Introduction

Precambrian lamproites are quite rare in the geological record compared to those recorded from the Phanerozoic. As a result, most of the current lamproite petrogenetic models are strongly influenced by studies on Phanerozoic lamproites (e.g., Leucite Hills, Smoky Butte, West Kimberley, Gaussberg, and the Mediterranean region; Murphy et al. 2002; Carlier and Lorand 2003; Coban and Flower 2006; Mirnejad and Bell 2006; Davies et al. 2006; Prelević et al. 2012). A significant number of Mesoproterozoic lamproites occur in the Eastern Dharwar, the Bastar and the Bundelkhand cratons of the Indian shield (Chalapathi Rao et al. 1996, 1999; Kumar et al. 2001; Osborne et al. 2011; Sahu et al. 2013) and they provide a unique opportunity to understand the tectono-magmatic processes when these cratons constituted a part of Proterozoic supercontinents such as Columbia and Rodinia. Furthermore, the spatial association of Mesoproterozoic lamproites, some of which are diamondiferous, with two of the major Proterozoic Purāna sedimentary basins of peninsular India – the Vindhyan (Chalapathi Rao 2005 and references therein; Masun et al. 2009) in central India and the Cuddapah (Bergman 1987; Chalapathi Rao 2007) in southern India – remains one of the puzzling aspects of Indian geology. Of all these lamproite occurrences, only those from the Vindhyan Basin (Majhgawan, Hinota and Bunder) are diamondiferous whereas the lamproites from the Cuddapah Basin (Chelima and Zangamarajupalle) are non-diamondiferous (Geological Survey of India 2011). This scenario is intriguing since the gravels of the rivers Krishna and Tungabhadra (flowing through the Cuddapah Basin) and the Banganapalle Conglomerate horizon (belonging to the Kurnool Group of sedimentary rocks) within the Cuddapah Basin, have yielded the world-famous historical diamonds such as the Koh-I-Noor, Pitt, and Orloff implying that the primary source of these diamonds still remains elusive despite centuries of search (see Radhakrishna 2007 and the references therein).

Against this backdrop, De Beers India Exploration company has discovered lamproitic dykes intruding the Tadpatri shales of the Lower Cuddapah Supergroup, located west of Banganapalle town near Garledinne (Joy et al. 2012) in the south-western part of the Cuddapah Basin. Only reconnaissance petrological and geochemical information of these lamproite dykes is available and petrogenetic details are lacking. The Garledinne lamproite dykes thus offer an excellent opportunity to investigate the sub-continental lithospheric mantle (SCLM) in an entirely new geological domain within the SW part of the Cuddapah Basin. The major objectives of this paper are to: (i) document the detailed petrology and geochemistry of the Garledinne lamproite dykes so as to expand the available data set on such rocks from the Cuddapah Basin and the Eastern Dharwar Craton (EDC; Fig. 1a); (ii) compare the new data with those of the previously well-studied Chelima and Zangamarajupalle lamproites within the Cuddapah Basin (hence forth termed as Cuddapah lamproites) and the EDC lamproites (Krishna and Ramadugu lamproite fields) so as to decipher their similarities and differences; (iii) infer the petrogenesis of the Garledinne lamproite dykes; and (iv) better understand the geodynamic evolution of the SCLM beneath the Cuddapah Basin.

Fig. 1
figure 1

a A generalised geological map of the Dharwar craton, southern India (after Naqvi and Rogers 1987) showing the Paleo-Mesoproterozoic Cuddapah basin and surrounding areas with the locations of kimberlite, lamproite and syenite occurrences. WKF= Wajrakarur kimberlite field; RKF= Raichur kimberlite field; NKF= Narayanpet kimberlite field; RLF= Ramadugu lamproite field; KLF= Krishna lamproite field; VLF= Vattikod lamproite field. Elchuru (E), Kotappakonda (K), Purimetla (P), Uppalapadu (U) and Errakonda (Ek). b Stratigraphic column of the Cuddapah basin (after Nagaraja Rao et al. 1987; Anand et al. 2003) showing the associated igneous rocks and their radiometric ages c Garledinne lamproite dyke of this study showing cross-cutting relationship with the Tadpatri shales

Full size image

Geological setting and background information

The crescent-shaped Cuddapah Basin, covering an area of 44,000 km2 (Fig. 1a), is one of the largest of the Purāna (Holland 1907) Paleo- to Meso-Proterozoic sedimentary basins of India. It has been extensively studied for well over a century with regard to its structure, tectonics, igneous activity, sedimentation, stratigraphy, geochronology and mineral wealth (e.g., King 1872; Nagaraja Rao et al. 1987; Anand et al. 2003; Chetty 2011; Patranabis-Deb et al. 2012; Tripathy and Saha 2013; Basu and Bickford 2013; Bickford et al. 2013; Collins et al. 2014; Deb and Pal 2014). Despite these studies, the origin and evolution of the Cuddapah Basin still remains unclear with diverse models in vogue viz.: (i) Mesoproterozoic rifting (Chaudhuri et al. 2002); (ii) astrobleme impact (Krishna Brahmam and Dutt 1992); (iii) thermal subsidence (Bhattacharji and Singh 1984; Drury 1984; Mall et al. 2008); (iv) separation of the Napier Complex (presently in Antarctica) from the southern Indian block (Mohanty 2011); and (v) a peripheral foreland basin that evolved through continent-continent collision (Singh and Mishra 2002).

The Cuddapah Basin, with an estimated in-fill thickness of 12 km, contains rocks belonging to the Cuddapah Supergroup that is overlain by the younger Kurnool Group (Fig. 1a and b). The latter are hosted in two sub-basins - the Kurnool and Palnad basins (Fig. 1a). Extrusive as well as intrusive igneous rocks are present throughout the Cuddapah Supergroup with the exception of the uppermost Nallamalai Group. Extensive sub-aerial and submarine lava flows, pillow lavas, ash beds, ignimbrites, agglomerates, dykes and sills occur predominantly in the Vempalle Formation and to a lesser extent in the Tadpatri Formation along the SW fringes of the basin (Sen and Narasimha Rao 1971; Srikantia 1984; Chatterjee and Bhattacharjee 1998; Sesha Sai 2014). Most of the lavas are essentially basalt to trachy-basalt with very few being picrite (Anand et al. 2003; Mallikharjuna Rao et al. 2007). Minor lava flows, including rare spilitised basalts, are reported from the Tadpatri Formation (Reddy 1988) but no volcanic rocks are reported in the rocks younger than Tadpatri Formation. Intrusive rocks in the form of sills and dykes are present throughout the Lower Cuddapah formations and are mostly dolerite/gabbro, rarely picrite and a few are lamproite. The Cumbum Formation is intruded by lamproites at Chelima and Zangamarajupalle, and riebeckite-bearing alkali syenites at Racherla, Idamakallu and Giddalur (Fig. 1a; King 1872; Madhavan et al. 1994). The Kurnool Group is devoid of any igneous activity except for felsic (rhyolitic) tuff beds (Saha and Tripathy 2012) that have been recently re-interpreted to be volcaniclastic sandy mudstone (Bickford et al. 2013). A stratigraphic column of the Cuddapah Basin is provided in Fig. 1b with all available geochronological information.

