siddanpalli area is also associated withsubduction-related boninitic magmatism during the Neo-archaean (Manikyamba et al. 2005). It would be of interest,therefore, to investigate whether subduction has had anyinfluence on the continental lithospheric mantle at the timeof eruption of the SKC.

The objectives of this contribution are three-fold: (i) todocument the petrology and geochemistry of the Siddanpallikimberlite cluster; (ii) to compare and contrast their mineralchemistry and geochemistry with that of well characterizedkimberlites from the WKF and the NKF; and (iii) to infer thepetrogenesis of the Siddanpalli kimberlites in light of recentpetrogenetic models developed for southern African (e.g.Becker and Le Roex 2006; Becker et al. 2007) and EasternDharwar Craton (e.g. Chalapathi Rao et al. 2004; Chalapa-thi Rao and Srivastava 2009) kimberlites.

Geological setting

Metabasalts, amphibolite, granitoids, quartz-sericite schist,banded-iron formation and boninitic rocks constitute themajor lithologies of the Gadwal granite-greenstone belt (e.g.

SK-1SK-2

SK-3Siddanpalli

Undraladoddi

Ghut

Raichur

Gadwal

Atmakur

10 km

R

R

77 19’ 77 49’16 24’

16 00’77 19’

16 00’77 49’

O

O

O

O

16 24’

O

O

O

O

120 km

Chelima

78 80

14

17

O O

O

O

O

KLF

WKF

NKF

Gadwal

RLF

Zangamarajupalle

Hyderabad

Siddanpalle

RKF

Cuddapah Basin

CuddapahBasin

(A) (B)

KimberlitesPink & Grey Granite/Granodiorite

Peninsular GneissicComplex

Gadwal Schist Belt Town/City

INDEX

R Quartz ReefR

Fig. 1 a Location map of the Cuddapah Basin in Eastern DharwarCraton, southern India showing the disposition of various kimberliteand lamproite fields. NKF Narayanpet kimberlite field, RKF Raichurkimberlite field, WKF Wajrakarur kimberlite field, KLF Krishnalamproite field, RLF Ramadugu lamproite field, RKF Raichur

kimberlite field, Lamproites within the Cuddapah Basin are locatedat Chelima and Zangamarajupalle. b Geological map of the Siddan-palle area showing the location of kimberlite pipes (after Sridhar et al.2004)

314 N.V. Chalapathi Rao et al.

Page 3

Srinivasan 1990; Ramam and Murthy 1997; Manikyamba etal. 2005), with quartz reefs and kimberlites of theSiddanpalli cluster (Sridhar et al. 2004) constituting theyoungest sequences. Three generations of folding arereported from the Gadwal belt (Matin 2001) and gravitymodeling suggests a synformal structure (Ramadass et al.2001).

The SK-1 pipe of the SKC, has been dated by Rb-Srphlogopite/whole rock method, and gave an age of1,093±4 Ma (Kumar et al. 2007). Even though SK-2 andSK-3 kimberlites of the Siddanpalli cluster have not beendated, it is likely that all of them share a similar age, since theiremplacement is structurally controlled by a NW-SE trendingmajor fault in the area (Sridhar et al. 2004). Limestone andcarbonate xenoliths, which are a characteristic feature ofthese kimberlites, are inferred to have been derived from theeroded Proterozoic sedimentary cover of Bhima-KurnoolPurana sediments (Dongre et al. 2008; Chalapathi Rao et al.2009). The geology of the individual Siddanpalli kimberlitesis briefly summarized below:

SK-1 (160 19’ 89”; 77 55’ 43”): This body is located2 km NE of Siddanpalli village and measures about 100×65 m. It is the best exposed of the three kimberlites in thiscluster. The kimberlite is melanocratic and belongs to thehard banke variety. Crustal xenoliths of country rockgranitoids and carbonate (12×10 cm) are observed on ornear the outcrop surface. The carbonate xenoliths occureither as irregular masses or as angular enclaves. Theydisplay very sharp contacts and petrographically arerecrystallzed carbonates.

SK-2 (160 19’ 27”; 77 55’ 60”): This body occurs1.7 km NE of Siddanpalli village and is mostly covered bysoil. Sridhar et al. (2004) notes that this kimberlite is a twinbody with the main body measuring 100×50 m and asatellite body measuring 10×10 m. Where exposed, SK-2 isweathered compared to SK-1 and is most competent instream cuttings, where a number of limestone xenolithswere recovered. These xenoliths are hard, competent and upto 12×10 cm. They display bands and laminae of chert.Their contact with the kimberlite is sharp without anygradation or reaction.

SK-3 (160 17’ 82”; 77 55’ 68”): This kimberlite islocated about 1.2 km NE of Siddanpalli village andmeasures about 26×14 m. It is exposed in a well and alsoin an adjacent stream cutting. SK-3 is less weathered thanSK-2 and is similar in appearance to the SK-1 pipe,containing country-rock granitoids and carbonate xenoliths.The carbonate xenoliths occur either as irregular enclavesor as rare recrystallized rhombohedral crystals of up to2 cm. However, samples from SK3 could not be accom-modated in this study.

From the mode of occurrence and distribution ofcarbonate/limestone xenoliths and the nature of the kim-

berlites it can be inferred that the pre-existing carbonatehorizon was wide-spread and its thickness was considerableat the time of eruption of the kimberlites and the diatremeand crater facies have been eroded. Detailed studies on thecarbonate xenoliths and their geochemistry are addressedseparately (Chalapathi Rao et al. 2009).

Sample preparation and analytical techniques

The freshest possible outcrop samples were obtained.Weathered surfaces were removed and all visible xenolithswere manually separated from kimberlites by hand pickingprior to crushing and powdering. However, inherentalteration in the samples due to the exposure of thekimberlites to tropical weathering could not be eliminated.Mineral grains were analyzed with a CAMECA SX-50electron microprobe (installed at the Institute of Diamonds,Russian academy of Natural Sciences, Moscow, Russia)operating in wavelength-dispersive mode with an acceler-ating voltage of 15 kV, a specimen current of 20 nA, andbeam diameter of 5µm. Some of the mineral analyses werealso done by one of us (NVCR) at the Mineral Resources,Technical University, Clausthal, Germany, using a CAMECASX100 electron microprobe fitted with 4 WDS spectrometersand by employing an acceleration voltage of 15 kV, beamcurrent of 20nA and beam diameter of 1µm. Natural andsynthetic standards were used in calibration. Geochemicalanalyses of the samples were carried out at the laboratoriesof the National Geophysical Research Institute (NGRI),Hyderabad, India. Major elements were determined by X-ray fluorescence spectrometry (XRF) using a PhilipsMAGIX PRO Model 2440. Trace, REE and HFSE wereanalyzed by Inductively Coupled Plasma Mass Spectrom-etry (ICP-MS) using a Perkin Elmer SCIEX ELAN DRC II.Two international kimberlitic reference materials namelySARM-39 ( South African Kimberlite) and MY-4 (RussianKimberlite) were used as reference materials following Royet al. (2007).

Petrography and mineral chemistry

Petrographic studies reveal that all three bodies are coherentkimberlites (previously termed hypabyssal kimberlites)(Cas et al. 2008) with an inequigranular texture impartedby rounded to anhedral olivine macrocrysts and subhedralto euhedral olivine phenocrysts. In general, olivine macro-crysts and phenocrysts have been altered to serpentine-group minerals and may be replaced by carbonates (Fig. 2aand b). The groundmass is dominated by serpentine-groupminerals, phlogopite, spinel, perovskite, calcite and apatite.Xenocrysts of chrome-diopside, magnesian ilmenite and

Petrology, geochemistry and genesis of kimberlites from southern India: subduction related magmatic sources? 315

Page 4

pyrope garnet are also present. Extensive carbonate enrich-ment is a characteristic feature of SKC bodies, possibly dueto the assimilation of platformal sedimentary carbonatehorizons. The representative mineral chemistry of individualphases (cores) is given in Tables 1 and 2, and the results aresummarized below.

