Late Devonian and Triassic Basalts from the Southern Continental Margin of the East European Platform, Tracers of a Single Heterogeneous Lithospheric Mantle Source

Home , Geology of Russia

Late Devonian and Triassic basalts from the southern continental margin of the East European Platform, tracers of a single heterogeneous lithospheric mantle source

Franc¸oise Chalot-Prat1,∗, Petr Tikhomirov2 and Aline Saintot3,4 1Centre de Recherches Petrographiques et Geochimiques – Nancy University, BP20, 15 rue Notre Dame des Pauvres, F-54501 Vandoeuvre-les-Nancy Cedex, France. 2Geological Faculty of Moscow State University; Vorobiovy Gory, 117311 Moscow, Russia. 3Vrije Universiteit, Instituut voor Aardwetenschappen, Tektoniek afdeling, De Boelelaan 1085, 1081 HV Amsterdam, Netherlands. 4Geological Survey of Norway, N-7491 Trondheim, Norway. ∗e-mail: [email protected]

In Late Devonian and Early-to-Late Triassic times, the southern continental margin of the Eastern European Platform was the site of a basaltic volcanism in the Donbas and Fore-Caucasus areas respectively. Both volcanic piles rest unconformably upon Paleoproterozoic and Late Paleozoic units respectively, and emplaced during continental rifting periods some 600 km away from expected locations of active oceanic subduction zones. This paper reports a comparative geochemical study of the basaltic rocks, and views them as the best tracers of the involved mantle below the Eastern European Platform. The Late Devonian alkaline basic rocks diﬀer from the calc-alkaline Triassic basic rocks by their higher alkali-silica ratio, their higher TiO2,K2O, P2O5 and FeO contents, their higher trace element contents, a higher degree of fractionation between the most and the least incompatible elements and the absence of Ta-Nb negative anomalies. These general features, clearly distinct from those of partial melting and fractional crystallization, are due to mantle source eﬀects. With similar Nd and Sr isotopic signatures indicating mantle-crust mixing, both suites would originate from the melting of a same but heterogeneous continental mantle lithosphere (refertilized depleted mantle). Accordingly the Nd model ages, the youngest major event associated with mantle metasomatism occurred during Early Neoproterozoic times (∼650 Ma).

1. Introduction Episodes of rifting destabilized the subsequent platform regime of the East European Craton, the The southern margin of the Eastern European first during the Riphean (Meso- to Neoprotero- Platform is structurally divided into two main zoic) and the second during the Late Devonian. parts: the Sarmatia segment of the East European The latter created the large Pripyat-Dniepr-Donets Craton and the Scythian Platform lying south of intra-cratonic rift basin (Bogdanova et al 1996; it (figure 1). Artemieva 2003). The Sarmatia cratonic area is formed by four or The origin and nature of the Scythian Plat- five Archean terranes, welded together at 2.3 and form basement (figure 1) remain unclear. It is 2.1 Ga (see Bogdanova et al 1996). overlain by a quite thick Phanerozoic sedimentary

Keywords. Basalts; East European Platform; Late Devonian; Triassic; Donbas; Fore-Caucasus; geochemistry; geodynamics; geology.

J. Earth Syst. Sci. 116, No. 6, December 2007, pp. 469–495 © Printed in India. 469 470 Fran¸coise Chalot-Prat et al

Figure 1. Late Paleozoic tectonic map of the southern East European craton and adjacent areas (modified after Tikhomirov et al 2004) in the hypothesis of a Variscan basement for the Scythian Platform (SP) (as for example in Zonenshain et al 1990). 1. Early Permian basins within the East European Craton; 2. Areas of Precambrian consolidation within the Late Paleozoic orogen; 3. Areas of Late Paleozoic consolidation; 4. Donets and Karpinsky basins, thought to be inverted during the Permian time; 5. Late Paleozoic and possibly younger thrusts; 6. Presumed shelf margin for the end of Permian; 7. Gissar-Donets fault zone; 8. Location of the Late Devonian (D) Donbas and Triassic (T) Fore Caucasus areas. A. Agrakhan–Guriev fault. cover (locally more than 10 km thick). Geophys- by episodes of rifting during the Late Devonian, ical data only allow characterizing the basement the Permo(?)-Triassic (Gaetani 2000), and further as few drill-holes reached at the deepest the Late southwards, Jurassic (Lordkipanidze et al 1989; Paleozoic succession. A classical model suggests Nikishin et al 1998a, b) and Cretaceous (Nikishin that the Scythian Platform is part of the Late et al 1998a, b). Variscan orogenic belt, between Western Europe Among the rift basins which formed on the and Urals (Arthaud and Matte 1977; Zonenshain Scythian Platform during Phanerozoic times et al 1990; Nikishin et al 1996 and presented as (figure 1), five were accompanied by significant such on figure 1). However, no strong penetrative eruptions and/or sub-surface intrusions of mantle- Late Paleozoic deformation is actually observed on derived magma: the Scythian Platform. Another hypothesis to test • the Late Devonian Dniepr-Donets-Donbas basin is that the Scythian Platform was part of the Late (Chekunov and Naumenko 1982; Wilson and Proterozoic Baikalian belt fringing Baltica (e.g., Lyashkevich 1996; McCann et al 2003); Saintot et al 2006). This Late Proterozoic orogenic • the Triassic Eastern Fore Caucasus basin event might have resulted from the amalgama- (Dubinski and Matsenko 1965; Burshtar et al tion/accretion of subduction-related arc/oceanic 1973; Nazarevich et al 1986; Tikhomirov et al complexes and micro-continental blocks (as the 2004); Beloretzk terrane along the Southern Proto-Urals, • the Jurassic Western Fore Caucasus basin (Lord- Glasmacher et al 1999; the Pechora-Barentsia ter- kipanidze et al 1989); ranes on the Timan-Pechora foldbelt, Nikishin et al • the Cretaceous Karkinit rift basin (western part 1996; see Torsvik and Rehnstrom 2001) along the of Scythian Platform) (Muratov 1969; Chekunov margin of the present-day northern and eastern et al 1976; Grigorieva et al 1981; Leschukh 1992; Baltica, without involving collision of large conti- Nikishin et al 1998a, b, 2001); nental masses. The Scythian Platform, considered • the Neogene Minvody basin (Stavropol High) herein as part of the Late Proterozoic Baikalian (Polovinkina 1960; Adamia and Lordkipanidze belt, became a renewed passive margin affected 1989). Late Devonian Triassic basalts East European Platform Margin 471