Four sets of lamproite dykes, intruding the Tadpatri shales of the lower Cuddapah Supergroup, were reported by the De Beers Company in three areas near Banganapalle town: (i) in a road material quarry 4.5 km west, (ii) 1.7 km to the SW; and (iii) in a valley 24 km due west and exposed along a stream bed (Joy et al. 2012) near Garledinne village. In this study, we sampled a Garledinne lamproite dyke (15°18“27”N; 78° 00“8”E; Fig. 1c) that trends WNW-ESE and corresponds to the 188 551/110/0003 lamproite recorded by Joy et al. (2012). The dyke shows pinching and swelling and extends for about 30 m along strike. Based on the age grouping of 1287–1370 Ma for some of the zircons recovered from heavy mineral stream samples in the Banganapalle waste dumps, Joy et al. (2012) infer the emplacement age of these lamproites to be Mesoproterozoic (1.28–1.37 Ga). We have obtained a slightly older whole-rock 40Ar/39Ar plateau age of 1433 ± 8 Ma (2 σ) for a sample collected from the Garledinne lamproite (written comm. Ray Burgess, Manchester).

Sampling and analytical work

Samples of the freshest looking surface material were collected from the Garledinne dyke along its strike. Petrographic studies were carried out by deploying an Olympus BX51 research petrological microscope at the Department of Geology, BHU, Varanasi. Minerals were analysed using a CAMECA-SX100 electron microprobe (EPMA) at Mineral Resources, Technical University of Clausthal, Germany. The analyses were carried out using wavelength-dispersive spectrometry. An accelerating voltage of 15 kV (for major elements) and 20 kV (for trace elements), a beam current of 50 nA and a beam diameter of 1 μm were used employing TAP, PET and LIF crystals and a PAP online correction program. Several in-house natural and synthetic standards were used for calibration. After repeated analyses it was found that the error on major element concentrations is <1 % whereas the error on trace elements varied between 5 and 10 %.

Whole-rock chemical analysis of five samples, carefully handpicked and free from xenoliths, was carried out at Activation Laboratories, Ancaster, Canada. After multi-acid digestion, ICP-OES (Model: Thermo-JarretAsh ENVIRO II) was used to analyse major elements and a few trace elements (Sc, V, Ba, Sr, Zr, and Y), whereas ICP-MS (Model: Perkin Elmer Sciex ELAN 6000) was used to determine the remaining trace and rare-earth element concentrations. STM1 MRG1, DNC1, W2, and SY3 were used as internal standards and the precision is approximately 5 and 5–10 % for the major oxides and trace elements, respectively, when reported at 100x detection limit. The analytical procedure is detailed by Gale et al. (1997) and available at the Activation Laboratories Ltd website (http://www.actlabs.com). The mineral chemistry data is presented in Tables 1, 2, 3, 4 and 5 and whole-rock geochemical data is provided in Table 6.

Table 1 Representative mineral chemistry composition (oxide wt%) of phlogopite from the Garledinne lamproite dykes. (R) represents rim composition whereas all other data are from cores
Full size table
Table 2 Mineral chemistry composition (oxide wt%) of phlogopite from the Garledinne lamproite dykes
Full size table
Table 3 Representative mineral chemistry composition (oxide wt%) of groundmass spinel (core) from the Garledinne lamproite dykes
Full size table
Table 4 Representative mineral chemistry composition (oxide wt%) of apatite (core) from the Garledinne lamproite dykes
Full size table
Table 5 Representative mineral chemistry composition (oxide wt%) of carbonate, chlorite and rutile from the Garledinne lamproite dyke
Full size table
Table 6 Bulk-rock geochemistry of the Garledinne lamproite dykes
Full size table

Petrography

Petrographic studies by optical microscopy and back-scattered electron imaging (Fig. 2a–f) reveal that the Garledinne lamproite has undergone varying degree of pervasive (deuteric or hydrothermal) silicification and carbonate alteration. However, the overall texture and some relict primary mineralogy are preserved and provide important insights into the original nature of the protolith. The lamproite is porphyritic to weakly porphyritic and contains olivine macrocrysts as well as microphenocrysts (identified by their habit) entirely pseudomorphosed either by carbonate or by silica. The presence of pseudomorphed olivine is further supported by the substantial amounts of Ni, Cr and V (see ‘Geochemistry’ section below) in these rocks. Highly pleochroic phlogopite is the most abundant pristine phase preserved in these rocks. It is present mostly in a size ranging from 150 to 500 μm as phenocrysts, microphenocrysts and very rarely as irregular poikilitic crystals in the matrix (Fig. 2a, b and c). Phlogopite is orange–yellow with a reddish tinge, which is widely regarded to be a characteristic feature of titanium-rich micas (Mitchell and Bergman 1991). Alteration of phlogopite to chlorite and serpentine is mostly seen at the margins and rare carbonatised laths are also present. The groundmass chromian spinel (up to ~50 μm), apatite and rutile constitute the accessory phases and carbonate (mostly calcite), chlorite, quartz and barite constitute the secondary phases present in varying amounts. Apatite is a common accessory phase and occurs in three modes: (i) inclusions in phenocrysts and microphenocrysts of phlogopite (Fig. 2d); (ii) subhedral prismatic grains or crystal groups in the groundmass (Fig. 2d); and (iii) acicular laths of up to 500 μm in length (Fig. 2f). Occasional oval to sub-rounded xenocrysts of carbonate (Fig. 2c), derived from the carbonate horizons of the Vempalle Formation with shaly intercalations (of the Tadpatri Formation), are also present.

