Geochemical, isotopic, and zircon (U-Pb, O, Hf isotopes) evidence for the magmatic sources of the volcano-plutonic Ollo de Sapo Formation, Central Iberia

Montero, P., Talavera, C., Bea, F., 2017. Geochemical, isotopic, and zircon (U-Pb, O, Hf isotopes)evidence for the magmatic sources of the volcano-plutonicOllo de Sapo Formation, Central Iberia. Geologica Acta, 15(4), 245-260, I. DOI: https://doi.org/10.1344/GeologicaActa2017.15.4.1

The Ollo de Sapo (Toad’s Eye) Formation is the largest accumulation of Cambrian-Ordovician magmatic rocks in the Iberian Massif. It comprises variably metamorphosed felsic peraluminous volcanic rocks and high-level granites. The Formation crops out in the core of a Variscan anticlinorium extending for about 600km from the Cantabrian coast to central Spain in the northern part of the Central Iberian Zone (Fig. 1; Parga-Pondal et al., 1964; Díez Montes et al., 2004, and references therein). Rocks of similar composition, structure, and age are also found in other localities: i) forming a batholithic complex in the Eastern part of the Spanish Central System (Vialette et al., 1987; Wildberg et al., 1989; Talavera et al., 2013), ii) scattered in the anatectic complexes of the Central Iberian Zone (Bea et al., 2003; Talavera et al., 2013), and iii) forming the Urra Formation at the southern boundary of the Central Iberian Zone with the Ossa Morena Zone (Fig. 1; Sola et al., 2008, and references therein).

The Ollo de Sapo gneisses have two notable characteristics. First, they are not obviously connected with any major tectonic or metamorphic event (Gutiérrez Marco et al., 2002). Second, they have an unusually high proportion of inherited zircons: in the 20 massifs we have studied so far, no less than 70-80%, and in some samples nearer 100%, of zircon grains contain pre-magmatic cores (Bea et al., 2006b; Montero et al., 2007; Montero et al., 2009; Talavera et al., 2013). This high degree of inheritance seems to be a common feature of the Cambrian- Ordovician rocks of Western Europe (e.g.Laumonier et al., 2004; Teipel et al., 2004; Helbing and Tiepolo, 2005) that undoubtedly reflects a fundamental fact of their formation.

Figure 1.
A) Distribution of the paleogeographic zones of Iberia; the gray area represents the Iberian Massif, and the black areas represent exposures of Cambrian-Ordovician rocks. CIZ = Central Iberian Zone; CXG-D = Schist-Graywacke Complex Domain; CZ = Cantabric Zone; GTOMZ = Galicia Trás os Montes Zone; OMZ = Ossa Morena Zone; OS-D = Ollo de Sapo Domain; SPZ = South Portuguese Zone; and WALZ = Western-Asturian Leonian Zone. B) The Ollo de Sapo Anticlinorium. Stars represent the three samples studied for zircon oxygen and hafnium isotopes. Their composition is given in Table 1.

The whole-rock geochemistry of the Ollo de Sapo gneisses is felsic, markedly peraluminous and alkaline- calcic —in the Frost et al. (2001) rather than in the Peacock (1931) sense— with trace element features that bear similarities with those shown by arc magmas (see below). Based on this, and despite the lack of petrological or regional evidence, some authors have proposed that the Ollo de Sapo magmas originated in a supra-subduction setting (e.g. Gebauer et al., 1993; Valverde-Vaquero and Dunning, 2000; Navidad and Bea, 2004; Del Greco et al., 2016). Other authors, however, disagreed and favored a rifting environment. Bea et al. (2007) for example, based on the abnormally high zircon inheritance, proposed that the Ollo de Sapo magmas resulted from extremely fast melting rates in the middle-lower crust followed by fast melt transportation to the upper crust in an extensional regime. Díez Montes et al. (2010) reached analogous conclusions based on field and structural data.

It is evident that the controversies about the petrogenesis and geodynamic significance of the Ollo de Sapo gneisses mainly arise from the lack of understanding of the nature of the magmas involved. This paper attempts to identify the magmatic sources which, we believe, is the first step in that direction. To this end we combined the inferences derived from exploitation of a large dataset of whole-rock element and Sr and Nd isotopes obtained by the authors over the years, with new oxygen and hafnium isotope data from the very characteristic Ollo the Sapo zircons. The results indicate that the Ollo de Sapo magmas derived from immature sediments of Late Ediacaran to Cambrian age and are, therefore, S-type felsic magmas generated in an extensional environment.