The two volcanic suites and associated sediments The Donbas rift belongs to the Prypiat-Dniepr- of interest herein, of the Donbas and of the East- Donets (PDD) mega-rift (figure 1). This Late ern Fore-Caucasus, were deposited in quite simi- Devonian structure was contemporaneous with lar tectonic settings. The Late Devonian event of (1) widespread peri- and intra-cratonic magmatism the Donbas and the Triassic event of the Eastern elsewhere on the whole Eastern European Platform Fore Caucasus are both related to continental rift- (about 4 million km2) from the Timan–Pechora ing of a previously deformed and peneplaned base- and East Barents Sea rift systems northwards to ment. The rifting was due to tensional lithospheric the Peri-Caspian Basin eastwards (Stephenson et al stresses in the Donbas (Saintot et al 2003), and 2006), (2) the accretion of an intra-oceanic vol- more likely transtensional in the Eastern Fore canic arc following the eastward subduction, there- Caucasus (Gaetani 2000). fore away from the Eastern European Platform of At Late Devonian times, subaerial basaltic lava (or part of) the Paleo-Ural Ocean (Puchkov 1997; flows were emplaced at the beginning of the Brown et al 1998, 2002; Brown and Spadea 1999), Donbas rift formation (figure 1) (McCann et al and (3) development of extensional basins on the 2003 and ref. therein), well after the final Protero- southern margin of the Scythian Platform (on the zoic consolidation of the Sarmatia basement (at Greater Caucasus area, Khain 1975). The back-arc 1500 Ma, according to Milanovsky 1996 and ref- position of such remains uncertain despite the fact erences therein, or at 2000–1800 Ma according to that a northward subduction zone of the Proto- Bogdanova et al 1996). Tethys Ocean has very often been adopted in From Early to Late Triassic times, three mafic plate kinematic models (e.g., Adamia et al 1981; volcanic suites, including some rhyolitic materi- Gamkrelidze 1986; Ziegler 1990; Zonenshain et al als, were emplaced successively in continental and 1990; Stampfli et al 2001; Stampfli and Borel 2002; marine sedimentary environments on the Eastern von Raumer et al 2003). The closest Paleozoic Fore Caucasus (figure 1) (Tikhomirov et al 2004 subduction zone, if any, could have been far to and ref. therein), some tens of millions years after the south, at 600 km minimum along the Greater the basement consolidation of the Scythian Plat- Caucasus, remnants of which are mafic and ultra- form if Variscan or some hundreds of millions years mafic rocks interpreted as an ophiolitic complex after, if Baikalian. Devonian in age (Khain 1975; Adamia et al 1981) Remark: The time gap between the last orogeny and emplaced prior to the late Visean (Khain 1975; and continental rifting is important to specify in as Adamia et al 1981). Whereas Zonenshain et al much as the petrotectonic setting of these continen- (1990) considered this ophiolitic series as the suture tal rifting-related eruptions is usually described as of a wide ocean basin, that is, “proto-Tethys” or “anorogenic or continental intra-plate”, and “post- a branch of the Iapetus–Tornquist oceanic system, orogenic or post-collisional” after either a long or Adamia et al (1981) suggested the closure of a a short time gap respectively. Apriori,thistime small back-arc oceanic basin with the main subduc- gap does not allow defining the depth, lithospheric tion zone south of the Transcaucasus area (Adamia or asthenospheric or even deeper, of mantle partial et al 1981; Gamkrelidze 1986). However, if a Pale- melting synchronous with the continental rifting ozoic subduction zone existed south of the Eastern process. European Platform, it seems unlikely to be respon- Two types of volcano-sedimentary environments sible for the widespread peri- and intra-cratonic developed, expressing the competition between magmatism of the immense Eastern European uplift and subsidence at surface. Whereas the Late Platform. Devonian and Late Triassic volcanoes were sub- The Early Mesozoic subduction of Paleo-Tethys aerial and associated with intra-continental basins Ocean beneath Eurasia is better constrained. It (active rifting with uplift prevailing over subsi- should have been just north of the Lesser Cau- dence), the Early to Middle Triassic volcanoes casus and in Turkey along the Pontides (Stampfli emplaced within marine basins upon continental et al 2001; Nikishin et al 2001; Dercourt et al 1993 crust (passive rifting with subsidence prevailing and 2000 and all references therein). Therefore, over uplift). the Eastern Fore Caucasus area was somewhat In time and at large scale, the Late Devonian and far (600 km minimum) from the subduction-related the Triassic periods were both the early stages of margin of the Paleo-Tethys. Furthermore, the Fore a continental crust thinning process leading to the Caucasus basin formed probably onto the north- formation of continental marine basins (Nikishin ern passive margin of a widespread back-arc basin et al 1996; Stovba and Stephenson 1999; Nikishin behind this active margin. The Late Triassic vol- et al 1998a, b). Nevertheless the geodynamic set- canic cycle ended along with a general uplift of ting of these basins is still under discussion espe- the Fore Caucasus area which lasted until the cially for the Late Devonian time slice. end of Jurassic. Then the overlying undisturbed 472 Fran¸coise Chalot-Prat et al marine Cretaceous cover attests to a stable Eastern Voronezh Massifs, to the west of the Donbas area, European Platform, overlapped by the worldwide underwent passive rifting with a shallow marine transgression. sedimentation (Nikishin et al 1996; Alekseev et al The compositional changes recorded in the 1996). During the Late Devonian, continental lithospheric continental mantle during orogenic rifting associated with basement uplift, fluviatile events, are mainly due to its interactions with the sedimentation and subaerial basaltic volcanism crust, oceanic and continental, during the subduc- occurred from the Pripyat-Dniepr-Donets system tion processes during the closure of an ocean and in the south to the Barents Sea in the north the collision between its margins (Doglioni et al (Chirvinskaya and Sollogub 1980; Wilson and 1999). As mafic eruptive rocks are the most impor- Lyashkevich 1996; Stovba and Stephenson 1999; tant source of information on mantle composi- McCann et al 2003). tions, we have undertaken a comparative study of Our sampling area of basalts is located on the mafic volcanics emplaced on the Eastern European southern margin of the Donbas Foldbelt (30 km to Platform margin during two successive continental the south of the city of Donetsk) (figure 1). The rifting periods. geology of the region is characterised by a series The aims of our petrological study of these of WNW-trending half grabens filled with Mid- basaltic (s.l.) rocks, based on petrographic, geo- dle Devonian to lower Visean volcano-sedimentary chemical and isotopic (Rb-Sr; Sm-Nd) data on deposits. The earliest eruptions are Late Devonian whole rocks, are: basaltic fissure eruptions. The basalts are inter- stratified with fluviatile deposits. According to • to decipher and compare the compositions of McCann et al (2003), uplift and fault activity pre- mantle-derived magmas emplaced on a continen- dated and went on during eruptions. The stud- tal lithospheric plate; ied samples come from lava flows (Df in table 1) • to deduce and compare the compositions of their and their possible feeder dykes (Dd in table 1) possible mantle sources; cross-cutting the Proterozoic basement to the • to relate these results to the regional geo- south. dynamics. 2.2 Eastern Fore-Caucasus area and 2. Detailed geological background Triassic volcanics of both eruptive areas The Eastern Fore Caucasus is located in the east- 2.1 Donbas Foldbelt and ern part of the Scythian Platform, between the Late Devonian volcanics East European Craton and the Great Caucasus Alpine belt (figure 1), few hundred kilometers The Donbas Foldbelt forms the inverted part of from the Donbas area. As previously mentioned, the 2000 km-long NW-SE-trending intracratonic the nature of the Scythian Platform basement Late Paleozoic Pripyat-Dniepr-Donets rift which is unknown and few are available on the Paleozoic developed on the south-western part of the East- succession. ern European Platform (figure 1) (Chekunov et al The accessible data come from few boreholes 1992; Stephenson et al 1993). The rift cuts across reaching the Permo?-Triassic rocks between 3 and the roughly N-S main structural grain of the Pre- 6 km. These deposits have a highly variable thick- cambrian basement, and has separated the lat- ness (up to 2 km). They unconformably overlie ter in two massifs, the Ukrainian Shield to the the older Paleozoic succession and are covered by north and the Voronezh Massif to the south (fig- Jurassic (Pliensbachian and younger) strata. The ure 1). The Donbas Foldbelt consists of a folded Paleozoic (Devonian to latest Carboniferous) suc- Middle Devonian to Upper Carboniferous volcano- cession shows greenschist facies metamorphism, sedimentary sequence up to 22 km thick (Chekunov and is intruded by Carboniferous to Early Permian et al 1993; Stovba et al 1996). The tectonic inver- granites (Letavin 1980). sion of the Donbas is Mesozoic (Latest Triassic According to the classical scheme of evolution and Latest Cretaceous) and not Permian (Stovba (Zonenshain et al 1990; Nikishin et al 2001), after and Stephenson 1999; Saintot et al 2003) as clas- the Late Paleozoic folding and granitic intrusions, sically described and presented as such in fig- the Fore-Caucasus area was uplifted during Late ure 1. The Permian tectonics of the Donbas is Permian and eroded while continental sedimenta- tensional leading to a rift reactivation associated tion occurred eastwards. with salt diapirism (Stovba and Stephenson 1999, Another interpretation could be as follows: the 2003). greenschist metamorphism was the result of a At the onset of lithospheric extension in Mid- deep burial of the Devonian to latest Carbonifer- dle Devonian time, the Proterozoic Priazov and ous succession before its exhumation and erosion Late Devonian Triassic basalts East European Platform Margin 473 during the widespread Permian uplift of the south- This volcano-sedimentary sequence is uncon- ern Eastern European Platform. This uplift was formably overlain by 300 m thick Ladinian (late synchronous with a widespread continental rift- Middle Triassic) subaerial (?) clastic rocks, rework- ing affecting not only the southern margin of the ing products of the volcanics. Eastern European Platform, but also the west- Throughout this Early to Middle Triassic period, ern Variscan Europe as well as the northern the tilting of strata at small and large scales attests Europe (Ziegler 1990). No Late Paleozoic pen- for a high degree of instability of the basement. etrative deformation is recorded in the succes- During the Early Triassic (T1), another erup- sion of the Scythian Platform (Stephenson et al tive zone located southwards to the East Manych 2004; Saintot et al 2006). Where high grade meta- zone would have been the source of rhyolitic ash morphism of the Scythian Platform basement is falls. observed, as in the Stavropol High region, it is Thus, T1 and T2 eruptions respectively occurred reported to be Baikalian in age (Belov 1981). during the opening (increase of the subsidence rate As the basement of the Pre-Caspian basin, lying during a passive syn-rifting process), then the dis- north of the area of interest, could contain relics appearance (decrease of the subsidence rate during of eclogites, Baikalian in age (Volozh 2003), the a passive post-rifting process) of a marine basin. last major orogenic event recorded by the Scythian That suggests a change/release of the lithospheric Platform could likely be Baikalian rather than stresses. Variscan. In Late Triassic times (T3), the entire Fore- During Triassic times, periods of continental and Caucasus area was first uplifted, tilted and marine sedimentation alternated with some hia- deeply eroded during the Carnian. Then during tuses in-between. The total rate of subsidence var- the Norian, a huge event of subaerial eruptions ied, being very low or even null in the central part occurred synchronously with a fluvio-lacustrine (the Stavropol High) relative to the western and sedimentation (Tikhomirov et al 2004 and ref. eastern parts. therein). This T3 volcano-sedimentary sequence, Three eruptive cycles, in Early (T1), Middle named the Nogai Formation (Dubinskii and Mat- (T2) and Late (T3) Triassic times, are recognized senko 1965), is up to 1500 m thick. A few (Tikhomirov et al 2004 and references therein). mafic lava flows (T3 in table 1) are interstrat- They correspond to two different sedimentation ified with enormous volumes of rhyolitic and epochs [Early to Middle Triassic = T1–T2] and dacitic ignimbrites and related fall deposits. The [Late Triassic = T3] separated by a period of sediments accumulated within a NW-SE strik- strong uplift and erosion. T1, T2 and T3 erup- ing linear depression superposed on the East tive cycles followed each other in time over about Manych zone, suggesting an intra-continental rift- 50 Ma. ing environment. The Nogai Formation was par- The Early (T1) and Middle (T2) Triassic tially overlain by the Rhaetian Zurmutinskaya sequences include both transgressive (Early continental suite (up to 300 m thick) includ- Triassic) and regressive (Middle Triassic) series. ing minor rhyolitic tuffs. The same environment, Continental clastics at the foot of Triassic strata but without volcanism, prevailed during Jurassic presumably fill linear NW-SE striking troughs. times. In time, the sedimentation area widened, and continental clastic sedimentation was gradually 3. Petrography of the basaltic rocks replaced by marine clastic and carbonate accumu- lation. The maximal transgression corresponds to 3.1 Late Devonian group the Olenekian times when the Scythian platform was almost completely immersed. A marine car- The Late Devonian (D) basaltic samples, either bonate sedimentation with bioherms dominated from lava flows or dykes, are microlitic (25 in the shallow central part of the Kayasula basin, to 50%), mostly porphyritic (1 to 8 mm in while two deeper basins of terrigeneous sedimenta- diameter; 1 to 20% in volume), and frequently tion developed both to the north (Karpinsky) and microvesiculated. Minerals are Ti-augite (with to the south. frequent sector-zoning and concentric zoning), Eruptions of basaltic submarine lava flows asso- and/or pseudomorphs of olivine (not as micro- ciated with some rhyolitic and dacitic flows and lites), and/or plagioclase (often as microlites domes occurred during both Early (T1) and Mid- only), magnetite, apatite (abundant in UK16) and dle (T2) Triassic periods at the northern limit sometimes brown hornblende. Mantle xenocrysts of the Kayasula area on the East Manych nar- (deformed and partially resorbed olivine and row zone (Nikishin et al 2001). They are inter- pyroxene) are sometimes observed (UK44A1). stratified with significant volumes of clastic marine More detailed descriptions are in McCann et al sediments. (2003). 474 Fran¸coise Chalot-Prat et al sic; T2: . 5.31 2.61 3.600.64 3.51 0.835.38 3.45 0.41 2.64 4.43 0.55 3.87 4.21 0.71 3.78 3.60 0.62 4.00 3.56 0.27 4.51 0.85 4.43 0.78 3.87 4.09 43.6112.7716.52 51.34 18.35 8.60 43.07 9.96 13.15 46.24 16.08 11.82 47.1044.19 14.1812.94 12.66 47.92 51.86 10.75 18.54 12.34 47.90 46.32 16.37 10.71 12.88 46.65 47.46 14.23 44.87 16.50 13.42 48.56 13.39 13.57 14.62 48.83 10.95 50.37 17.21 47.76 14.57 45.90 13.70 Major and trace element (when available) compositions of basaltic suites from the Fore-Caucasus Triassic and Donbas Devonian areas. T1: Early Trias tot. 16.74 8.69 14.14 12.13 13.05 12.58 13.54 13.74 13.88 3 3 3 3 5 2 2 O 2.33 3.49 2.17 3.34 3.49 3.56 3.17 4.27 3.55 2 2 O O O O 2 O 2.47 5.07 1.61 2.32 2.15 1.77 1.99 1.81 1.76 O 2 2 2 2 2 2 Middle Triassic; T3 Late Triassic; Df: Late Devonian lava flows; Dd: Late Devonian dikes P Al LOITotalRecalculated on an anhydrous basis SiO 1.71 100.39Al 100.43 1.43 99.94 6.95 100.13 2.70 100.02 3.03 99.96 1.83 99.94 4.84 99.99 100.33 2.31 2.57 Table 1. LocationSampleAgeDfDfDfDfDfDfDdDdDd SiO UK11A Donbas UK16 Donbas UK44A1 UK45 Donbas UK49B2 Donbas UK50 Donbas UK22 UK59 Donbas UK60 Donbas Donbas Donbas FeO––––––––– MnO 0.22 0.20 0.15 0.20 0.22 0.20 0.16 0.20 0.23 K TiO Fe FeO––––––––– MnOMgOCaONa 0.22 5.12 9.69 0.20 3.38TiO 5.13 0.14Fe 7.50 11.38 0.19 5.04 8.34 0.21 5.58 7.46 0.20 6.07 10.47 0.15 4.38 3.78 0.20 5.10 7.37 0.22 6.50 9.12 Late Devonian Triassic basalts East European Platform Margin 475 L.D. L.D. < < L.D. L.D. 0.4292 < < L.D. 0.51 L.D. L.D. < < < L.D. L.D. L.D. L.D. 1.80 0.38 0.31 < < < < L.D. L.D. L.D. L.D. 0.13 0.11 0.12 0.11 < < < < L.D. L.D. 0.32 L.D. < < < L.D. L.D. L.D. 23.59 L.D. < < < < L.D. L.D. L.D. < < < 0.65 0.84 0.44 0.56 0.73 0.63 0.28 0.87 0.80 L.D. L.D. L.D. < < < 5 O 2.36 3.53 2.33 3.43 3.60 3.63 3.33 4.37 3.63 2 O 2.50 5.12 1.73 2.38 2.22 1.80 2.09 1.85 1.80 O 2 2 P MgOCaONa 5.19 9.82 3.41 5.18 8.07 12.24 5.17 8.56 5.75 7.69 6.19 10.67 4.61 3.97 5.22 7.55 6.65 9.33 K As 0.63 0.59 0.81 1.27 BaBeBi 1155 1.38 3133 3.26 552 1.92 1748 2.25 959 1.84 783 3.37 927 2.16 907 2.03 843 1.92 Cd CeCoCrCs 126 52 50 1.30 224 20 27 0.59 83 52 438 109 40 45 152 40 10 141 35 277 73 45 19 129 39 83 129 45 266 CuDyErEuGaGd 124Ge 6.19Hf 2.48Ho 3.67In 23.7 47 10.07La 7.58Lu 3.26 1.45 5.11 8.80 32.2 1.04 12.43 101 5.29 2.25 1.32 59 12.70 2.66 0.24 19.1 1.29 7.29 5.78 94 2.22 1.50 6.07 112 3.16 0.36 22.3 0.92 7.86 6.85 74 2.78 1.39 3.93 7.20 0.27 40 26.6 0.95 9.96 6.98 110 2.67 1.39 3.85 9.65 0.26 15.0 10.55 52 1.18 4.46 1.99 28 1.64 2.64 9.19 0.34 25.4 6.79 1.13 6.75 73 2.68 1.98 63 3.93 5.43 24.4 0.30 9.72 0.76 6.58 66 2.56 1.66 9.05 3.77 72 0.23 23.5 9.56 1.14 35 1.78 8.42 0.30 1.13 58 0.29 58 476 Fran¸coise Chalot-Prat et al L.D. < L.D. 0.68 < L.D. 0.12 < L.D. < (Continued) Table 1. LocationSampleAgeDfDfDfDfDfDfDdDdDd Mo UK11A DonbasNbNd UK16Ni DonbasPb UK44A1Pr 1.66Rb UK45 Donbas 64Sb 66 UK49B2 70 1.25 Donbas 2.82 UK50 15.9 101 44 95 Donbas 0.12 UK22 1.77 6.15 17 25.6 UK59 40 Donbas 108 0.11 44 2.51 6.33 160 UK60 Donbas 10.4 55 0.16 27 0.58 Donbas 53 4.23 48 13.2 Donbas 69 1.34 64 3.72 71 18.1 26 66 0.85 4.14 38 68 17.0 101 1.30 14.17 35 32 39 9.5 11 11.86 1.76 64 183 16.5 69 44 5.42 58 16.2 44 66 63 47 SmSnSrTaTb 11.78ThTm 2.69U 897V 5.02 16.58W 1.26Y 3.69 5.75 0.332Yb 2152 8.53Zn 7.49 1.52 1.60Zr 12.43 427 2.09 0.445 0.36 340 28.5 3.09 9.77 1.86 2.76 0.91 3.63 182 0.310 184 2.66 0.78 364 1315 35.3 13.05 4.27 2.63 0.55 1.02 0.278 4.87 161 2.75 332 0.59 566 867 12.94 24.9 5.41 1.84 1.55 1.35 0.389 6.94 3.07 118 277 0.40 256 8.29 484 25.4 5.15 1.91 1.41 0.343 1.41 6.30 2.19 133 0.40 303 13.05 304 30.8 377 2.64 2.28 1.59 0.262 0.86 3.84 2.70 158 0.42 12.48 313 380 31.1 786 4.86 0.345 2.30 4.23 1.26 6.14 2.61 0.77 205 424 22.1 371 4.48 0.329 829 1.68 1.64 1.25 5.16 0.77 188 289 30.1 207 1.94 1.32 0.72 168 318 29.1 388 1.81 154 349 Late02 Devonian Triassic basalts East European Platform Margin 477 sic; T2: . -Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus 1.34 1.46 1.220.23 1.12 0.47 1.761.46 0.19 1.41 1.59 0.19 2.27 1.27 0.27 1.17 2.18 0.21 1.81 0.34 2.14 1.45 0.35 2.33 0.32 2.25 2.20 Major and trace element (when available) compositions of basaltic suites from the Fore-Caucasus Triassic and Donbas Devonian areas. T1: Early Trias 45.9116.7810.98 49.04 17.18 46.72 9.13 16.46 46.65 9.64 15.95 49.57 10.1550.03 15.49 51.06 10.4918.28 53.25 16.38 50.78 8.76 18.66 48.66 15.93 17.14 12.27 51.61 48.81 15.46 16.69 50.89 11.79 52.52 15.90 14.98 52.51 11.73 16.85 52.11 16.35 53.28 15.96 54.11 15.43 tot. 11.96 9.91 10.04 10.62 10.77 9.01 12.59 12.17 12.09 3 3 3 3 5 2 2 O 3.91 2.86 2.66 3.11 3.88 4.82 5.49 5.16 4.30 2 2 O O O O 2 O 1.20 2.09 0.98 0.75 1.70 0.78 0.62 0.68 0.90 O 2 2 2 2 2 2 e––– – – – FeO–––––– K P Table 1 (suite). LocationDrillhole Fore-CaucasusSample Fore-CaucasusAgeT3T3T2T2T2T2T2T2T2 Arbali-11 Fore-Caucasus Fore-CaucasusSiO Fore 73375 Arbali-11 Ilmenskaya-1 73360 Ilmenskaya-1 Ilmenskaya-1 Ilmenskaya-1 78270 Svetloyarskaya-102 Svetloyarskaya-102 Svetloyarskaya-1 78273 78277 78290 82602 82616 82603 Middle Triassic; T3: Late Triassic; Df: Late Devonian lava flows; Dd: Late Devonian dikes e––– – – – TiO Fe FeO–––––– MnOMgOCaONa 0.18 6.42 4.82 0.08 4.47 5.31 0.15TiO 9.05 8.94Fe 0.20 10.06 7.40 0.20 5.86 8.18 0.14 4.84 8.83 0.22 3.98 5.54 0.19 3.98 5.47 0.22 4.01 5.94 Al LOITotalRecalculated on an anhydrous 8.10 basis SiO 99.87Al 7.79 99.88 99.89 3.88 99.89 4.31 99.87 2.47 99.84 2.61 99.87 2.43 99.9 3.03 99.87 2.81 478 Fran¸coise Chalot-Prat et al 02 L.D. L.D. < < L.D. L.D. 0.116 L.D. < < < L.D. L.D. < < L.D. 0.110 L.D. L.D. < < < L.D. L.D. 1.07 1.66 3.13 0.91 L.D. -Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus < < < L.D. L.D. 0.338 L.D. L.D. < < < < L.D. L.D. L.D. 1.36 1.35 1.10 1.03 1.58 1.53 L.D. < < < < L.D. L.D. L.D. < < < 0.25 0.51 0.20 0.20 0.28 0.22 0.35 0.36 0.33 L.D. L.D. L.D. < < < (Continued) 5 O 4.26 3.11 2.77 3.25 3.98 4.96 5.63 5.33 4.43 2 O 1.31 2.27 1.02 0.78 1.75 0.80 0.64 0.70 0.93 O 2 2 CeCoCrCs 15.8CuDy 46Er 536 2.58Eu 59.3Ga 75 2.86Gd 35 221Ge 1.87 10.22 0.90Hf 24.0 19.4Ho 52 4.06 3.23In 48 2.03 420 7.82 0.99 1.58 20.5 2.70 16.6 0.65 3.59 5.16 43 2.15 2.65 394 1.54 44 1.34 35.3 3.85 17.0 0.82 3.50 4.17 60 2.18 1.18 3.01 157 1.27 28.8 33 2.54 16.5 0.75 5.38 3.78 0.22 2.94 1.26 109 55 1.80 31 2.44 40.9 17.8 0.75 4.28 6.10 2.66 1.61 34 1.88 1.53 10 3.63 17.8 31 1.08 5.02 39.8 6.14 1.82 3.56 3.09 2.03 32 2.30 0.92 23.4 11 6.50 39.3 29 5.80 1.88 3.63 4.60 2.01 1.27 2.22 35 22.8 6.50 7 5.61 1.90 28 3.40 4.24 1.87 1.31 22.0 35 6.04 2.04 4.17 1.17 Cd K P Table 1. LocationDrillhole Fore-CaucasusSample Fore-CaucasusAgeT3T3T2T2T2T2T2T2T2 Arbali-11 Fore-Caucasus Fore-CaucasusMnO Fore 73375MgO Arbali-11CaO Ilmenskaya-1Na 0.20 73360 Ilmenskaya-1 7.00 Ilmenskaya-1 5.25 Ilmenskaya-1 78270 0.09 Svetloyarskaya-102 Svetloyarskaya-102 4.85 Svetloyarskaya-1 5.77 78273 0.16 9.43 9.31 78277 0.21 10.53 7.74 78290 0.21 6.02 82602 8.40 0.14 4.98 9.08 82616 0.23 4.08 5.69 82603 0.20 4.11 5.65 0.23 4.13 6.12 BaBe 63 3.19 225 1.84 224 319 239 148 174 174 274 Bi As 39.89 4.98 0.61 Late Devonian Triassic basalts East European Platform Margin 479 L.D. < L.D. < L.D. < L.D. 0.56 0.47 1.78 1.11 1.88 < L.D. 0.84 0.51 < NbNdNi 3.26 10.8 158 20.41 28.9 124 3.83 15.1 142 4.22 12.7 142 5.67 22.3 4.64 53 18.1 33 7.84 25.1 7.48 23.9 7.27 22.8 PbPrRbSb 18.92Sm 2.29Sn 41.3Sr 0.34Ta 9.95 2.77Tb 6.86 1.32Th 63.4Tm 161 1.10 0.281 3.55U 5.68 0.456V 3.25 2.08 0.63 22.4W 0.290 1.500 410Y 0.13 2.44 0.654 0.254 3.83Yb 2.81Zn 1.36 209 4.92 19.3 0.271 1.062Zr 0.327 0.25 320 4.67 0.627 1.092 17.4 3.25 1.97 4.62 0.95 0.294 0.480 1.79 134 183 42.0 0.350 113 0.14 0.548 0.507 5.30 253 21.3 5.70 1.93 3.88 0.289 0.196 1.58 1.41 0.462 18.5 199 94 0.850 0.326 0.15 169 416 20.5 3.96 4.54 1.98 0.432 0.254 1.31 0.381 5.36 2.49 226 0.728 24.6 0.627 84 111 21.0 180 0.61 2.14 0.391 0.268 6.08 4.56 2.16 0.643 2.06 0.566 268 1.002 5.34 81 100 30.2 27.0 0.378 2.88 499 0.553 0.49 5.78 3.10 0.926 292 4.29 0.649 2.14 101 26.0 0.937 155 5.07 2.51 0.513 28.6 0.531 497 0.53 5.83 311 1.092 3.01 0.621 126 81 37.8 2.01 0.923 3.53 0.384 0.506 455 0.900 196 305 118 3.01 35.3 3.36 0.327 184 116 295 34.9 3.15 183 129 LaLuMo 5.9 0.291 28.6 0.312 0.321 10.0 0.306 8.6 0.427 14.4 0.388 12.0 0.553 17.2 0.519 16.0 0.495 16.0 480 Fran¸coise Chalot-Prat et al sic; re-Caucasus - Svetloyarskaya- . -Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus Fo 84 1 1 1 1 102 102 102 102 102 1.69 1.65 1.540.22 1.50 0.25 1.801.77 0.23 1.32 1.75 0.24 1.36 1.60 0.23 1.100.23 1.57 0.20 0.26 2.44 1.87 0.19 0.24 2.24 1.49 0.14 0.25 1.43 0.40 0.24 1.16 0.36 0.23 2.53 0.20 2.30 0.15 0.41 0.37 45.1615.4510.58 47.77 15.60 49.17 9.91 14.56 48.90 9.95 15.95 48.19 11.57 16.2347.19 41.38 11.4416.15 15.69 50.57 48.74 2.43 16.51 16.65 51.14 49.26 2.49 15.14 16.70 51.18 51.00 16.69 2.29 16.02 50.18 51.71 16.90 5.17 46.59 15.20 17.67 11.80 51.20 17.49 51.87 17.58 52.89 16.61 53.06 15.60 Major and trace element (when available) compositions of basaltic suites from the Fore-Caucasus Triassic and Donbas Devonian areas. T1: Early Trias tot. 11.06 10.49 10.35 12.11 11.91 2.74 2.62 2.41 5.36 12.11 3 3 3 3 5 5 2 2 O 4.71 5.20 4.07 4.55O 4.29 4.92 4.50 5.50 4.68 4.23 4.20 4.76 3.70 4.47 5.61 5.07 4.92 4.42 3.84 5.76 2 2 O O O O 2 2 O 1.08 0.21 0.73 0.85O 0.96 1.13 0.08 0.22 0.22 0.76 0.42 0.89 1.10 1.00 0.75 0.09 0.23 0.44 1.14 0.77 O O 2 2 2 2 2 2 2 2 SampleAge 78552 T3? 78285 T2 78292 78293 T2 80202 T2 82598 T2 82599 T2 82618 82631 T2 AN-94-5 T2 T2 T2 FeOMnOMgOCaONa – 0.18 7.93 9.45 0.21 – 7.30 7.17 0.17 8.33 – 8.04 0.30 6.66 – 5.59 0.22 8.21 5.00 – 0.19 4.97 14.11 6.87 0.18 7.43 6.76 7.55 0.17 7.71 7.90 6.19 0.28 4.33 6.57 6.03 0.22 4.26 5.57 – K K Al LOITotalRecalculated on an anhydrous basis SiO 3.64 99.33Al 5.38 99.84 99.48 3.33 99.26 3.71 99.92 3.89 96.9 8.09 97.57 2.37 96.9 1.93 97.87 1.44 99.74 2.28 P P Table 1 (suite). LocationDrillhole Fore-Caucasus Fore-Caucasus Svetloyarskaya- Fore-Caucasus Ilmenskaya- Fore-Caucasus Fore Ilmenskaya-SiO Ilmenskaya- Ilmenskaya- Svetloyarskaya- Svetloyarskaya- Svetloyarskaya- Svetloyarskaya T2: Middle Triassic; T3: Late Triassic; Df: Late Devonian lava flows; Dd: Late Devonian dikes TiO Fe FeOMnOMgOCaONa – 0.17 7.59 9.04 0.20 – 6.90 6.77 0.16TiO 8.01 – 7.73Fe 0.29 6.36 – 5.34 0.21 7.88 4.80 – 0.17 4.41 12.53 6.10 0.17 7.07 6.44 7.19 0.16 7.32 7.50 5.88 0.27 4.18 6.34 5.81 0.21 4.15 5.43 – Late Devonian Triassic basalts East European Platform Margin 481 sic; a- Svetloyarskaya- . -Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus Fore-Caucasus 2.30 1.98 1.320.37 1.48 0.382.37 2.08 0.22 2.08 1.72 0.26 1.41 1.860.38 0.29 1.56 1.70 0.40 0.27 2.20 1.52 0.23 0.28 1.80 0.27 0.30 1.94 0.31 0.23 1.75 0.28 1.59 0.29 0.31 0.24 102 102 102 102 102 102 102 102 102 51.7815.3911.63 52.60 15.97 2.59 46.42 15.71 47.28 2.32 15.55 47.22 2.3453.30 16.4315.84 49.32 2.00 55.24 15.00 16.77 49.40 49.52 1.71 15.79 16.76 48.76 49.83 1.67 16.82 16.39 47.96 49.90 1.46 18.52 17.36 51.65 1.47 15.71 51.47 16.45 50.19 17.31 50.17 19.37 Major and trace element (when available) compositions of basaltic suites from the Fore-Caucasus Triassic and Donbas Devonian areas. T1: Early Trias tot. 11.97 2.72 2.47 2.47 2.11 1.79 1.74 1.50 1.54 3 3 3 3 5 5 2 2 O 5.69 3.94 2.64 3.76O 5.86 4.24 4.14 3.56 2.82 3.66 3.96 2.85 4.48 3.44 3.73 3.81 2.93 3.60 2 2 O O O O 2 2 O 0.63 0.38 0.28O 0.10 0.65 0.18 0.40 0.40 0.30 0.20 0.11 0.30 0.19 0.42 0.42 0.21 0.31 0.44 O O 2 2 2 2 2 2 2 2 P P FeOMnOMgOCaONa – 0.23 3.93 5.47 0.18 8.31 5.32 4.44 0.22 8.10 11.30 6.87 0.20 7.00 9.58 8.64 0.21 7.96 8.28 7.00 0.24 7.42 8.10 8.86 0.27 7.19 7.89 8.73 0.28 8.05 6.46 10.90 0.12 8.38 7.28 7.28 K K SampleAgeT2T1T1T1T1T1T1T1T1 AN-94-6 82652 82656 82660 82661 82667 82672 82676 82677 Table 1 (suite). LocationDrillhole Fore-Caucasus Svetloyarskaya- Fore-Caucasus Svetloyarskaya- Svetloyarskaya- Fore-Caucasus Svetloyarskaya- Fore-Caucasus Svetloyarskaya-SiO Svetloyarskaya- Fore Svetloyarskaya- Svetloyarskay TiO Fe FeOMnOMgOCaONa – 0.22 3.82 5.31 0.17 7.91 5.07 4.23TiO 0.21 7.59 10.59Fe 6.44 0.19 6.64 9.09 8.20 0.20 7.53 7.83 6.62 0.23 7.09 7.73 8.46 0.26 6.90 7.57 8.38 0.27 7.82 6.28 10.59 0.11 8.01 6.96 6.96 Al LOITotalRecalculated on an anhydrous basis SiO 2.13 99.27Al 1.42 96.64 95.75 2.01 98.44 3.55 96.24 1.62 97.4 1.91 97.13 1.16 98.4 1.25 96.77 1.17 T2: Middle Triassic; T3: Late Triassic; Df: Late Devonian lava flows; Dd: Late Devonian dikes 482 Fran¸coise Chalot-Prat et al