Fig. 2
figure 2

Photomicrographs and back scattered electron images (BSE) of the Garledinne lamproites a Overall texture of the lamproites; Note the alteration due to carbonatization and abundance of phlogopite in the matrix (NVC/GDL/CB; plane light); b A higher magnification image showing the variably sized deep-orange laths of phlogopite with reddish margins (NVC/GDL/A; plane light); c A large (~2 mm sized) carbonate xenolith with shaly intercalations at its left-hand edge entrained in the lamproite; note the presence of rutile and phlogopite in the groundmass (NVC/GDL; plane light); d BSE image showing partially carbonated matrix phlogopite and associated euhedral apatite (NVC/GDl); e BSE image showing variably sporadic distribution of spinel in a matrix dominated by carbonate and phlogopite (NVC/KVS/CB-1); and f Apatite (up to 200 μm) also occurs in acicular habit in the groundmass ((NVC/KVS/CB-2). Abbreviations: Cb carbonate, Ph phlogopite, Ru rutile, Sp spinel, Ap apatite, Xen carbonate xenolith, Sh shaly intercalations, Ba barite, Ch chlorite, Qt quartz

Full size image

Mineral chemistry

The chemical compositions of phlogopite (Tables 1 and 2) spinel (Table 3), apatite (Table 4) and carbonate, chlorite and rutile (Table 5) in the Garledinne lamproite are briefly summarized here. EPMA studies (Tables 1 and 2) reveal that the phlogopite exhibits a wide compositional range in its TiO2 (5.83–7.77 wt%), Al2O3 (8.76–10.41 wt%) and FeOT (6.12–10.63 wt%) content. In Mg/(Mg+Fe) and Si (atoms per formula unit) space (Fig. 3a) the samples are compositionally similar to phlogopite from the other Cuddapah and Ramadugu lamproites and are distinct from those of the Krishna lamproites which are more evolved and some of which are biotite. In terms of TiO2 and Al2O3 composition (Fig. 3b), phlogopite in the Garledinne lamproite is similar to that from the: (i) Cuddapah, Krishna and Ramadugu lamproites and (ii) Leucite Hills lamproites (U.S.A), West Kimberley lamproites (Australia), and the Mediterranean lamproites (Fig. 3b). With increasing titanium content, the Garledinne phlogopite displays a moderate decrease in alumina (Fig. 3b) similar to other Indian lamproites, but lacks strong Al depletion characteristic of some lamproites (Mitchell and Bergman 1991).

Fig. 3
figure 3

a Si (atoms per formula unit) versus Mg/(Mg + Fe) mica classification diagram (after Reider et al. 1998). Data sources: Krishna lamproites (Paul et al. 2007; Chalapathi Rao et al. 2010); Cuddapah lamproites (Chelima and Zangamarajupalle) (Chalapathi Rao et al. 2004; Chalapathi Rao 2007); Ramadugu lamproites (Chalapathi Rao et al. 2014a). b TiO2 vs Al2O3 (wt%) variation plot for groundmass phlogopite. Data symbols and sources are the same as in Fig. 3a. Fields of kimberlites, Leucite Hills, West Kimberley, Smoky Butte and Mediterranean lamproites are taken from Gibson et al. (1995 and the references therein) and Fritschle et al. (2013)

Full size image

Spinel from the Garledinne lamproite is chromite and rarely magnesiochromite (Table 3; Fig. 4a). When projected on to the front face of a reduced Fe spinel prism, the Garledinne lamproite spinels do not display a marked titanian magnetite (T2) trend that is similar to the trend displayed by spinel from lamproites and orangeites (Fig. 4b; Mitchell 1995). However, the range in Fe2+/(Fe2++Mg) at relatively constant high Cr# does appear to be distinctive of lamproites (cf. Barnes and Roeder 2001). Chromian spinel is absent from other Cuddapah lamproites and the wide range of chemical variation displayed by the Garledinne chromian spinel is analogous to that of the chromian spinel from the Ramadugu lamproites (Fig. 4b). However, the Ti content of the Garledinne chromian spinel is markedly lower than that reported from the chromian spinel of Late Cretaceous ultrapotassic rocks from the Damodar valley, eastern India, also shown for comparison in Fig. 4b.

Fig. 4
figure 4

a Mg# [Mg/Mg + Fe2+] versus Cr# [Cr/Cr + Al] (after Pe-Piper et al. 2014) showing the compositional range of the spinel in this study. Open triangles represent spinel composition from 117 Ma ultrapotassic rocks from Damodar valley, eastern India (Mitchell and Fareeduddin 2009; Chalapathi Rao et al. 2014b) shown for comparison. Symbols and sources for the other lamproite spinel data are the same as in Fig. 3a. b Compositional variation of spinel from the Garledinne lamproites of this study projected onto the front face of a reduced spinel prism. Compositional fields and trends for spinel from kimberlites (T1) and lamproites (T2) are taken from Mitchell (1995). Data symbols and their source are the same as in Fig. 4a

Full size image

Strontium enrichment (SrO up to 3.44 wt%) is a characteristic feature (Table 4) of apatite from the Garledinne lamproite and in terms of SrO and CaO contents it is indistinguishable from the apatite compositions reported from the Cuddapah, Ramadugu, Damodar valley and West Kimberley lamproites (Fig. 5); however, all of them are quite distinct from apatite of the Krishna lamproites which is much impoverished in SrO (Fig. 5). The abundance of light rare earth elements (viz., La2O3 up to 0.58 wt% and Ce2O3 up to 1.31 wt%) implies that apatite is the main REE host in the Garledinne lamproite dykes.

Fig. 5
figure 5

CaO (wt%) versus SrO (wt%) of apatite from the samples under study. Data symbols and their source are the same as in Fig. 4a. Note the similarity in composition of Garledinne apatite with that from Ramadugu

Full size image

Compositions of the carbonate component (including that occurring in carbonate xenoliths), chlorite and rutile are presented in Table 5. The data show little variation. Rutile is characteristically niobian (Nb2O5 up to 2.56 wt%) and iron bearing (FeO up to 1.65 wt%) (Table 5). Such niobian rutiles have been reported from the Chelima lamproite (Chalapathi Rao 2011) and also from ultrapotassic rocks from the Damodar Valley, eastern India (see Chalapathi Rao et al. 2014b and references therein).