The Ollo de Sapo Formation crops out in the core of a Variscan anticlinorium extending for about 600km from the west Cantabrian Coast to central Spain (Fig. 1; see an overview in Díez Montes et al., 2004). It consists of metavolcanic and metagranitic rocks that are overlain by a clastic succession topped by the Armorican Quartzite, and underlain by another succession of Early to Middle Cambrian rocks that were locally highly metamorphosed (Díez Montes et al., 2010).

The metavolcanic rocks, which originally consisted of dacitic to rhyolitic ignimbrites and tuffs, are currently represented by augen-gneisses with large, and locally rapakivi, megacrysts of K-feldspar, within a fine- to medium-grained and strongly foliated felsic peraluminous groundmass. These gneisses may be grouped into two facies, one coarse-grained and the other fine-grained. The coarse-grained facies usually crops out in the lower part of the formation. It is mainly composed of K-feldspar megacrysts (up to 15cm) and plagioclase and quartz phenocrysts (up to 3cm) embedded in a fine-grained groundmass that, in most cases, consists of quartz, K-feldspar, muscovite, biotite, rare albite, chlorite and sericite, with apatite, zircon, monazite, xenotime, ilmenite and Fe-Cu sulfides as accessories. The fine-grained facies is commonly found in the upper part of the formation and consists of small phenocrysts of K-feldspar (up to 2cm) and quartz (up to 1.5-2cm) in a fine-grained matrix similar to the groundmass of the coarse-grained facies.

The metagranites, which originally consisted of high- level granites, are currently represented by coarse-grained augen-gneisses with intrusive contacts and recognizable aplopegmatitic dykes and enclaves. They crop out as sills or laccoliths in the cores of large antiforms and intruding the Early to Middle Cambrian sequences underlying the Ollo de Sapo (González Lodeiro, 1981; Iglesias Ponce de Leon and Ribeiro, 1981; Díez Montes et al., 2010). The metagranites are composed of large K-feldspar megacrysts embedded in a coarse-grained groundmass of quartz, plagioclase, K-feldspar, biotite, muscovite that may be accompanied of rare tourmaline, cordierite and garnet. The accessories assemblage consists of apatite, zircon, monazite, rare xenotime and Fe-Ti oxides.

For this work we have used 32 samples collected all throughout the Ollo de Sapo Formation. All samples were analyzed for major and trace elements, and 28 of them were also analyzed for Sr and Nd isotopes. To avoid inter-laboratory effects all analyses, except Hf isotopes, have been undertaken in the authors’ laboratories at the Centro de Instrumentación Científica of the University of Granada (CIC-UGR). Zircons from 11 samples have been dated previously using single-grain evaporation and SIMS techniques (Montero et al., 2007; 2009; Talavera et al., 2013). For this work we have separated zircons from three new samples representing the most common facies of the Ollo de Sapo rocks: one metagranite (R19) and two metavolcanic rocks, one coarse-grained (R32) and the other fine-grained (R31; Fig. 1). Using the SHRIMP, we dated them and determined their oxygen isotope composition. Hf isotopes were determined using a MC-LA-ICPMS system.

Whole-rock major-element and Zr determinations were done by XRF after fusion with lithium tetraborate. Typical precision was better than ±1.5% for an analyte concentration of 10 wt.%. Trace-element determinations, except Zr, were done by ICP-MS using Rh as an internal standard. Samples were dissolved in HNO3 + HF at a pressure of 180p.s.i. in a microwave digestor, dried and redissolved in 4% HNO3. Precision, estimated from the analyses of 10 replicates of one sample, was better than ±2% and ±5% for analyte concentrations of 50 and 5ppm, respectively.

Samples for Sr and Nd isotope studies were digested in a clean room using ultra-clean reagents and analyzed by TIMS in a Finnigan Mat 262 spectrometer after chromatographic separation with ion-exchange resins. Normalization values were 86Sr/88Sr=0.1194 and ¹⁴⁶Nd/¹⁴⁴Nd=0.7219. Blanks were 0.6 and 0.09ng for Sr and Nd, respectively. The external precision (2σ), estimated from the results of the last 10 replicates of the standard WS-E (Govindaraju et al., 1994), which is routinely analyzed each 10 unknown samples, was better than 0.003% for ⁸⁷Sr/⁸⁶Sr, and 0.0015% for ¹⁴³Nd/¹⁴⁴Nd. ⁸⁷Rb/⁸⁶Sr and ¹⁴⁷Sm/¹⁴⁴Nd were directly determined by ICP-MS (Montero and Bea, 1998) with a precision, estimated by analyzing 10 replicates of the standard WS-E, better than 1.2% and 0.9% (2σ), respectively.