Olivine Olivine macrocrysts and phenocrysts are complete-ly replaced by serpentine-group minerals. The serpentine-group minerals that form pseudomorphs after olivine haveMg/(Mg+Fe) ranging from 88–90. These values fall withinthe range of Mg/(Mg+Fe) of 84–92, reported for freshmacrocrysts, as well as phenocrysts of olivine from the

(C)

Ol

Cb

(A)

2 mm

Ol

2 mm

(B)

Ol

Cb

Ol

Ol

Ol

(D)

(H)

Pv

(G)

Pv

(F)

Ilm

(E)

Ol

Fig. 2 a (SK–1A) Inequigranu-lar texture imparted by subhe-dral to anhedral olivine (Ol)macrocrysts. Note that both theolivines have been altered toserpentine-group minerals andreplaced by carbonates; polar-ised light; uncrossed nicols; (b;SK–2B) Carbonate veins arequite common and the black(opaque) phases in the ground-mass are the oxide phases;polarised light; uncrossed nic-ols; (c; SK1–B ) Back ScatteredElectron (BSE) image depictingextensive carbonation (Cb) ofserpentinised olivine macro-crysts and phenocrysts; bright(white) grains are various oxidephases; (d; SK–2C) BSE show-ing carbonation (Cb) of thematrix as well; (e; SK–2B) BSEdepicting thorough carbonationof serpentinised olivines; notethat some of the serpentinisedolivine cores are free from car-bonation; (f; SK–2B) BSE ofgroundmass ilmenite (Ilm)grains (~20 μm) showing zon-ing; BSE depicting two modesof occurrence of perovskite(Pv); as discrete grains (g; SK–2A) and as necklaces aroundphenocrystal olivines(h; SK–2B)

316 N.V. Chalapathi Rao et al.

Page 5

Tab

le1

EPMA

analyses

(wt%

)of

theselected

mineralsfrom

theSKC

Oxides(w

t%)

SK1

Phlog

opite

SK1

Serpentine

SK1

Serpentine

SK1

Calcite

SK1

Calcite

SK1

Spinel

SK2a

Spinel

SK2a

Spinel

SK2a

Ilmenite

SK2a

Ilmenite

SK2a

Ilmenite

SK2a

Ilmenite

SiO

239

.53

41.48

42.23

––

0.57

0.81

0.49

0.07

0.05

0.10

0.35

TiO

20.69

0.13

0.06

––

29.95

14.59

14.34

50.35

51.79

51.87

49.93

Al 2O3

15.17

3.27

1.29

––

1.46

8.95

6.02

0.00

0.00

0.00

0.00

Cr 2O3

1.33

0.16

0.16

––

1.71

11.15

5.31

0.27

0.07

0.02

0.28

FeO

3.08

8.26

7.03

0.01

0.04

49.99

39.62

31.06

44.66

43.50

41.42

43.53

Fe 2O3

––

––

–6.74

19.12

37.32

––

––

MnO

0.10

0.07

0.09

0.02

0.03

5.54

1.57

1.11

3.78

3.54

3.94

2.04

MgO

23.95

35.24

38.37

0.01

0.00

2.08

3.47

2.34

0.05

0.09

0.07

2.11

CaO

0.10

0.10

0.21

55.35

55.04

0.63

0.75

0.59

0.45

0.17

0.81

0.72

Na 2O

0.17

0.08

0.10

––

0.19

0.03

0.35

0.02

0.01

0.04

0.03

K2O

8.81

0.01

0.00

––

–0.00

0.00

0.00

0.00

P2O5

–―

––

––

––

––

BaO

3.16

0.01

0.00

0.09

0.18

––

––

–

SrO

––

–0.67

0.80

––

––

–

Total

96.09

88.81

89.54

55.48

55.29

98.86

100.06

98.93

99.65

99.22

98.27

98.99

Oxy

gens

224

44

44

66

66

Si

5.75

31.115

1.12

30.02

10.02

70.01

80.00

30.00

30.00

50.01

8

Ti

0.07

60.00

30.00

10.82

80.36

40.38

81.93

91.98

62.00

11.91

8

Al

2.60

20.10

40.04

00.06

30.35

00.25

50.00

00.00

00.00

00.00

0

Cr

0.16

20.00

30.00

30.05

00.29

20.15

10.011

0.00

30.00

10.011

Fe2

+0.37

50.18

60.15

61.53

60.98

80.84

11.91

21.85

51.77

61.85

9

Fe3

+0.00

00.00

00.00

00.18

60.53

01.12

3–

––

–

Mn

0.01

20.00

20.00

20.114

0.04

40.03

40.16

40.15

30.17

10.08

8

Mg

5.19

61.41

21.52

00.02

50.17

10.12

50.05

00.00

70.00

50.16

1

Ca

0.01

60.00

30.00

60.01

40.02

70.02

30.45

00.00

90.04

50.03

9

Na

0.04

80.00

40.00

5–

0.00

20.02

40.00

10.00

10.00

20.00

1

K1.63

50.00

10.00

0–

0.00

00.00

00.00

00.00

00.00

00.00

1

Ba

0.33

0–

––

––

––

––

Sr

––

––

––

––

––

Total

16.205

2.83

12.85

72.83

72.80

12.98

34.05

54.01

34.00

14.07

6

aanalyses

done

atTU,Clausthal

Petrology, geochemistry and genesis of kimberlites from southern India: subduction related magmatic sources? 317

