Author's personal copy

Geology, petrochemistry, and genesis of the bimodal lavas of Osham Hill,Saurashtra, northwestern Deccan Traps

Hetu C. Sheth a,⇑, Ashwini Kumar Choudhary b, Ciro Cucciniello c, Sudeshna Bhattacharyya a,Ramesh Laishram b, Trupti Gurav a

a Department of Earth Sciences, Indian Institute of Technology Bombay (IITB), Powai, Mumbai 400 076, Indiab Institute Instrumentation Center, Indian Institute of Technology Roorkee (IITR), Roorkee 248 001, Indiac Dipartimento di Scienze della Terra, Universitá di Napoli Federico II, Via Mezzocannone 8, 80134 Napoli (Naples), Italy

a r t i c l e i n f o

Article history:Received 26 May 2011Received in revised form 30 August 2011Accepted 5 September 2011Available online 16 September 2011

Keywords:Deccan TrapsSaurashtraOsham HillRhyolaitePitchstonePetrogenesis

a b s t r a c t

The Saurashtra region in the northwestern Deccan continental flood basalt province (India) is notable forcompositionally diverse volcano-plutonic complexes and abundant rhyolites and granophyres. A lavaflow sequence of rhyolite–pitchstone–basaltic andesite is exposed in Osham Hill in western Saurashtra.The Osham silicic lavas are Ba-poor and with intermediate Zr contents compared to other Deccan rhyo-lites. The Osham silicic lavas are enriched in the light rare earth elements, and have eNd (t = 65 Ma) valuesbetween �3.1 and �6.5 and initial 87Sr/86Sr ratios of 0.70709–0.70927. The Osham basaltic andesiteshave initial eNd values between +2.2 and �1.3, and initial 87Sr/86Sr ratios of 0.70729–0.70887. Large-ion-lithophile element concentrations and Sr isotopic ratios may have been affected somewhat by weath-ering; notably, the Sr isotopic ratios of the silicic and mafic rocks overlap. However, the Nd isotopic dataindicate that the silicic lavas are significantly more contaminated by continental lithosphere than themafic lavas. We suggest that the Osham basaltic andesites were derived by olivine gabbro fractionationfrom low-Ti picritic rocks of the type found throughout Saurashtra. The isotopic compositions, and thesimilar Al2O3 contents of the Osham silicic and mafic lavas, rule out an origin of the silicic lavas by frac-tional crystallization of mafic liquids, with or without crustal assimilation. As previously proposed forsome Icelandic rhyolites, and supported here by MELTS modelling, the Osham silicic lavas may have beenderived by partial melting of hot mafic intrusions emplaced at various crustal depths, due to heating byrepetitively injected basalts. The absence of mixing or mingling between the rhyolitic and basaltic andes-ite lavas of Osham Hill suggests that they reached the surface via separate pathways.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Most continental flood basalt (CFB) provinces of the world con-tain at least some silicic (rhyolite–dacite–trachyte) magmatism.These silicic magmas have been variably interpreted as products ofcrystal fractionation of mafic magmas, partial melting of underplat-ed mafic rocks or the deeper parts of the CFB lava pile, combinedassimilation and fractionation processes, or anatexis of the olderbasement crust (e.g., Lightfoot et al., 1987; Sheth and Ray, 2002; Mel-luso et al., 2008, 2009; McCurry et al., 2009; Cucciniello et al., 2011).

In the �65 million year old Deccan Traps flood basalt province,covering an area of 500,000 km2 in western and central India, rhyo-litic and trachytic rocks are locally abundant (e.g., Subba Rao, 1971).Their main outcrops are located in the Mumbai, Pavagadh, Rajpiplaand Chhota Udaipur areas, as well as the Saurashtra peninsula in thenorthwestern Deccan Traps (Fig. 1a) (e.g., Krishnamurthy and Cox,

1980; Lightfoot et al., 1987; Gwalani et al., 1993; Chatterjee andBhattacharji, 2001, 2004; Sheth et al., 2011a; Kshirsagar et al., inpress; Zellmer et al., in press). A small but significant rhyolite–pitchstone–basaltic andesite lava flow sequence is exposed atOsham Hill in Saurashtra (Fig. 1a). Maithani et al. (1996) have pre-sented some major and trace element (including rare earth element)data on the Osham rhyolites, but not the associated basaltic ande-sites. Isotopic data on Deccan rhyolitic rocks are scarce, and arenot available on the Osham rhyolites. Here we present a geochemi-cal, mineral chemical, and Nd–Sr isotopic study of the Osham silicicand mafic lavas. We present a model for their genesis, and discussthe petrogenetic implications of the data with particular referenceto the rhyolitic rocks in the Deccan Traps.

2. Geology of Osham Hill

The relatively flat and low-lying Saurashtra peninsula is coveredlargely by the Deccan lavas (Fig. 1b), except along its fringes whereTertiary and Quaternary sediments (limestone and alluvium) cover

1367-9120/$ - see front matter � 2011 Elsevier Ltd. All rights reserved.doi:10.1016/j.jseaes.2011.09.008

⇑ Corresponding author. Tel.: +91 22 25767264; fax: +91 22 25767253.E-mail address: hcsheth@iitb.ac.in (H.C. Sheth).

Journal of Asian Earth Sciences 43 (2012) 176–192

Contents lists available at SciVerse ScienceDirect

Journal of Asian Earth Sciences

journal homepage: www.elsevier .com/locate / jseaes

Author's personal copy

the Deccan rocks. In the northern part Mesozoic sedimentary rocksare exposed. No pre-Mesozoic rocks are known in Saurashtra fromoutcrops, borings, or xenoliths. Borings in northeastern Saurashtrahave encountered intercalated basalt and picritic flows and somepyroclastic deposits totalling �400 m thickness under the alluviumcover (West, 1958; Peng and Mahoney, 1995). The Saurashtra re-gion has several features that make it strikingly different fromthe main Deccan Plateau in west-central India, composed of thick,extensive flood tholeiites (e.g., Najafi et al., 1981). These featuresinclude Saurashtra’s various volcano-plutonic complexes, a greatcompositional diversity, and an abundance of rhyolite and grano-phyre (e.g., De and Bhattacharya, 1971; De, 1981; Melluso et al.,

1995; Sheth et al., 2011a; Kshirsagar et al., in press). Comparedto the our knowledge of the flood basalts of the Deccan Plateauand the Western Ghats escarpment (Fig. 1a), our knowledge ofDeccan magmatism in Saurashtra is very rudimentary. The presentcontribution on the Osham Hill lava sequence expands our knowl-edge base of Deccan magmatism in Saurashtra and its place in theoverall framework of Deccan flood volcanism.

Osham Hill, located in the Survey of India toposheet 41 K/6(1:50,000 scale), is a table mountain situated between the Girnarand the Barda-Alech complexes (Fig. 1b). The hill is a few kilome-ters in lateral extent, is surrounded by soil cover and Miliolite lime-stone, and exposes a �225 m thick Deccan lava sequence (Figs. 1c

80 m

314 m

Chichodvillage264 m

176 m

Patanvavvillage

Osham Hill

temple

staircase

21o 38'

21o 39'

70o 16' E 70o 17'

OSH1OSH3 ro

ad

21o 39'N

50 km

DECCANTRAPS

INDIA

Mumbai

ARABIANSEA

Saurashtra

Kachchh Gulf ofKachchh

ARABIAN SEA

Rajkot

CHOGAT-CHAMARDI

GIRNAR

OSHAM

BARDA ALECH

S A U R A S H T R A

CAM

BAY

RIF

T

Gul

f of C

amba

y (K

ham

bhat

)

Western G

hats

MesozoicDeccan Traps

Tertiary & Quaternary

Diu

Scarp

General contour outline

Cambay

Rajula

Palitana

Shihor

Pavagadh

Rajpipla

Junagadh

Dhandhuka

Mot

i Mar

ad 1

0 km

Dhor

aji 2

0 km

eastern spur

OSH2

OSH4

OSH6

OSH5

Miliolitelimestone

Basaltic andesite

Pitchstone

Rhyolite

Soil cover

Chhota Udaipur

b

a

c

300 kmTo

Uple

ta (1

5 km

)

500 m

Pachmarhi

Shahdol

Dhule

Bhuj

Wadhwan

Botad

Fig. 1. Simplified geological maps of the Deccan Traps (a) and of Saurashtra (b), with the important central complexes and locations in Saurashtra marked (after De, 1981), aswell as localities mentioned in the text. Dhandhuka, Botad, and Wadhwan are where basalts and picritic basalts have been encountered in boreholes (West, 1958). Ellipticalarea in (a) is the general northwestern Deccan region, with great compositional diversity. (c) Geological map of Osham Hill (from Maithani et al., 1996) showing the locationsof the samples of this study, and the Quaternary Miliolite limestone and soil cover.

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 177

Author's personal copy

and 2a). The hill’s summit (314 m) can be easily reached along astaircase leading to a Hindu temple near the top (Figs. 1c and2a). The uppermost flow is a thick (>100 m) rhyolite flow withcolumnar jointing and spectacular flow layering and flow folding(Fig. 2b–d; sample number OSH1 taken at the summit). Belowthe rhyolite flow is a thick (tens of meters) flow of black vitreouspitchstone (Fig. 2e and f; sample OSH2) which is partly layeredor banded at millimeter- to centimeter-scale (sample OSH3), aswell as locally folded, and contains small (millimeter-size) spheru-lites. By pitchstone we mean silicic volcanic glass which containsmuch more H2O (up to 6 wt.%) and crystalline material thanobsidian (<1 wt.% H2O), and which therefore has a waxy or resin-ous lustre unlike the vitreous lustre (and characteristic conchoidalfracture) of obsidian (Hatch et al., 1983; Kshirsagar et al., in press).

From the absence of basal tuffs, the rhyolite and pitchstoneflows appear to have been entirely subaerial, effusive eruptions.Wakhaloo (1967) interpreted the Osham rhyolite and pitchstoneas ignimbrites based on apparent vitroclastic textures which wehave not observed. We are aware that high-temperature rhyoliteignimbrites behave like lavas and can undergo strong and

spectacular rheomorphic deformation just like rhyolite lavas (e.g.,Wolff and Wright, 1981; Andrews and Branney, 2001). We havenot observed basal autobreccias that may distinguish rhyolite lavasfrom high-temperature ignimbrites (Henry and Wolff, 1992). Onthe other hand, the absence of typical features of ignimbrites, suchas pyroclastic or vitroclastic textures, fiamme, and gas elutriationpipes (Henry and Wolff, 1992) in the Osham Hill rhyolite andpitchstone flows does not support their ignimbritic origin. Wetherefore consider the Osham rhyolite and pitchstone to be lavaflows. Open, isoclinal flow folds observed in the rhyolite (e.g.,Fig. 2b and c) may have developed in hot, low-viscosity lava duringthe early stages of its emplacement, and angular folds (Fig. 2d) la-ter, in cooler, higher-viscosity lava. The pitchstones may be earliereruptive lava units than the overlying rhyolite, or may representthe lower chilled base of the rhyolite lava flow. Highly weatheredand amygdaloidal basaltic andesite (sample OSH4) is exposed atthe base of the Osham Hill, below the rhyolite and pitchstoneflows. From the outermost hillock of �176 m height on the easternspur of the Osham Hill (Figs. 1c and 2a), near Chichod village, wesampled two fine-grained, mesocratic lava flows (OSH5, OSH6).