Zircon was separated using panning, first in water and then in ethanol. After eliminating the magnetic fraction from the concentrates with a neodymium magnet, zircons were handpicked under a binocular microscope and put in a SHRIMP megamount (Ickert et al., 2008). Once mounted and polished, zircon grains were studied by optical and cathodoluminescent imaging, coated with a 10nm thick gold layer, and analyzed for U-Th-Pb and Oxygen isotopes using a SHRIMP IIe/mc ion microprobe at the IBERSIMS laboratory of the Centro de Instrumentación Científica of the University of Granada (CIC-UGR). The SHRIMP U-Th-Pb analytical method roughly followed the described by Williams and Claesson (1987), and is described in detail in www.ugr.es/ibersims. Uranium concentration was calibrated using the SL13 reference zircon (U: 238ppm). U/Pb ratios were calibrated using the TEMORA-II reference zircon (417Ma; Black et al., 2003) which was measured every 4 unknowns. When required, common lead was corrected from the measured 204Pb/206Pb, using the model of terrestrial Pb evolution of Cumming and Richards (1975). Point- to-point errors (95% C.I.) on the age standard were ±0.23% for ²⁰⁶Pb/²³⁸U and ±0.46% for ²⁰⁷Pb/²⁰⁶Pb. Data reduction was done with the SHRIMPTOOLS software (downloadable from www.ugr.es/~fbea) using the STATA™ programming language.

Once analyzed for U-Th-Pb, the megamounts were cleaned, re-polished, and coated with a 30nm thick gold layer for oxygen isotope analyses. To this end the SHRIMP primary ion optics was set with a 120µm Kohler aperture to produce a ~18µm diameter spot on the mount surface. The Cs gun was set to yield a ~8nA Cs+ beam. The e-gun to neutralize Cs ions on non-conductive materials was set to an intensity of about 1µA. Spots to be analyzed were burned for about 5 minutes before measurements. During this time the secondary beam and the e-gun were fully optimized to maximize the16O signal. Measurements were done in two sets of 10 scans each. The scans were of 10 seconds each so that the real data collection time was 200 seconds per spot. The EISIE (electron induced secondary ion emission) background was recorded during 10s before and after each set, and subtracted from the ¹⁸O and ¹⁶O counts. As a standard we used the TEMORA-II zircon measured every four unknowns and cross-checked again the 91500 zircon every 20 unknowns. The reproducibility of the standards was excellent: ∂¹⁸O=8.17±0.34 (2s) for the TEMORA-II and ∂¹⁸O=9.98±0.26 (2s) for the 91500, respectively. Data reduction was done with the POXY program developed by P. Lanc and P. Holden at the Australian National University.

Hf isotopes were done at the Geochronology and Isotope Geochemistry-SGIker facility of the University of the Basque Country using a Thermo-Fisher Scientific Neptune MC-ICP-MS coupled to the New Wave Research UP-213 laser system with a SuperCell laser cell. Data were collected in static mode during 50s of ablation with a spot size of 40μm. Masses 171, 173, and 175 were simultaneously monitored during each step to correct the isobaric interferences of Lu and Yb isotopes on mass 176. ¹⁷⁶Yb and ¹⁷⁶Lu were calculated assuming a ¹⁷⁶Yb/¹⁷³Yb of 0.796179 and a ¹⁷⁶Lu/¹⁷⁵Lu of 0.02655 (Chu et al., 2002) and taking into account the instrumental mass fractionation of each individual analysis. To correct Instrumental Mass Fractionation (IMF) Yb isotope ratios were normalized to ¹⁷³Yb/¹⁷¹Yb=1.132685 and Hf isotope ratios to ¹⁷⁹Hf/¹⁷⁷Hf to 0.7325 (ibid.) using an exponential law. The Lu IMF was assumed to follow that of Yb. Data reduction was performed using the Iolite 2.5 software package for deconvolution of time resolved data (Paton et al., 2011).