Page 6

Tab

le2

EPMA

analyses

(wt%

)of

apatite

andperovskite

from

theSKC

Oxides

(wt%

)SK1

Apatite

SK1

Apatite

SK1

Apatite

SK1

Apatite

SK1

Apatite

SK2

Apatite

SK2

Apatite

SK1

Perov

skite

SK1

Perov

skite

SK1

Perov

skite

SK1

Perov

skite

SK2

Perov

skite

SK2

Perov

skite

SiO

20.92

0.84

0.85

0.82

0.86

1.64

1.75

–0.18

0.02

0.13

––

TiO

2–

––

––

––

58.07

56.43

57.41

56.90

59.01

58.10

Al 2O3

––

––

––

––

0.28

0.24

0.21

––

Cr 2O3

––

––

––

––

0.02

0.05

0.01

––

FeO

0.07

0.05

0.10

0.10

0.07

0.12

0.18

0.98

1.74

1.04

1.04

0.71

0.88

MnO

0.02

0.00

0.00

0.03

0.03

0.01

0.03

–0.18

0.00

0.04

––

MgO

0.13

0.08

0.12

0.10

0.05

0.08

0.12

–1.28

0.07

0.12

––

CaO

53.32

52.83

52.85

52.96

53.14

53.83

54.18

37.59

36.45

37.26

38.08

37.32

37.88

Na 2O

0.14

0.11

0.20

0.08

0.19

0.41

0.26

0.54

0.81

0.73

0.69

0.39

0.74

P2O5

40.38

40.29

40.34

40.19

40.09

40.69

39.49

––

––

––

Nb 2O5

––

––

––

–0.51

0.43

0.46

0.74

0.58

0.68

La 2O3

0.02

0.07

0.07

0.00

0.03

0.03

0.03

0.46

0.61

0.44

0.02

0.62

0.72

Ce 2O3

0.06

0.09

0.19

0.00

0.06

0.13

0.14

1.09

1.40

1.40

1.10

1.65

2.48

BaO

0.11

0.04

0.00

0.08

0.00

0.03

0.10

––

––

––

SrO

0.61

0.59

0.61

0.53

0.53

1.03

1.00

––

––

––

ThO

2–

––

––

––

0.03

0.04

0.03

0.00

0.01

0.00

ZrO

2–

––

––

––

0.13

0.16

0.08

0.08

0.22

0.19

Cl

0.05

0.03

0.00

0.03

0.04

0.04

0.04

––

––

––

F1.65

1.81

2.67

3.12

3.12

1.90

2.37

––

––

––

Total

97.48

96.83

98.00

98.04

98.20

99.84

99.79

99.40

100.01

99.23

99.41

100.51

99.32

318 N.V. Chalapathi Rao et al.

Page 7

WKF kimberlites (Chalapathi Rao et al. 2004; ChalapathiRao and Srivastava 2009), and suggest that compositions offresh olivine from the WKF and serpentinized olivine fromthe SKC are comparable. Carbonation of the macrocrysticand phenocrystic serpentinised olivines is also prominentlyseen (Fig. 2b–e).

Phlogopite Even though phlogopite is an essential mineralin SKC samples its modal abundance is very low; perhapsdue to their highly altered nature. Only one acceptableanalysis could be obtained for pristine groundmass mica(see Table 1) and the data showing low totals are notpresented. These SKC mica is extremely low in TiO2 (0.69wt.%) and substantially aluminous (up to 15 wt.%)compared to those from the WKF and the NKF in theEastern Dharwar Craton (Chalapathi Rao et al. 2004;Chalapathi Rao and Srivastava 2009). K2O content is~8.81 wt.%, which is close to the pristine values (Deer etal. 1979). Mica is also barium-rich having BaO ~3.16 wt.%.

Spinel Spinel constitutes an important groundmass oxidephase (Fig. 2a–e). It occurs either as dispersed grains in thegroundmass or as aggregates. Compositionally the spinel ofthe SKC are relatively impoverished in MgO compared tothose from the southern African kimberlites (Fig. 3a), andeven other Eastern Dharwar Craton kimberlites. Mitchell(1986) delineated two trends amongst groundmass kimber-lite spinels: a magnesio-ulvöspinel-magnetite trend (trend1), unique to spinels in kimberlites and a Ti-magnetite trend(trend 2) that is similar to zoning of spinel in basalts (seeRoeder and Schulze 2008). Spinel of the Eastern DharwarCraton have extreme compositional variation and displayboth these trends, and are also substantially Fe-enriched(Fig. 3b).

Perovskite Perovskite is a common phase in SKC samples.It forms at the final stages of magmatic crystallization andoccurs either as discrete grains dispersed throughout thegroundmass (Fig. 2g), or as necklaces surrounding theearlier formed olivine grains (Fig. 2h). The perovskite ofthis study are compositionally similar (Table 2), with CaOvarying from 36.45–3.08 (wt %) and TiO2 from 56.90–59.01 (wt.%). Their FeOT contents display slight variationfrom 0.71–1.44 wt.% and are comparable to the perovskitefrom the NKF (0.85–0.98 wt.%) and WKF kimberlites(0.92–2.22 wt.% FeOT; Chalapathi Rao et al. 2004) andthose of Group I kimberlites (1–2 wt.%; Mitchell 1995, p.221). The SKC perovskite La2O3 (0.21–0.61 wt.%), Ce2O3

(up to 2.48 wt.%) and Nb2O5 (0.43–0.74 wt.%) contents aresimilarly to those reported for the other Eastern DharwarCraton kimberlites (Chalapathi Rao 1998; Chalapathi Raoet al. 2004). Minor amounts of thorium (ThO2: 0.04 wt.%)and zirconium (ZrO2: up to 0.22 wt.%) are also present.

Fluorapatite Fluorapatite is a late-stage magmatic phase andoccurs as euhedral prisms or as sprays of acicular crystals inthe mesostasis. It displays a tight compositional range (CaO:52.83–54.18 wt.%, P2O5: 39.49–40.38 wt.%; Table 3) withsubstantial fluorine (up to 3.12 wt.%) and minor strontium(SrO: 0.53–1.03 wt.%). La2O3 (0.02–0.07 wt.%) and Ce2O3

(0.06–0.14 wt.%) contents in perovskite are conspicuouslylow compared to those in SKC pervoskites.

Ilmenite Groundmass ilmenite is particularly abundant inSK-2 where it occurs as discrete grains of up to 30μm sizewhich are occasionally zoned (Fig. 2f). Compositionally(Table 1) ilmenite is manganese-bearing (up to 3.94 wt%MnO) and displays a wide-range in magnesium content(MgO : 0.05–2.11 wt%) and very low Cr2O3 contents

0,00

5,00

10,00

15,00

20,00

0,00 5,00 10,00 15,00 20,00

MgO ( wt%)

Al 2

O3(

wt%

)

(A)

SouthAfrican kimberlitesDharwar

cratonkimberlites

0,00

0,50

1,00

0,20 0,40 0,60 0,80 1,00 1,20

Fe2+/(Fe2+ + Mg)

Ti/(

Ti+

Cr+

Al)

Fieldof Dharwar cratonkimberlite spinels

Trend-2

Trend-1

(B)

Fig. 3 a MgO (wt%) vs Al2O3 (wt%) of spinels from SKC. The datafor Southern African kimberlite spinels is from Scott-Smith andSkinner (1984) and for Eastern Dharwar Craton kimberlites is fromChalapathi Rao et al. (2004). b Fe2+/(Fe2++Mg2+) vs Ti/(Ti+Cr+Al)(mol fraction) for groundmass kimberlite spinels projected onto thefront face of the ‘reduced’ spinel prism. Trends 1 and 2 exhibited bysouthern African spinels are from Mitchell (1986). Field of DharwarCraton kimberlites is from Chalapathi Rao et al. (2004) andChalapathi Rao and Srivastava (2009)

Petrology, geochemistry and genesis of kimberlites from southern India: subduction related magmatic sources? 319

Page 8

Table 3 Whole rock geochemistry of SKC samples under study

SK-1A SK-1B SK-1C SK-1D SK-1E SK-2A SK-2B SK-2C

SiO2 30.51 31.53 35.05 32.76 32.49 31.62 34.17 32.61

TiO2 1.17 1.28 1.21 1.53 1.54 1.46 1.41 0.85

Al2O3 1.53 1.63 1.75 1.53 1.48 1.63 2.58 3.63

Fe2O3 6.23 6.89 7.89 8.09 7.62 6.11 7.10 5.48

MnO 0.10 0.09 0.12 0.13 0.14 0.11 0.11 0.07

MgO 30.38 31.64 32.34 34.17 32.40 27.11 28.67 24.90

CaO 15.60 13.58 10.75 11.74 12.61 15.82 13.73 17.39

Na2O 0.01 0.02 0.01 0.01 0.02 0.03 0.01 0.01

K2O 0.09 0.11 0.14 0.03 0.05 0.03 0.04 0.05

P2O5 1.25 1.28 1.05 1.58 1.33 1.25 0.93 0.61

LOI 13.96 12.45 10.24 9.20 9.9 14.21 10.78 13.66

Total 100.83 100.50 100.55 100.77 99.58 99.38 99.53 99.26

CI 1.05 1.05 1.13 1.00 1.05 1.23 1.28 1.45

Ilm.I 0.24 0.26 0.28 0.28 0.28 0.28 0.3 0.25

Trace Elements (ppm)

