The Jurassic meta-ophiolitic rocks of Cape Steno, Andros, Greece: a high-pressure/low-temperature mélange with Pelagonian affinity in the Cycladic Blueschist Unit?

This study aims at clarifying the relationship between the Cape Steno mélange, southern Andros, and the main tectonic units of the Attic-Cycladic Crystalline Belt. Jurassic protolith ages and geochemical characteristics indicate a Pelagonian affinity and point to a correlative relationship with the Tsiknias Ophiolite on Tinos Island. However, jadeitites and high-Si phengite in the gneisses clearly indicate a high-pressure metamorphic overprint that is unknown from the Tsiknias outcrop and other occurrences of the Upper Cycladic Unit. A correlation with the Cycladic Blueschist Unit (CBU) is an obvious assumption, but initially seemed difficult to reconcile with the Cretaceous protolith ages of meta-ophiolitic rocks from the CBU and distinct geochemical characteristics of associated jadeitites. The Jurassic ages of the Cape Steno rock suite either document a broader spectrum of source rocks than previously known from the CBU, or the existence of a distinct tectonic unit. We assume that the geological and tectono-metamorphic evolution of the Cape Steno occurrence is similar to that of the Makrotantalon Unit of NW Andros, which represents a Pelagonian subunit in the nappe stack of the CBU, with abundant slices of serpentinites, rare meta-gabbro and a metamorphic history comprising both Cretaceous and Eocene HP/LT episodes.

A poorly understood aspect of the complex structural architecture of the ACCB concerns the importance of Jurassic meta-ophiolitic rocks that are exposed at the southern promontory of Andros Island (Cape Steno, Fig. 1b;Mukhin 1996;Bröcker and Pidgeon 2007;Bulle et al. 2010). Field observations led to the conclusion that this occurrence was affected by high-pressure/low-temperature (HP/LT) metamorphism and thus can be correlated with the CBU of NW Tinos (e.g. Bulle et al. 2010). However, such a relationship has not been clearly established yet and it is uncertain whether the block-inmatrix sequence of NW Tinos, located directly across from Cape Steno Bulle et al. 2010), represents a lateral equivalent of the Cape Steno occurrence, or was formed at different times and by different processes. Ion microprobe U-Pb zircon dating of Cape Steno metagabbros and gneisses yielded Jurassic protolith ages (ca. 174-156 Ma;Bröcker and Pidgeon 2007;Bulle et al. 2010) suggesting a relationship to the ophiolites of the larger Balkan region (e.g. Robertson 2002;Lamont et al. 2020a, and references therein). Jurassic meta-ophiolitic rocks were also described from the Upper Unit of the Tsiknias area on Tinos ( Fig. 1c; Lamont et al. 2020a), whereas meta-gabbros in mélanges of the CBU on Tinos, Syros and Samos yielded only Late Cretaceous protolith ages (Keay 1998;Tomaschek et al. 2003;Bulle et al. 2010;Bröcker and Keasling 2006;Bröcker et al. 2014). The relationship of the Cape Steno serpentinite-meta-gabbro-gneiss association to ultramafic rocks exposed in NW Andros is unclear (Fig. 1b;Papanikolaou 1978a, b;Gerogiannis et al. 2019).
This study attempts to clarify the status of the Cape Steno occurrence within the regional context. We combine field observations with new and existing mineralogical, geochemical and geochronological (U-Pb, Rb-Sr) data to unravel litho-or tectonostratigraphic relationships between the Cape Steno occurrence and the meta-ophiolitic rocks of NW Andros and Tinos. An extensive data set for metagabbros from both islands is already available (Bulle et al. 2010;Bröcker et al. 2014;Lamont et al. 2020a) but no corresponding information for associated serpentinites. To close this gap, we systematically determined the bulk rock geochemistry of ultramafic rocks from Andros and Tinos and evaluated the mineral chemistry of chromian spinels as indicator of the tectonic environment. The bulk rock composition of serpentinites records the influence of protolith geochemistry, mineral assemblage, fluid-rock interaction as well as later alteration during submarine and subaerial weathering but original REE and trace element abundances are often considered to have been largely preserved (e.g. Deschamps et al. 2013, and references therein;Cooperdock et al. 2018).
The focus of our study is placed on regional aspects. Petrogenetic considerations resulting from this data will be discussed elsewhere. We will show that the originally assumed correlation of the meta-ophiolitic Cape Steno rock suite with the blueschist sequences of NW Tinos is most likely wrong and that a relationship to the Makrotantalon Unit of NW Andros is instead more likely.