The major element compositions of the Ollo de Sapo gneisses (Table 1) correspond to variably silicic granitoids that are magnesian (Fig. 2A), alkaline-calcic to calc-alkaline (Fig. 2B) and markedly peraluminous: the aluminum saturation index (ASI=mol. Al₂O₃/(CaO+Na₂O+K₂O)) varies from 1.16 to 2.07 with the most common values around 1.4 (Fig. 3). The proportions of normative feldspars correspond to granites sensu stricto (Fig. 4A) and the proportions of normative albite, orthoclase and quartz are identical to ordinary corundum-normative (peraluminous) granites (Fig. 4B). We have found no meaningful compositional differences between metavolcanic and metaplutonic facies.

Table 1.
Major and trace elements of the Ollo de Sapo gneisses. See text

Table 1. (cont.)

The major elements are notably decoupled; TiO₂, Al₂O₃, MgO, and FeO_total are well correlated with SiO₂ but CaO, Na₂O, K₂O and P₂O₅ are not. In contrast to differentiated calc-alkaline granitoids, in these rocks the saturation in alumina decreases with increasing silica sothat the less silicic samples are notably more aluminous than the most silicic ones (Fig. 3). Noticeably, these lack such geochemical features as low K/Rb, low Zr/Hf, etc., that characterize highly differentiated granite rocks (e.g.Taylor et al., 1956; Bea et al., 2006a).

Figure 2.
Frost et al. (2001) plots for the classification of graniterocks. The Ollo de Sapo rocks are magnesian (A) and alkaline-calcicto calc-alkaline (B).

Chondrite-normalized REE patterns are parallel for LREE and MREE but diverge considerably for the heaviest REE (Fig. 5). LaN is typically between 100 and 200, decreasing to SmN between 30 and 60, with a small negative Eu anomaly between Eu/Eu*≈0.53-0.65 and LuN between 2 and 20. LaN/LuN is between 7 and 60.

Figure 3.
Aluminium Saturation Index (ASI=mol. Al₂O₃/CaO +Na₂O + K₂O) vs. SiO₂. The area corresponding to calc-alkalinesubduction-related granites was drawn using more than 1000analyses from different plutons of the Urals and the circum-Pacificbatholiths of America. The curve along the left Y axis is the densitydistribution of the analyzed samples. Whereas in calc-alkalinegranites the A.S.I. increases with silica, in the Ollo the Sapo gneissesit decreases. This suggests that the major element chemistry ofthese rocks is controlled by restite unmixing. See text for discussion.

NMORB-normalized spidergrams (Fig. 6) show a marked enrichment with increasing incompatibility, especially for the mica-hosted elements Rb and Cs, with notable troughs in Nb, Ti and Sr, and peaks in K and Pb. They however, lack the enrichment in Sr and moderate depletion in Zr characteristic of true arc-magmas.

Figure 4.
Proportions of normative (CIPW) components of Ollo de Sapo gneisses A) Ab-Or-An plot; fields are those delimited by O’Connor (1965).Note how all the gneisses plot in the field of granites s.s. B) Ab-Or-Q plot; phase boundaries are those determined by Luth et al. (1964) at PH₂O=5bar.The gray area corresponds to corundum-normative granites as calculated by the same authors. The Ollo de Sapo gneisses plot almost exactly in thisgray area.

The Sr and Nd isotope compositions of the Ollo de Sapo (Table 2) are clearly crustal: ⁸⁷Sr/⁸⁶Sr_485Ma is within the 0.7094 to 0.7113 interval (at 95% confidence) and clusters at 0.7103; the εNd_485Ma is within the -4.76 to -4.23 interval (at 95% confidence) and clusters around -4.5 (Fig. 7A). It is worth mentioning that in a⁸⁷Rb/⁸⁶Sr vs.⁸⁷Sr/⁸⁶Sr plot most points scatter along the 485Ma reference line (Fig. 7C) so that, excluding the sample with the highest ⁸⁷Rb/⁸⁶Sr, the remaining 27 samples fit an errorchron at 493±27Ma which nearly matches the average U-Pb zircon age (see next section). The goodness- of-fit of the errorchron despite the large sampled area and variable metamorphic grade indicates that the Ollo de Sapo rocks were not heavily disturbed during the Variscan metamorphism.

Figure 5.
Chondrite-normalized REE patterns of the Ollo deSapo gneisses. They are parallel for LREE and MREE but divergeconsiderably for the heaviest REE.

Figure 6.
N-MORB-normalized spidergrams. Note the markedenrichment with increasing incompatibility, the troughs in Nb, Ti andSr, and the peaks in K and Pb. These patterns lack the enrichment inSr and moderate depletion in Zr characteristic of true arc-magmas.