3.2 Triassic groups isotope data were determined by isotopic dilution on the ICP-MS Elan 6000. The Early (T1) and Middle (T2) Triassic basic rocks do not differ from each other. They comprise 4.2 Major element compositions and basalts, dolerites, and gabbros forming a series preliminary interpretations about mantle of lava flows and cognate subsurface intrusions. source compositions and genetic relationships Basalts and basaltic andesites are sparsely por- between volcanics of each group phyritic (1 to 3% of olivine and rare plagioclase, 1–3 mm in size), massive or vesicular (up to 10% Major element data from Donbas and Fore- by volume). Groundmass often has a quenched Caucasus samples are listed in table 1. On the texture with hollow ‘box-shaped’ plagioclase and whole diagrams, Late Devonian (D), Early (T1), skeletal pyroxene, typical of subaqueous basic Middle (T2) and Late (T3) Triassic groups are lavas. Dolerites and gabbros are massive, mainly identified. Each of them is successively character- poikilitic or equigranular, with large (up to 5 mm) ized then compared to each other. brown augite (40–50%) including numerous weakly The Late Devonian (D) samples,eitherlava zoned plagioclase An40−60 (55–45%), ilmenite and flows (Df) or feeder dykes (Dd) include, after Le Ti-magnetite (up to 5%). Maitre classification (1989), mainly trachybasalts The Late (T3) Triassic basic rocks are found and rare basanites, basalts and basaltic trachy- as lava flows and subsurface intrusions. Basalts andesites (figure 2 – data recalculated on an anhy- are somewhat more crystal rich than T1 and drous basis). The LOI (lost on ignition) values for T2. The phenocrysts (up to 10–15% by vol- whole rocks (1.5 to 3% on average; up to 7%) are ume; 1–5 mm in size) include plagioclase, corroded not related to significant loss or gain in SiO2,K2O olivine, augite and sometimes brown amphibole, and Na2O (figure 3 – untreated data). Neverthe- scattered within an intersertal matrix. Dolerites less only data recalculated on an anhydrous basis are usually equigranular, never poikilitic, with the are plotted in the binary diagrams where MgO same mineralogical assemblage than T1 and T2. in abscissa is considered as the best discrimina- More detailed descriptions are in Tikhomirov et al tor (figure 4). First, no distinction exists between (2004). flows and dykes. On the whole, volcanics are rather In summary, there are no significant petro- silica-poor (44 to 49% in average; up to 52%) and graphic differences between these groups of alkaline alkali-rich (4 to 6%; up to 8%) with a moderate basic rocks, except that the augite is much more MgO content (8 to 5.5%; down to 3.5%) and a titaniferous in the Late Devonian group, and a CaO content between 12 and 4%. They belong to brown amphibole appears in the Late Triassic (T3) the group of alkali basalts (s.l.). Like alkali basalts, group. The most massive and less altered samples were chosen for the geochemical study.