Sr 1921.28 1018.27 1835.20 1603.82 1393.47 1895.36 1199.19 951.18

Rb 12.23 20.34 15.65 6.48 7.73 4.11 7.04 6.69

Zn 62.41 38.91 36.98 44.49 50.03 38.69 40.53 31.90

Cu 66.82 74.37 48.49 90.91 48.32 54.68 79.24 54.40

Ni 897.76 1110.87 959.82 1065.94 1002.76 870.72 976.65 738.90

Cr 959.52 1051.38 958.24 1065.50 1077.46 974.33 1077.03 665.68

Co 79.71 89.15 82.26 90.63 88.81 82.60 83.65 81.22

V 127.54 115.48 134.44 144.75 125.39 111.38 110.85 76.14

Sc 10.58 10.54 10.80 12.95 12.34 11.05 10.76 8.86

Y 34.77 33.56 36.21 41.30 38.98 36.65 34.43 29.62

Zr 365.47 356.11 379.20 436.85 415.49 365.47 417.01 274.07

Nb 174.69 172.02 204.16 211.49 204.53 182.69 188.15 109.65

Cs 0.38 1.76 0.43 0.26 0.53 1.07 2.83 2.23

Ba 800.37 394.46 858.21 546.96 741.66 1638.46 1220.34 745.28

Hf 7.86 7.53 8.18 9.58 8.82 7.73 9.07 6.49

Ta 17.46 17.89 23.03 21.75 21.59 19.48 21.84 13.80

Pb 15.98 13.49 17.98 13.73 14.10 13.89 23.68 17.76

Th 21.97 20.96 22.98 26.67 24.35 22.72 21.76 24.01

U 4.93 5.12 4.74 6.80 5.01 5.01 5.38 4.38

REE (ppm)

La 193.97 185.75 200.88 229.37 213.02 207.60 173.25 113.89

Ce 503.20 485.04 525.31 631.36 575.50 562.02 482.05 226.63

Pr 43.49 41.31 44.93 54.08 48.47 45.95 40.95 26.85

Nd 165.77 156.62 171.63 204.73 184.98 173.92 156.28 104.31

Sm 25.02 24.03 26.10 30.97 27.93 26.32 23.96 16.54

Eu 6.12 5.64 6.36 7.29 6.76 6.34 5.37 3.90

Gd 18.07 17.35 18.89 22.43 20.35 19.15 17.16 11.83

Tb 2.17 2.05 2.24 2.63 2.40 2.26 2.07 1.50

Dy 9.17 8.74 9.47 11.11 10.24 9.33 8.96 6.84

Ho 1.32 1.26 1.36 1.62 1.47 1.38 1.31 1.05

Er 3.09 2.95 3.23 3.80 3.44 3.29 3.03 2.50

Tm 0.35 0.34 0.37 0.43 0.39 0.38 0.35 0.34

Yb 1.87 1.77 1.95 2.23 2.02 2.02 1.83 1.78

Lu 0.23 0.21 0.23 0.28 0.25 0.25 0.23 0.22

320 N.V. Chalapathi Rao et al.

Page 9

(<0.28 wt%). Such Mn-bearing groundmass ilmenites arereported from southern African Group I and II kimberlites.

Carbonate Calcite is the only carbonate observed and itoccurs as an abundant secondary phase, as irregular patches inthe mesostatis, and also forming pseudomorphs after olivine(Fig. 2a–e). Veins of calcite are also observed locally(Fig. 2b). The composition of calcite is provided in Table 1

Garnet Only one garnet xenocryst is analysed in this study.Its composition is dominated by the pyrope end-member(70%). High Mg/(Mg+Fe) content (~0.71) negate a crustalorigin and suggest an affinity with sub-continental litho-spheric mantle (SCLM) (Schulze 2003). Low Cr2O3

content (<0.34 wt.%) and low Ca/(Ca+Mg+Fe) ratios(<0.3) suggest derivation either from a peridotitic or a Cr-poor megacrystic source.

Geochemistry

Whole-rock, major-element geochemical data for the SKCsamples is presented in Table 3. The contamination index(C.I.) of Clement (1982) [where C.I.=(SiO2+Al2O3+Na2O)/ (MgO+K2O)] is used to assess the role of crustalassimilation on the bulk chemistry of samples. Kimberliteswith a C.I.<1.4 are generally regarded as uncontaminatedor fresh. The C.I. for majority of the SKC samples is <1.28(Table 3 and Fig. 4) and thus, their contamination by crustalgranitoid material could be minimal. Relatively high C.I.values seen in case of SK-2 samples (Table 3) are due tocarbonate enrichment and alteration and which has resultedin lowering the MgO content and hence increased C.I. Thisis further supported by all samples from the SKC havingSiO2 contents between 30.5 and 35.1 wt%. The IlmeniteIndex (Ilm. I) of Taylor et al. (1994) is also widely used toidentify kimberlites that may have accumulated ilmenitemegacrysts and xenocrysts. This index is defined as: Ilm.I=(FeOT+TiO2)/(2K2O+MgO). Samples with an Ilm. I<0.52 are regarded as uncontaminated. The Ilm. I of theSKC samples are <0.30, suggesting that they have thelowest degree of ilmenite contamination index yet recog-nized in kimberlites (Fig. 4) known from the EasternDharwar Craton. Potassium is the most mobile element inthe Ilm. I and its leaching due to alteration should insteadhave resulted in a higher index. As many of the high-field

strength elements in the SKC show good correlation witheach other (see below) titanium is inferred to have beenimmobile and we do not favour lower Ilm resulting from themobility of titanium under tropical conditions of weatheringand pervasive alteration.

The SKC samples are all silica-undersaturated (SiO2<35wt.%), with some of them displaying the lowest silicacontent amongst the EDC kimberlites (Fig. 5a) possibly dueto carbonate contamination: their CaO contents are uni-formly high (10.75 to 17.39 wt.%; Fig. 5b). The SKCsamples are the most magnesian-rich kimberlites yetrecorded from the Eastern Dharwar Craton with MgO upto 34.17 wt% (Fig. 5a). The K2O content of the kimberlitesis low (0.03 to 0.14 wt.%), which is possibly related to thedegree of alteration of phlogopite in the samples. However,K2O/Na2O ratios of all samples are invariably >1, therebydisplaying their potassic nature (cf. Foley et al. 1987). Withincreasing MgO, the samples display a general decrease ofSiO2 (Fig. 5a), CaO (Fig. 5b), Al2O3 (Fig. 5c), and anincrease in Fe (Fig. 5d), suggesting fractionation of theparent magma. A good correlation between MgO and CaOin Fig. 5b indicate contamination with carbonate and alsosuggests that some of the carbonate to be primary in origin.