Fig. 2. Field geology of the Osham Hill. (a) Panoramic view of Osham Hill, looking approximately southwest. Note prominent ‘‘bench’’ formed by upper, gently dippingrhyolite flow. (b) Open folds in rhyolite at summit; sample OSH1 was taken here. Pen is 15 cm long. (c) Small fold in larger fold. (d) Chevron folds in rhyolite. Pen is 15 cmlong. (e) Contact (white dashed line) between the rhyolite above and pitchstone below. (f) A hand specimen of the Osham pitchstone. White spots are spherulites, and thescale is in centimeters.

178 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192

Author's personal copy

Maithani et al. (1996) have depicted rhyolites capping this easternspur.

No intrusive igneous activity is observed within the Osham Hill,but a small (0.5 km long and low-lying) outcrop of fresh olivinegabbro is found at the village of Moti Marad, 10 km northeast ofOsham Hill on the road to Dhoraji town (Fig. 1c). This has no

obvious relationship to the Osham Hill, from which it is separatedby soil cover. Major and trace element data and spinel composi-tions for this gabbro are available (sample D30 of Melluso et al.,2010a). From the northeasterly trend of this olivine gabbro bodywe consider it as one of the many arcuate dykes surrounding themain intrusive focus of the Girnar complex (Auden, 1949). Gabbros

Fig. 3. Petrographic features of the Osham Hill lavas. Sample numbers are marked on individual panels. (a) A train of microspherulites (sph) in rhyolite OSH1. (b) Perliticcracks in pitchstone OSH2. (c) Flow layers, microspherulite (sph), and perlitic cracks in banded pitchstone OSH3. Note how the perlitic cracks run across the layering. (dthrough h) Features of the basaltic andesites. ol is olivine, cpx is clinopyroxene, and pl is plagioclase. Note in (f) a prominent layer with concentrated Fe-Ti oxides (tiny blackgrains) running from the upper right to the lower left. Photomicrographs (a), (b), (c) and (f) were taken over the polarizer, and the rest between crossed polars.

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 179

Author's personal copy

are also known to outcrop in the low grounds between Osham Hilland Mount Girnar (Wakhaloo, 1967).

The Osham rhyolite and pitchstone have important counterpartsin the Deccan, namely rhyolites in the Rajula–Palitana–Shihor–Cho-gat–Chamardi strip in eastern Saurashtra (Chatterjee and Bhatta-charji, 2001, 2004; Sheth et al., 2011a; Kshirsagar et al., in press),rhyolites in the Barda and Alech complexes in western Saurashtra(De and Bhattacharya, 1971; De, 1981), and rhyolites in the Mumbai,Rajpipla, and Pavagadh areas (Lightfoot et al., 1987; Krishnamurthyand Cox, 1980; Sheth and Melluso, 2008) (Fig. 1a and b).

3. Petrography and mineral chemistry

Fig. 3 shows characteristic textures of the rock samples col-lected from the Osham Hill lava sequence (the rock names usedhere are based on collective field, petrographic, and whole-rockgeochemical criteria discussed below). The upper rhyolite (OSH1)is very fine grained and aphyric, and shows a microscopic inter-growth of quartz and alkali feldspar. Devitrification features(spherulites) are common (Fig. 3a). Magnetite is a ubiquitous phasein the groundmass, and zircon is a common accessory mineral andmay display compositional zoning (Fig. 4). The pitchstones, bothnon-banded (OSH2) and banded (OSH3) varieties, contain abun-

dant perlitic cracks due to hydration and tiny scattered spherulites(Lofgren, 1971; Kshirsagar et al., in press) up to a centimeter across(Fig. 3b and c); the latter also can be seen in hand specimen(Fig. 2f). The basaltic andesite OSH4 (Fig. 3d) is highly weatheredand shows phenocrysts of clinopyroxene and plagioclase, and a fi-ner groundmass essentially overprinted by secondary zeolites andcalcite. The basaltic andesite sample OSH5 from the eastern spur isvery fine-grained and almost aphyric, with sparse micro-phenocrysts of clinopyroxene and highly altered olivine (Fig. 3e).Basaltic andesite sample (OSH6) from this location shows sub-hor-izontal, millimeter-thick laminations in hand specimen. Under themicroscope the layering is seen to be produced by abundant tinygrains of Fe–Ti oxides concentrated along parallel planes (Fig. 3f).Sample OSH6 is weakly porphyritic with microphenocrysts of pla-gioclase and clinopyroxene set in a fine-grained groundmass ofplagioclase, clinopyroxene, pigeonite, opaque oxides and rare alkalifeldspars (Fig. 3g). The plagioclase microphenocrysts show intrigu-ing zoning patterns (Fig. 3g and h).

The minerals were analysed using an Oxford InstrumentsMicroanalysis Unit equipped with an INCA X-act detector and aJEOL JSM-5310 microscope at the CISAG (Centro Interdipartimen-tale Strumentazioni per Analisi Geomineralogiche), University ofNaples. An accelerating voltage of 15 kV and a filament currentof 50–100 mA was used for all the analyses. Measuring times were50 s. For calibration both natural and synthetic standards wereused (for full details see Melluso et al., 2010b).

Augite (Ca36Mg46Fe18 to Ca28Mg43Fe29; Mg# = 72–60) andpigeonite (Ca8Mg57Fe35 to Ca13Mg34Fe53; Mg# = 62–39) are thetwo types of pyroxenes observed in the basaltic andesite OSH6 (Ta-ble 1, Fig. 5a). TiO2 and Al2O3 contents of augites range from 0.2 to0.9 wt.% and from 1.3 to 3.0 wt.%, respectively (Table 1). Tempera-ture values based on two-pyroxene geothermometry (Lindsley,1983) range from �1100 �C to �1000 �C. Similar temperatureshave been obtained with whole-rock geothermometers (e.g., Groveand Juster, 1989; Toplis and Carroll, 1995).

Plagioclase phenocrysts in basaltic andesite range from An58 toAn46 (Table 2, Fig. 5b). Both normal and reverse zoning have beenobserved (Table 2). The iron content (as FeOt) ranges from 0.4 to1.9 wt.% (Table 2). Similar values have been also found in otherDeccan volcanic rocks (Melluso and Sethna, 2011). K-feldspar israre and occurs in the groundmass (Table 2, Fig. 5b).

Alkali feldspar in rhyolite OSH1 is sanidine with compositionbetween An4Ab42Or53 and An2Ab37Or61 (Table 2, Fig. 5b). The com-positional range is restricted, when compared to the range of alkalifeldspars of other evolved rocks of the Deccan Traps (Melluso and

Table 1Representative chemical analyses (in wt.%) of Osham pyroxenes.

Sample SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O Cr2O3 V2O3 Sum Ca Fe� Mg Mg#

OSH6 gm 50.08 0.25 2.11 17.96 0.58 15.37 14.20 0.30 0.05 0.11 101.01 28.4 28.9 42.7 60OSH6 gm 49.41 0.56 1.71 21.15 0.56 10.39 15.29 0.52 0.03 0.23 99.85 32.7 36.3 31.0 46OSH6 gm 49.40 0.48 2.13 20.06 0.55 12.46 15.43 0.32 0.00 0.19 101.01 31.6 32.9 35.5 52OSH6 gm 49.60 0.58 1.79 22.00 0.55 12.06 14.17 0.34 0.04 0.21 101.34 29.2 36.3 34.5 49OSH6 gm 50.45 0.49 1.30 23.60 0.54 10.14 14.31 0.69 0.00 0.06 101.59 30.3 39.9 29.8 43OSH6 gm 49.88 0.88 1.79 16.14 0.31 13.86 17.27 0.17 0.31 0.01 100.62 35.0 26.0 39.0 60OSH6 core 50.83 0.66 2.99 12.78 0.30 14.79 17.60 0.20 0.00 0.21 100.35 36.4 21.1 42.5 67OSH6 rim 50.31 0.32 2.61 11.48 0.16 14.86 18.00 0.14 0.08 0.25 98.20 37.7 19.0 43.3 69OSH6 core 52.57 0.21 2.06 11.01 0.29 16.28 18.08 0.32 0.02 0.02 100.86 36.5 17.8 45.7 72OSH6 gm 47.62 0.55 0.86 32.82 0.61 11.40 5.09 0.12 0.03 0.00 99.10 10.8 55.5 33.7 38OSH6 gm 51.11 0.21 1.04 25.78 0.46 18.59 5.20 0.32 0.00 0.00 102.72 10.1 39.7 50.2 56OSH6 gm 49.46 0.34 0.73 30.62 0.70 11.43 5.78 0.06 0.00 0.00 99.13 12.5 53.0 34.5 39OSH6 gm 51.19 0.16 0.99 27.50 0.56 14.76 5.12 0.11 0.14 0.00 100.52 10.8 46.1 43.2 48OSH6 gm 50.59 0.04 0.69 29.09 0.56 12.94 5.34 0.10 0.00 0.00 99.36 11.5 49.8 38.7 44OSH6 gm 53.51 0.15 1.28 21.63 0.55 20.18 4.13 0.26 0.00 0.21 101.91 8.3 35.0 56.7 62

Note: OSH6, basaltic andesite; gm = groundmass. Ca.Fe� (Fe + Mn) and Mg in mol%.Mg# = 100 �Mg/(Mg + Fe + Mn). Analytical uncertainties on the values are within 1%.

Fig. 4. Back-scattered electron (BSE) image of the groundmass of Osham rhyolitesample OSH1: zrn is zircon, mgt is magnetite, and alk feld is alkali feldspar.

180 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192

Author's personal copy

Sethna, 2011; Fig. 5b). The iron content is relatively low (<0.6 wt.%;Table 2).

Opaque oxides are ilmenite and magnetite in the basaltic andes-ite (Table 3, Fig. 6). Ilmenites have low Al2O3 (0.3–0.4 wt.%) andMgO (1.4–2.2 wt.%) contents. Magnetites have variable TiO2 (3.3–16.5 wt.%; 50–10 mol% ulvöspinel) and low Al2O3 (<2.2 wt.%) con-tents (Table 3). The only Fe–Ti oxide observed in the rhyolite OSH1is magnetite. It has low ulvöspinel (<7 mol%) and MgO (<0.6 wt.%)contents (Table 3).

4. Whole-rock chemical and isotopic compositions

Major and trace elements (including rare earth elements) in theOsham rock samples were analyzed on an inductively coupledplasma atomic emission spectrometer (ICP-AES, instrument: JobinYvon Ultima-2) at the Department of Earth Sciences, Indian Insti-tute of Technology Bombay. Nd-Sr isotopic analyses were per-formed using a fully automatic Thermo Fisher TRITONmulticollector thermal ionization mass spectrometer (TIMS), lo-cated at the National Facility for Isotope Geology and Geochronol-ogy, Indian Institute of Technology Roorkee. Sample dissolutionprocedures and other analytical details are as described by Shethet al. (2011a). The elemental and isotopic data are reported in Ta-bles 4 and 5, respectively.