4. Chemical and isotopic compositions of basaltic rocks

4.1 Analytical methods

Both major and trace elements have been analysed respectively by ICP-AES and ICP-MS (SARM; CRPG-CNRS Nancy) on available samples of the Late Devonian and Triassic volcanics. Among the Triassic samples, if major elements are available for mafic rocks of each period, no one Early Triassic (T1) sample was available for trace element and isotopic analysis at CRPG. Besides Figure 2. Variations of total Alkali vs. SiO2 (data recal- major element data alone are given for samples culated on an anhydrous basis) and classification of mag- analysed by wet chemistry in 1983–84 at the chem- matic whole rocks after Le Maitre (1989). Circle: Late ical lab of “Ukrchermetgeologia” state enterprise Devonian (D) Donbas volcanics; white: as lava-flow; black: (Khar’kov, Ukraine) and in 1996 at the Leeds Uni- as dyke; square and triangle: Triassic Fore Caucasus vol- versity analytical lab (Great Britain). Sr and Nd canics; black square: Early Triassic (T1) lavas; white square: Middle Triassic (T2) lavas; black triangle: Late isotope data were obtained after standard chemi- Triassic (T3) lavas. 1. Picrobasalt; 2. tephrite-basanite; cal separation techniques on a Finnigan MAT 262 3. basalt; 4. trachybasalt, 5. basaltic andesite; and 6. basaltic (CNRS-CRPG; Nancy). Rb and Sm concentration trachyandesite. Late Devonian Triassic basalts East European Platform Margin 483

Figure 3. Variations of SiO2,Na2O, CaO and K2O with Loss On Ignition (LOI) on whole-rocks (untreated data). Same symbols as in ﬁgure 2.

they are significantly rich in TiO2 (3.5 to 4.5% some T1 and T2 samples contain more than 10% in average; up to 5.4%), FeO (11 to 15% in aver- MgO, and are relatively primitive rocks. On the age), P2O5 (0.5 to 1%), and rather poor in Al2O3 whole, SiO2,TiO2,Na2OandP2O5 are inversely (10 to 15%). Positive correlations exist between correlated with MgO, whereas K2O, CaO, FeO and MgO, FeO, CaO, and TiO2 on one hand, and SiO2, Al2O3 do not show significant variations with MgO Al2O3,P2O5,Na2OandK2O on the other. decrease. Data for the Early (T1) and Middle (T2) Tri- The Late Triassic T3 samples are mostly trachy- assic samples do not overlap in the classification basalts (figure 2), slightly more potassic than T1 diagram (figure 2 – data recalculated on an anhy- and T2 (figure 4). The small number of available drous basis), T2 having a higher alkali/silica ratio. samples does not allow evaluation of positive corre- T1 includes basalts and rare andesites, whereas lation of their rather high LOI contents with K2O T2 comprises basalts, trachy-basalts and trachy- and SiO2 and negative correlation with Na2Oand andesitic basalts and rare basaltic andesites. Both CaO (figure 3 – untreated data). In figure 4, the groups show a positive correlation between alkali T3 compositional field overlaps more or less with and silica. The LOI (lost on ignition) contents of the T2 and T1 ones. Nevertheless T3 shows signif- T1 samples (figure 3 – untreated data) are rather icant and sometimes opposite variations between low (up to 2%) and do not show any correlation major elements compared to those existing within with silica, alkali and CaO. In turn, the some- T1 and T2. In T3, SiO2,K2OandAl2O3 are neg- what higher LOI contents (up to 4%) of T2 sam- atively correlated with MgO, whereas CaO, FeO ples are correlated with more or less significant loss and TiO2 are positively correlated with MgO. Note in SiO2 and Na2O. As for the Donbas samples, that T3 shows a positive correlation between MgO only data recalculated on an anhydrous basis are and Na2O and the strongest opposite correlation plotted against MgO in the binary diagrams (fig- between Na2OandK2O. Nevertheless both fea- ure 4). T1 and T2 data overlap in most diagrams, tures could be due to some loss in Na2Owith except that T2 rocks have slightly lower P2O5 and alteration. higher K2O contents than T1 rocks. These differ- Preliminary interpretations: The Late Devonian encescouldbeprimaryinasmuchasitisafea- (D) and Triassic (T) basic rocks are both rather ture of the whole group and there is no correlation alkali-rich with a more or less high alkali/silica between both oxides and LOI variations. Besides ratio, an indication of mantle sources of rather 484 Fran¸coise Chalot-Prat et al

Figure 4. Variations of major oxides vs. MgO (data recalculated on an anhydrous basis). Same symbols as in figure 2. The scatter in the data for each group is evidenced. similar compositions. Nevertheless each group samples display the characteristics of rather K- has its own specificity. The D samples display poor and calc-alkaline silica-saturated series (after the whole features of alkaline and titaniferous Wilson 1989), somewhat poorer in TiO2 and total mafic series (after Wilson 1989), whereas the T alkali, but richer in SiO2 than the D group. Late Devonian Triassic basalts East European Platform Margin 485

Figure 5. Chondrite-normalized Rare Earth Element spidergrams and Primitive Mantle-normalized Trace Element spidergrams for Late Devonian (D), Middle (T2) and Late (T3) Triassic samples. The spidergrams of N-MORB, E-MORB and OIB (ref. http://earthref.org/GERM/reservoirs), a calc-alkaline basalt from the Central Volcanic Zone of the Andes (CVZ– CAB) (in Wilson 1989) and continental tholeiitic basalts from Siberia (S-CTB) and Deccan (D-CTB) traps (in Lightfoot et al 1990a and b) are shown for comparison.