Whole-rock geochemistry

The whole-rock chemistry of the Garledinne lamproites is presented in Table 6 and is evaluated here in conjunction with such data generated by Joy et al. (2012). The Garledinne lamproites have restricted ranges of SiO2 (44.64–47.42 wt%), MgO (8.3–11.4 wt%) and Mg# (75–77; Table 6) which are indistinguishable from those of the Cuddapah, Ramadugu and Krishna lamproites (Fig. 6a). Decreasing MgO is accompanied by an increase in SiO2 implying crystal fractionation. Potassium (K2O: 0.09–0.44 wt%) contents are low, consistent with the altered nature of some of the phlogopite crystals and the absence of K-feldspar (sanidine), but nevertheless far exceed that of sodium (Na2O: 0.02–0.04 wt%) identifying the rock as distinctly ultrapotassic (Table 6). In the K2O/Al2O3 versus SiO2 (wt %) diagram for classifying ultrapotassic rocks (Fig. 6b), the Garledinne lamproites show a marked affinity with those from the Roman province and some samples plot in the field of kamafugites. Their lower K2O/Al2O3 ratio, compared to other Cuddapah lamproites (Fig. 6b) is due to the altered nature of their micas (above), but despite their silicification their whole-rock SiO2 contents are comparable with those of other southern Indian lamproites (Fig. 6b). Generally, Al2O3 and TiO2 constitute the least mobile of the major elements during alteration, and in their bi-variate plot (Fig. 6c) the Garledinne lamproites display strong similarities to the other Cuddapah, EDC and Western Australian lamproites. However, their high TiO2 contents, a hallmark of lamproites (Foley et al. 1987) and some ultramafic lamprophyres (Tappe et al. 2004), are markedly higher than those from the subduction-related lamproites from Spain and Italy (Fig. 6c). The Fe2O3 (T) contents of the Garledinne lamproites (6.5–7.47 wt%; Table 6) are lower than those from the Ramadugu (5.93–11.01 wt%; Chalapathi Rao et al. 2014a), Krishna (10.06–27.82 wt%; Paul et al. 2007; Chalapathi Rao et al. 2010) and Cuddapah lamproites (7.08–10.42 wt%; Chalapathi Rao et al. 2004) and reflect the absence of ilmenite and paucity of other iron oxides. Both CaO contents and Loss on Ignition (LOI) of the Garledinne lamproites are >10 wt% (Table 6) and reflect the carbonate and other volatile components. P2O5 varies from 1.07 to 1.62 wt% (Table 6) and generally is hosted by apatite.

Fig. 6
figure 6

Major oxide variation (wt%) diagrams of a MgO vs SiO2, b SiO2 vs K2O/Al2O3, c Al2O3 vs TiO2 d MgO vs CaO/ Al2O3, e K2O vs CaO/Al2O3 and f Al2O3 vs CaO for the Garledinne samples (including those reported by Joy et al. 2012) compared with those from RLF (Chalapathi Rao et al. 2014a); KLF (Paul et al. 2007; Chalapathi Rao et al. 2010); Cuddapah lamproites (Chalapathi Rao et al. 2004; Chalapathi Rao 2007). Various compositional fields shown in (a) and (c) are from Foley et al. (1987) and Altherr et al. (2004) respectively. The vectors shown in (e) and (f) are from Kirchenbaur et al. (2012). Data symbols remain the same as in (a) for all plots

Full size image

In the bi-variate plots involving MgO, Al2O, K2O, and CaO (Fig. 6d, e and f), the Garledinne lamproites show systematic trends, that might indicate olivine fractionation, and exclude clinopyroxene fractionation as displayed by some of the EDC lamproites. However, the trend shown in Fig. 6d could be a reflection of increasing secondary carbonation, and the trend shown in Fig. 6e very likely indicates increasing secondary carbonate and decreasing K2O with alteration. Similarly, Al2O3 remains fairly constant in Fig. 6f and the increase in CaO probably is related to the secondary carbonate content. Therefore, the trend displayed in this diagram (Fig. 6f) for these Garledinne samples bears no relationship to the illustrated vector, contrary to the apparent trend seen in other Indian lamproites. In order to further evaluate the effects of secondary carbonation and silicification on the whole-rock geochemistry of the Garledinne lamproites, various bi-variate plots involving critical major and trace elements have been additionally deployed. A good correlation is displayed between (i) various major oxides and trace elements [MgO vs Ni (Fig. 7a), MgO vs Cr (Fig. 7b), K2O vs Rb (Fig. 7c)] and (ii) various LILE (large ion lithophile elements) and HFSE (high field strength elements) including fluid-mobile trace elements such as U and Th [Zr vs Hf (Fig. 7d); Nb vs Ta (Fig. 7e); La vs Zr (Fig. 7f); La vs Ba (Fig. 7g); U vs Th (Fig. 7h)] demonstrating that despite being subjected to secondary alteration, the bulk-rock composition of the Garledinne lamproites, apart from K2O and CaO, remained largely unaffected and of petrogenetic significance.

Fig. 7
figure 7

Bi-variate plots between a MgO (wt%) vs Ni (ppm), b MgO (wt%) vs Cr(ppm), c K2O (wt%) vs Rb (ppm), d Zr vs Hf (ppm), e Nb vs Ta (ppm), f La vs Zr (ppm), g La vs Ba (ppm) and Th vs (U (ppm) for the Garledinne lamproite samples of this study. Data, where available, of Joy et al. (2012) are also shown (open diamonds) for comparison. Note the varying degree of good correlation displayed in all plots

Full size image

HFSE such as Zr and Nb are considered to be least mobile during low-temperature hydrothermal alteration, and hence are widely used to discriminate between lamproites and kimberlites (e.g. Paton et al. 2009). In such a plot (Fig. 8), the Garledinne samples show stronger affinities to the Ramadugu and Krishna lamproites than to other Cuddapah lamproites. The zirconium contents of the Garledinne lamproites are also elevated implying the possible presence of a zirconium-bearing phase such as zircon, which, however, could not be identified during this study.