Trace-element data for the SKC is provided in Table 3.The Sc content of the kimberlites (8–12 ppm), which ishosted by phlogopite, is significantly lower in comparisonto other EDC kimberlites (13–27 ppm) (Chalapathi Rao et

Table 3 (continued)

SK-1A SK-1B SK-1C SK-1D SK-1E SK-2A SK-2B SK-2C

La/Yb 103.56 104.89 102.86 102.81 105.30 102.87 94.52 64.09

0

2

4

6

8

0.00 0.20 0.40 0.60 0.80 1.00

Ilmenite Index

Co

nta

min

atio

n In

dex

Olivine accumulation

Phlogopite

Crustal contamination

Entrained Ilmenite

Fig. 4 Contamination index (Clement 1982) versus the Ilmenite index(Taylor et al. 1994). Black squares SKC kimberlites, black circlesNarayanpet kimberlites and white circles Wajrakarur kimberlites. Datafor Narayanpet and Wajrakarur kimberlites are from Chalapathi Rao etal. (2004). The shaded area represents the field of the SKC pipes

Petrology, geochemistry and genesis of kimberlites from southern India: subduction related magmatic sources? 321

Page 10

30

32

34

36

38

40

42

44

10 20 30 40MgO (wt%)

SiO

2(w

t%)

7

9

11

13

15

17

19

10 15 20 25 30 35 40

MgO (wt%)

CaO

(w

t%)

1

2

3

4

22 27 32 37

MgO (wt%)

Al 2

O3

(wt%

)

5

6

7

8

9

25 27 29 31 33 35

MgO (wt%)

Fe 2

O3

(wt%

)

(A) (B)

(C) (D)

0

50

100

150

200

250

200 250 300 350 400 450

Zr (ppm)

La

(pp

m)

(G)

0

2

4

6

8

200 250 300 350 400 450

Zr (ppm)

U (

pp

m)

(H)

0

200

400

600

800

1000

1200

1400

10 15 20 25 30 35 40

MgO (wt%)

Ni (

pp

m)

(E)

0

2

4

6

8

10

12

200 250 300 350 400 450

Hf

(pp

m)

(F)

Zr (ppm)

Fig. 5 Variation diagrams of MgO vs other oxides and elements (a–e) and Zr vs other elements (f–h) for Siddanpalli kimberlites. Data forNarayanpet and Wajrakarur kimberlites are from Chalapathi Rao et al. (2004). Symbols are same as in the Fig. 4

322 N.V. Chalapathi Rao et al.

Page 11

al. 2004). Vanadium in kimberlites is hosted primarily byphlogopite and spinel. All SKC samples have lower Vabundances (76–144 ppm) relative to those from other EDCkimberlites (up to 355 ppm) (Chalapathi Rao et al. 2004;Chalapathi Rao and Srivastava 2009). Nickel in kimberliteis principally hosted by olivine, and hence its abundance isproportional to the macrocryst olivine content. However,despite their high Mg contents the SKC contain less Ni(738 – 1.110 ppm; Fig. 5e) and Cr (665–1.077 ppm) thanmany EDC kimberlites (Ni: 480–1.450 ppm and Cr: 330–1.510 ppm).

In contrast to their compatible trace-element concen-trations, the data for the SKC show that they haveincompatible trace-element abundances that are eithercomparable to, or higher than, the WKF and NKFkimberlites from the Eastern Dharwar Craton (ChalapathiRao et al. 2004). In general, the HFSE elements displaygood correlations amongst themselves (Fig. 5f and g), andalso with the fluid-mobile LILE elements such as U(Fig. 5h). The SKC whole-rock data has strong geochem-ical affinities towards the Group I kimberlites (Fig. 6a–d)from southern Africa in this aspect. The SKC samples showhighly fractionated chondrite-normalized REE distributionpatterns, with La abundances being 500 to 900 timeschondrite (Fig. 7a), and HREE abundances of 10 to 30times chondrite. All the SKC samples share similar LREEand MREE patterns, but of variable magnitude suggestingthat different degrees of partial melting were involved in theirgenesis. This aspect is evaluated further in “Petrogenesis”section below. The La/Yb ratios (64–105) of the SKC samplesare considerably less than those of other EDC kimberlites(108–145; Chalapathi Rao et al. 2004; Chalapathi Rao andSrivastava 2009) owing primarily to their high HREEabundances. This is illustrated by comparison of theirchondrite-normalized REE patterns with those of the NKFand WKF kimberlites (Fig. 7b and c). The La/Sm and La/Ybratios of the SKC samples are similar to, or lower than, theWKF and NKF kimberlites and all of them appear toconstitute a well defined array (Fig. 7d), probably reflectingvariable degrees of melting.

Normalized multi-element plots (Fig. 8) demonstrate thattrace elements in the SKC samples are highly enriched overthe primitive mantle, with moderate negative spikes at Sr,Nb, Hf and conspicuous negative spikes at K and Ti. Suchnegative anomalies either reflect hydrothermal alteration orthe presence of residual phases in the melt source regions.Negative depletions at Nb, Hf and Ti in kimberlites arecommonly interpreted to be subduction-related signatures

0

20

40

60

80

Ce/

Pb

0 20 40 60 80Ba/Nb

(A)

Ce/Pb=22

Field of on- and off-cratonof Group IIkimberlites ofSouth Africa

Ba/

Nb=

11.5

Field of on- and off-cratonGroup I kimberlites of

South Africa

0

0,1

0,2

0,3

0,4

Th

/Nb

0 0,5 1 1,5 2 2,5La/Nb

Th/Nb=0.135

La/N

b=1.

1

Field of on- and off-cratonof Group IIkimberlites ofSouth Africa

Field of on- and off-cratonGroup I kimberlites of

South Africa

(B)

0

5

10

15

20

25

La/

Th

0 5 10 15 20 25 30Nb/Th

MORB

Karoolavas

OIB

Group IkimberlitesSouth Africa

(D)

Avg. crust

1

10

100

Ce

/Pb

1 10 100 1000Ce (ppm)

(C)

Continentalcrust Group II

kimberlitesSouth Africa

OIMORB B

Group IkimberlitesSouth Africa

Fig. 6 a Ba/Nb vs Ce/Pb b La/Nb vs Th/Nb c Ce (ppm) vs Ce/Pb andd Nb/Th vs La/Th of the SKC kimberlites compared with those of theNarayanpet and Wajrakarur kimberlites. Symbols are same as in theFig. 4. Various fields are taken from Harris et al. (2004), Le Roex etal. (2003) and Becker and Le Roex (2006)

b

Petrology, geochemistry and genesis of kimberlites from southern India: subduction related magmatic sources? 323

Page 12

(e.g. Coe et al. 2008). Relatively low contamination indicesand the apparent immobility of various LILE and HFSEsuggest that these negative anomalies are likely to be sourcerelated.

Petrogenesis

Nature of the mantle source region, and depth and degreeof partial melting

Whilst all the SKC bodies experienced carbonate contam-ination, the good positive correlations displayed by theirHFSE demonstrate the relative immobility of these ele-ments. Therefore their abundances and ratios are exploitedto infer the petrogenesis. Fractionated HREEs coupled withextremely low Al2O3 (1.48–3.63 wt%) contents infer adeeper mantle sources for the SKC. Figure 9a reveals themto have been derived by melting of a similar source to thekimberlites of Wajrakarur and Narayanpet fields in theDharwar Craton, southern India (data sources: ChalapathiRao et al. 2004; Chalapathi Rao and Srivastava 2009).However, LREE enrichment (La/Sm)N in case of SKC isrelatively much weaker and may relate to (i) larger degreesof melting or (ii) their generation from a less metasomatisedsource. The degree of partial melting involved in theproduction of mantle-derived igneous rocks can be estimat-ed from the SiO2 content of a melt and the respective Nb/Yratio, since the latter encompasses the range of elementincompatibility in a garnet-bearing peridotite, and has beenfound to be not greatly affected by metasomatic processes(Rogers et al. 1992; Beard et al. 1998). Fig. 9b demon-strates that the SKC rocks: (i) are smaller-degree meltproducts than Ocean Island Basalts (OIB) and Mid OceanicRidge Basalts (MORB); and (ii) the amount of partialmelting undergone by their source regions is much higherthan any of the other EDC kimberlites (WKF and NKF),and also those from southern Africa (not shown). There-fore, the relatively lower La/Yb ratios of the SKC reflectsthe larger degrees of partial melting in their source rocks(cf. Mitchell 1995). The relatively unradiogenic initital87Sr/86Sr ratio of 0.70340 for the SK1 kimberlite (AnilKumar et al. 2007) is similar to the Group I kimberlitesfrom southern Africa.