These specificities suggest compositional varia- Ridge Basalts (N-MORB) (Salters and Stracke tions in both proportions and compositions of 2004), Enriched Mid-Oceanic Ridge Basalts alkali- and titanium mantle minerals (metasomatic (E-MORB) and Oceanic Island Basalts (OIB) (ref. hydrated phases as amphibole and/or phlogopite) http://earthref.org/GERM/reservoirs), of a calc- from one source to another. In that case, the alkaline basalt from the Central Volcanic Zone of sources were all within volatile-enriched mantle the Andes (CVZ-CAB) (in Wilson 1989) and of zones. continental tholeiitic basalts from Siberia (S-CTB) Besides, T1 and T2, and T3 and D appear to and Deccan (D-CTB) traps (in Lightfoot et al differ in their evolutionary trends. Correlations 1990a and b) are plotted in figure 5. between MgO and [CaO, FeO, TiO2], negative or The trace element composition of each mag- insignificant in T1 and T2, are positive in T3 matic group is studied in order to decipher the and D; whereas the correlations between MgO and composition of its mantle source on one hand and [SiO2,Al2O3,K2O], somewhat insignificant in T1 on another hand the genetic relationships (partial and T2, are negative in T3 and D. P2O5 always melting and/or fractional crystallization) existing displays a negative correlation with MgO, signif- among samples of any group. Our approach is as icantlyhigherinDandT3thaninT1andT2. follows. First, their main compositional features According to petrographical observations, as no (# magmatic affinity) are evidenced by the trace evidence of magma mixing is observed, these vari- element total contents, the PM-normalized ratios ations should originate from fractional crystalliza- between elements either among the most incom- tion processes. patible or with close incompatibility during mantle partial melting [Thn/Ta n,Thn/Lan,Lan/Smn, 4.3 Trace element compositions and Nbn/Zrn,Smn/Zrn] and between the most and the preliminary interpretations about mantle less incompatible [Thn/Ybn,Tan/Ybn,Lan/Ybn]. source compositions and genetic relationships These ratios, preserved on the whole during between volcanics of each group magmatic differentiation, are tracers of not only the mantle source composition but also the par- Trace element data of Donbas (Late Devonian- tial melting degree which can increase if it is low, D) and Fore-Caucasus (Middle-T2 and Late-T3 or decrease if it is high, the primary composi- Triassic) samples are listed in table 1. Trace tional features of the mantle. Then partial melting element data for Early Triassic (T1) volcanics and fractional crystallization trends are evidenced do not exist as the samples were no more in correlating between them contents, or ratios, available. On Chondrite (CH)-normalized Rare between some trace and major elements which Earth Element and Primitive Mantle (PM)- have contrasted behaviors from one process to normalized trace element spidergrams (figure 5), another, i.e., incompatibility of elements during partial melting of mantle increases towards the left. For (1) [Thn,Tan,Lan,SiO2] which decrease but not comparison, spidergrams of Normal Mid-Oceanic at the same rate with mantle melting increase, 486 Fran¸coise Chalot-Prat et al

and which increase at the same rate during intermediate between those of OIB and E-MORB. basic magma crystallization; On the PM-normalized trace element spidergrams (2) [Ni, Cr] and [MgO] which respectively barely (figure 5) negative anomalies, rather strong in Ta or strongly increase with mantle melting andNbandmoderateinBaandTi,existaswell increase; as a moderate positive anomaly in Sr. These fea- (3) [Thn/Ta n,Lan/Ta n] which slightly decrease tures fit with a calc-alkaline affinity as defined with with mantle melting increase, and remain con- major elements. Their PM-normalized total trace stant or slightly increase during crystallization element content (sumTE: 400 ppm in average; fig- of a basic magma. ure 6) is much lower than in the Late Devonian group. Thn/Lan (1.3 to 1.5), Lan/Smn (1.6 to 1.8), / 4.3.1 Late Devonian (D) group Smn Zrn (0.8 to 0.9) ratios are higher than in the Late Devonian group, whereas Thn/Ybn (5 to 5.5), / / No difference exists between feeder dyke and flow Lan Ybn (3.4 to 3.7) and Smn Ybn (1.9 to 2.15) samples. On CH-normalized rare earth element spi- ratios are less high. dergrams (figure 5), the patterns are similar, paral- Preliminary interpretations: As for the Late lel, generally more enriched than OIB with similar Devonian group, the source of these Middle Triassic to much higher Lan/Ybn (mostly 15 to 23; up to (T2) volcanics could be located in the lithospheric 30). The same is observed on PM-normalized trace mantle, the composition of which already included element spidergrams (figure 5), except that some Ta and Nb anomalies and had a lower content samples can show more or less accentuated unusual in trace elements and a less high fractionation anomalies, positive in Ba and Sr, negative in U and between the most and the less incompatible ele- Ti. The PM-normalized total trace element content ments than for the Late Devonian group. Another in each sample is rather high (sumTE: 1200 ppm hypothesis would be that currently invoked to in average; figure 6). The ratios Thn/Ta n (0.55 explain the calc-alkaline feature of basic magmas. to 1), Thn/Lan (0.7 to 0.9; up to 1.3), Lan/Smn The source would be an asthenospheric residual (3 to 3.6; up 4.3), Smn/Zrn (0.7 to 1) are rather mantle contaminated by partial melts generated steady within the group. None Ta and Nb negative during subduction of an oceanic lithosphere, these anomalies exist. The ratios Thn/Ybn (10 to 20), melts including Ta and Nb anomalies. As pre- Ta n/Ybn (10 to 30) and Lan/Ybn (15 to 25), even sented in Introduction, that geodynamic context if they vary within the group, remain higher than is aprioriprecluded accounting for the great dis- in all the other groups. Such patterns are typically tance of the oceanic slab (600 km minimum) on the those of alkaline and titaniferous basic volcanics as plate tectonics reconstruction (Adamia et al 1981; already defined with major elements (after Wilson Gamkrelidze 1986; Zonenshain et al 1990; Stampfli 1989). et al 2001) and the tectono-sedimentary evolution itself of the Fore-Caucasus area. Preliminary interpretations: The features of the Besides both partial melting and fractional crys- Late Devonian (D) volcanics could be interpreted tallization trends can be defined among the Middle in two ways: Triassic (T2) samples when Thn and MgO are cor- (1) they attest to a mantle source already enriched related with Ni, Cr and Thn/Ta n (figure 6). in most of trace elements with a rather high fractionation between the most and the less 4.3.3 Late Triassic (T3) group incompatible elements and without any Ta and Nb depletion; or The two analyzed samples of this group, T3-A (2) they attest to a very low partial melting degree and T3-B, show really different CH- and PM- of a mantle composition close to that of the normalized patterns (figure 5). The fractionation primitive mantle. between the most and less incompatible elements and the total trace element content (figure 6) are When Thn and MgO are correlated with Ni, higher for T3-A and close to that of Late Devonian Cr and Thn/Ta n (figure 6), most of samples show patterns, but rather low for T3-B and close to those fractional crystallization relationships, but partial of Middle Triassic (T2) patterns. However, the frac- melting relationships also exist between the richest tionations between the most incompatible elements samples in Ni and Cr. are close to that in T2. Besides both samples display slight negative Ta and Nb anomalies, hardly 4.3.2 Middle Triassic (T2) group more visible in T3-A (Thn/Ta n =1.6) than in T3-B (Thn/Tan =1.1). On CH-normalized rare earth element spidergrams (figure 5), their patterns display an enrichment Preliminary interpretations: The similarities of in the most incompatible (Lan/Ybn =2.8to3.7), trace element patterns, either with Late Devonian Late Devonian Triassic basalts East European Platform Margin 487

Figure 6. Variations of total trace element, Ni and Cr contents vs. MgO; variations of Thn/Ta n,Ni,Crvs. Thn.Same symbols as in ﬁgure 2: Possible Partial Melting (PM) and Fractional Crystallization (FC) trends. See text for explanation.