4.1. Major and trace elements

The pitchstone OSH2 and the basaltic andesite OSH4 have highvalues of LOI (loss on ignition), reflecting their weathered and hy-drated nature. Table 4 also presents the CIPW norms for the Oshamrocks computed using the SINCLAS program (Verma et al., 2002)which recalculates the major oxide data on a LOI-free basis, com-putes the Mg Number and provides a rock name consistent withthe IUGS Subcommission on the Systematics of Igneous Rocks (LeBas et al., 1986). We used the option of Middlemost (1989), offeredby this program, for dividing total iron into Fe+2 and Fe+3. All rocksare subalkalic and quartz-normative in character, the amount of

1000 o C1100 o C

diopside

Anorthite

Albite Orthoclase

sanidineanorthoclase

oligoclase

andesine

labradorite

bytownite

OSH6OSH1Chogat-Chamardi silicic rocksChogat-Chamardi mafic rocks

a

b

Mg2Si2O6 Fe2Si2O6

CaMgSi2O6 CaFeSi2O6

ferrosiliteenstatite

pigeonite

augite

hedenbergite

1200 o C

900 o C

Fig. 5. (a) Pyroxene compositions projected in the Ca–Mg–Fe diagram for Osham hilllavas (dark grey diamonds). The pyroxenes of the Chogat–Chamardi complex (Shethet al., 2011a) are shown as open circles (silicic rocks) and open squares (mafic rocks).The pyroxene compositions of Deccan volcanic rocks (Melluso and Sethna, 2011 andreferences therein, and L. Melluso, unpubl. data) are also shown: asterisks aretholeiitic basalts, dark grey dashes are mildly alkaline basalts, and black crosses areevolved tholeiitic rocks. Temperature curves (dashed black lines) of Lindsley (1983)are also shown. (b) Feldspar compositions observed in the Osham Hill lavas. Feldsparcompositions of Deccan volcanic rocks and the Chogat-Chamardi complex (Shethet al., 2011a) are also shown. Symbols and data sources are as in Fig. 5a.

Table 2Representative chemical analyses (in wt.%) of Osham feldspars.

Sample SiO2 Al2O3 FeO CaO Na2O K2O BaO SrO Sum An Ab Or

OSH6 core 53.15 28.21 1.89 11.06 4.60 0.35 – 0.27 99.54 55 42 2OSH6 rim 53.33 28.45 1.16 11.07 4.78 0.34 – 0.21 99.33 55 43 2OSH6 gm 54.57 28.64 1.48 11.14 4.74 0.30 0.12 – 101.00 55 43 2OSH6 gm 56.74 27.20 1.06 9.31 5.96 0.33 0.08 – 100.67 45 53 2OSH6 gm 68.92 17.45 0.78 0.98 0.12 13.74 0.09 0.02 102.09 6 1 93OSH6 gm 53.61 28.91 1.29 11.61 4.98 0.23 – – 100.63 56 43 1OSH6 gm 54.08 29.12 1.18 11.42 4.67 0.30 0.10 0.43 101.29 56 42 2OSH6 core 56.94 26.77 0.51 9.65 5.92 0.28 – 0.63 100.69 46 53 2OSH6 rim 53.68 29.25 1.60 12.05 4.71 0.28 0.05 0.54 102.16 57 42 2OSH6 rim 53.59 29.08 1.44 12.23 4.67 0.29 – 0.65 101.96 57 41 2OSH6 core 57.00 27.92 0.38 10.84 5.46 0.45 0.10 0.19 102.33 51 47 3OSH6 rim 54.37 28.80 1.15 12.34 4.86 0.32 0.33 0.58 102.75 56 42 2OSH6 rim 55.87 28.65 0.71 10.91 5.29 0.23 – 0.41 102.06 52 47 1OSH6 core 54.30 29.30 0.87 12.38 4.61 0.32 0.19 0.43 102.41 58 40 2OSH6 rim 54.16 28.27 0.99 11.60 4.70 0.37 – 0.30 100.40 56 42 2OSH6 gm 53.65 27.68 1.20 11.38 4.63 0.31 – 0.13 98.98 56 42 2OSH1 gm 67.48 18.60 0.50 0.32 3.84 10.89 0.15 – 101.77 2 34 64OSH1 gm 67.15 18.45 0.29 0.42 4.17 10.43 0.26 – 101.16 2 37 61OSH1 gm 66.89 19.02 0.26 0.62 4.42 10.31 – 0.14 101.67 3 39 59OSH1 gm 67.11 18.74 0.41 0.37 3.85 11.13 0.16 0.76 102.54 2 35 63OSH1 gm 68.00 18.47 0.31 0.55 3.94 9.78 0.11 0.21 101.37 3 37 60OSH1 gm 68.05 18.61 0.31 0.64 4.55 9.80 – 0.10 102.06 3 40 57OSH1 gm 66.03 18.35 0.18 0.37 4.13 10.59 0.01 0.39 100.04 2 37 61OSH1 gm 65.99 18.61 0.25 0.88 4.69 9.27 0.17 0.39 100.26 4 42 54OSH1 gm 66.01 17.38 0.33 0.23 3.52 10.75 – 0.44 98.67 1 34 65OSH1 gm 66.52 17.99 0.56 0.39 4.68 9.60 0.20 0.54 100.48 2 42 56

Note: OSH6, basaltic andesite; OSH1, rhyolite; gm = groundmass. An, Ab and Or in mol%. Analytical uncertainties on the values are within 1%.

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 181

Author's personal copy

normative quartz varying from 7.87 (OSH5) to 38.20 wt.% (OSH2).Based on the LOI-free recalculated silica and alkali values, samplesOSH1, OSH2 and OSH3 are named as rhyolites, samples OSH4 andOSH5 as basaltic andesites, and sample OSH6 as an andesitethough its recalculated SiO2 value (57.57 wt.%) only marginally ex-ceeds the lower SiO2 limit for andesites on the total alkali-silicadiagram (Le Bas et al., 1986; Fig. 7). Its petrographic features aresimilar to those of the basaltic andesites OSH4 and OSH5, and itlacks features of typical andesites such as amphibole. We thereforeprefer to name it basaltic andesite. Data for the Deccan ‘‘flood ba-salt’’ lavas of the Western Ghats region, which are basalts andbasaltic andesites in composition, are also shown in Fig. 7. Alsoplotted are data for rhyolite-bearing rock suites from various areasof the Deccan (Saurashtra, Mumbai, Rajpipla, and Pavagadh). Therarity of intermediate (andesitic) rock compositions in the DeccanTraps is evident.

The Osham basaltic andesites have low TiO2 contents (0.62–1.17 wt.%), and are similar to the low-Ti tholeiites described fromSaurashtra by Melluso et al. (1995, 2006). Note that these low

TiO2 concentrations cannot be caused by alteration, as Ti is analteration-resistant high-field-strength (HFS) element (e.g., Rollin-son, 1993).

The basaltic andesite sample OSH4 also has a very low Zr contentof 46 ppm compared to the range in the Deccan tholeiites (generally100–250 ppm, but as low as 79 ppm, Beane, 1988). We note thatsome southern Saurashtra low-Ti tholeiites analyzed by Mellusoet al. (1995) also have very low Zr contents of 55–60 ppm. Zr isnotably also a HFS element resistant to alteration and weathering,and Ba is moderately so (e.g., Mahoney et al., 2000). We have plot-ted the SiO2 contents of various Deccan rhyolitic rocks vs. their Zrcontents in Fig. 8a, and Ba/Zr ratios vs. Zr contents in Fig. 8b forcomparison. We find that the Osham data overlap with data forthe rhyolitic suites of Saurashtra, whereas data for the Pavagadhand Rajpipla rhyolites partly do so. The Mumbai rhyolites are how-ever quite distinct and well separated from all other Deccan rhyo-lites, with substantially higher Zr contents.

Chondrite-normalized rare earth element (REE) patterns for theOsham rhyolites and pitchstones are shown in Fig. 9a. The patternsdepict relative enrichment of the light rare earth elements (LREE)over the heavy rare earth elements (HREE). The patterns for the rhy-olite and pitchstone samples analyzed (OSH1, 2, 3) approximatelycoincide with the patterns of the Osham rhyolites studied by Maith-ani et al. (1996). However, the patterns of the rhyolites of Maithaniet al. (1996) show large, anomalous terbium peaks and dysprosiumtroughs, which are absent in the smooth patterns observed for sam-ples OSH1, 2 and 3. Both sets of patterns nonetheless show largenegative europium anomalies. In Fig. 9b, the REE pattern of theOsham rhyolite (OSH1) is compared to patterns of other Deccanrhyolites. The Osham rhyolite has lower REE concentrations thanthe Mumbai rhyolites (Lightfoot et al., 1987), but higher REE con-centrations (except Eu) than the eastern Saurashtra rhyolites (Chat-terjee and Bhattacharji, 2001). The Osham and Mumbai rhyolites’patterns are however parallel and are all LREE-enriched, havingsmall to large negative europium anomalies. The Osham Hill basal-tic andesites are also LREE-enriched (Fig. 9c), though they havemuch lower REE concentrations than the Osham rhyolites.

4.2. Nd-Sr isotopic ratios

The neodymium and strontium isotopic ratios (Table 5) havebeen age-corrected to 65 Ma. Initial 87Sr/86Sr ratios of the OshamHill lavas are all higher than 0.7071 but lower than 0.7093. Initial

Table 3Representative chemical analyses (in wt.%) of Osham oxide phases.

Sample TiO2 Al2O3 Fe2O3 FeO MnO MgO Cr2O3 NiO V2O3 Sum Ulv Ilm

MagnetiteOSH6 gm 3.30 2.21 60.49 34.00 0.16 0.65 0.21 – 1.10 102.13 10.0 –OSH6 gm 5.72 1.97 56.13 36.08 0.46 0.55 – – 1.26 102.17 17.1 –OSH6 gm 5.02 0.49 58.03 31.00 0.40 2.37 – 0.34 0.34 98.00 13.1 –OSH6 gm 16.54 1.27 33.78 43.93 0.71 1.01 – – 1.26 98.50 49.5 –OSH6 gm 6.71 2.19 50.37 35.56 0.26 0.63 – – 1.40 97.12 21.5 –OSH6 gm 10.20 1.82 45.2 39.27 0.63 0.31 0.11 0.24 0.90 98.70 31.70 –OSH1 gm 2.30 0.20 63.29 30.12 1.77 0.48 – 0.15 – 98.30 6.27 –OSH1 gm 2.18 0.18 63.59 30.11 1.63 0.54 0.07 0.02 0.10 98.41 5.95 –OSH1 gm 2.41 0.43 63.26 30.59 1.60 0.50 – 0.12 – 98.90 6.62 –OSH1 gm 1.71 0.44 64.68 30.42 1.59 0.28 0.07 – – 99.20 4.77 –OSH1 gm 1.44 0.38 64.92 30.15 1.63 0.17 0.05 0.30 – 99.05 4.01 –OSH1 gm 1.68 0.45 64.69 29.95 1.51 0.55 – 0.15 0.01 99.00 4.61 –OSH1 gm 0.03 0.52 67.87 30.58 0.37 – 0.03 – – 99.40 0.07 –

IlmeniteOSH6 gm 43.59 0.36 16.09 35.83 0.55 1.58 0.05 – 1.06 99.10 – 83.8OSH6 incl 40.38 0.32 23.60 32.24 0.44 2.04 – – 0.49 99.50 – 76.3OSH6 gm 40.26 0.32 23.73 31.83 0.45 2.21 – – 1.29 100.08 – 76.1OSH6 gm 44.28 0.25 13.85 36.84 0.49 1.40 0.01 – 1.07 98.20 – 86.0

Note: OSH6, basaltic andesite; OSH1, rhyolite; gm = groundmass; incl = inclusion. Ulv = ulvöspinel (mol%); Ilm = ilmenite (mol%). Analytical uncertainties on the values arewithin 1%.

OSH6OSH1

Chogat-Chamardi silicic rocksChogat-Chamardi mafic rocks

Fig. 6. Fe–Ti–Mn–Mg (atomic) diagram for the oxide minerals in the Osham Hilllavas. Symbols are as in Fig. 5.

182 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192

Author's personal copy

eNd values range from +2.2 (basaltic andesite OSH4) to �6.5 (rhyo-lite OSH1). The basaltic andesites (OSH4, 5, 6) have higher initialeNd values (+2.2 to �1.3) than the rhyolite and pitchstones, thoughthey have the same range of Sr isotopic ratios.