Chondrite-normalized REE distribution patterns of theSKC samples have significantly higher HREE abundances

1

10

100

1000R

OC

K/C

HO

ND

RIT

E

1

10

100

1000

RO

CK

/CH

ON

DR

ITE

NP-5 (Narayanpet)

MD/11A (Maddur)

KK11 (Kotakonda)

PDF/1B (Padiripahad)

1

10

100

1000

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb LuREE

RO

CK

/CH

ON

DR

ITE

6/PR1-B (Wajrakarur)

7/PR2 (Wajrakarur)

P3A (Lattavaram)

P5 (Muligiripalle)

P6/5 (Wajrakarur)

G/P7 (Venkatampalle)

P8/10 (Lattavaram)

P9/11 (Lattavaram)

CH-10 (Chigicherla)

P11/2A (Chigicherla)

WKF(C)

(B) NKF

(A) Siddanpalli

0

50

100

150

200

0 5 10 15 20

La/Sm

La/

Yb

(D)

Fig. 7 Chondrite-normalized (after Evensen et al. 1978) REEdistribution patterns for kimberlites of SKC (a) Narayanpet kimberlitefield (b) and Wajrakarur kimberlite field (c). Dashed line illustrates thehigh HREE in the SKC samples. La/Sm vs La/Yb (d) of the SKCcompared with those from WKF and NKF. Dashed curve represents apossible trend. Symbols are the same as in Fig. 4. The data for NKFand WKF are from Chalapathi Rao et al. (2004)

R

324 N.V. Chalapathi Rao et al.

Page 13

than those from the WKF and NKF of the Eastern DharwarCraton (Fig. 7a). Higher HREE abundances can be eitherinterpreted as: (i) derivation from relatively shallowermantle depths with lesser control of residual garnet; (ii)contamination by country rock granitoids; or (iii) due to theincorporation of a subducted crustal component in themantle source region (e.g. Le Roex et al. 2003; Becker andLe Roex 2006). As mantle-derived garnet xenocrysts arepresent in the SKC and contamination indices (C.I.)involving major elements are low, we would prefer therole of a possible subducted component in producing higherHREE contents. Boninites are high siliceous, magnesium-rich magmas widely accepted to have had been formed byhydrous melting of metasomatised mantle above subductionzones (e.g. Arndt 2003). The reported occurrence ofNeroarchaean boninitic magmas in the vicinity of theSKC (Manikyamba et al. 2005) provide evidence for thesubduction experienced in this domain. It has been amplydemonstrated that the subducted oceanic lithosphere canpersist up to depths of >2,000 km beneath the continentallithosphere (Grand et al. 1997) and inheritance of such longterm memories of such subduction by the continentallithosphere and are reflected in the subsequently eruptedmagmas in various cratons (e.g. Goodenough et al. 2002;Srivastava et al. 2009). The presence of negative depletionsat Nb, Hf and Ti in the multi-element spidergrams (above)further supports this proposal for subduction. This earlierepisode involving extraction of boninitic magma could alsohave contributed to their relatively low Ni, Cr and Cocontents than some of the EDC kimberlites. Kumar et al.(2007) have related the widespread kimberlite magmatism inthe Eastern Dharwar Craton at 1100Ma to a global period ofenhanced short lived plume activity and/ or major changeand re-organization of mantle convection regimes at thattime. However, the presence of subduction-related signaturesin the SKC samples suggest the role of mantle plume to bethat of a predominantly heat, but not melt, contributor.

0,00

5,00

10,00

15,00

20,00

25 30 35 40 45 50 55

SiO2 (wt%)

Nb

/Y

MORB

MelilititesOIB

(B)Field ofWKF & NKFkimberlites

1,5

2

2,5

3

3,5

4

2 4 6 8 10 12(La/Sm)N

(Ho

/Lu

) N

Kimberlites (WKF)

Kimberlites (NKF)

Siddanpalli kimberlites

(A)

Fig. 9 a (La/Sm)N vs (Ho/Lu)N of the SKC compared with those ofNKF and WKF. b Nb/Y vs SiO2 (wt%) plot illustrating the degree ofpartial melting in the kimberlites of this study. Data of WKF and NKFis from Chalapathi Rao et al. (2004) and Chalapathi Rao andSrivastava (2009). The other fields are taken from Beard et al.(1998). Symbols are the same as in Fig. 9a

Fig. 8 Primitive mantlenormalized (values from Sunand McDonough 1989) multi-element patterns for thekimberlites of SKC

Petrology, geochemistry and genesis of kimberlites from southern India: subduction related magmatic sources? 325

Page 14

Therefore a highly refractory mantle that was alreadydepleted in transition elements due to the extraction ofhigh-magnesian boninitic magmas could have influencedkimberlite source region(s). A recent study of the geo-chemistry of Group II kimberlites from southern Africarevealed that they have characteristics similar to calc-alkaline magmas and are related to lithospheric mantlemetasomatized melts or fluids associated with ancientsubduction events unrelated to mantle plume upwelling(Becker and Le Roex 2006) which also finds support fromsome plate-tectonic models (e.g. Helmstaedt and Gurney1984; McCandless 1999). Likewise, the transitional kim-berlites of southern Africa are inferred to be have beenderived from sources that has subducted- as well as- plumecomponents (Becker et al. 2007). Therefore, the geochem-ical characteristics displayed by the SKC rocks could bedue to long term, subduction-related modification experi-enced by the lithospheric mantle in that domain. Furtherstudies, involving Os and Hf isotopic systematics, areclearly required to test this suggestion.

Geochemical comparison of the SKC with South Africanmodels for kimberlite petrogenesis

In recent years a variety of models, based essentially onwhole-rock geochemistry, have been developed to constrainthe petrogenetic processes involved in the generation andevolution of kimberlites (e.g. Tainton and McKenzie 1994;Chalapathi Rao et al. 2004; Le Roex et al. 2003; Beckerand Le Roex 2006). We have compared the observed REEratios of kimberlites from the SKC, as well as those ofothers from the WKF and NKF fields (the latter data weretaken from Chalapathi Rao et al. 2004, and Chalapathi Raoand Srivastava 2009), with the melting trajectories ofinferred southern African Group I and II kimberlite sourceregions presented by Becker and Le Roex (2006), in aneffort to test whether the southern African model isapplicable for the EDC kimberlite samples.

Results presented in Fig. 10 show that a simple meltingtrajectory of assumed Group I kimberlite can account forthe observed REE compositions of the SKC. However, theSKC samples display a greater range in the degree ofmelting (2 % to 6 %) than those from the EDC and the on-craton occurrences from South Africa. A relatively greaterdegree of melting experienced by the SKC source region isalso additionally supported by the relationship betweenSiO2 and Nb/Y (Fig. 9).

Variation in the degree of partial melting of Group I andII kimberlite sources cannot explain the observed variationof all the EDC kimberlite samples (Fig. 10). Instead, acomplex interplay between Group I and II type kimberlitesources involving mantle heterogeneity is implied, similarto that invoked for the generation of transitional kimberlites

of southern Africa (Becker et al. 2007). This is suggestiveof the predominant derivation of the SKC melt fractionfrom a non-convective mantle source such as the sub-continental lithospheric mantle (SCLM), (e.g. Agashev etal. 2008; Francis and Patterson 2009). It may be recalledthat Haggerty and Birckett (2004, p. 543) also attribute thevariations in the EDC kimberlites to varying degrees ofpartial melting and mixing of fertile, depleted and enrichedsources.