(D) or Middle Triassic (T2), as for the major ele- Triassic volcanic groups are source effects, clearly ments, argue in favor a mantle source rather close distinct from partial melting and fractional crys- in composition either with that of D or that of T2, tallization effects which are identified within each thus located in the same mantle at large scale. This group. On the whole Late Devonian and Triassic is also supported by the fact that T3-A and T3-B volcanic groups show many common points but plot on the partial melting trend of both other also some significant differences. In detail, Late groups (figure 6). So T3-A and T3-B would origi- Triassic (T3) appears to be transitional between nate from successive and increasing melting of the Late Devonian (D) and Middle Triassic (T2). In same mantle source, and this mantle source had a that hypothesis, all the volcanics come from partial global composition close to the source of the D or melting of distinct sources located within a same T2 basalts. heterogeneous mantle, an already trace element enriched mantle with a ratio between the most and Summary of preliminary interpretations: Most of the less incompatible elements always above 1. The observations lead to think that a number of D source differs from the T2 source in having a trace element features of both Late Devonian and much higher total trace element content, a higher 488 Fran¸coise Chalot-Prat et al degree of fractionation between the most and the Such initial isotopic values imply that they less incompatible elements, but a lower degree of plot on mixing trends between depleted mantle fractionation between the highly incompatible ele- (DM) and continental crust (CC) end-members. ments themselves. The T3 source has intermediate The crustal end member could be replaced by an features between both the two others. Besides Ta enriched mantle (EM), already metasomatized by and Nb negative anomalies, absent in the D man- crustal material. Indeed positive εNdi values inher- tle source, are slight in the T3 mantle source but ent to most studied samples suggest a depleted significant in the T2 mantle source. In each group, mantle end-member, whereas higher than CHUR 87 86 samples are genetically linked by partial melting ( Sr/ Sr)i values suggest that this depleted man- and above all crystal fractionation processes. Both tle was metasomatized by melts or aqueous flu- T3 samples could be representative of successive ids with a dominant continental crust signature. partial melts from a mantle source closer in terms The limited scattering of data of D and T groups of composition to the mantle source of D. can be explained by percolation, within a rather homogeneous depleted mantle (asthenosphere, as 4.4 Sr and Nd isotopic compositions defined for the MORB source), of contaminant melts with similar isotopic ratios but with slightly Rb-Sr and Sm-Nd isotopic data (table 2) have been different Nd/Sr ratios and in different proportions obtained on the two most primitive samples in each (see Faure 1986 for explanations). This gave birth group. As Rb data obtained via inductive coupled to small scale mantle heterogeneity, whereas D plasma-mass spectrometer (ICP-MS) or isotopic and T mantle sources (200–300 km far from each dilution (ID) differ from each other, two distinct other) could belong to a same mantle at large (87Sr/86Sr)i ratios for each sample was calculated. scale. In any case, the isotopic signatures are typi- The eruption times for each group are taken from cal for a trace element enriched mantle as found in McCann et al (2003) and Tikhomirov et al (2004). the continental lithospheric mantle (Wilson 1989; Nevertheless as the relative distribution of data King and Anderson 1995; Wilson and Lyashkevich from the three groups on the εNdi vs. (87Sr/86Sr)i 1996; Chalot-Prat and Boullier 1997 and references diagram does not change depending on the used Rb therein; Furman and Graham 1999; Weinstein et al value, the demonstration will be based on isotopic 2006 and references therein). ratios using Rb ICP-MS values. The limited scat- On a diagram illustrating the evolution of Nd tering of data for both Late Devonian (D) and Tri- isotopic composition of mantle sources for each assic (T) groups and the fact that they superpose in sample (figure 7b), all trends cross the evolution part to the isotopic data of Late Devonian basalts line of the depleted mantle (Salters and Stracke from the Pripyat-Dniepr-Donets (PDD) (Wilson 2004) within a rather straight age interval (around and Lyashkevich 1996) rules out the possibility 650 Ma ± 50; up to 850 Ma; early Neoproterozoic that the high 87Sr/86Sr ratio may reflect a leach- III). The evolution of the primitive mantle residue ing problem. Nevertheless for some data, a slight was thus disturbed during this period. The causes subhorizontal shift towards elevated 87Sr/86Sr due could be partial melting combined with, or fol- to secondary alteration cannot be excluded and it lowed by, major metasomatic events during this will be taken into account for the discussion about period. The evolution of the Sr isotopic composi- the Sr model age. tions (figure 7c) gives a larger range of values (700 On the εNdi vs. (87Sr/86Sr)i diagram (figure 7a), to 2700 Ma). They do not contradict the Nd model D and T basalts occupy a very restricted field ages, but cannot be further discussed as Rb and Sr within the north-east quadrant with εNdi between are believed to be much more mobile than the rare +1.80 and +4.28 and (87Sr/86Sr)i between 0.70434 earth elements during alteration processes. and 0.70592. It is not excluded however that this contami- Our isotopic data partly superpose to the nated mantle belonged to the much deeper low published isotopic data of Late Devonian mafic velocity transitional zone interlayered between the and ultramafic volcanics and subsurface intru- upper and the lower mantles and in which crust is sives from the same Pripyat-Dniepr-Donets mega- continuously recycled (Ivanov and Balyshev 2005). rift (in Wilson and Lyashkevich 1996). Also they This would fit with the hypothesis of Wilson and totally superpose to those of Late Devonian Lyashkevich (1996) for an origin of Late Devonian kimberlites and melilitites from the Archangelsk basalts from a deep mantle plume. This hypothe- region northwards (Mahotkin and Juravlev’s data, sis about the very great mantle melting depth was 1993, in Wilson and Lyashkevich 1996). So based upon the parameterization of melting exper- our data are coherent with previously pub- iments on anhydrous mantle peridotite. It does not lished ones. Also the 87Sr/86Sr subhorizontal shift, match with the temperature and lithospheric depth which could be assumed for our data, is likely location of hydrous peridotite solidus in the Falloon slight. and Green diagram (1990), as also underlined by Late Devonian Triassic basalts East European Platform Margin 489 initial eSr initial eNd initial eSr initial Nd/ initial Sr/ Nd 143 Sr 87 144 86 initial 1.20.7 0.512454 0.5124671.22.9 1.8 0.512381 2.6 0.512305 4.3 2.8 Sr/ Nd . ε − − − − Sr 87 86 Sr ID ICP-MS ID ICP-MS ε . Nd (2s) Sr (2s) 144 86 Nd/ Sr/ 143 87 Sr 86 Nd 144 Rb/ 87 Sm/ 147 Sr 86 Rb/ 87 Rb Sm-Nd isotopic data for whole-rock basalts from the Fore-Caucasus Triassic and Donbas Devonian areas Rb-Sr isotopic data for whole-rock basalts from the Fore-Caucasus Triassic and Donbas Devonian areas Table 2 (suite). Fore-CaucasusFore-Caucasus 73375 73360 T3-B 215 Ma T3-A 215 Ma 11.73 28.66 2.78 5.75 0.144 0.122 0.512769 0.512625 14 15 1.6 0.512566 4.0 Fore-CaucasusFore-Caucasus 78270Donbas 78273Donbas T2 235 Ma T2 UK 235 Ma 11 A 16.00 UK 50 Df 14.75 370 Ma 3.11 Df 3.86 370 Ma 64.95 0.118 69.30 10.60 0.159 10.85 0.512649 0.099 0.512706 0.095 18 0.512621 12 0.512535 0.4 18 13 0.512461 2.5 Location Sample Age Nd ID Sm ID Table 2. LocationFore-Caucasus 73375Fore-Caucasus 73360 SampleFore-Caucasus T3-B 215 78270 MaFore-Caucasus T3-A 215 78273 Ma AgeDonbas – T2 56.94Donbas 235 Ma T2 Rb ID 63.36 41.35 235 Ma UK 11 ICP-MS – A 436 174 UK 50 Sr Df – 370 0.378 22.44 Ma 0.633 Df 19.30 ID 40.59 310 370 Ma 0.421 329 0.688 43.99 29.62 0.193 ICP-MS 953 0.706171 0.156 0.708025 32.19 25 0.210 27 528 0.123 0.170 24 50 0.705457 0.162 0.705015 0.134 0.706243 21 0.704885 – 23 14 0.176 0.705678 8.6 25 0.705922 32 0.705274 – 17 – 3 – 6.7 0.705029 0.704756 11 0.704974 0.705676 0.704419 21.4 – 9.7 0.704345 – 1.0 8.9 5.0 18.1 0.0 490 Fran¸coise Chalot-Prat et al

Wilson (1993). Also the exclusive involvement of the continental lithospheric mantle both for the D and T sources easily explains the similarity of the isotopic signatures of basaltic eruptions occur- ring after an interval of 110 to 150 Ma and not so far from each other (200–300 km) on the East European Platform margin.

5. Hypotheses in favor of mantle sources within the same continental lithosphere at the southern margin of the East European Platform

As explained in the Introduction, the aim of our study is to improve our knowledge of the mantle below the southern margin of the Eastern European Platform at time of the Late Devonian and Triassic eruptions. Considering all observations gathered successively on the mantle magmas emplaced during each period, the main features of mantle sources can be summarized as follows. From Sr and Nd isotopic signatures, the involved mantle is rather homogeneous on a large scale with a composition resulting from mixing between depleted mantle and continental crustal end- members. According to its Nd isotopic evolution, this mantle was disturbed during the Early Neo- proterozoic III (around 650 ± 50 Ma) and thus recorded the same history on the whole. No trace of a Variscan event is detected. Mantle partial melting and/or only mantle metasomatism by percolation of crustal liquids could be the causes of this major disturbance. At small scale, slight compositional variations and proportions of both end-members involved during the mixing event gave birth to a heterogeneous metasomatized mantle like in any continental lithosphere. From trace elements, the large-scale homogene- ity concerns the high trace element content with a systematic positive ratio between the most and the less incompatible elements. The heterogeneity concerns: • the total trace element contents, much lower in Triassic than in Late Devonian sources; • the fractionation between the most and the Figure 7. (a) Initial Nd and Sr isotopic signatures of Late less incompatible elements decreasing from Late Devonian (D), Middle (T2) and Late (T3) Triassic basalts; same symbols as in figure 2. Also plotted for comparison, the Devonian (D) to Middle Triassic (T2) sources, initial Nd and Sr isotopic signatures of the Late Devonian the Late Triassic (T3) source having mixed mafic and ultramafic rocks (Wilson and Lyashkevich 1996) features. from the same Pripyat-Dniepr-Donets mega-rift (PDD) than the D basalts. DM: depleted mantle; CC: continental crust; At last in the Triassic sources and above all EM: enriched mantle. (b) Nd model ages for mantle sources in the T2 source, the Ta and Nb contents are of Late Devonian (D), Middle (T2) and Late (T3) Trias- lower than expected considering their high incom- (c) sic basalts. Sr model ages for mantle sources of Late patibility degree during mantle partial melting. Devonian (D), Middle (T2) and Late (T3) Triassic basalts. DM: Depleted Mantle with DM isotopic values after Salters All these characteristics cannot originate from and Stracke (2004). Vertical dotted lines in b and c indicate the trace element depleted mantle end-member, minimum and maximum model ages. but from the composition of contaminant melts Late Devonian Triassic basalts East European Platform Margin 491 more or less trace element enriched and fraction- have been the late Proterozoic Baikalian devel- ated and specifically Ta- and Nb-poor in the T opment of an accretionnary belt (Saintot et al sources. These melts, without or with Ta and Nb 2006). negative anomalies, could have come from par- In such a model, the partial melting events of tial melting of suboverthrusted slices of continen- this continental mantle, synchronous with a major tal crust/metasomatized mantle during the last continental crust fracturing for magma ascent up major orogenic event before eruptions, early Neo- to the surface, during Late Devonian then Triassic proterozoic III in age according the Nd model times were a response to the large-scale stretch- ages. ing and uplift of the continental lithosphere dur- From major elements, the mantle was rich in ing the rifting phases, as already proposed in alkali with a rather high alkali-silica ratio, which similar environments by other authors (Wilson can be explained with the existence of mantle and Downes 1992; Wilson 1993 and references metasomatic amphibole and/or phlogopite and/or therein; Anderson 1994; Sheth 1999a, b; Chalot- even diopside (Wilson 1989; Wilson and Downes Prat and Girbacea 2000). A mantle-crust decou- 1991, 1992; Best and Christiansen 2001; Weinstein pling, as considered by Chalot-Prat and Girbacea et al 2006 and references therein). Higher TiO2, (2000) is a likely possibility. The melting would K2O, P2O5 and FeO contents in the Late Devonian be induced by dramatic decompression of man- mantle source mean necessarily a higher content tle just below the Moho discontinuity, whereas of these elements in metasomatic minerals as phl- major but transient fractures would open through- ogopite, diopside but also apatite, rutile and some out the continental crust to enable the magma other accessory minerals (Foley 1992; Chalot-Prat to erupt at surface. This decoupling could also and Arnold 1999; Best and Christiansen 2001, and occur along rheological discontinuities within the references therein). heterogeneous lithospheric mantle itself, which The presence or not of Ta and Nb negative would explain why mantle sources slightly var- anomalies in Late Devonian and Triassic basic ied in composition with time during the Trias- rocks being source effects, they were necessar- sic, and from the Late Devonian to the Triassic. ily included in the metasomatic minerals, intersti- Another explanation for triggering mantle melt- tial or in veins within the peridotite, and due to ing during lithospheric extension could be shear- the signature of contaminant crustal melts (Lloyd ing along the rheological discontinuities within the et al 1985; Chalot-Prat and Boullier 1997; Pecer- mantle (Doglioni et al 2005). Lithospheric plate illo 1999, 2002; Pecerillo and Panza 1999; Chalot- motions relative to the underlying asthenosphere, Prat and Girbacea 2000; Pecerillo et al 2001). This according to the demonstrations of Doglioni (1990, variety of mantle signatures at small scale is typ- 2003 and references therein), were likely the ical of the continental lithosphere (Johnson et al main motor of these decoupling/shearing and 1978; Wilson 1989; Foley 1992; Wilson and Downes fracturing processes, while the rheological het- 1991, 1992; Serri et al 1993; Smith 1993; Chalot- erogeneity of the lithosphere and of the mantle Prat and Boullier 1997; Sheth 1999a, b; Best and itself helped to localize deformation and partial Christiansen 2001; Neumann et al 2002; Pecerillo melting. and Lustrino 2005). The Eastern European Platform continental mantle, successively involved after an interval from 6. Conclusions 110–150 Ma in areas at least two/three hundred kilometers far from each other, has a common his- This comparative study of basic volcanics, tory of partial melting and metasomatism at least emplaced after an interval of 110 to 150 Ma in areas since the early Neoproterozoic III. As no trace of at least two/three hundred kilometers far from each a Variscan event is detected, that would confirm other in distinct continental rifting contexts on the that the Permian event was not related to plate southern margin of the East European Platform, collision but to crustal transtensional reactivation shows that alkaline and cal-alkaline basic magmas (Gaetani 2000; Saintot et al 2003, 2006), synchro- could originate from partial melting of a similar nous with widespread regional uplift affecting the continental lithospheric mantle, heterogeneous at southern margin of the Eastern European Plat- small scale. form (as the Priazov massif in Stovba and Stephen- From the major and trace element characteristics son 1999). This uplift brought to the surface the of the volcanics, the whole mantle source is sup- buried greenschist Paleozoic succession and was posedtohavebeenmoreorlessrichinalkaliand likely related to regional rifting as in the Donbas TiO2, with a positive ratio between the most and (Stovba and Stephenson 1999; Saintot et al 2003; the less incompatible trace element. These com- Shymanovskyy et al 2004). In that case, the lat- positional features are known to come from melt- est orogenic event of the Scythian Platform might ing of some main metasomatic mantle minerals 492 Fran¸coise Chalot-Prat et al such as amphibole and/or phlogopite and/or diop- Reviews by A Ivanov and an anonymous reviewer side and/or apatite and/or rutile, etc. The rather were of great help. This is the CRPG contribution similar Nd and Sr isotopic signatures of volcanics n◦ 1694. of both ages match with that of a mantle-crust mixing source such as the continental metasoma- References tized mantle. The crustal component could have come from either continental crust (old oceanic Adamia S A, Chkhouta T, Kekelia M, Lordkipanidze M B, crust?) or already metasomatized mantle type- Shavishvili I D and Zakariadze G S 1981 Tectonics of EMII, while the mantle component was a residue the Caucasus and adjoining regions: implications for the evolution of the Tethys ocean; J. Struct. Geol. 3(4) after melting of the Primitive Mantle. From the 437–447. Nd model ages, the youngest major event (with Adamia Sh A and Lordkipanidze M B 1989 An outline of metasomatism at least) recorded by the man- Georgian Geological Structure; In: Evolution of the north- ± ern margin of Tethys (eds) RakuˇsM,DercourtJand tle sources occurred about 650 50 Ma ago, a ◦ long time before both Late Devonian and Triassic Nairn A E M, IGCP project n 198 Mém. Soc. Géol. 154 eruptions. France, Paris Nouvelle Série II 63–66. Alekseev A S, Kononova L I and Nikishin A M 1996 The In detail, the Late Devonian mantle source dif- Devonian and Carboniferous of the Moscow Syneclise fers from the Triassic mantle sources by its higher (Russian Platform): stratigraphy and sea-level changes; alkali-silica ratio, its higher TiO2,K2O, P2O5 and In: EUROPROBE: Intraplate Tectonics and Basin FeO contents, the lower fractionation between the Dynamics of the Eastern European Platform; (eds) most incompatible elements, the higher fractiona- Stephenson R A, Wilson M, De Boorder H and Starostenko V I, Tectonophys. 268 149–168. tion between the most and the less incompatible Anderson D L 1994 The sublithospheric mantle as the source elements and the absence of Ta and Nb negative of continental flood basalts; the case against the continen- anomalies. These differences are linked to the type, tal lithosphere and plume head reservoirs; Earth Planet. the composition and the proportion of metasomatic Sci. Lett. 123 269–280. mantle minerals, themselves dependent on both the Artemieva I M 2003 Lithospheric structure, composition, and thermal regime of the East European Craton: impli- composition and volume of the contaminant melts cations for the subsidence of the Russian platform; Earth injected within the mantle during the last major Planet. Sci. Lett. 213(3–4) 431–446. orogenic event at least. Arthaud F and Matte P 1977 Late Paleozoic strike-slip So the Late Devonian and Triassic eruptions, faulting in southern Europe and northern Africa; result registered within tectono-sedimentary events after of a right-lateral shear zone between the Appalachi- ans and the Ural; Geol. Soc. Amer. Bull. 88(9) an interval of 110–150 Ma and in areas at least 1305–1320. two–three hundred kilometers far from each other, Belov A A 1981 Tectonic Development of the Alpine Fold involved the continental lithospheric mantle which Area in the Paleozoic, Nauka, Moscow (in Russian). underlay the southern continental margin of the Best M G and Christiansen E H 2001 Igneous Petrology; East European Platform. The causes of mantle Blackwell Science, 458 pp. Bogdanova S V, Pashkevich I K, Gorbatschev R and melting and magma eruptions are believed to be a Orlyuk M I 1996 Riphean rifting and major Palaeopro- direct consequence of plate motions. They are terozoic crustal boundaries in the basement of the East attributed to decoupling and/or shearing along the European Craton: geology and geophysics; Tectonophys. rheological discontinuities between crust and man- 368 1–21. tle or even within the lithospheric mantle itself, Burshtar M S, Myshkova Yu F and Shvemberger Yu N 1973 Sedimentary-volcanic complex (Upper Triassic– associated to a major but transient fracturing of Lower Jurassic) of the East Fore-Caucasus, and perspec- the overlying continental lithosphere. tives of its oil and gas productivity; Litologiya i poleznye iskopaemye 6 58–67 (in Russian). BrownD,JuhlinC,Alvarez-MarronJ,Perez-EstaunAand Acknowledgements Oslianski A 1998 Crustal-scale structure and evolution of an arc-continent collision zone in the Southern Urals; Russia; Tectonics 17(2) 158–171. The study was carried on in the framework of Brown D and Spadea P 1999 Processes of forearc and EUROPROBE’s Georift/INTAS 97-0743 project. accretionary complex formation during arc-continent col- PT warmly thanks the CRPG for their welcome lision in the Southern Ural Mountains; Geology (Boulder) and all the facilities he benefited from as a visitor 27(7) 649–652. Brown D, Juhlin C and Puchkov V (eds) 2002 Mountain scientist during spring seasons of 2001 and 2002. Building in the Uralides: Pangea to the Present;AGU FCP and PT sincerely thank Laurie Reisberg and publications, Geophysical Monograph Series 132 288 pp. Catherine Spatz for helping PT to perform the Sm- Chalot-Prat F and Boullier A M 1997 Metasomatic events in Nd and Rb-Sr isotopic analysis at CRPG/CNRS the subcontinental mantle beneath the Eastern Carpathi- and for help in reducing and finalizing the data. ans (Romania): new evidences from trace elements; Con- trib. Mineral. Petrol. 129 284–307. We are very grateful to H Sheth and R Stephen- Chalot-Prat F and Arnold M 1999 Immiscibility between son for their pertinent remarks and constructive calcio-carbonatitic and silicate melts and related wall reviews of a preliminary version of the manuscript. rock reactions in the upper mantle: a natural case Late Devonian Triassic basalts East European Platform Margin 493