Sr isotopic ratios are known to change, often significantly, asa result of subaerial weathering, whereas Nd isotopic ratios areunaffected by weathering (e.g., Mahoney et al., 2000; Shethet al., 2011b). Thus, rhyolite OSH1 and banded pitchstoneOSH3 have initial Nd isotopic ratios identical within error (of�6.5 and �6.2, respectively). Their corresponding Sr isotopic ra-tios are 0.70895 and 0.70709, suggesting that the higher ratio ofthe former may be a result of weathering. The Nd isotopic dataare much more reliable noting the alteration present in therocks, and these help resolve the question of whether the pitch-stones are the lower part of the top rhyolite flow, or separatefrom it. The initial eNd values for the rhyolite OSH1 (�6.5), andthe pitchstone (OSH2) immediately below it (�3.1) are distinctlydifferent, and therefore the pitchstone cannot represent thechilled base of the overlying rhyolite but must represent an old-er eruptive flow unit genetically unrelated to and independent of

the rhyolite. Also, the pitchstone represented by sample OSH2and the banded pitchstone represented by sample OSH3 aretwo geochemically and petrogenetically distinct eruptive units.

5. Petrogenesis

5.1. The Osham basaltic andesites

Low-Ti picrites, picritic basalts, and tholeiites (with TiO2 < 1.8wt.%) are widespread in the Saurashtra region, as are high-Ti pi-crites and picritic basalts (Melluso et al., 1995, 2006; Peng andMahoney, 1995). The high-Ti lavas of Saurashtra (and Pavagadh)have ocean-island-basalt-like geochemical characteristics andmay have been derived from clinopyroxene-rich mantle sources(Melluso et al., 1995; Peng and Mahoney, 1995). On the otherhand, the low-Ti picritic basalts of Saurashtra were modelled byMelluso et al. (2006) with non-modal fractional melting of dom-inantly spinel peridotite with some contribution from garnetperidotite.

Table 4Major oxide (wt.%) and trace element (ppm) data and wt.% CIPW norms (in italics) for the Osham Hill lavas.

Rock type Rhyolite Rhyolite (pitchstone) Rhyolite (Banded pitchstone) Basaltic andesite Basaltic andesite Basaltic andesite Ref. Meas.Sample OSH1 OSH2 OSH3 OSH4 OSH5 OSH6 BCR-2 BCR-2

SiO2 71.60 71.80 73.78 51.90 50.63 55.82 54.10 55.48TiO2 0.12 0.10 0.13 0.62 0.89 1.17 2.26 2.29Al2O3 13.49 12.17 13.62 13.66 13.16 12.62 13.50 13.73Fe2O3T 2.96 2.76 2.94 12.92 13.46 13.78 13.80 12.88MnO 0.02 0.05 0.08 0.21 0.21 0.21 0.196 0.19MgO 0.15 0.07 0.16 4.47 7.50 4.70 3.59 3.43CaO 0.01 0.08 0.56 5.31 8.59 6.74 7.12 7.54Na2O 3.04 3.89 3.91 3.08 1.79 2.13 3.16 3.42K2O 6.11 3.14 4.55 0.10 0.03 0.65 1.75 1.92P2O5 0.02 0.02 0.02 0.06 0.14 0.19 0.35 0.32LOI 1.34 5.18 2.90 6.73 2.89 1.70Total 98.85 99.26 102.65 99.06 99.29 99.71Mg# 12.7 6.74 13.5 46.5 58.4 47.0

Q 30.00 38.20 31.23 10.95 7.87 16.59Or 37.10 19.76 27.01 0.65 0.18 3.96Ab 26.43 35.06 33.24 28.54 15.89 18.59An – 0.28 2.66 25.35 29.14 23.67C 1.93 2.42 1.30 – – –Di – – – 2.55 11.96 7.91Hy 2.90 2.70 2.94 26.15 28.48 21.59Mt 1.37 1.32 1.33 4.36 4.35 4.93Il 0.23 0.20 0.25 1.29 1.77 2.29Ap 0.02 0.05 0.05 0.15 0.34 0.45

BCR-2 BCR-2Ba 22.5 12.0 37.2 359 79.8 143 683 668Sr 12.0 18.0 24.4 1525 108 135 346 317Pb 16 9 14 nd nd nd 11 12

BHVO-2 BHVO-2Zr 436 410 437 46.0 101 162 172 158Y 136 148 137 21.4 28.6 44.6 26 24.2

GSP-2 GSP-2La 104.3 125.4 110.8 6.41 8.71 12.1 180 156.0Ce 249.0 271.6 253.7 6.54 19.2 28.5 410 401.3Pr 21.1 25.5 23.0Nd 89.3 107 94.4 4.98 10.2 13.6 200 171Sm 21.2 25.7 22.7 1.47 2.97 3.81 27 28.5Eu 0.36 0.44 0.40 0.28 0.63 0.80 2.3 2.02Gd 15.5 19.4 16.6 1.40 2.84 3.54 12 9.47Tb 1.92 2.51 2.18 0.41 0.48 0.50Dy 16.7 19.8 18.8 2.71 4.02 4.70 6.1 5.09Ho 3.51 4.00 3.89 0.51 0.81 1.00 1.0 1.00Er 9.83 11.4 10.9Tm 1.35 1.48 1.45Yb 8.85 9.83 9.37 1.45 2.16 2.61 1.6 1.00Lu 1.33 1.48 1.36

Notes: ‘‘Ref.’’ are the recommended values (Wilson, 1997, 1998, 2000) and ‘‘Meas.’’ the measured values on USGS reference materials. nd = not detectable. Mg# = [atomic Mg/(Mg + Fe2+)] � 100, where Fe2+ and Fe3+ are computed using the Middlemost (1989) criteria, as well as CIPW norms and rock type names, using the SINCLAS program.

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 183

Author's personal copy

The low-Ti tholeiites of Saurashtra (Melluso et al., 1995), whichthe Osham basaltic andesites resemble, were probably derived byfractional crystallization of the low-Ti picritic basalts involvingolivine, clinopyroxene and plagioclase, and this was accompaniedby some incorporation of continental crust or enriched mantle(Melluso et al., 2006). Such a mechanism involving olivine gabbrofractionation is consistent with the petrographic assemblages ob-served in the Osham basaltic andesites, their low magnesium num-bers (46–58), and the small negative europium anomalies in theirchondrite-normalized rare earth element patterns.

Nd–Sr isotopic ratios of the Saurashtra low-Ti rocks are alsobroadly similar to the isotopic ratios of the Osham basalticandesites. The low-Ti rocks have eNdt = +2.1 to �2.1 (with onesample having eNdt as low as �6.3), and (87Sr/86Sr)t values rangefrom 0.70658 to 0.70976 (Melluso et al., 2006). The Osham basaltic

andesites have eNdt = +2.2 to �1.3 and (87Sr/86Sr)t = 0.70729–0.70887. In particular, Saurashtra low-Ti picritic basalt D130 withMgO = 12.40 wt.%, Sr = 243 ppm, eNdt = +2.1, and (87Sr/86Sr)t =0.70706 (Melluso et al., 2006) is an appropriate choice as a parentalmagma of the Osham basaltic andesites. Major element mass bal-ance calculations using the program XLFRAC (Stormer and Nicholls,1978) show that the basaltic andesite OSH6 can be derived by 81%fractional crystallization of the Saurashtra low-Ti picritic basaltD130, the fractionation assemblage comprising olivine (23.3%),clinopyroxene (30.9%), plagioclase (42.3%) and magnetite (3.4%)(Table 6). A similar result was obtained with fractional crystalliza-tion calculations using the MELTS code of Ghiorso and Sack (1995),which show that at under anhydrous conditions at the QFM oxygenbuffer and low pressure (3 kbar), the basaltic andesite OSH6 is pro-duced by 69% fractional crystallization of picritic basalt D130, the

Table 5Sr and Nd isotopic and isotope dilution data for the Osham Hill lavas.

Rhyolite Pitchstone Banded pitchstone Basaltic andesite Basaltic andesite Basaltic andesiteSample OSH1 OSH2 OSH3 OSH4 OSH5 OSH6

Rb ppm 356.1 368.0 293.8 6.80 4.26 12.74Sr ppm 12.30 16.57 14.06 1311 117.4 135.6Rb/Sr 28.948 22.206 20.891 0.005 0.036 0.09487Rb/86Sr 83.76 64.25 60.45 0.015 0.105 0.272(87Sr/86Sr)p 0.78629 0.76860 0.76291 0.70888 0.70748 0.70754± 0.00002 0.00002 0.00002 0.00003 0.00002 0.00002(87Sr/86Sr)t 0.70895 0.70927 0.70709 0.70887 0.70738 0.70729Nd ppm 110.4 115.1 113.3 5.09 10.66 17.07Sm ppm 22.47 23.88 22.82 1.76 3.02 4.52Sm/Nd 0.2035 0.2074 0.2014 0.3458 0.2833 0.2648147Sm/144Nd 0.1230 0.1254 0.1218 0.2091 0.1713 0.1601(143Nd/144Nd)p 0.512275 0.512447 0.512287 0.512754 0.512687 0.512554± 0.000040 0.000040 0.000040 0.000080 0.000080 0.000060(143Nd/144Nd)t 0.512223 0.512394 0.512235 0.512665 0.512614 0.512486eNdp �7.1 �3.7 �6.8 +2.3 +1.0 �1.6eNdt �6.5 �3.1 �6.2 +2.2 +1.2 �1.3

Notes: The total procedure blank in the laboratory is less than 8 ng of Sr and less than 1 ng for Nd at present. The reported 87Sr/86Sr and 143Nd/144Nd ratios for the samples arethe mean of about 350 ratios and the errors are standard error on the mean. The 87Sr/86Sr ratios were normalized to 88Sr/86Sr = 0.1194. Mean value for NIST Sr standard SRM-987 was 87Sr/86Sr = 0.710248 ± 10 (1 se) during the period of analysis. The 143Nd/144Nd ratios were normalized to 146Nd/144Nd = 0.7219. Mean value for the Japanese Ndstandard JNdi-1 was 0.512980 ± 10 (1 se) during the period of analysis. eNd values have been calculated for the rocks using a chondritic average values of143Nd/144Nd = 0.512638 and 147Sm/144Nd = 0.1967 (present-day), which give initial (at 65 million years) chondritic average value of 143Nd/144Nd = 0.512555. Errors on the Srand Nd isotopic ratios are 2 se.

SiO2 (wt.%)

Phonolite

TrachyteTrachydacite

Picro-basalt Basalt

Basaltic andesite

Andesite

Dacite

Rhyolite

Basaltic trachyandesite

Trachy-basalt

Trachyandesite

TephriteBasanite

Tephriphonolite

ALKALIC

SUBALKALIC

35 40 45 50 55 60 65 70 75 800

2

4

6

8

10

12

All Saurashtra (M95)Shihor-Palitana-Rajula (CB01)Chogat-Chamardi (S11a)

ULTRABASIC BASIC INTERMEDIATE ACID

14

Osham (M96 and this study)Mumbai silicics (L87)Rajpipla (M85)Pavagadh (SM08)

FoiditePho

notep

hrite

Na 2

O +

K2O

(wt.%

)

Fig. 7. The Osham Hill rock data on the total alkalis-silica (TAS) plot (Le Bas et al., 1986). Boundary lines between the alkalic and subalkalic fields proposed by Macdonald andKatsura (1964, short heavy line) and Irvine and Baragar (1971, curved dashed line) are shown. Also shown for comparison are data for 624 samples of the Western Ghats lavas(lightly shaded field), and the known rhyolitic suites from the Deccan Traps. Data sources are Beane, 1988 (Western Ghats), Chatterjee and Bhattacharji, 2001 (Rajula–Palitana–Shihor), Lightfoot et al., 1987 (Mumbai), Mahoney et al., 1985 (Rajpipla), Sheth and Melluso, 2008 (Pavagadh), and Sheth et al., 2011a (Chogat–Chamardi). TheSaurashtra rhyolite data are from Melluso et al. (1995).