Fig. 8
figure 8

Zr and Nb (ppm) contents of the Garledinne lamproites compared with those from other lamproites from EDC and Cuddapah basin. The dotted line depicts the field of ultrapotassic rocks with very low Nb (<50 ppm) which are considered to be characteristic of subduction-related settings (taken from Sheppard and Taylor 1992) together with the other fields shown for comparison. The data symbols and their sources are the same as in Fig. 6

Full size image

Highly fractionated chondrite-normalised REE distribution patterns with a marked enrichment of LREE over HREE (La/Yb up to 347) are a characteristic feature of the Garledinne lamproites (Fig. 9a). Whereas their LREE contents are similar, their HREE display a considerable range suggesting a varying degree of partial melting experienced by their source region. The REE distribution patterns of the Garledinne lamproites parallel the distribution patterns of lamproites from the Cuddapah, Krishna and Ramadugu (Fig. 9a) suggesting an overall similarity in their genesis.

Fig. 9
figure 9

a Chondrite-normalized (after Evenssen et al. 1978) rare earth element distribution patterns (solid lines) for lamproites from Garledinne compared with those from representative samples (labelled in the legend) of Ramadugu (Chalapathi Rao et al. 2014a), Chelima and Zangamarajupalle (Chalapathi Rao et al. 2004) and Krishna (Chalapathi Rao et al. 2010) lamproites. Lower (depicted by a dotted line) and variable HREE content is a hall-mark of Garledinne lamproites. b Primitive mantle normalized (Palme and O’Neill 2003) multi-element patterns for Garledinne samples of this study compared with those from other lamproites from Cuddapah basin and EDC (data sources are the same as in Fig. 9a, as well as ultrapotassic rocks from Damodar valley, eastern India (data source: Chalapathi Rao et al. 2014b). The crosses represent calculated trace element patterns for near-solidus partial melts of carbonated peridotite at 8.6 GPa (Dasgupta et al. 2009). The model patterns were obtained by using the experimentally determined bulk peridotite/melt partition coefficients and a 0.1 % batch partial melting model adapted from Tappe et al. (2013)

Full size image

Primitive-mantle normalised incompatible trace-element patterns of the Garledinne lamproite samples are provided in Fig. 9b. Their incompatible element abundances predominate over those of the primitive mantle with LILE showing a large enrichment in comparison to HFSE. Negative anomalies at Rb, K, Pb and Sr occur in all Garledinne samples whereas such anomalies for U are displayed by only two samples. A number of plausible reasons exist for the anomalies in the multi-element normalized patterns of potassic-ultrapotassic rocks, which attribute them either to: (i) the residual phases in the source; (ii) hydrothermal and/or secondary alteration; or (iii) fractionating phases (see Becker and Le Roex 2006; Coe et al. 2008). Whilst fluid-mobile behaviour of some of these elements (Rb, K, Pb and U) could have partially contributed to their negative anomalies, good correlations amongst many of these elements (Fig. 7) compel us to infer such anomalies to be source-related. The effect of olivine fractionation could also have contributed to the varying depletion (Fig. 9b). An overall similarity, especially for the incompatible trace elements (Sr to Lu) in the multi-element patterns (Fig. 9b) for the Garledinne and other Cuddapah and EDC lamproites as well as ultrapotassic rocks from the Damodar valley, eastern India, additionally suggest similarity in their petrogenetic evolution.

Discussion

The Garledinne lamproite dykes have undergone secondary alteration but their textural aspects (porphyritic to weakly porphyritic), relict minerals and their mineral compositions (Ti-rich phlogopite, chromian spinel, Sr-rich apatite and niobian rutile) and bulk-rock geochemistry (impoverishment of sodium, abundance of LILE, HFSE, high LREE and depleted HREE) are similar to those of other lamproites both in southern Indian and elsewhere (e.g., West Kimberley, Leucite Hills, and Smoky Butte). The Garledinne lamproite strikingly resembles the Chelima and Zangamarajupalle lamproites located elsewhere in the Cuddapah Basin (see Chalapathi Rao et al. 2004; Chalapathi Rao 2007) in having abundant secondary carbonate and a lack of other well-preserved ferro-magnesian phases apart from phlogopite.

There are some conspicuous mineralogical differences between the Cuddapah Basin lamproites (including the Garledinne lamproite) and the EDC lamproites. The Cuddapah Basin lamproites, in general, lack sanidine, clinopyroxene, potassic richterite, and titanite – all of which have been recorded from the EDC lamproites (see Chalapathi Rao et al. 2004, 2010, 2014a). We relate this to either mantle heterogeneity in their source regions or to the difference in the evolution of their magmas. On the other hand minerals such as leucite, priderite, wadeite and perovskite, reported from lamproites worldwide are not found in any of the southern Indian lamproites and they resemble Mediterranean lamproites in this aspect (see Akal 2008). Interestingly, a recent Re-Os isotopic study involving samples from the Krishna lamproite field (KLF; at the NE margin of the Cuddapah Basin) show them to have very radiogenic γOs values (+56 to +355), indicating high Re/Os material – possibly a subducted component – in their source, that are strikingly similar to the compositions of subduction-related lamproites of the Italian peninsula (Chalapathi Rao et al. 2013b). Based on Sr-Nd-Hf-Pb isotopes and major- and trace-element ratios, the involvement of Archaean carbonate-metasomatized subducted komatiite in the source regions of the Krishna lamproites has also been inferred (Chakrabarti et al. 2007). Therefore the mineralogical similarity of the cratonic southern Indian lamproites with that of the Mediterranean lamproites is likely due to similarity in the ancient (Archaean) subduction-related events experienced by the source regions of the former. This aspect merits further evaluation using multi-element isotope systematics involving samples from all the lamproite fields and clusters.

Hydrothermal alteration and the resulting carbonatization and silicification of the Garledinne lamproites can account for their low (<1) perpotassic and peralkaline indices. In fact, extensive secondary alteration appears to be a characteristic feature of all lamproites from the Cuddapah Basin implying a widespread hydrothermal imprint, after lamproite emplacement, across the basin. On the other hand, strong correlations between various major and trace elements and their ratios in the, Garledinne lamproites imply that the geochemical trends (barring those that involve K and Ca) are of primary nature and remained intact despite alteration. Following the modal classification of lamproites (Mitchell and Bergman 1991) the Garledinne dykes correspond to the olivine-phlogopite lamproite variety.