The Nd TDM model age distribution is asymmetric; it peaks a 1.5Ga and shows a large tail towards higher values with a minor peak at 1.8Ga (Fig. 7B). The 1.5Ga value is very common in Iberia and northern Africa (e.g. Bea et al., 2010) but finds no correspondence with the U-Pb zircon ages. Bea et al. (2011) interpreted this value as a “mixed” Nd model age that is characteristic of most Pan-African granitoids generated throughout the Saharan metacraton

Table 2.
Isotope composition of the Ollo de Sapo gneisses. See text

As mentioned in the Introduction, some authors have proposed a supra-subduction setting for the Ollo de Sapo and compared these rocks, either implicitly or explicitly, with I-type calc-alkaline subduction-related granite- rhyolite series. This hypothesis is based solely on the general enrichment in the most incompatible trace elements and depletion in Nb and Ti. It, however, is inconsistent with major element and Sr and Nd isotopes.

The lack of correlation between CaO, Na₂O, K₂O and P₂O₅ with SiO₂ rules out that the Ollo de Sapo rocks might represent variably differentiated members of magma batches with a similar initial composition. Of particular relevance are the elevated peraluminosity of the less silicic samples and the negative correlation between ASI and SiO₂. These features, which are never seen in arc-related calc-alkaline granite-rhyolite series (Fig. 3), suggest different degrees of restite unmixing in anatectic magmas derived from peraluminous sources. Moreover, Sr and Nd isotopes reveal that the magmatic sources contained no detectable juvenile components and so must have resided in the crust long before the Cambrian-Ordovician melting event.

It might be argued that the original chemical and isotopic features of the Ollo de Sapo rocks were profoundly disturbed by the Variscan metamorphism, or that the spatial heterogeneity of the magmatic sources masked the inter-element correlations characteristic of felsic arc-magmas. However, none of these ideas fined support from Sr isotopes as shown by the excellent match between the twenty-seven samples Rb-Sr errorchron age (493±27Ma; Fig. 7) and the SHRIMP zircon U-Pb ages (Fig. 8) mentioned before.

Geochemical evidence, therefore, is inconsistent with a supra-subduction setting but rather points to an intra- continental origin. The magmatic arc-like trace element features can be attributed to inheritance from the magma source which consisted of rocks with abundant Ediacaran zircons. The question is now to understand whether these sources consisted of Ediacaran igneous rocks with some older zircon components, or younger immature sediments derived from them, a question first raised by Montero et al. (2007).

The answer comes from the zircon oxygen isotope compositions. All rims without exception have an elevated ∂¹⁸O of around 10. This implies a source with whole-rock ∂¹⁸O close to 12 which is a typical “sedimentary” value absolutely uncharacteristic of fresh igneous rocks from magmatic arcs (Valley, 2003, and references therein). The oxygen isotope composition of the 600-615Ma cores, in contrast, is around ∂¹⁸O≈6, thus compatible with Ediacaran magmas being arc-derived. Accordingly, the contrasting oxygen isotope composition of Cambrian-Ordovician rims and Ediacaran cores indicates that the Ollo de Sapo magmatic sources likely consisted of detrital sediments mainly derived from Ediacaran igneous rock rather than the Ediacaran igneous rocks themselves. Considering zircon ages and composition (Talavera et al., 2012) the most likely candidates are the graywackes of the Schist- Graywacke-Complex, the dominant Central Iberian meta-sedimentary formation (e.g.Rodríguez Alonso et al., 2004, and references therein).

A sedimentary source explains the large variability of U-Pb ages and Hf isotopes in the Ediacaran cores (Figs. 8 and 10). If we accept that the Ollo de Sapo magmas were swiftly produced and emplaced high in the crust, as suggested by the abnormally high zircon inheritance (Bea et al., 2007), we must also accept that the so produced magmas could not have mixed over large volumes. Had the source been an igneous rock, the Hf isotope composition of their zircons would have been nearly uniform. If, on the other hand, the source consisted of sediments in which zircons were randomly accumulated as detrital particles coming from isotopically different Ediacaran rocks, the magmas generated from them would have had isotopically heterogeneous pre-magmatic zircons. However, as a result of homogenization of the melt phase, they would have crystallized nearly isotopically homogenous rims with a Hf composition close to the core average. This can explain the distribution of εHfT in rims and cores of zircons from the Ollo de Sapo rocks depicted in Figure 11.