Geological background
On Andros, the metamorphic succession can be subdivided into three tectonic units, the Upper Unit, the Makrotantalon Unit and the Lower Unit (Papanikolaou 1978a, b). The Upper Unit (UU), which in the regional context is correlative with the UCU, is poorly preserved (Fig. 1b) and comprises an ultramafic breccia overlain by greenschists and serpentinites that are separated from the structurally lower sequences of the CBU by an extensional detachment (Mehl et al. 2007). Samples from the structurally lower rock sequences, collected close to the tectonic contact, yielded Oligocene Rb-Sr dates (29-25 Ma) of unclear geological significance that either indicate the time of shear zone activity (Huyskens and Bröcker 2014) or incomplete resetting of the isotope system. The CBU is represented by two subunits, which are described in the regional literature as Makrotantalon Unit and Lower Unit, respectively. The Makrotantalon Unit (MU; up to 600 m thick; Fig. 1b) lies structurally on top of the LU (Papanikolaou 1978a, b;Bröcker and Franz 2006). The MU mainly consists of dolomitic marbles, various types of metabasic and metasedimentary schists and serpentinites (Papanikolaou 1978a, b;Gerogiannis et al. 2019). Fossils in MU marbles yielded Permian ages (Papanikolaou 1978a, b), whereas U-Pb zircon dating of meta-igneous and clastic metasedimentary rocks of the LU indicate Triassic to Early Cretaceous protolith or maximum sedimentation ages (Bröcker and Pidgeon 2007;Bröcker et al. 2016). This oldover-young relationship suggests that the tectonic contact originated as a thrust fault. However, there is no structural evidence for a later reactivation as a low-angle normal fault, as assumed by Huyskens and Bröcker (2014). Instead, Gerogiannis et al. (2019) showed that the original contact was folded during exhumation and transposed by NE-directed thrust-sense shear zones.
Ar-Ar and Rb-Sr geochronological data indicate that the MU records a polymetamorphic history including an Early Cretaceous HP/LT event, a Late Cretaceous greenschist-to amphibolite-facies episode, Eocene blueschist-facies metamorphism and Miocene greenschist-facies retrogression (Bröcker and Franz 2006;Huyskens and Bröcker 2014; Gerogiannis et al. 2019). The preservation of pre-Eocene blueschists led Huet et al. (2015) to suggest a Pelagonian affinity. Other studies interpreted the MU as an integral part of the CBU because this tectonic slice was also affected by Eocene HP/LT metamorphism (Huyskens and Bröcker 2014;Gerogiannis et al. 2019).
The Cape Steno mélange at the southern tip of Andros ( Fig. 1b) mainly includes non-deformed to variably sheared meta-gabbros, felsic gneisses, meta-basalts, and serpentinites, which are squeezed in between a marble-schist sequence and slivers of schists (Papanikolaou 1978a, b;Mukhin 1996;Bröcker and Pidgeon 2007;Bulle et al. 2010). Mineral assemblages indicate lowto medium-pressure metamorphic conditions except for a quartz-free jadeitite with a presumed HP/LT mode of formation (Buzaglo-Yoresh 1995). U-Pb zircon dating of metagabbros and gneisses yielded Jurassic protolith ages (ca. 174-156 Ma;Bröcker and Pidgeon 2007;Bulle et al. 2010). These rocks were interpreted to represent SSZ-type (suprasubduction zone) ophiolites, linked to the Vardar Ocean or a different coeval oceanic basin (Bröcker and Pidgeon 2007). Fu et al. (2015) questioned a SSZ origin and used oxygen and hafnium isotope data to show that these rocks may be related to partial melting in a metasomatized mantle wedge with significant assimilation of supracrustal material. The marble-schist sequence below the Cape Steno mélange represents the topmost part of the LU on Andros (Papanikolaou 1978a, b; for a contrasting view see Mukhin 1996).
On Tinos (Fig. 1c), the metamorphic succession can be subdivided into at least three tectonic subunits: the Akrotiri Unit, the Upper Unit, and the Lower Unit (Melidonis 1980;Okrusch and Bröcker 1990). The Akrotiri Unit (300-350 m thick) mainly consists of epidote-bearing amphibolites and quartzo-feldspathic gneisses which either represent a tectonic slice of the Upper Unit or a distinct tectonic unit that is unrelated to other subunits on this island (Patzak et al. 1994;Katzir et al. 1996;Lamont et al. 2020a). K-Ar hornblende dating yielded Cretaceous dates (ca. 77-66 Ma; Patzak et al. 1994).