Crustal contamination

With the exception of a solitary carbonate xenolith, no other crustal or mantle xenoliths were observed in the studied Garledinne lamproite samples that contain low silica (SiO2: 44.64–47.71 wt%), alumina (Al2O3: 5.98–6.65 wt%) and particularly low sodium (Na2O <0.05 wt% Table 6). The Tadpatri Shale of the Cuddapah Supergroup and the granitoids of the EDC, which constitute the basement of the Cuddapah Basin, appear to have had little influence on the composition of the Garledinne lamproites. Assimilation of shale, and felsic minerals such as quartz and feldspar derived from the granitoids, would have resulted in a marked elevation of silica, alumina and sodium that is not the case here. Besides, systematic variation trends displayed between the bulk-rock major and trace elements cannot be explained by assimilation. Furthermore, the abundance of compatible trace elements such as Cr (up to 690 ppm), Ni (up to 480 ppm) and Co (up to 42 ppm), as well as incompatible trace elements such as Sr (up to 1522 ppm), Zr (up to 1940 ppm), Nb (up to 158 ppm) and La (up to 334 ppm) in the Garledinne lamproites is several orders higher in magnitude compared to the continental crust. High Mg # (75–79), highly fractionated REE patterns, low HREE, Y contents and absence of positive Eu and Pb spikes, additionally argue against significant assimilation of crustal material.

The Ce and Nb contents and Ce/Pb and Nb/U ratios of the Garledinne lamproites are: (i) much higher (Fig. 10a and b) than those of average upper- and lower- continental crust material; (ii) close to the supposedly uncontaminated mantle reservoirs such as MORB and OIB; and (iii) strikingly similar to those of other lamproites from the Cuddapah Basin, EDC (especially Ramadugu), West Kimberley and Smoky Butte. The Hf/Sm ratios of the Garledinne lamproites slightly overlap with the field of subducted continental crust-derived ultrapotassic rocks from East China but are clearly higher compared to the pelagic material enriched mantle source regions of ultrapotassic rocks from southern Tibet (Fig. 10c). This again excludes significant assimilation of upper crustal and pelagic material either by contamination or by subduction. This inference is also supported by the lack of Nb-Ta-Ti depletions in the primitive mantle normalized multi-element diagrams (Fig. 9 above) and is consistent with the findings from several previous studies on lamproites which revealed that crustal contamination has limited influence on their extremely high incompatible trace element enrichments (e.g. Nelson 1992; Murphy et al. 2002; Davies et al. 2006; Mirnejad and Bell 2006; Chakrabarti et al. 2007).

Fig. 10
figure 10

a Ce (ppm) vs Ce/Pb and b Nb (ppm) vs Nb/U bi-variate plots for the Garledinne samples compared with those from other lamproites from EDC and Cuddapah basin. The data symbols and their sources are the same as in Fig. 6. The compositional fields of lower continental crust, upper continental crust, MORB and OIB (the range is shown by a gray field and median by a broken line) are also shown for comparison and are taken from Sun et al. (2014). Various lamproite fields are taken from Davies et al. (2006) and Yilmaz (2010). c Nd/Hf vs Hf/Sm plot (after Wang et al. 2014) for the Garledinne lamproites compared with those from other lamproites from EDC and Cuddapah basin. The data symbols and their sources are the same as in Fig. 6. The field of continental subduction-related potassic and ultrapotassic rocks from East China and pelagic material enriched source rocks from southern Tibet are also shown for comparison (Wang et al. 2014). d Zr/Hf vs Nb/Ta (after Tappe et al. 2008) of the Garledinne lamproites compared with other EDC and Cuddapah basin lamproites. Data symbols and their sources are the same as in Fig. 6. Arrow indicates the trend of a melt formed in the presence of residual rutile in Bulk Silicate Earth-like (BSE) material. Note that Garledinne lamproites exclude crustal contamination and are distinct in their composition to the data of KLF and other Cuddapah lamproites but are strikingly similar to that of RLF

Full size image

Mantle source regions

Mantle-derived xenoliths or xenocrysts were not found in the Garledinne lamproites in this study. However, Joy et al. (2012) report a small population of garnets from crushing ~1000 kg of Garledinne lamproite material. Compositionally, these garnets belong to the fertile lherzolitic variety in contrast to the low-temperature depleted harzburgitic variety recorded from the kimberlites of WKF, towards the western margin of the Cuddapah Basin (Sastry et al. 2005). Geochemical characteristics such as high Mg#, abundance of compatible elements (Ni, Cr and Co), and low HREE contents imply a refractory mantle source that had experienced previous melt extraction (e.g. Mitchell and Bergman 1991; Beard et al. 2000; Mirnejad and Bell 2006) in the genesis of the Garledinne lamproites. Extensive extrusive and intrusive mafic magmatic activity (ca. 1.9 Ga) recorded as dykes, sills and volcanic rocks in the Vempalle and Tadpatri Formations of the Cuddapah Supergroup provide evidence for such previous melt extraction (Anand et al. 2003; French et al. 2007). However, the high concentration of incompatible trace elements, e.g., LILE and HFSE (above), highlights the subsequent enrichment of the mantle source that needs to be confirmed by isotopic evidence. Thus, a previously depleted as well as subsequently enriched mantle source, similar to that found in garnet peridotite nodules (Beard et al. 2007; Gibson et al. 2008) and commonly assumed as the source region of lamproites and kimberlites (e.g. Mitchell and Bergman 1991; Tainton and McKenzie 1994; Becker and Le Roex 2006), is also inferred for the Garledinne lamproites. The presence of xenocrystic garnet (Joy et al. 2012) and high La/Yb and low HREE contents also imply generation of the Garledinne lamproite magma by small degrees of partial melting from an enriched (metasomatised) mantle source within the garnet stability field.