Lastly, a sedimentary source may also explain why the Hf T_DM of the zircons (1.3Ga) is significantly younger than the whole-rock Nd T_DM (1.5Ga). This effect suggest that the source, in addition to detrital materials coming from Ediacaran igneous rocks, also contained a small fraction of older materials with higher REE/Zr the remnants of which are the Mesoproterozoic to Archean zircon cores of the Ollo de Sapo. Nevertheless, this suggestion must be considered with caution because calculating Hf T_DM model ages of zircons requires an assumption regarding average crust composition, and this introduces a difficult-to-evaluate uncertainty factor (see Vervoort and Kemp, 2016, and references therein). In the present case we tuned the calculations for the average composition of the pre-Variscan Iberian crust (Lu/Hf=0.045; submitted) which is remarkably uniform, but even using an average crust composition (Lu/Hf=0.09; Rudnick and Fountain, 1995) the Hf T_DM yields 1.35Ga, which is still younger than the Nd T_DM.

Accordingly, we infer that the Ollo de Sapo magmas were produced by melting of (meta) sedimentary sources that mostly, but not solely, derived from 600-615Ma polygenic igneous rocks. Currently, no examples of such materials are known in Central Iberia. The only Ediacaran rocks precisely dated are the 580±3Ma Aljucén gabbrodiorites (Talavera et al., 2008) and, close to the Cambrian boundary, the 543±6 Almohalla granodiorites (Bea et al., 2003) so that the main component of the Ollo de Sapo magmatic sources remains hidden to direct observation. Nonetheless, considering the chemical and isotopic composition of the Ollo de Sapo rocks and the ages of the inherited zircons, we propose that this component would have not be very different to the Pan-African granitoids of North-East Africa from which Central Iberia detached during the Cambrian-Ordovician (see Bea et al., 2010; 2011, and references therein).

The Ollo de Sapo gneisses are Cambrian-Ordovician S-type granitoids that contain an abnormally elevated proportion of zircons with premagmatic cores, most of which are Ediacaran.

The gneiss major element compositions correspond to peraluminous felsic melts in which the saturation in alumina decreases with increasing silica. This is the opposite of what occurs in fractionated calc-alkaline series and indicates that the chemical variability of the Ollo de Sapo gneisses mainly resulted from variable unmixing of peraluminous restites from felsic near-haplogranitic melts.

The gneiss Sr and Nd isotope compositions (⁸⁷Sr/⁸⁶Sr_485Ma=0.7104±0.00008, εNd_485Ma=-4.50±0.26)) are decidedly crustal, indicating their sources contained little, if any juvenile material.

The trace element compositions bear some similarities to high-K calc-alkaline rocks from subduction settings such as the enrichment in the most incompatible elements, depletion in Nb and Ti, and peaks in K and Pb but lack the enrichment in Sr and moderate depletion in Zr characteristic of true arc-magmas.

The Ollo de Sapo zircons are markedly bimodal with respect to oxygen isotopes. Whereas the Cambrian- Ordovician rims yielded unimodal distributions that cluster around ∂18O=10, typical of S-type magmas formed from melting of altered crust, the Ediacaran cores cluster around

∂¹⁸O=6.5, typical of arc-magmas. Rims and cores have the same average Hf isotope composition but the former are much more homogeneous.

The elevated ∂¹⁸O of the Cambrian-Ordovician zircon rims reveals that the Ollo de Sapo gneisses are in fact S-type granites and rhyolites that resulted from anatexis of younger-than-600Ma immature sediments. These sediments mostly, but not solely, derived from different Ediacaran igneous rocks with a wide range of Hf isotope composition. Despite the lack of direct field evidence, zircon data suggest that the metasedimentary sources of the Ollo de Sapo were the graywackes that form part of the Schist-Graywacke Complex.

The abundance of inherited zircons indicates that melting and emplacement occurred with little fractionation and discharge of suspended solids. This indicates that the arc-like trace element features were inherited from the sedimentary sources formed from erosion of Ediacaran arc-like igneous rocks.

Our data preclude that the Ollo de Sapo rocks might have resulted from arc-related magmatism and lend support to the idea of generation in an extensional regime as proposed by Bea et al. (2007) and Díez Montes et al. (2010).

This paper has been financed by the Spanish Grant CGL2013- 40785-P, and the Andalusian Grant P12.RNM.2163. This is the IBERSIMS publication nº 39.