The Upper Unit (UU) comprises lenses and fragments (up to several hundred meters in size) of serpentinites, meta-gabbros, meta-plagiogranites, ophicalcites and listvenites that are embedded in or associated with mostly metabasic phyllites (Melidonis 1980;Katzir et al. 1996;Bröcker and Franz 1998;Zeffren et al. 2005;Lamont et al. 2020a;Mavrogonatos et al. 2021). The UU does not record any evidence of a HP/LT metamorphic event and is considered to belong to the Upper Cycladic Unit (e.g. Katzir et al. 1996;Bröcker and Franz 1998;Zeffren et al. 2005). The meta-ophiolitic rock sequence on Tinos records amphibolite-facies metamorphism followed by a greenschist-facies event (Katzir et al. 1996;Bröcker and Franz 1998;Zeffren et al. 2005). U-Pb zircon dating of a plagiogranitic sill and a meta-gabbro from the Tsiknias area ( Fig. 1c) yielded Jurassic protolith ages of ca. 162 Ma and ca. 144 Ma, respectively, and suggests a relationship to the Pelagonian ophiolites of mainland Greece (Lamont et al. 2020a). Amphibolites interpreted to represent the metamorphic sole yielded Cretaceous U-Pb zircon ages between ca. 64 and 113 Ma (Lamont et al. 2020a). Tectonic juxtaposition of the UU onto the LU was achieved by a lowangle normal fault (e.g. Avigad and Garfunkel 1989;Brichau et al. 2007) and probably occurred during a regional greenschist-facies episode at ca. 21 Ma (Bröcker and Franz 1998).
The lowermost part of the metamorphic sequence is exposed in NW Tinos near Panormos (Fig. 1c). Here, a tectonic contact separates calcite-rich marbles (> 50 m) intercalated with thin bands of quartzites from a discontinuous horizon of phyllites and quartzites (< 2 m thick), which are underlain by dolomite marbles (> 100 m; Avigad and Garfunkel 1989). Melidonis (1980) and Bröcker and Franz (2005) interpreted the basal sequence as part of the LU, whereas Avigad and Garfunkel (1989) interpreted the dolomites and phyllites as part of a para-autochthonous Basal Unit, as also described from Samos and Evia (Ring et al. 1999(Ring et al. , 2001Shaked et al. 2000).

Sampling and analytical methods
Newly collected samples from Cape Steno represent serpentinites and mica schists from the mélange as well as clastic metasedimentary rocks, calcschists and greenschists from the underlying marble-schist sequence. Ultramafic rocks were also collected from various serpentinite bodies of the MU and LU in NW Andros (Fig. 1b). On Tinos, ultramafic rocks were collected from four major occurrences of the UU (Tsiknias, Marlas, listvenite and ophicalcite areas; Fig. 1c). Furthermore, meta-gabbros, schists and gneisses were taken from outcrops at the NW coast close to Aghios Theodoros (Gavalas area; Fig. 1c). Thin sections of previous studies (Bröcker and Pidgeon 2007;Bulle et al. 2010) were re-examined for indications of HP/LT metamorphism. GPS coordinates are reported in Online Resource 1. Field images are shown in Fig. 2.
The mineral assemblages of most ultramafic rocks consist of serpentine polymorphs (mostly antigorite as confirmed by XRD analysis of representative samples) with minor amounts of chlorite, talc, carbonates, chromian spinel and magnetite. Chlorite is the dominant silicate phase in two samples from NW Andros (8077, 8091) and four serpentinites from Cape Steno (8116, 8117, 8118, 8120). Some samples contain carbonates (calcite, magnesite). Most ultramafic rocks are completely serpentinized. Relic pyroxene and olivine were only found in ultramafic rocks from the Tsiknias area (8129, 8135). Variably altered chromian spinels occur as disseminated grains in the serpentinitic matrix. Magnetite forms rims around chromian spinel, and also occurs in irregular networks and as fine grains.
The Cape Steno mélange includes a lensoid block of jadeitite . The mineral assemblage of sample 5100 from this occurrence mainly consists of two types of sodic clinopyroxene (colourless jadeite, greenish omphacite; > 85 vol.% pyroxene), plagioclase, epidote and white mica. Omphacite and albite occur as secondary phases. Titanite and apatite are accessory minerals. Sample 5100 was previously U-Pb dated by Bulle et al. (2010) and incorrectly described as gneiss but jadeitite or Jd-Omp granofels are more appropriate rock names. The complex age range of the zircon population led to the interpretation that sample 5100 is of metasedimentary origin but a metaigneous origin with zircon crystals recording inheritance is more likely.
Samples 8123 and 8170 were collected from the marbleschist sequence below the Cape Steno mélange near Aghios Stephanos (Fig. 1b). Sample 8123 is a calcschist that was selected for Rb-Sr dating. The mineral assemblage consists of calcite, quartz, phengite, epidote, plagioclase and titanite. Sample 8170 is a calcite-rich mica schist that was used for U-Pb zircon dating. The mineral assemblage comprises quartz, calcite, phengite, chlorite and plagioclase. Titanite, rutile, tourmaline and zircon are present as accessory phases.