184 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192

Author's personal copy

extract comprising olivine (22%), clinopyroxene (29%), plagioclase(39%) and spinel (10%).

5.2. The Osham rhyolites

Various petrogenetic models have been proposed for rhyoliticsuites in the Deccan Traps. Lightfoot et al. (1987) argued that therhyolites and trachytes of Mumbai (Fig. 1a) were the products offractional crystallization of basaltic magmas or of partial meltingof deeper parts of the basaltic pile or gabbroic underplates. Gwalaniet al. (1993) proposed a similar origin for the oversaturated andperaluminous trachytes of Siriwasan-Dugdha, near Chhota Udaipur(Fig. 1a). On the other hand, concurrent assimilation and fractionalcrystallization (AFC, DePaolo, 1981) mechanisms have been in-voked for the Deccan rhyolites (and trachytes) of Mumbai (Shethand Ray, 2002), for the Pavagadh rhyolites (Sheth and Melluso,2008), and for the Chogat–Chamardi granophyres and rhyolites(Sheth et al., 2011a). They have also been invoked for rhyolitic rocksin Madagascar and elsewhere (e.g., Cucciniello et al., 2011).

As noted, the initial 87Sr/86Sr ratios for the Osham rhyolite andpitchstone flows overlap with those of the associated mafic rocks,but the initial 143Nd/144Nd ratios of the silicic rocks are systemati-cally lower, whereas their Nd contents are higher (Fig. 10). There-fore, the Osham rhyolites cannot be the products of simpleclosed-system fractional crystallization of the Osham basalticandesites or their parental mafic magmas.

Sheth et al. (2011a) modelled the Chogat–Chamardi granophyresand rhyolites by AFC using a mafic magma and two real, different

granitic contaminants. They concluded that the production of someof the Chogat–Chamardi silicic rocks involved moderate to consider-able inputs from basement crust, whereas others could be nearlyclosed-system fractionation products of the associated mafic rocks.However, a key argument against both closed-system fractionalcrystallization and AFC models for the Osham rocks is that whereasplagioclase is a major phase in the Osham basaltic andesites (Fig. 3d–h), plagioclase cannot have formed any substantial proportion of thecrystal extract because the Osham rhyolites have the same Al2O3

contents as the basaltic andesites. Any plausible petrogenetic modelfor the rhyolites must satisfy this major oxide relationship.

Rhyolites in Iceland have been interpreted as products of exten-sive fractional crystallization of basaltic magmas, sometimes withcrustal assimilation, or as products of partial melting of unalteredor hydrothermally altered old crustal basalts or their silicic differ-entiates (e.g., Carmichael, 1964; O’Nions and Grönvold, 1973;Lacasse et al., 2007; Martin and Sigmarsson, 2007). Interestingly,Zellmer et al. (2008) observed similar Al2O3 contents in the basalticand rhyolitic lavas of the Torfajökull–Veidivötn area of Iceland, andconsidered the production of the rhyolites incompatible with feld-spar-dominated fractional crystallization of mafic liquids. Usingthe MELTS code they found that partial melting of K-metasoma-tized mafic protoliths in the Icelandic crust produces dacitic melts,from which a little plagioclase feldspar fractionation can producethe rhyolites.

We have obtained similar results for the Osham silicic rocksbased on MELTS modelling. Modal batch melting (�25%) of a sourcewith the composition of the basaltic andesite OSH6 (mineralogically,

Rajula-Palitana-Shihor rhyolites (CB01)Pavagadh rhyolites (SM08)Mumbai rhyolites (L87)Rajpipla rhyolites (M85)

Osham rhyolites (M96 and this study)

0.5

1.0

1.5

2.0

2.5

3.0

Ba/

Zr

60

62

64

66

68

70

72

74

76

78

80

0

0 200 400 600 800 1000 1200 1400 1600 1800 2000

200 400 600 800 1000 1200 1400 1600 1800Zr ppm

SiO

2(w

t.%)

Chogat-Chamardi microgranites, granophyres and rhyolites (S11a)

a

3.5

4.0

4.5b

2000

Zr ppm

Fig. 8. Binary trace element plots of (a) SiO2 (LOI-free recalculated values) vs. Zr, and (b) Ba/Zr vs. Zr for the Osham and other Deccan rhyolites. Data sources are as in Fig. 7.

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 185

Author's personal copy

46.6 wt.% plagioclase, 13.7 wt.% clinopyroxene, 13.2 wt.% orthopy-roxene, 10.0 wt.% magnetite, and 16.5 wt.% quartz as computed withthe MINSQ program, Herrmann and Berry, 2002) can generate rhyo-litic melts with high Al2O3 contents that closely resemble the Oshamrhyolite (Fig. 11). In particular, a 25% batch melt of OSH6 produced at�800 �C and 3 kbar has a major oxide composition that plots veryclose to or within the fields defined by partial melts of various hy-drated basaltic lithologies at varied crustal pressures (Beard and Lof-gren, 1991; Nakajima and Arima, 1998; Shukuno et al., 2006). This isnot to say that the source protolith of the Osham rhyolite OSH1 wasprecisely or specifically like OSH6. These two have somewhat differ-ent Nd–Sr isotopic compositions, requiring some crustal input intothe rhyolite. However, because it is energetically favourable for ahot, intracrustal basalt to assimilate crust than for an evolved andsubstantially cooler rhyolite, we propose that the source mafic

protolith of the Osham rhyolite may already have had the isotopiccomposition of the rhyolite before partial melting.

Concentrations of the alteration-resistant trace elements (Zrand Y) are consistent with �25% modal batch melting of OSH6 togenerate OSH1 as inferred from MELTS major element modelling.(Ba and Sr are unreliable as they are significantly affected byweathering.) Thus, taking the MINSQ program-derived modal min-eral assemblage in OSH6, and Kd values for Zr and Y for these min-erals for basaltic and basaltic andesite liquids compiled byRollinson (1993), we calculate that 25% and 30% modal batch meltsof OSH6 have Zr = 536 ppm and 465 ppm Zr, respectively, the latterclose to the value of 436 ppm Zr in OSH1. Similarly, the 25% partialmelt has 116 ppm Y, and a 20% partial melt has 130 ppm Y (close tothe value of 136 ppm in OSH1). It should be noted that the resultsof calculations such as these are strongly dependent on the choice

OSH3 (banded pitchstone)

1

10

100

1000

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

1

10

100

1000

3

10

100OSH4 (Osham basaltic andesite)OSH5 (Osham basaltic andesite)OSH6 (Osham basaltic andesite)

rock

/cho

ndrit

ero

ck/c

hond

rite

rock

/cho

ndrit

e

OSH/3 OSH/4 OSH/7 OSH/8 OSH/16

Maithani et al. 1996

D89 (Pavagadh pitchstone, SM08)Set138 (Mumbai rhyolite, L87)Set54 (Mumbai rhyolite, L87)Set104 (Mumbai rhyolite, L87)OSH1 (Osham rhyolite, this study)

This studyOSH1 (rhyolite)OSH2 (pitchstone)

a

b

c

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

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

Fig. 9. (a) CI chondrite-normalized rare earth element patterns of the Osham rhyolite and pitchstones, plotted using data of this study and those of Maithani et al. (1996, graydashed lines). (b) REE pattern for Osham rhyolite OSH1 compared to some other Deccan rhyolites. Data sources are as in Figs. 7 and 8. The many grey patterns are the Rajula-Palitana-Shihor rhyolite data of Chatterjee and Bhattacharji (2001). (c) REE patterns for the Osham basaltic andesites. Normalizing values are from Sun and McDonough (1989).

186 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192

Author's personal copy

of partition coefficients used, but the broad agreement with the ob-served data is encouraging.

The large negative europium anomalies in the Osham rhyoliticrocks (Fig. 9a) need to be explained, as these apparently requiresignificant plagioclase fractionation which would not keep Al2O3

contents elevated. Europium anomalies and their magnitude de-pend on multiple, unconstrained parameters, such as plagioclasecomposition, bulk magma composition, and oxygen fugacity (e.g.,Rollinson, 1993). We believe that the Eu anomalies in the rhyolitesmay, at least partly, be a feature of their mafic protoliths, acquiredby the protoliths during their own derivation from picritic magmasby olivine gabbro fractionation. If so, the negative europium anom-alies are a feature largely ‘‘inherited’’ by the rhyolites.

6. Discussion

6.1. Isotopic compositions of the Osham lavas compared to those ofother Deccan flood basalts and rhyolitic rocks

Fig. 12a shows Nd–Sr isotopic data for the Osham lavas,compared to those for the Deccan flood basalts of the Western Ghatssequence as well all analyzed Deccan rhyolites. The kilometers-thickWestern Ghats sequence has been divided into three stratigraphicsubgroups and eleven formations. Magmas of the Ambenali Forma-tion in the Western Ghats are the most isotopically primitive Deccan

magma type, with eNd of up to +8. They show the least continentallithospheric influence and are the presumed parental magma to allother Western Ghats flood basalt lavas (e.g., Najafi et al., 1981;Mahoney et al., 1982). In Fig. 12a, the elogated arrays defined bythe various stratigraphic formations of the Western Ghats sequenceemanate from the Ambenali Formation field and trend towards var-ious lithospheric (including crustal) end members. These arrayshave been interpreted to result from mixing (Mahoney et al.,1982; Peng et al., 1994; Vanderkluysen et al., 2011).

Fig. 12a underscores the scarcity of combined Nd–Sr isotopicdata for Deccan rhyolites; many more Deccan rhyolites have beenanalyzed for Sr isotopic ratios alone, and their ranges are shown bygrey bands below the plot. The higher end of initial 87Sr/86Sr ratiosfor the Pavagadh, Mumbai and Osham rhyolites is about the same.However, the Girnar silicic porphyries (Paul et al., 1977) have veryhigh initial 87Sr/86Sr ratios of up to 0.7281, and may represent stillgreater proportions of ancient Rb-rich granitic basement crust.

Of the few Deccan rhyolites analyzed for both Nd and Sr isotopicratios, the lowest initial eNd values (at 65 Ma) are seen in the Cho-gat–Chamardi rhyolites (as low as �13.9). In comparison, the low-est eNd value of the Osham silicic lavas is �6.5 (sample OSH1). Ndor Sr isotopic data are not available on the eastern Saurashtra rhy-olites of the Shihor–Palitana–Rajula belt, south of the Chogat–Chamardi complex, which are thought to contain considerable con-tributions from old basement crust based on trace element evi-dence (Chatterjee and Bhattacharji, 2001, 2004). In fact theRajula rhyolites contain zircons and monazites with cores inher-ited from basement crust (Chatterjee and Bhattacharji, 2004),which is important evidence for incorporation of the Saurashtrabasement crust into Deccan magmas. The zoning observed in thezircon from the Osham rhyolite (Fig. 4) suggests the intriguing pos-sibility that the zircons may have cores from the basement crustovergrown by later additions; however, more work is required toevaluate this possibility.