study from Romanian mantle xenoliths; Lithos 46(4) Glasmacher U A, Reynolds P, Alekseyvev A A, 627–659. Puchkov V N, Taylor K, Gorozhanin V and Walter R 40 39 Chalot-Prat F and Girbacea R 2000 Partial delamination 1999 Ar/ Ar thermochronology west of the Main of continental mantle lithosphere, uplift-related crust- Uralian fault, southern Urals, Russia; Geologische Rund- mantle decoupling, volcanism and basin formation: a schau 87 515–525. new model for the Pliocene-Quaternary evolution of the Grigorieva V A, Kamenetsky A E, Pavlyuk M I, Palin- southern East-Carpathians, Romania; Tectonophys. 327 scky R V and Plakhotny L G 1981 Formations and oil and 83–107. gas potential of the Cretaceous sediments of the southern Chekunov A V, Vecelov A A and Guilkman A I 1976 Struc- Ukraine; Kiev, Naukova Dumka, 140 pp (in Russian). tural geology and history of Black Sea basin development; Ivanov A V and Balyshev S O 2005 Mass flux across Kiev, Naukova Dumka, 163 pp (in Russian). the lower-upper mantle boundary: vigorous, absent, or Chekunov A V and Naumenko V V 1982 Relationship limited? In: Plates, Plumes, and Paradigms (eds) Foul- between the Earth’s crust’s deep rearrangement, tec- ger G R, Natland J H, Presnall D C, Anderson D L, Geo- tonic movements, magmatism, metamorphism and metal logical Society of America Special Paper 388 327–346. content in the Dniepr–Donets Paleorift; J. Geophys. 4 Johnson R W, Mackensie D E and Smith I E M 1978 Delayed 25–34. partial melting of subduction-modified mantel in Papua Chekunov A V, Garvish V K, Kutas R I and Ryabchun L I New Guinea; Tectonophys. 46 197–216. 1992 Dniepr-Donets palaeorift; Tectonophys. 208(1–3) Khain V E 1975 Structure and main stages in the tectono- 257–272. magmatic development of the Caucasus: an attempt at Chekunov A V, Kaluzhnaya L T and Ryabchun L I 1993 the geodynamic interpretation; American Journal of Sci- The Dniepr-Donets paleorift, Ukraine, deep structures ence 275-A 131–156. and hydrocarbon accumulations; J. Petroleum Geology King D L and Anderson D L 1995 An alternative mechanism 16 183–196. of flood basalt formation; Earth Planet. Sci. Lett. 136 Chirvinskaya M V and Sollogub V B 1980 Deep structure 269–279. of the Dniepr-Donets Aulacogen from geophysical data; Le Maitre R W (ed.) 1989 A classification of igneous rocks Naukova Dumka, Kiev, 178 pp. (in Russian). and glossary of terms; Oxford, Blackwell Scientific. Dercourt J, Ricou L-E and Vrielynck B 1993 Atlas Tethys, Leschukh R Y 1992 Lower Cretaceous of the western and Palaeoenvironmental maps; Gauthier-Villars, Paris, 307, southern Ukraine; Naukova Dumka, Kiev, 208 pp (in 14 maps, 1 pl and explanatory notes. Russian). Dercourt J, Gaetani M, Vrielynck B, Barrier E, Biju- Letavin A I 1980 Basement of the Young Platform of the Duval B, Brunet M F, Cadet J P, Crasquin S and San- Southern Part of the USSR; Nauka Moscow, 153 pp (in dulescu M 2000 Atlas Peri-Tethys, Palaeogeographical Russian). maps; CCGM/CGMW Paris 269 24 maps and explana- Lightfoot P C, Hawkesworth C J, Devey C W, Rogers N W tory notes. and Van Calsteren P W C 1990a Source and differen- Doglioni C 1990 The global tectonic pattern; J. Geodynam- tiation of Deccan Trap lavas: implications of geochem- ics 12(1) 21–38. ical and mineral chemical variations; J. Petrol. 31(5) Doglioni C, Harabaglia P, Merlini S, Mongelli F, Peccer- 1165–1200. illo A and Piromallo C 1999 Orogens and slabs vs. their Lightfoot P C, Naldrett A J, Gorbachev N S, Doherty W direction of subduction; Earth-Sci. Rev. 45 167–208. and Fedorenko V A 1990b Geochemistry of the Siberian Doglioni C 2003 Rift asymmetry and continental uplift; Tec- Trap of the Noril’sk area, USSR, with implications for the tonics 22(3) 1–13. relative contributions of crust and mantle to flood basalt Doglioni C, Green D H and Mongelli F 2005 On the shal- magmatism; Contrib. Mineral. Petrol. 104 631–644. low origin of hotspots and the westward drift of the Lloyd F E, Arima M and Edgar A D 1985 Partial melting lithosphere; In: Plates, Plumes, and Paradigms (eds) of a phlogopite-clinopyroxenite nodule from south-west Foulger G R, Natland J H, Presnall D C and Ander- Uganda: an experimental study bearing on the origin of sonDL,Geological Society of America Special Paper 388 highly potassic continental rift volcanics; Contrib. Min- 735–749. eral. Petrol. 91 321–329. Dubinski A Ya and Matsenko N A 1965 Volcano- Lordkipanidze M B, Meliksetian B and Djarbashian R sedimentary sequence in the basement of the sedimentary 1989 Mesozoic-Cenozoic Magmatic Evolution of the cover of the Scythian Platform eastern part; Sovetskaiya Pontian-Crimean-Caucasian Region; In: Evolution of the Geologiya 8 151–157 (in Russian). northern margin of Tethys (eds) RakuˇsM,DercourtJ ◦ Falloon T and Green D H 1990 Solidus of carbonated fertile and Nairn A E M, IGCP project n 198; Mémoire peridotite under fluid-saturated conditions; Geology 18 SociétéGéologique France, Paris, Nouvelle Série 154(II) 195–199. 103–124. Faure G 1986 Principles of isotope geology (New York: John McCann T, Saintot A, Chalot-Prat F, Kitchka A, Fokin P, Wiley & Sons) 589 pp. Alekseev A and EUROPROBE-INTAS RESEARCH Foley S 1992 Vein-plus-wall-rock melting mechanisms in the TEAM 2003 Evolution of the southern margin of the Don- lithosphere and the origin of potassic alkaline magmas; bass (Ukraine) from Devonian to Early Carboniferous Lithos 28 435–453. times; In: Tracing Tectonic Deformation Using the Sed- Furman T and Graham D 1999 Erosion of lithospheric imentary Record (eds) McCann T and Saintot A, Geol. mantle beneath the East African Rift system: geochemi- Soc. London Spec. Publ. 208 117–135. cal evidence from the Kivu volcanic province; Lithos 48 Milanovsky E E 1996 Geology of Russia and Adjacent Areas 237–262. (Northern Eurasia); Moscow University Press, Moscow, Gaetani M 2000 Wordian; In: Atlas Peri-Tethys, Palaeo- 448 pp (in Russian). geographical maps – Explanatory notes, S. Crasquin Muratov M V 1969 Geology of the USSR. Volume VIII, (coord.), CCGM/CGMW Paris 19–25. Crimea. Part 1, Geology, Nedra, Moscow, 576 pp (in Gamkrelidze I P 1986 Geodynamic evolution of the Cauca- Russian). sus and adjacent areas in Alpine time; Tectonophys. 127 Nazarevich P B, Nazarevich I A and Shvydko N I 1986 261–277. The Upper Triassic Nogai volcano-sedimentary formation 494 Fran¸coise Chalot-Prat et al