High Ce and Ce/Pb contents of the SKC samples(Fig. 6c) are unlike that of a source similar to a convectivemelt like MORB. Phlogopite and carbonate, inferred in thisstudy to be essential melting assemblages in their mantlesource, are also known to be unstable at temperatures ofconvecting asthenospheric mantle (~1,480°C; McKenzieand Bickle 1988) and can persist only at the P-T conditionsof cold, continental lithospheric mantle (<1,400°C; e.g.,Ulmer and Sweeney 2002). This further constrains the SKCsource region to lie within the lithospheric portion of thecratonic mantle. Studies involving trace-element analysis ofthe garnet xenocrysts (e.g. Griffin et al. 1999) andradiogenic isotope systematics (e.g. Re-Os; Hf-Lu; Nowellet al. 2004) need to be undertaken to further evaluate thesource regions.

Conclusions

This study provides the first comprehensive petrologicaland geochemical information on the Siddanpalli kimber-

La/Sm

Gd

/Yb

0.1%0.5%

1%

2%

4%

10%

Source regions

Kimberleykimberlites(GrpI)

Swartrugens &Star kimberlites(Grp II)

Grp I kimberlite source

Grp ll kimberlite source

Eastern Dharwar cratonkimberlites(WKF and NKF)

20150

5 10

5

10

15

20

Fig. 10 La/Sm vs Gd/Yb for the kimberlites of study as well as thosefor the other kimberlites from the WKF and NKF (data taken fromChalapathi Rao et al. 2004; Chalapathi Rao and Srivastava 2009).Illustrated curves (from Becker and Le Roex 2006) represent meltingtrajectories of inferred Group I and II kimberlite source regions havingresidual mineralogy as follows: Grp I, ol:opx:cpx:gt =0.67:0.26:0.04:0.03; Grp II, ol:opx:cpx:gt = 0.67:0.26:0.06:0.01.Numbers shown represent the degree of melting. Fields for Kimberley(Grp I) and Swartrugens and Star (Grp II) kimberlites are from Beckerand Le Roex (2006)

326 N.V. Chalapathi Rao et al.

Page 15

lites. The SKC are the most magnesian, as well as highlysilica-undersaturated kimberlites amongst such rocksstudied so far from the Eastern Dharwar Craton. Theyalso possess the lowest granitoid and ilmenite contami-nation indices. The kimberlites of the SKC incorporateabundant carbonate xenoliths, presumably derived fromthe Proterozoic platformal cover that has now beeneroded. Despite their highly magnesian nature, the SKCpossess lower abundances of compatible elements (e.g.Ni, Cr, Co) compared to many other EDC kimberlites.On the other hand, their incompatible trace-elementconcentrations are similar or relatively much higher.Their La/Yb ratios (64–104) are considerably less thanthose of other EDC kimberlites (La/Yb=108–145) pri-marily owing to their high HREE abundances. As thereis no evidence of crustal contamination by country rockgranite, relatively higher HREE abundances and negativespikes at Nb, Hf and Ti on multi-element spidergramsare attributed to the influence of a subducted crustalcomponent in their mantle source region. A comparison ofthe REE geochemistry of the kimberlites of the SKC with theresults of REE based, semi-quantitative forward modelling ofbatch melting of southern African Group I and II kimberlitesource compositions, involving a metasomatized garnetlherzolite (Becker and Le Roex 2006), demonstrate that theSKC displays a wide range in the degree of melting thanthose in many kimberlites from the EDC and southernAfrica. The unusual geochemistry of the SKC, compared toother EDC kimberlites, may be explained by a combinationof factors involving: (i) higher degrees of partial melting; (ii)the involvement of a possible subducted component in theirmantle source region; and (iii) previous extraction ofboninitic magmas from their geological domain.

Acknowledgements V.Balaram and N.N.Murthy (NGRI, Hyderabad)are thanked for their help in the geochemical analyses of samples.Special thanks should go to V. Balaram for kindly arranging the re-analyses of trace element contents in the samples using the newkimberlite standards developed by his laboratory. Roy Eccles (Geolog-ical Survey of Canada) provided a very constructive review of an earlierversion of this manuscript. We are much thankful to the threeanonymous journal reviewers for their thorough reviews and helpfulcomments. Editorial suggestions by Kathryn Moore are greatlyappreciated. NVCR thanks Alexander von Humboldt Stiftung, Bonn,for awarding him a Humboldt Post Doctoral Fellowship during thetenure of which this version of the MS was finalized.

References

Agashev AM, Pokhilenko AP, Takazawa E, Macdonald JA, VavilovMA, Watanabe T, Sobolev NV (2008) Primary melting sequenceof a deep (>250km) lithospheric mantle as recorded in thegeochemistry of kimberlite-carbonatite assemblages, Snap Lakedyke system, Canada. Chem Geol 255:317–328

Arndt N (2003) Komatiites, kimberlites and boninites. J Geophys Res108(B6):2293. doi:10.1029/2002J13002157

Beard AD, Downes H, Hegner E, Sablukov SM, Vetrin VR, Balogh K(1998) Mineralogy and geochemistry of Devonian ultramaficminor intrusions of the southern Kola peninsula, Russia:implications for the petrogenesis of kimberlites and melilitites.Contrib Mineral Petrol 130:288–303

Becker M, Le Roex AP (2006) Geochemistry of South African on-and off-craton Group I and II kimberlites: petrogenesis andsource region evaluation. J Petrol 47:673–703

Becker M, Le Roex AP, Class C (2007) Geochemistry andpetrogenesis of South African transition kimberlites located onand off the Kaapvaal Craton. South Afr J Geol 110:631–646

Cas RAF, Porritt L, Pittari A, Hayman P (2008) A new approach tothe kimberlite facies terminology using a revised generalapproach to nomenclature for all volcanic rocks: descriptive togenetic. J Volcan Geotherm Res 174:226–240

Chalapathi Rao NV (1998) Light rare earth elements (LREE) inperovskite from kimberlites of Andhra Pradesh, India. J Geol SocInd 51:741–746

Chalapathi Rao NV, Srivastava RK (2009) Petrology and geochem-istry of diamondiferous Mesoproterozoic kimberlites fromWajrakarur Kimberlite Field, Eastern Dharwar Craton, SouthernIndia: genesis and constraints on mantle source regions. ContribMineral Petrol 157:245–265

Chalapathi Rao NV, Gibson SA, Pyle DM, Dickin AP (2004)Petrogenesis of Proterozoic lamproites and kimberlites from theCuddapah Basin and Dharwar Craton, southern India. J Petrol 45(5):907–948

Chalapathi Rao NV, Anand M, Dongre A, Osborne I (2009) Carbonatexenoliths hosted by the Mesoproterozoic Siddanpalli KimberliteCluster (Eastern Dharwar Craton): implications for the geo-dynamic evolution of southern India and its diamond anduranium metallogenesis. Int Jour Earth Sci. doi:10.1007/s00531-009-0484-7

Clement CR (1982) A comparative geological study of some majorkimberlite pipes in northern Cape and Orange Free State.Unpublished Ph.D. Thesis, University of Cape Town, SouthAfrica

Coe N, le Roex AP, Gurney JJ, Pearson G, Nowell G (2008)Petrogenesis of the Swartruggens and Star Group II kimberlitedyke swarms, South Africa: constraints from whole rockgeochemistry. Contib Mineral Petrol 156:627–652

Deer WA, Howie RA, Zussman J (1979) An introduction to therock forming minerals. ELBS Longman Group Ltd., London, p528 pp