Interestingly, basaltic andesite sample OSH4 plots in the arraycovering the Rewa and Pachmarhi tholeiitic dykes and some of theChogat–Chamardi granophyres. This may indicate that there is ayet unknown crustal end member affecting all these rocks, over alarge (600–900 km) distance (see Sheth et al., 2011a). However, amuch more likely explanation is that the severely weathered basalticandesite OSH4 (LOI = 6.73%) had its 87Sr/86Sr ratio increased whilehaving its 143Nd/144Nd ratio essentially unchanged, which thereforeshifted its data point horizontally on the Nd–Sr isotopic plot.

Fig. 12b provides a comparison between various mafic rocksanalyzed from Saurashtra, including the high-Ti picritic basaltsand basalts encountered in the eastern Saurashtra boreholes (Pengand Mahoney, 1995) and the low-Ti picritic basalts from surface

-8

-6

-4

-2

0

2

4

0 25 50 75

Nd (ppm)

Osham silicic rocksOsham mafic rocks

100 125

Fig. 10. Variation of initial Nd isotopic ratios with Nd concentration in the Oshammafic and silicic rocks. Analytical uncertainties are smaller than the size of thesymbols.

Table 6Major element mass balance calculation for the transition from picrite D130 to basaltic andesite OSH6.

From To Mineral compositions MELTS

D130 OSH6 ol cpx pl mt Res2 OSH6 calc.

SiO2 wt.% 48.07 57.75 38.90 50.81 49.73 – % solids removed 0.002 57.37TiO2 1.17 1.21 – 0.92 – 21.14 �80.8 0.016 0.58Al2O3 14.00 13.06 – 3.48 31.07 1.70 0.003 12.52FeOt 10.82 12.83 19.49 9.05 1.20 75.51 0.007 13.35MnO 0.18 0.22 0.32 0.26 – 0.39 0.000 –MgO 12.55 4.87 40.93 15.44 – 1.98 0.000 4.75CaO 10.92 6.98 0.25 19.11 14.03 – 0.001 7.10Na2O 1.77 2.21 – 0.25 3.28 – 0.027 2.70K2O 0.40 0.67 – – 0.19 – 0.045 1.28P2O5 0.11 0.20 – – – – 0.005 0.35Mineral proportion 23.3% 30.9% 42.3% 3.4%

PRes2 0.106

Notes: The whole-rock compositions are recalculated to 100% on a LOI-free basis. Whole-rock and mineral compositions for D130 are from Melluso et al. (1995).P

Res2 = sumof squared residuals in the calculation using the XLFRAC program; ol, olivine; cpx, clinopyroxene; pl, plagioclase; mt, magnetite. The last column shows the calculated OSH6composition using MELTS fractional crystallization model for D130.

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 187

Author's personal copy

outcrops (Melluso et al., 1995, 2006). Also shown are data for theOsham basaltic andesites (this study) and the Chogat–Chamardigabbros and mafic (basaltic andesite) dykes (Sheth et al., 2011a).As seen the data for the Osham basaltic andesite OSH6 and a fewother samples overlap with the elongated arrays for the boreholelavas. The Chogat–Chamardi and Osham samples have not beenanalyzed for Pb isotopes unlike the other lavas shown, but theNd–Sr isotopic data point to a broadly similar petrogenetic historyfor all these rocks. Fig. 12b also helps visualize the vast range in

composition covered by the Deccan (Saurashtra region included)silicic rocks (Fig. 12a), compared to the range of the Saurashtra ma-fic rocks. The Nd–Sr isotopic range of the Western Ghats strati-graphic formations, itself considerable, is significantly extendedby the central Deccan dolerite dyke with initial eNd of �20.2, thelowest yet known in the Deccan Traps, and initial 87Sr/86Sr of0.72315 (Chandrasekharam et al., 1999). The Chogat–Chamardi si-licic rocks, though not as low in Nd isotopic ratio as the centralDeccan dyke, attain considerably higher Sr isotopic values.

Partial melt of basaltic andesite OSH6Osham silicic rocksOsham mafic rocks

a b

c d

e f

Fig. 11. Comparison of the Osham rhyolite compositions with experimental liquids produced by partial melting of hydrated basaltic rocks, greenstones, and amphibolites.Fields enclose the experimental data of Wolf and Wyllie (1994) at 10 kbar, Nakajima and Arima (1998) at 10 kbar, Beard and Lofgren (1991) at 1, 3 and 6.9 kbar, and Shukunoet al. (2006) at 900–1000 �C and 3 kbar. The grey star represents the batch partial melt of basaltic andesite OSH6 at 3 kbar, �800 �C, H2O = 0.8 wt.% and fO2 buffer = QFM, asmodelled with MELTS (Ghiorso and Sack, 1995; Smith and Asimow, 2005). The actual (LOI-free recalculated) and MELTS-derived major oxide values of rhyolite OSH1 are,respectively: SiO2 = 73.58, 73.94; TiO2 = 0.12, 0.14; Al2O3 = 13.86, 16.79; Fe2O3 = 0.94, 0.20; FeO = 1.89, 0.85; MnO = 0.02, 0.00; MgO = 0.15, 0.13; CaO = 0.01, 1.93,Na2O = 3.12, 0.79; K2O = 6.28, 1.90; P2O5 = 0.02, 0.17. The MELTS-derived partial melt also has an H2O content of 3.16%. The model melt has higher Al2O3 and CaO and lowerNa2O and K2O than the actual OSH1 composition, suggesting that some plagioclase fractionation from a model-composition melt can generate a residual liquid similar toOSH1 (see text).

188 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192

Author's personal copy

6.2. A possible physical scenario for the Osham Hill lava sequence

Annen and Sparks (2002) and Annen et al. (2006) have modelledthe effects of repetitive, periodically emplaced basaltic intrusionson melt generation in the continental crust, particularly at middleand lower crustal levels which they describe as ‘‘deep crustal hotzones’’. In their model basalt intrusions are injected into the crustat depths between 20 and 30 km, under geothermal gradients of10–30 �C/km. The earliest basalt intrusions incubate and solidify,but if the basalt intrusion rate exceeds 50 m thickness per10,000 years, melt generation starts 0.5–1 million years after initialintrusion. Later intrusions can heat and partially melt both earliersolidified basalt intrusions and the continental crust depending onthe temperature and amount of volatile-bearing minerals in thecrust. For wet, cool basalt (1100 �C, 2 wt.% H2O), the main meltsproduced are the residual silicic liquids from basalt crystallization,whereas for dry and hot basalt (1300 �C, 0.3 wt.% H2O) emplaced

into crustal rocks, melts are dominantly crustal, though this obvi-ously depends on the fertility and temperature of the intrudedcrustal rocks. With a considerable diversity in the temperatureand water contents of the basalt intrusions, and the temperatureand lithology of the crustal rocks, a range of melt types can be gen-erated by basalt crystallization and crustal melting, and thesemelts, with both mantle and crustal signatures, can mix during as-cent. High-level, upper crustal magma chambers can develop dueto density effects, lithospheric stress variations or lithological orstructural discontinuities, and different ascending magmas can un-dergo further mixing and differentiation in them. The concept isillustrated in Fig. 13.

Simple field stratigraphic relationships, as observed at OshamHill, Mumbai, and particularly Pavagadh (Lightfoot et al., 1987;Sheth and Melluso, 2008), show that rhyolite magmas in the DeccanTraps have typically erupted during the advanced to late stages ofthe flood basalt magmatism, i.e., at a time when the crustal magma

+5

-5

-10

-15

0

-20

+10

Igatpuri-Jawhar

Ambenali

Neral

Khandala

Bushe

Mahabaleshwar

Bhimashankar

Thakurvadi

Poladpur

0.7200.705 0.710 0.715 0.725 0.730

1

2

3

45

6

D129

Mumbai rhyolites (L87)Rajpipla rhyolite (M85)

Osham rhyolite & pitchstone (this study)

Chogat-Chamardi silicic rocks (S11a)

Rewa tholeiitic dykes (L11)Pachmarhi tholeiitic dyke PMD6 (S09)

Osham basaltic andesites (this study)

Boradi (central Deccan) dolerite dyke (C99)

Saurashtra low-Ti picritic basalts (M06)

(87Sr/86Sr)tPavagadh rhyolites:

0.7064-0.7089

Girnar silicic porphyries: 0.7269-0.7281

Mumbai rhyolites: 0.7037-0.7074

Panhala

0.702 0.7100.704 0.706 0.708

+8

-2

0

+10

+12

+6

+4

+2

-4

-6

-8

(87Sr/86Sr)t

Ambenali

45

6

D129Reunion

Bhuj(central

Kachchh)

most Indian N-MORB

Saurashtra Trend 1

Saurashtra Trend 2

Chogat-Chamardi mafic rocks (S11a)

Legend

a

b

area of (b)

D130

D130

Fig. 12. (a) Nd-Sr isotopic plot for the Osham Hill rocks (numbered 1–6, without the prefix OSH to avoid cluttering). Data are also plotted for hitherto analyzed Deccanrhyolites (data sources as above, and Paul et al. (1977) for Mount Girnar silicic porphyries) and for the Western Ghats stratigraphic formations (Peng et al., 1994, modified byVanderkluysen et al., 2011 with their latest data added). Data are also shown for a central Deccan dyke in the Dhule region, a dyke from Pachmarhi, and the Rewa dykes nearShahdol (Fig. 1a) (Chandrasekharam et al., 1999; Sheth et al., 2009; Lala et al., 2011), and for the Saurashtra low-Ti picritic basalts D129 and D130 (Melluso et al., 2006), thelatter taken as a parental magma in the XLFRAC model. All data are initial values for 65 Ma. (b) Nd-Sr isotopic plot with data for various Saurashtra mafic rocks plotted. Therocks include the Osham basaltic andesites (this study), and the Chogat-Chamardi gabbros and mafic dykes, the low-Ti picrites and picritic basalts (samples D129 and D130are marked), and the ‘‘Trend 1’’ and ‘‘Trend 2’’ high-Ti picritic basalts and basalts encountered in boreholes at Dhandhuka, Wadhwan and Botad (Fig. 1b). Fields for Indian N-MORB, Bhuj area mafic alkalic rocks (the central Kachchh monogenetic volcanic field, Kshirsagar et al., 2011), and Réunion lavas, are also shown for comparison. Data sourcesare Peng and Mahoney (1995), Melluso et al. (2006) and references therein, and Sheth et al. (2011a). Note that panels (a) and (b) are drawn at the same scale, and arevertically aligned at the same 87Sr/86Sr values. This allows an easy visual comparison between the isotopic range of the Saurashtra silicic rocks (a) and the Saurashtra maficrocks (b).

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 189

Author's personal copy

chambers were well established and the heat flow was elevated.The Saurashtra region forms a part of the western Indian rifted mar-gin, as does the Mumbai area with its abundant rhyolites whichLightfoot et al. (1987) interpreted as partial melts of gabbroic crus-tal sill complexes or fractionation derivatives of basaltic melts. Asillustrated in Fig. 13, hot, dry basalt (and picritic) intrusions mayhave been repeatedly injected into the Saurashtra basement crust,forming sill complexes (cf. Cox, 1980; Annen and Sparks, 2002).Fractional crystallization of picritic melts in such sill complexeswould produce evolved basalts and basaltic andesites, and theseevolved melts could rise towards the surface, in some cases withshallow-crustal storage en route, forming high-level magma cham-bers in which conditions for further differentiation to rhyolitic liq-uids would obtain. Crustal melting can also occur under suitableconditions and siliceous melts can be produced from partial meltingof basement granites and amphibolites (cf. Peng et al., 1994). Re-peated injections of hot mafic and picritic magmas would also heatthe earlier mafic intrusions residing at all levels of the crust, espe-cially where earlier mafic sills are sandwiched between later ones(Annen and Sparks, 2002). Partial melting of these mafic intrusionswould generate rhyodacitic or rhyolitic melts which being buoyantwould be able to rise to the surface rapidly. Where dykes transport-ing mafic melts intersect silicic magma reservoirs, mingling andlimited mixing and hybridization between the mafic and silicicmagmas will occur, as shown by Zellmer et al. (in press) from theMumbai area. The absence of such phenomena at Osham Hill sug-gests that the basaltic andesite and rhyolitic melts reached the sur-face via separate plumbing systems.