The exact source of the melts and fluids responsible for an enriched (metasomatised) mantle source of potassic–ultrapotassic rocks remains unclear and a variety of models currently exist: (i) subducted lithosphere (e.g., Ersoy et al. 2014; Garza et al. 2013); (ii) convective (asthenospheric) mantle (e.g., McKenzie 1989); (iii) deep-mantle plumes (e.g., Torsvik et al. 2010); and (iv) cratonic sub-continental lithospheric mantle (e.g., McKenzie 1989; Foley 1992; Tappe et al. 2008). The composition of the melts involved in mantle metasomatism can also vary from carbonate and alkali-rich silicate melts to H2O- and CO2-rich fluids (e.g., Coltorti et al. 1999; Ionov et al. 2002). The primitive mantle normalized multi-element patterns (Fig. 9b) of the Garledinne lamproites exclude subduction-related signals and their Nb/Pb, and Ce/Pb contents are too high to be products of melts/fluids derived from subducted pelagic sediments (Fig. 10). Negative anomalies at K and Rb in multi-element plots (Fig. 9b) imply the presence of amphibole or phlogopite in the source (see Sato et al. 1997), although alteration has played an important role in leaching of these mobile elements and thereby enlarging the magnitude of these anomalies. Residual clinopyroxene (Tappe et al. 2004) or a source that had previously undergone depletion in Sr due to melt extraction (Mitchell 1995) can account for the depletions in Sr for the Garledinne lamproites (Fig. 9b). Likewise, the slight negative anomalies for P in some of the Garledinne samples suggest the presence of a phosphate in the residue (Mitchell and Edgar 2002). Available experimental evidence also favours phlogopite, potassic richterite, diopside and apatite to be some of the essential phases in lamproite source regions, very likely occurring as metasomatic veins within a lherzolitic or harzburgitic source (e.g., Edgar and Vukadinovic 1992; Foley 1992; Konzett et al. 1997).

The Garledinne lamproites have super-chondritic Zr/Hf (43–46) as well as sub-chondritic Nb/Ta (15–18) (Fig. 10d). An enriched mantle source, with the involvement of an asthenospheric component, is attributed to the super-chondritic Zr/Hf whereas a depleted sub-continental lithosphere is held responsible for sub-chondritic Nb/Ta (see Weyer et al. 2003). The Garledinne lamproites are thus inferred to have been derived from mantle sources that experienced depletion and subsequent enrichment. Zr/Hf and Nb/Ta contents of the Garledinne lamproites (Fig. 10d) are: (i) indistinguishable from those of the Ramadugu lamproites; (ii) markedly dissimilar to those of the other Cuddapah Basin lamproites and a majority of the Krishna lamproites; (iii) lack the effect of residual rutile in the BSE-like source; (iv) exclude signatures of continental crust; and (v) suggest involvement of an asthenospheric (oceanic basalt) component. The lack of Ti depletion in the primitive mantle normalized multi-element plot (Fig. 9b) additionally provides support for the absence of titanite in the residue. Recently, Tappe et al. (2013) calculated trace-element patterns for partial melting of a carbonated peridotite source between 6 and 10 GPa utilizing a fixed 0.1 % batch partial melting model (McKenzie 1985; Gudfinnsson and Presnall 2005) and the source trace-element contents set at primitive mantle values (Palme and O’Neill 2003). The ‘bulk’ peridotite/melt partition coefficients, reported by Dasgupta et al. (2009), were used for this purpose. The calculated trace-element compositions remarkably match those from the Lac de Gras kimberlites in Canada whose derivation was inferred to be from low-degree partial melting of carbonated peridotite under convective mantle conditions (Tappe et al. 2013). The melting model of Tappe et al. (2013), when adapted for the Garledinne lamproite (Fig. 9b), displays a very good fit for many of the incompatible trace elements suggesting the involvement of carbonated peridotite in their petrogenesis. A similarly good fit has also been observed in the case of other EDC and Cuddapah Basin lamproites and ultrapotassic rocks from Damodar valley (Fig. 9b).

Genesis

The ultimate site where the lamproite source regions have undergone partial melting remains controversial and ranges from the SCLM (e.g., Mitchell 2006; Davies et al. 2006; Avanzinelli et al. 2009) to the transition zone (e.g., Murphy et al. 2002). The poorly defined correlation of La/Yb with La (not shown) implies that partial melting together with fractional crystallization played an important role in the compositional variation of the Garledinne lamproites. Petrography as well as whole rock major-element chemistry have indeed demonstrated (above) olivine fractionation in the Garledinne lamproite dykes. Relatively lower Nb/Y with increasing SiO2 (wt%) contents (Fig. 11a) suggest that the Garledinne lamproites (together with other lamproites from the Cuddapah Basin and EDC) have been products of much greater degrees of partial melting than the kimberlites from the WKF and NKF. This is likely due to the location of the EDC and Cuddapah Basin lamproites at the margin of the craton where the lithosphere is relatively thin compared to its ‘core’ (e.g. Priestly and McKenzie 2006; Chalapathi Rao et al. 2010). The Garledinne lamproites have (La/Sm)N and (Ho/Lu)N contents indistinguishable from those in the other Cuddapah lamproites and kimberlites from WKF and NKF (not shown) implying the involvement of a similar mantle source in their genesis. The involvement of an asthenospheric (MORB and OIB) component in the genesis of the Garledinne lamproites and other lamproites from the Cuddapah Basin and EDC is apparent from their Nb/U, Ce/Pb, Zr/Hf, and Nb/Ta ratios (Figs. 10a–d). Thus, we infer the involvement of multiple reservoirs, viz., SCLM and asthenospheric sources, in the genesis of the Garledinne lamproites. Recent studies of potassic–ultrapotassic rocks have also invoked multiple reservoirs in their genesis (Le Roex et al. 2003; Tappe et al. 2011). The Th/Nb ratio, normalized to Zr (to eliminate the effects of fractional crystallization) can constitute a sensitive discriminator between orogenic and anorogenic tectonic settings of ultrapotassic rocks, with the former being characterized by a higher Th/Nb (see Kullerud et al. 2011). In such a plot (Fig. 11b) the Garledinne lamproites, along with the other lamproites from the Cuddapah Basin and a great majority of EDC are strikingly confined to the anorogenic lamproite field and are similar to the well-known extension-related ultrapotassic lavas of the eastern Virunga and western Anatolia volcanic fields in this aspect.