Dongre A, Chalapathi Rao NV, Kamde G (2008) Limestone xenolithin Siddanpalli kimberlite, Gadwal granite-greenstone terrain,Eastern Dharwar Craton, Southern India: remnant of Proterozoicplatformal cover sequence of Bhima/Kurnool age? J Geol116:184–191

Evensen NM, Hamilton PJ, O’Nions RK (1978) Rare earth abundan-ces in chondritic meteorites. Geochim Cosmochim Acta42:1199–1212

Foley SF, Venturelli G, Green DH, Toscani L (1987) The ultra-potassic rocks: characteristics, classification and constraints forpetrogenetic models. Earth Science Rev 24:81–134

Francis D, Patterson M (2009) Kimberlites and aillikites as probes ofthe continental lithospheric mantle. Lithos 109:72–80

Goodenough KM, Upton BGJ, Ellam RM (2002) Long-term memoryof subduction processes in the lithospheric mantle: evidence fromthe geochemistry of basic dykes in the Gardar province of SouthGreenland. J Geol Soc Lond 159:705–714

Grand SP, van der Hilst RD, Widiyantoro S (1997) Global seismictomography: a snapshot of convection in the Earth. GSA Today7:1–7

Petrology, geochemistry and genesis of kimberlites from southern India: subduction related magmatic sources? 327

Page 16

Griffin WL, Doyle BJ, Ryan CG, Pearson NJ, O’Reilly SY, Davies R,Kivi K, Van Achterbergh E, Natapov LM (1999) Layered mantlelithosphere in the Lac de Gras area, Slave Craton: composition,structure and origin. J Petrol 40:705–727

Haggerty SE, Birckett T (2004) Geological setting and chemistry ofkimberlite clan rocks in the Dharwar Craton, India. Lithos76:535–554

Harris M, Le Roex AP, Class C (2004) Geochemistry of theUintiejsberg kimberlite, South Africa: petrogenesis of an off-Craton, group I kimberlite. Lithos 74:149–165

Helmstaedt HH, Gurney JJ (1984) Kimberlites of southern Africa- arethey related to subduction processes ? In: Kornprobst J (ed)Kimberlites: 1 Kimberlites and related rocks, Developments inpetrology, vol 11B. Elsevier, Netherlands, pp 424–438

Kumar A, Heaman LA, Manikyamba C (2007) Mesoproterozoickimberlites in south India: a possible link to ~1.1 Ga globalmagmatism. Precambrian Res 15:192–204

Le Roex AP, Bell DR, Davis P (2003) Petrogenesis of Group Ikimberlites from Kimberley, South Africa: evidence from bulk-rock geochemistry. J Petrol 44:2261–2286

Lynn M (2005) The discovery of kimberlites in the Gulburga andRaichur districts of Karnataka. Proc. Group Discussion onKimberlites and related rocks of India. Geol Soc India 48–49

Manikyamba C, Naqvi SM, Subba Rao DV, Ram Mohan M, KhannaTC, Rao TG, Reddy GLN (2005) Boninites from the Neo-archaean Gadwal greenstone belt, Eastern Dharwar Craton, India:implications for Archaean subduction processes. Earth Planet SciLett 230:65–83

Matin A (2001) Structure of the Gadwal schist belt, Eastern DharwarCraton, Mahbubnagar and Kurnool district, Andhra Pradesh.Indian J Geol 73:199–205

McCandless TE (1999) Kimberlites: mantle expressions of deep-seated subduction. In: Gurney JJ, Gurney JL, Pascoe MD,Richardson SH (eds) Proceedings of the VIIth Internationalkimberlite conference (PH Nixon volume), pp 545–549

McKenzie D, Bickle MJ (1988) The volume and compaction of meltgenerated by extension of lithosphere. J Petrol 29:625–679

Mitchell RH (1986) Kimberlites: mineralogy, geochemistry andpetrology. Plenum, New York, p 442 pp

Mitchell RH (1995) Kimberlites, orangeites and related rocks.Plenum, New York 410 pp

Nowell GM, Pearson DG, Bell DR, Carlson RW, Smith CB, KemptonPDM, Noble SR (2004) Hf isotope systematics of kimberlites andtheir megacrysts: new constraints on their source regions. J Petrol45:1583–1612

Ramadass G, Ramaprasada Rao IB, Srinivasulu N (2001) Gravityinferred subsurface structure of Gadwal schist belt, AndhraPradesh. J Earth System Sc 110:25–32

Ramam PK, Murthy VN, (1997) Geology of Andhra Pradesh.Geological Society of India Bangalore, 245 p.

Ravi S, Satyanarayana SV (2007) Discovery of kimberlites inChagapuram area, Mahaboobnagar district, Andhra Pradesh. JGeol Soc Ind 70:689–692

Roeder PL, Schulze DJ (2008) Crystallization of groundmass spinel inkimberlite. J Petrol 49:1473–1495

Rogers NW, Hawkesworth CJ, Palacz ZA (1992) Phlogopite in thegeneration of melilitites from Namaqualand, South Africa andimplications for element fractionation processes in the uppermantle. Lithos 28:347–365

Roy P, Balaram V, Kumar A, Sathyanarayan M, Gnaneshwara Rao T(2007) New REE and trace element data on two kimberliticreference materials by ICP-MS. Geostand Geoanal Res 31:261–273

Schulze DJ (2003) A classification scheme for mantle-derived garnetsin kimberlite: a tool for investigating the mantle and exploringfor diamonds. Lithos 71:195–213

Scott-Smith BH, Skinner EMW (1984) A new look at the PrairieCreek, Arkansas. In: Kornprobst J (ed) Kimberlites 1: Kimber-lites and related rocks. Proceedings of the Third InternationalKimberlite Conference, Developments, vol 1. In Petrology,Elsevier, Amsterdam, pp 255–283

Shivanna S, Srivastava SK, Nambiar AR (2002) Kimberlite occur-rence in Raichur area, Karanataka. J Geol Soc India 59:269–271

Sridhar M, Chowdhary VS, Nayak SS, Augustine PF (2004)Discovery of kimberlite pipes in Gadwal area, Mahbubnagardistrict, Andhra Pradesh. J Geol Soc India 63:95–99

Srinivasan KN (1990) Geology of the Veligallu and Gadwal schistbelts. Records of Geological Survey of India 123:Pt 5

Srivastava RK, Chalapathi Rao NV, Sinha AK (2009) Cretaceouspotassic intrusives with affinities to aillikites from Jharia area:magmatic expression of metasomatically veined and thinnedlithospheric mantle beneath Singhbhum craton. Eastern IndiaLithos. doi:10.1016/j.lithos.2009.05.005

Sun SS, McDonough WF (1989) Chemical and isotopic systematics ofoceanic basalts: implications for mantle composition and pro-cesses. In: Saunders AD, Norry MJ (eds) Magmatism in oceanbasins, vol 42. Geological Society of London Special Publica-tion, pp 313–345

Tainton KM, McKenzie D (1994) The generation of kimberlites,lamproites and their source rocks. J Petrol 35:787–817

Taylor WR, Tompkins LA, Haggerty SE (1994) Comparativegeochemistry of West African kimberlites: evidence for amicaceous kimberlite end member of sub lithospheric origin.Geochim Cosmochim Acta 58:4017–4037

Ulmer P, Sweeney RJ (2002) Generation and differentiation of GroupII kimberlites: constraints from a high-pressure experimentalstudy to 10 GPa. Geochim Cosmochim Acta 66:2139–215

328 N.V. Chalapathi Rao et al.