7. Conclusions

The abundance of silicic rocks (rhyolites and granophyres) inthe Saurashtra region of the northwestern Deccan Traps is notable.The Osham Hill in western Saurashtra exposes a small but signifi-cant sequence of rhyolite, pitchstone, and basaltic andesite lavaflows. Major and trace element (including rare earth element)and Nd–Sr isotopic data presented here for the Osham lavas donot support the derivation of the rhyolites as residual liquids offractional crystallization (with or without crustal incorporation)of the associated or related mafic melts. Instead, the rhyolites ap-pear to have formed by partial melting of Deccan-age basalticintrusions previously emplaced at various levels in the crust, dueto repetitive and long-continued basalt injection and crustal heat-ing. In this, the Osham rhyolites resemble rhyolites of the Tor-fajökull–Veidivötn area in Iceland (Zellmer et al., 2008). On theother hand, the Osham rhyolites differ notably from the rhyoliticrocks of Chogat–Chamardi and elsewhere in eastern Saurashtra,which have been argued to originate by assimilation-fractionalcrystallization processes involving moderate to large amounts ofancient granitic crust (Chatterjee and Bhattacharji, 2001, 2004;Sheth et al., 2011a). It thus appears that various rhyolitic magmasin the Deccan Traps have evolved by different petrogenetic mech-anisms (Fig. 13), an unsurprising result given the sheer scale ofDeccan magmatism. Rhyolites and granophyres dominate in thelarge and little-studied Barda and Alech complexes of western Sau-rashtra (De and Bhattacharya, 1971; De, 1981); their petrogenesisis an exciting topic for future work..

5

10

15

20

25

30

kmrhyolite

basaltic andesite

basalt

picrite

picrite

ancient granitic pluton

rhyolitic residual melt in crystallized mafic sill

sedimentary basin

picritic lava

rhyolite melt from fractionation of

mafic liquids

mingled basalt-rhyolite lava and

plug

rhyolite melt from anatexis of

old basement

flood basalts with variable crustal

assimilation

Moho

ancient metamorphic rocks

ancient metamorphic rocks

gabbro cumulates

dunite cumulates

arrested dyke

mantle undergoing partial melting

Upper crust

Middle crust

Lower crust

basalt

basalt

sill complex

sill complex

crustal anatexis

dyke

dyke

granite

crustal anatexis

rhyolitic partial melt of mafic sill

partial melting of mafic sill

Fig. 13. Model for the crustal substructure of a typical continental flood basalt province with rhyolitic magmas (such as the Deccan Traps) at an advanced stage of provincedevelopment. Elliptical shapes represent magma chambers (not to scale) which may have been of the nature of sills and sill complexes; their progressively lighter shades ofgrey indicate progressively more evolved liquid compositions. The model shows the various ways in which rhyolitic melts (the lightest grey) can be generated: anatexis of oldbasement crust, fractional crystallization of mafic melts, residual melt in cooling, crystallizing mafic sills, and partial melting by heating of mafic sills due to influx of newbatches of hot basalts or picrites. Mingling and mixing can occur, and the flood basalts that erupt contain a wide range of ‘‘mantle-like’’ to ‘‘crust-like’’ geochemical signatures.Thin vertical lines show magma transport in dykes. The horizontal dimension of the figure could be from a few hundred to a thousand kilometers. The model has featuresadapted from Cox (1980), Annen et al. (2006), and Zellmer et al. (2008).

190 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192

Author's personal copy

Acknowledgements

This work was supported by the Department of Science andTechnology (DST, Project Grant SR/FTP/ES-19/2007) as well as theIndustrial Research and Consultancy Centre, IIT Bombay (ProjectGrant 09YIA001) to Sheth. Funds for EPMA analyses by Cucciniellowere provided by Italian MIUR (PRIN Grants 2008 to Leone Mellu-so). We thank Dipak Gosain for field assistance, Pooja Kshirsagarand Shilpa Netrawali for help with the analytical work, and S.Viswanathan, Kanchan Pande, Jyotiranjan Ray, Leone Melluso, Sou-myajit Mukherjee, and Chris Talbot for discussions and commentson various aspects of the work. Reviews of earlier versions of themanuscript by Nilanjan Chatterjee, Lalou Gwalani, Talat Ahmad,Brian McConnell, and Leone Melluso are appreciated. The presentmanuscript was much improved by excellent, thorough reviewsby Georg Zellmer, Loÿc Vanderkluysen, and an anonymousreviewer.

References

Andrews, G.D.M., Branney, J.M., 2001. Emplacement and rheomorphic deformationof a large, lava-like rhyolitic ignimbrite: Grey’s Landing, southern Idaho.Geological Society of America Bulletin 123, 725–743.

Annen, C., Sparks, R.S.J., 2002. Effects of repetitive emplacement of basalticintrusions on thermal evolution and melt generation in the crust. Earth andPlanetary Science Letters 203, 937–955.

Annen, C., Blundy, J.D., Sparks, R.S.J., 2006. The genesis of intermediate and silicicmagmas in deep crustal hot zones. Journal of Petrology 47, 505–539.

Auden, J.B., 1949. Dykes in western India – a discussion of their relationships withthe Deccan Traps. Transactions of National Academy of Sciences of India 3, 123–157.

Beane, J.E., 1988. Flow Stratigraphy, Chemical Variation and Petrogenesis of DeccanFlood Basalts from the Western Ghats, India. Ph.D. Dissertation, WashingstonState University, USA.

Beard, J.S., Lofgren, G.E., 1991. Dehydration melting and water-saturated melting ofbasaltic and andesitic greenstones and amphibolites at 1, 3 and 6.9 kb. Journalof Petrology 32, 365–401.

Carmichael, I.S.E., 1964. The petrology of Thingmuli, a Tertiary volcano in easternIceland. Journal of Petrology 5, 435–460.

Chandrasekharam, D., Mahoney, J.J., Sheth, H.C., Duncan, R.A., 1999. Elemental andNd-Sr-Pb isotope geochemistry of flows and dykes from the Tapi rift, Deccanflood basalt province, India. Journal of Volcanology and Geothermal Research93, 111–123.

Chatterjee, N., Bhattacharji, S., 2001. Origin of the felsic and basaltic dykes and flowsin the Rajula-Palitana-Sihor area of the Deccan Traps, Saurashtra, India: ageochemical and geochronological study. International Geology Review 43,1094–1116.

Chatterjee, N., Bhattacharji, S., 2004. A preliminary geochemical study of zirconsand monazites from Deccan felsic dykes, Rajula, Gujarat, India: implications forcrustal melting. In: Sheth, H.C., Pande, K. (Eds.), Magmatism in India throughTime. Proceedings of the Indian Academy of Sciences (Earth and PlanetarySciences), vol. 113, 533–542.

Cox, K.G., 1980. A model for flood basalt vulcanism. Journal of Petrology 21, 629–650.

Cucciniello, C., Conrad, J., Grifa, C., Melluso, L., Mercurio, M., Morra, V., Tucker, R.D.,Vincent, M., 2011. Petrology and geochemistry of Cretaceous mafic and silicicdykes and spatially associated lavas in central-eastern coastal Madagascar. In:Srivastava, R.K. (Ed.), Dyke Swarms: Keys for Geodynamic Interpretation.Springer Verlag, pp. 345–375.

De, A., 1981. Late Mesozoic – Lower Tertiary magma types of Kutch and Saurashtra.In: Subbarao, K.V., Sukheswala, R.N. (Eds.), Deccan Volcanism, vol. 3. GeologicalSociety of India Memoir, pp. 327–339.

De, A., Bhattacharya, D., 1971. Phase-petrology with special reference to pyroxenesof the acid igneous complex of Barda Hills, western Saurashtra (Gujarat).Bulletin of Volcanology 35, 907–929.

DePaolo, D.J., 1981. Trace element and isotopic effects of combined wall rockassimilation and fractional crystallization. Earth and Planetary Science Letters53, 189–202.

Ghiorso, M.S., Sack, R.O., 1995. Chemical mass transfer in magmatic processes – IV.A revised and internally consistent thermodynamic model for the interpolationand extrapolation of liquid-solid equilibria in magmatic systems at elevatedtemperatures and pressures. Contributions to Mineralogy and Petrology 119,197–212.

Grove, T.L., Juster, T.C., 1989. Experimental investigations of low-Ca pyroxenestability and olivine-pyroxene-liquid equilibria at 1-atm in natural basaltic andandesitic liquids. Contributions to Mineralogy and Petrology 103, 287–305.

Gwalani, L.G., Rock, N.M.S., Chang, W.-J., Fernandez, S., Allegre, C.J., Prinzhofer, A.,1993. Alkaline rocks and carbonatites of Amba Dungar and adjacent areas,Deccan igneous province, Gujarat, India: 1. Geology, petrography andpetrochemistry. Mineralogy and Petrology 47, 219–253.

Hatch, F.H., Wells, A.K., Wells, M.K., 1983. Petrology of the Igneous Rocks, 13th ed.Thomas Murby, 551 p.

Henry, C.D., Wolff, J.A., 1992. Distinguishing strongly rheomorphic tuffs fromextensive silicic lavas. Bulletin of Volcanology 54, 171–186.

Herrmann, W., Berry, R.F., 2002. MINSQ–a least squares spreadsheet method forcalculating mineral proportions from whole rock major element analyses.Geochemistry: Exploration, Environment, Analysis 2, 361–368.

Irvine, T.N., Baragar, W.R.A., 1971. A guide to the chemical classification of thecommon rocks. Canadian Journal of Earth Sciences 8, 523–548.

Krishnamurthy, P., Cox, K.G., 1980. A potassium-rich alkalic suite from the DeccanTraps, Rajpipla, India. Contributions to Mineralogy and Petrology 73, 179–189.

Kshirsagar, P.V., Sheth, H.C., Seaman, S.J., Shaikh, B., Mohite, P., Gurav, T.,Chandrasekharam, D., in press. Spherulites and thundereggs from pitchstonesof the Deccan Traps: geology, petrochemistry, and emplacement environments.Bulletin of Volcanology. doi:10.1007/s00445-011-0543-3.

Kshirsagar, P.V., Sheth, H.C., Shaikh, B., 2011. Mafic alkalic magmatism in centralKachchh, India: a monogenetic volcanic field in the northwestern Deccan Traps.Bulletin of Volcanology 73, 595–612.

Lacasse, C., Sigurdsson, H., Carey, S.N., Johannesson, H., Thomas, L.E., Rogers, N.W.,2007. Bimodal volcanism at the Katla subglacial caldera, Iceland: insight intothe geochemistry and petrogenesis of rhyolitic magmas. Bulletin of Volcanology69, 373–399.

Lala, T., Choudhary, A.K., Patil, S.K., Paul, D.K., 2011. Mafic dykes of Rewa Basin,central India: implications on magma dispersal and petrogenesis. In: Srivastava,R.K. (Ed.), Dyke Swarms: Keys for Geodynamic Interpretation. Springer Verlag,pp. 141–166.

Le Bas, M.J., Le Maitre, R.W., Streckeisen, A., Zanettin, P., 1986. A chemicalclassification of volcanic rocks based on the total alkali-silica diagram. Journalof Petrology 27, 745–750.