of Eastern Fore-Caucasus: composition, constitution, and Puchkov V N 1997 Structure and geodynamics of the relations to earlier- and later-formed volcanics; In: The Uralian orogen; In: Orogeny trough times (eds) Burg J-P formations of sedimentary basins (eds) Timofeev P P and and Ford M, Geological Society Special Publication 121 Boulin Yu K, Moscow, 67–86 (in Russian). 201–236. Neumann E-R, Dunworth E A, Sundvoll B A and Tollef- von Raumer J F, Stampfli G M and Bussy F 2003 srud J I 2002 B1 basaltic lavas in Vestfold-Jeloya area, Gondwana-derived microcontinents – the constituents of central Oslo rift: derivation from initial melts formed by the Variscan and Alpine collisional orogens; Tectonophys. progressive partial melting of an enriched mantle source; 365 7–22. Lithos 61 21–53. Saintot A, Stephenson R, Brem A, Stovba S and Pri- Nikishin A M, Ziegler P A, Stephenson R A, Cloet- valov V 2003 Paleostress field reconstruction and inghSAPL,FurneAV,FokinPA,ErshovAV,Bolo- revised tectonic history of the Donbas fold and thrust tov S N, Korotaev M V, Alekseev A S, Gorbachev V I, belt (Ukraine and Russia), Tectonics 22(5) 1059, Shipilov E V, Lankreijer A, Bembinova E Yu and Shal- doi:10.1029/2002TC001366. imov I V 1996 Late Precambrian to Triassic history of the Saintot A, Stephenson R, Stovba S, Brunet M-F, Yegorova T East European Craton: dynamics of sedimentary basin and Starostenko V 2006 The south margin of the East evolution; Tectonophys. 268 23–63. European continent: its evolution during the Palaeozoic Nikishin A M, Cloetingh S A P, Bolotov S N, Bara- and Early Mesozoic; In: European Lithosphere Dynamics boshkin E Yu, Kopaevich L F, Nazarevich B P, Panov D I, (eds) Gee D and Stephenson R, Memoir of the Geological Brunet M-F, Ershov A V, Il’ina V V, Kosova S S and Society of London, London 32 482–505. Stephenson R A 1998a Scythian platform, Chronostratig- SaintotA,BrunetM-F,YakovlevF,Sébrier M, Stephen- raphy and polyphase stage of tectonic history; In: Stratig- son R, Ershov A, Chalot-Prat F and McCann T 2006 The raphy and Evolution of the Peri-Tethyan Platforms (eds) Mesozoic-Cenozoic tectonic evolution of the Greater Cau- Crasquin-Soleau S and Barrier E, Peri-Tethys Memoir 3 casus; In: European Lithosphere Dynamics (eds) Gee D Memoires du Museum National d’Histoire Naturelle Paris and Stephenson R, Memoir of the Geological Society of 177 151–162. London, London 32 277–289. Nikishin A M, Cloetingh S A P L, Baraboshki E Salters V J M and Stracke A 2004 Composition of the Yu, Bolotov S N, Kopaevich L F, Nazarevich B P, depleted mantle; Geochemistry, Geophysics, Geosystems Panov D I, Brunet M-F, Ershov A V, Kosova S S, 5 5. Il’ina V V and Stephenson R A 1998b Scythian plat- Serri G, Innocenti F and Manetti P 1993 Geochemical form, Caucasus and Black Sea region: Mesozoic-Cenozoic and petrological evidence of the subduction of delam- tectonic history and dynamics; In: Stratigraphy and Evo- inated Adriatic continental lithosphere in the genesis lution of the Peri-Tethyan Platforms (eds) Crasquin- of the Neogene-Quaternary magmatism of central Italy; Soleau S and Barrier E, Peri-Tethys Memoir 3 Memoires Tectonophys. 223 117–147. du Museum National d’Histoire Naturelle Paris 177 Sheth H C 1999a A historical approach to continental flood 163–176. basalt volcanism: insights into pre-volcanic rifting, sedi- Nikishin A M, Ziegler P A, Panov D I, Nazarevich B P, mentation, and early alkaline magmatism; Earth Planet. Brunet M-F, Stephenson R A, Bolotov S N, Koro- Sci. Lett. 168 19–26. taev M V and Tikhomirov P L 2001 Mesozoic and Sheth H C 1999b Flood basalts and large igneous provinces Cenozoic evolution of the Scythian Platform-Black Sea- from deep mantle plumes: fact, fiction, and fallacy; Caucasus domain; In: Peri-Tethyan Rift/Wrench Basins Tectonophys. 311 1–29. and Passive Margins (eds) Ziegler P A, Cavazza W, Shymanovskyy V A, Sachsenhofer R F, Izart A and Li Y Robertson A H F and Crasquin-Soleau S, Peri-Tethys 2004 Numerical modelling of the thermal evolution of the Memoir 6 Memoires du Museum National d’Histoire northwestern Dniepr–Donets Basin (Ukraine); Tectono- Naturelle Paris. phys. 381 61–79. Pecerillo A 1999 Multiple mantle metasomatism in central- Smith A 1993 The continental mantle as a source for hotspot southern Italy: geochemical effects, timing and geody- volcanism; Terra Nova 5 452–460. namic implications; Geology 27(4) 315–318. Stampfli G, Mosar J, Faure P, Pillevuit A and Vannay J-C Pecerillo A 2002 Plio-Quaternary magmatism in central- 2001 Permo-Mesozoic evolution of the Western Tethys southern Italy: a new classification scheme for volcanic realm, the Neotethys East Mediterranean basin con- provinces and its geodynamic implications; Bulletin Soci- nection; In: Peri-Tethyan Rift/Wrench Basins and Pas- ety Geologica Italiana 1 113–127. sive Margins (eds) Ziegler P, Cavazza W, Robert- Pecerillo A and Lustrino M 2005 Compositional variations son A H F, Crasquin-Soleau S; Peri-Tethys Memoir 6, of Plio-Quaternary magmatism in the circum-Tyrrhenian Memoires du Museum National d’Histoire Naturelle 186 area: Deep versus shallow mantle processes; In: Plates, 51–108. Plumes, and Paradigms (eds) Foulger G R, Natland J H, Stampfli G and Borel G D 2002 A plate tectonic model for Presnall D C, Anderson D L, Geological Society of Amer- the Paleozoic and Mesozoic constrained by dynamic plate ica Special Paper 388 421–434. boundaries and restored synthetic oceanic isochrons; Pecerillo A and Panza G F 1999 Upper mantle domains Earth Planet. Sci. Lett. 196 17–33. beneath central-southern Italy: petrological, geochemical Stephenson R A & the Europrobe Intraplate Tectonics and geophysical constraints; Pure and Applied Geophysics and Basin Dynamics Dniepr–Donets and Polish Trough 156 421–443. Working Groups 1993 Continental rift development in Pecerillo A, Poli G and Donati C 2001 The plio-quaternary Precambrian and Phanerozoic Europe: EUROPROBE magmatism of southern Tuscany and northern Latium: and the Dniepr–Donets Rift and Polish Trough basins; compositional characteristics, genesis and geodynamic Sedim. Geol. 86 159–175. significance; Ofioliti 26(2a) 229–238. Stephenson R A, Mart Y, Okay A, Robertson A, Saintot A, Polovinkina Yu I 1960 Magmatisme mésozo¨ıque et Stovba S and Khriachtchevskaia O 2004 TRANSMED cénozo¨ıque (tome II fascicule 4); In: (ed.) Markovsky A P Transect VIII: Eastern European Craton – Crimea – Structure géologique de l’URSS,Paris,CNRS(Moscow, BlackSea–Anatolia–Cyprus–LevantSea–Sinai– 1958) 2(4). Red Sea; In: The TRANSMED Atlas: The Mediterranean Late Devonian Triassic basalts East European Platform Margin 495

Region from Crust to Mantle (eds) Cavazza W, Roure F, wander and Cambrian palaeogeography; J. Geol. Soc. Spakman W, Stampfli G M and Ziegler P A (Berlin: London 158(2) 321–329. Springer-Verlag). Volozh Yu A, Antipov M P, Brunet M-F, Garagash I A, Stephenson R A, Yegorova T, Brunet M-F, Stovba S, Wil- Lobrovsky L I and Cadet J-P 2003 Pre-Mesozoic geody- son M, Starostenko V, Saintot A and Kusznir N 2006 namics of the Peri-Caspian Basin (Kazakhstan); Sedim. Late Palaeozoic intra- and pericratonic basins on the Geol. 156 35–58. East European Craton and its margins; In: European Weinstein Y, Navon O, Altherr R and Stein M 2006 The Lithosphere Dynamics; Gee D and Stephenson R (eds) role of lithospheric mantle heterogeneity in the generation Geological Society London Memoir 32 463–479. of Plio-Pleistocene alkali basaltic suites from NW Harrat Stovba S M and Stephenson R A 1999 The Donbas Fold- Ash Shaam (Israel); J. Petrol. 47(5) 1017–1050. belt: its relationships with the uninverted Donets seg- Wilson M 1989 Igneous Petrogenesis: a global tectonic ment of the Dniepr-Donets Basin, Ukraine; Tectonophys. approach, Unwin Hyman, London, 466 pp. 313(1–2) 59–83. Wilson M 1993 Magmatism and the geodynamics of basin Stovba S M, Stephenson R A and Kivshik M 1996 Struc- formation; Sedim. Geol. 86 5–29. tural features and evolution of the Dniepr–Donets Basin, Wilson M and Downes H 1991 Tertiary-Quaternary Ukraine, from regional seismic reflection profiles; In: extension-related alkaline magmatism in Western and Intraplate Tectonics and Basin Dynamics of the Eastern Central Europe; J. Petrol. 32(4) 811–849. European Platform (eds) Stephenson R A, Wilson M, De Wilson M and Downes H 1992 Mafic alkaline magma- Boorder H and Starostenko V I, EUROPROBE: Tectono- tism associated with the European Cenozoic rift system; phys. 268 127–147. Tectonophys. 208 173–182. Stovba S M and Stephenson R A 2002–2003 Style and timing Wilson M and Lyashkevich Z M 1996 Magmatism and of salt tectonics in the Dniepr-Donets Basin (Ukraine): the geodynamics of rifting of the Pripyat-Dniepr- implications for triggering and driving mechanisms of salt Donets rift, East European Platform; Tectonophys. 268 movement in sedimentary basins; Marine and Petroleum 65–81. Geology 19 1169–1189. Ziegler P A 1990 Geological Atlas of Western and Cen- nd Tikhomirov P, Chalot-Prat F and Nazarevitch B P 2004 tral Europe.2 edn. Shell International Petroleum, Mij. Triassic Volcanism in the Eastern Fore-Caucasus: Evolu- B.V., Geological Society of London, 239 pp. tion and Geodynamic Interpretation; Tectonophys. 381 Zonenshain L P, Kuzmin M I and Natapov L M 1990 119–142. Geology of the USSR: a plate tectonics synthesis (ed.) Torsvik T and Rehnström E F 2001 Cambrian palaeo- Page B M, American Geophysical Union, Washington magnetic data from Baltica: implications for true polar DC, Geophysics Geodynamics Series 21 242 pp.

MS received 13 February 2007; revised 18 June 2007; accepted 21 June 2007