Fig. 11
figure 11

a Nb/Y vs SiO2 (wt%) diagram (Beard et al. 1998) illustrating the degree of partial melting in the Garledinne lamproites. The data (x) of EDC kimberlites (WKF and NKF) is also shown for comparison and is from Chalapathi Rao et al. (2004) and Chalapathi Rao and Srivastava (2009). Rest of the data symbols and their sources are the same as in Fig. 6b Nb * 100/Zr vs Th * 100/Zr (after Yilmaz 2010) of the Garledinne lamproites depicting their anorogenic geochemical character. Data symbols and their sources are the same as in Fig. 6. Other fields are taken from Yilmaz (2010) and Kullerud et al. (2011)

Full size image

Diamond potential

Experimental studies show that only olivine lamproites, which originate from depths of at least 150–160 km, can contain diamond whereas high-silica lamproites can be generated from mica harzburgites at fairly low pressures of less than 3 GPa and are essentially non-diamondiferous (Foley 1993). A recent synthesis of global lamproites and their diamond prospectivity lends support to these experimental findings (Birkett 2008). The olivine lamproitic nature of the Garledinne lamproites assumes significance in this context. Even though the Garledinne lamproites are inferred to be sources for the diamonds in the nearby Banganapalle conglomerates, no diamond has actually been recovered in their processing (Joy et al. 2012). However, Coulson (1933) recorded that several sills (their locations and mafic/ultramafic nature were not published) that intruded the lower Cuddapah Supergroup were diamondiferous. Therefore, the diamondiferous nature of the Garledinne (and hitherto unknown) lamproite dykes and sills in the Lower Cuddapah formations remains an open question.

Geodynamic implications

The eruption ages of the lamproites and kimberlites constitute time capsules in a geodynamic framework and are of tectono-magmatic significance (e.g., Jelsma et al. 2009; Foley et al. 2012). Available radiometric ages of the Cuddapah Basin and EDC lamproites show them to be of Mesoproterozoic age with a great majority emplaced between 1.25 and 1.45 Ga (Chalapathi Rao et al. 1999; Kumar et al. 2001; Osborne et al. 2011). These ages are strikingly similar to: (i) the emplacement age (1326 ± 73 Ma) of the Racherla alkali syenite within the Cuddapah Basin (Fig. 1a for location; Chalapathi Rao et al. 2012); (ii) the spectrum of alkaline and sub-alkaline syenite complexes and lamprophyres (1250–1350 Ma) emplaced elsewhere within the Eastern Ghats Mobile Belt, towards the eastern margin of the Cuddapah Basin (Fig. 1a for location; Upadhyay et al. 2006; Kumar et al. 2007); (iii) the whole rock depleted-mantle (TDM) Nd model ages (1.38–1.47 Ma) of the ≥620 Ma Tokapal kimberlite system, Bastar Craton, central India (Lehmann et al. 2007); (iv) the perovskite depleted-mantle (TDM) Nd model ages (1.3–1.5 Ga; Chalapathi Rao et al. 2013a) of the 1.1 Ga NKF and WKF kimberlites, EDC, southern India; and (v) depleted-mantle TDM Nd model ages (1.4–1.5 Ga) of the diamondiferous Majhgawan diatreme of the Bundelkhand Craton, northern India (Lehmann et al. 2006), thereby implying involvement of regional geotectonic controls, across these cratons and mobile belts, in all kimberlite and lamproite generations.

Available depleted mantle (TDM) Nd model ages (which are minimum ages) indicate that the source regions of the Cuddapah and EDC lamproites were enriched due to a regional metasomatic event at ~2 Ga (Chalapathi Rao et al. 1998, 2004; Chakrabarti et al. 2007). The source regions of alkaline and sub-alkaline syenites and lamprophyres located towards the eastern margin of the Cuddapah Basin, in the Eastern Ghats Mobile belt, also had been affected by this regional metasomatic event (Upadhyay et al. 2006; Kumar et al. 2007) together with the tholeiitic intrusives in the Napier Complex, Enderby Land, Antarctica (Chatterjee and Bhattacharjee 1998), thereby implying that this regional metasomatic event was widespread across the Proterozoic eastern Gondwana. In fact, the long-lived (~2.6 Ga) nature of this regional enrichment event has been demonstrated recently by depleted mantle (TDM) Nd model ages for apatite from the Racherla alkali syenite from within the Cuddapah Basin (Chalapathi Rao et al. 2012). It is very likely that the source region of the Garledinne lamproite has also undergone this regional metasomatic enrichment, and this needs to be confirmed by isotopic studies.

Interestingly, the emplacement ages (1.25–1.45 Ga) of the majority of the lamproites from the Cuddapah Basin and the EDC, and the depleted mantle (TDM) Nd model ages of the kimberlites from the EDC, Bastar and Bundelkhand cratons of India (above), coincide with the break-up of the Paleo-Mesoproterozoic supercontinent of Columbia (Rogers and Santosh 2009; Meert 2012). We consider this coincidence to be of geodynamic significance since the long-term extensional events associated with the break-up of Columbia could have constituted the regional tectonic control capable of triggering not only lamproite eruptions at 1.25–1.45 Ga, but also the enrichment of source regions of kimberlites located at some distance across the various cratons. The geochemical similarity of the Garledinne lamproites to the well-known extension-related ultrapotassic lavas of eastern Virunga and western Anatolia (Fig. 11b) assumes significance in this respect. In fact, a small amount of lithospheric extension, without the need for a mantle plume, has been invoked in the genesis of the kimberlites and lamproites from the EDC and the Cuddapah Basin (Chalapathi Rao et al. 2004, 2010, 2014a), as well as the alkaline rocks occurring within and towards the east of the Cuddapah Basin in the Eastern Ghats Mobile belt (Upadhyay et al. 2006; Chalapathi Rao et al. 2012).

Conclusions

The Garledinne lamproites were subjected to varying degrees of pervasive alteration and are characterized by widespread carbonatization and silicification. They have a porphyritic to weakly porphyritic texture comprising pseudomorphs after olivine, Ti-rich phlogopite microphenocrysts, chromian spinel, Sr-rich apatite and niobian rutile. The Garledinne (and other Cuddapah Basin) lamproites are mineralogically distinct from the Eastern Dharwar Craton lamproites in their lack of sanidine, clinopyroxene, potassic richterite, and titanite. Deuteric alteration and crustal contamination have had limited effect on the whole-rock geochemistry of the Garledinne lamproites and olivine fractionation played a dominant role in their evolution. The Garledinne lamproites probably formed as small-degree partial melts derived from a refractory peridotitic mantle source that was previously enriched by carbonate-rich fluids/melts within the garnet stability field. Both the sub-continental lithospheric mantle and asthenosphere are inferred in the genesis of the Garledinne lamproite dykes. The emplacement of the Garledinne lamproites probably was related to extensional events during the break-up of the Columbia supercontinent.