Lightfoot, P.C., Hawkesworth, C.J., Sethna, S.F., 1987. Petrogenesis of rhyolites andtrachytes from the Deccan Trap: Sr, Nd, and Pb isotope and trace elementevidence. Contributions to Mineralogy and Petrology 95, 44–54.

Lindsley, D.H., 1983. Pyroxene thermometry. American Mineralogist 68, 477–493.Lofgren, G., 1971. Spherulitic textures in glassy and crystalline rocks. Journal of

Geophysical Research 76, 5635–5648.Macdonald, G.A., Katsura, T., 1964. Chemical composition of Hawaiian lavas. Journal

of Petrology 5, 82–133.Mahoney, J.J., Macdougall, J.D., Lugmair, G.W., Murali, A.V., Sankar Das, M., Gopalan,

K., 1982. Origin of the Deccan Trap flows at Mahabaleshwar inferred from Ndand Sr isotopic and chemical evidence. Earth and Planetary Science Letters 60,47–60.

Mahoney, J.J., Macdougall, J.D., Lugmair, G.W., Gopalan, K., Krishnamurthy, P., 1985.Origin of contemporaneous tholeiitic and K-rich alkalic lavas: a case study fromthe northern Deccan Plateau, India. Earth and Planetary Science Letters 72, 39–53.

Mahoney, J.J., Sheth, H.C., Chandrasekharam, D., Peng, Z.X., 2000. Geochemistry offlood basalts of the Toranmal section, northern Deccan Traps, India:implications for regional Deccan stratigraphy. Journal of Petrology 41, 1099–1140.

Maithani, P.B., Goyal, N., Banerjee, R., Gurjar, R., Ramchandran, S., Singh, R., 1996.Rhyolites of Osham hills, Saurashtra, India: a geochemical study. GondwanaGeological Magazine Special Volume 2, 213–224.

Martin, E., Sigmarsson, O., 2007. Crustal thermal state and origin of silicic magma inIceland: the case of Torfajökull, Ljósufjöll and Snaefellsjökull volcanoes.Contributions to Mineralogy and Petrology 153, 593–605.

McCurry, M., Hayden, K.P., Morse, L.H., Mertzman, S., 2009. Genesis of post-hotspot,A-type rhyolite of the Eastern Snake River Plain volcanic field by extremefractional crystallization of olivine tholeiite. Bulletin of Volcanology 73, 361–383.

Melluso, L., Sethna, S.F., 2011. Mineral compositions in the Deccan igneous rocks ofIndia: an overview. In: Ray, J.S., Sen, G., Ghosh, B. (Eds.), Topics in IgneousPetrology. Springer, Heidelberg, pp. 135–160.

Melluso, L., Beccaluva, L., Brotzu, P., Gregnanin, A., Gupta, A.K., Morbidelli, L.,Traversa, G., 1995. Constraints on the mantle sources of the Deccan Traps fromthe petrology and geochemistry of the basalts of Gujarat State (western India).Journal of Petrology 36, 1393–1432.

Melluso, L., Mahoney, J.J., Dallai, L., 2006. Mantle sources and crustal input asrecorded in high-Mg Deccan Traps basalts of Gujarat (India). Lithos 89, 259–274.

Melluso, L., Cucciniello, C., Petrone, C.M., Lustrino, M., Morra, V., Tiepolo, M.,Vasconcelos, L., 2008. Petrogenesis of Karoo volcanic rocks in the southernLebombo monocline, Mozambique. Journal of African Earth Sciences 52, 139–151.

Melluso, L., Sheth, H.C., Mahoney, J.J., Morra, V., Petrone, C.M., Storey, M., 2009.Geochemical correlations between silicic volcanic rocks of the St. Mary’s Islands(southwestern India) and eastern Madagascar: a tool for India-Madagascarreconstructions for the Late Cretaceous. Journal of Geological Society of London166, 283–294.

Melluso, L., de Gennaro, R., Rocco, I., 2010a. Compositional variations ofchromiferous spinel in Mg-rich rocks of the Deccan Traps, India. Journal ofEarth System Science 119, 343–363.

Melluso, L., Srivastava, R.K., Guarino, V., Zanetti, A., Sinha, A.K., 2010b. Mineralcompositions and magmatic evolution of the Sung Valley ultramafic- alkaline-carbonatitic complex (NE India). Canadian Mineralogist 48, 205–229.

Middlemost, E.A.K., 1989. Iron oxidation ratios, norms and the classification ofvolcanic rocks. Chemical Geology 77, 19–26.

H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192 191

Author's personal copy

Najafi, S.J., Cox, K.G., Sukheswala, R.N., 1981. Geology and geochemistry of thebasalt flows (Deccan Traps) of the Mahad-Mahabaleshwar section, India. In:Subbarao, K.V., Sukheswala, R.N. (Eds.), Deccan Volcanism, vol. 3. GeologicalSociety of India Memoir, pp. 300–315.

Nakajima, K., Arima, M., 1998. Melting experiments on hydrous low-K tholeiite:implications for the genesis of tonalitic crust in the Izu-Bonin-Mariana arc.Island Arc 7, 359–373.

O’Nions, R.K., Grönvold, K., 1973. Petrogenetic relationships of acid and basic rocksin Iceland: Sr-isotopoes and rare-earth elements in late and postglacialvolcanics. Earth and Planetary Science Letters 19, 397–409.

Paul, D.K., Potts, P.J., Rex, D.C., Beckinsale, R.D., 1977. Geochemical and petrogeneticstudy of the Girnar igneous complex, Deccan volcanic province, India.Geological Society of America Bulletin 88, 227–234.

Peng, Z.X., Mahoney, J.J., 1995. Drillhole lavas from the northwestern Deccan Traps,and the evolution of Réunion hotspot mantle. Earth and Planetary ScienceLetters 134, 169–185.

Peng, Z.X., Mahoney, J., Hooper, P., Harris, C., Beane, J., 1994. A role for lowercontinental crust in flood basalt genesis? Isotopic and incompatible elementstudy of the lower six formations of the western Deccan Traps. Geochimica etCosmochimica Acta 58, 267–288.

Rollinson, H., 1993. Using Geochemical Data: Evaluation, Presentation,Interpretation. Longman, Harlow, Essex, 352 p..

Sheth, H.C., Ray, J.S., 2002. Rb/Sr-87Sr/86Sr variations in Bombay trachytes andrhyolites (Deccan Traps): Rb-Sr isochron, or AFC process? International GeologyReview 44, 624–638.

Sheth, H.C., Melluso, L., 2008. The Mount Pavagadh volcanic suite, Deccan Traps:geochemical stratigraphy and magmatic evolution. Journal of Asian EarthSciences 32, 5–21.

Sheth, H.C., Ray, J.S., Ray, R., Vanderkluysen, L., Mahoney, J.J., Kumar, A., Shukla, A.D.,Das, P., Adhikari, S., Jana, B., 2009. Geology and geochemistry of Pachmarhidykes and sills, Satpura Gondwana Basin, central India: problems of dyke-sill-flow correlations in the Deccan Traps. Contributions to Mineralogy andPetrology 158, 357–380.

Sheth, H.C., Choudhary, A.K., Bhattacharyya, S., Cucciniello, C., Laishram, R., Gurav,T., 2011a. The Chogat-Chamardi subvolcanic complex, Saurashtra, northwesternDeccan Traps: geology, petrochemistry, and petrogenetic evolution. Journal ofAsian Earth Sciences 41, 307–324.

Sheth, H.C., Ray, J.S., Kumar, P.S., Duraiswami, R.A., Chatterjee, R.N., Gurav, T., 2011b.Recycling of flow-top breccia crusts into molten interiors of flood basalt lavaflows: field and geochemical evidence from the Deccan Traps. In: Ray, J., Sen, G.,Ghosh, B. (Eds.), Topics in Igneous Petrology. Springer, Heidelberg, pp. 161–180.

Shukuno, H., Tamura, Y., Tani, K., Chang, Q., Suzuki, T., Fiske, R.S., 2006. Origin ofsilicic magmas and the compositional gap at Sumisu submarine caldera, Izu-Bonin arc, Japan. Journal of Volcanology and Geothermal Research 156, 187–216.

Smith, P.M., Asimow, P.D., 2005. Adiabat_1ph: a new public front-end to the MELTS,pMELTS, and pHMELTS models. Geochemistry, Geophysics, Geosystems 6.doi:10.1029/ 2004GC000816.

Stormer Jr., J.C., Nicholls, J., 1978. XLFRAC: a program for interactive testing ofmagmatic differentiation models. Computers and Geosciences 4, 143–159.

Subba Rao, S., 1971. Petrogenesis of acid rocks of the Deccan Traps. Bulletin ofVolcanology 35, 983–997.

Sun, S.-s., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanicbasalts: implications for mantle composition and processes. In: Saunders, A.D.,Norry, M.J. (Eds.), Magmatism in the Ocean Basins, vol. 42. Magmatism in theOcean Basins, pp. 313–345.

Toplis, M.J., Carroll, M.R., 1995. An experimental study of the influence of oxygenfugacity on Fe–Ti oxide stability, phase relations, and mineral–melt equilibria inferro-basaltic systems. Journal of Petrology 36, 1137–1170.

Vanderkluysen, L., Mahoney, J.J., Hooper, P.R., Sheth, H.C., Ray, R., 2011. The feedersystem of the Deccan Traps (India): insights from dyke geochemistry. Journal ofPetrology 52, 315–343.

Verma, S.P., Torres-Alvarado, I.S., Sotelo-Rodriguez, Z.T., 2002. SINCLAS: Standardigneous norm and volcanic rock classification system. Computers andGeosciences 28, 711–715.

Wakhaloo, S.N., 1967. On the nature of volcanic eruption and of differentiation ofthe Girnar igneous complex, Junagarh, Kathiawar peninsula, India. Proceedingsof Symposium on Upper Mantle Project: Hyderabad 430, 449.

West, W.D., 1958. The petrography and petrogenesis of forty eight flows of DeccanTrap penetrated by borings in western India. Transactions of National Instituteof Sciences of India 4, 1–56.

Wilson, S.A., 1997. Data Compilation for USGS Reference Material BCR-2, ColumbiaRiver Basalt. U.S. Geological Survey Open File Rep.

Wilson, S.A., 1998. Data Compilation for USGS Reference Material GSP-2,Granodiorite, Silver Plume Colorado. U.S. Geological Survey Open File Rep.

Wilson, S.A., 2000. Data Compilation for USGS Reference Material BHVO-2,Hawaiian Basalt. U.S. Geological Survey Open File Rep.

Wolf, M.B., Wyllie, P.J., 1994. Dehydration-melting of amphibolite at 10 kbar; theeffects of temperature and time. Contributions to Mineralogy and Petrology115, 369–383.

Wolff, J.A., Wright, J.V., 1981. Rheomorphism of welded tuffs. Journal of Volcanologyand Geothermal Research 10, 13–34.

Zellmer, G.F., Rubin, K.H., Grönvold, K., Jurado-Chichay, Z., 2008. On the recentbimodal magmatic processes and their rates in the Torfajökull-Veidivötn area,Iceland. Earth and Planetary Science Letters 269, 388–398.

Zellmer, G.F., Sheth, H.C., Iizuka, Y., Lai, Y.-J., in press. Remobilization of granitoidrocks through mafic recharge: evidence from basalt-trachyte mingling andhybridization in the Manori-Gorai area, Mumbai, Deccan Traps. Bulletin ofVolcanology. doi:10.1007/s00445-011-0498-4.

192 H.C. Sheth et al. / Journal of Asian Earth Sciences 43 (2012) 176–192