Analytical methods (electron microprobe, whole rock geochemistry, U-Pb and Rb-Sr geochronology) are described in Online Resource 2. Analytical data is summarized in Online Resources 4, 5, 6.

Mineral chemistry
Serpentine compositions are dominated by SiO 2 (38.9-45.3 wt%) and MgO (32.9-42.1 wt%). Variable FeO and Al 2 O 3 contents range from 0.83 to 10.8 wt% and 0.10 to 3.8 wt%, respectively. The serpentine minerals contain up to 0.30 wt% TiO 2 and 0.59 wt% NiO. Only the Cr 2 O 3 concentrations vary among samples from Andros and Tinos. The highest amounts of Cr 2 O 3 were detected in samples from NW Andros (up to 3.9 wt%), the lowest in ophicalcites from Tinos (0.11-0.46 wt%). The Cr 2 O 3 concentrations in serpentine from Cape Steno and Mt. Tsiknias ultramafic rocks range from 0.10-0.89 wt% and 0.13-0.77 wt%, respectively.
For samples from the MU on Andros, the Al 2 O 3 vs. TiO 2 tectonic discrimination diagram (Fig. 4a) indicates the existence of two compositional groups. Type II Cr-spinels display SSZ affinity and plot in the high-Cr# area of the forearc field (Fig. 5b). Type I chromites have higher Al 2 O 3 concentrations and lower Cr# values which are both compatible with a similar tectonic setting but overlap with the MORB peridotite field (Fig. 4). The type I chromites from Cape Steno show similar characteristics but with a stronger trend towards lower Mg# values, probably recording alteration (Fig. 4b). In samples from Tinos, two groups with different geotectonic affinities can be distinguished. Type I and II chromites of ophicalcites as well as some Tsiknias samples (Tsik02) have compositions in between the values of the Andros serpentinites, but samples from Tinos also include a distinct group of type I Cr-spinel (Tsik01) with a MORB affinity (Fig. 4).
White mica in the Cape Steno gneisses is phengite with Si values in the range from 3.27 to 3.65 but mostly display values > 3.5 (Fig. 5a-e). X Mg [Mg/(Mg + Fe + Mn)] varies between 0.58 and 0.82 and X Na [Na/(Na + K + Ca)] is < 0.1. Phengites of the jadeitite 5100 are characterized by Si-contents of 3.38-3.55. X Mg is in the range from 0.42 to 0.68 and X Na < 0.1. The white mica population of sample 8123 comprises phengites with Si-values of 3.38-3.5. X Mg and X Na values are 0.54-0.71 and < 0.1, respectively.

Bulk rock geochemistry of serpentinites
A total of 61 serpentinite samples were selected for bulk rock geochemical studies. Analytical data is summarized    . Black line represents the "terrestrial array", indicating the trend from a primitive mantle to a harzburgitic composition (Jagoutz et al. 1979;Hart and Zindler 1986). PM value is from Sun and McDonough (1989). Data points in the circle plot are outside the displayed axis values moderate to strong enrichment in LREE compared to MREE (1.21 < La N /Sm N < 5.27) as well as depletion in MREE compared to heavy rare earth elements (HREE) with Gd N /Yb N as low as 0.19 (Fig. 7a) (Fig. 7c). REE compositions are almost chondritic, whereas the second group (Tsik02) shows a weak depletion from HREE to LREE (Gd N /Yb N = 0.44 and 0.62; La N /Sm N = 0.49 and 0.84, respectively). Both groups display small Eu anomalies (Eu N /Eu* = 0.89-1.26). PM normalized trace element patterns of Tsik01 samples indicate strong enrichments in Cs and Ba (Fig. 7d). HFSE and Sr are slightly to moderately depleted compared to neighboring elements. Tsik02 samples are depleted compared to PM values and display negative anomalies in HFSE as well as positive anomalies in Cs, Ba and U.
Ultramafic rocks from the Marlas area are characterized by weakly pronounced concave upward REE patterns with strong negative Ce anomalies (Ce N /Ce* = 0.16-0.08) (Fig. 7e). Trace element diagrams display moderate to strong enrichments in Cs, U, Pb, Zr and Hf as well as negative anomalies in Ce and Sr (Fig. 7f).
REE patterns of serpentinites from the listvenite area show strong depletions from HREE to MREE (Gd N / Yb N = 0.14-0.37) and positive Eu anomalies (1.39-2.87; Fig. 7g). Two samples are characterized by moderate to strong, negative Ce anomalies (Ce N /Ce* = 0.29 and 0.61). PM normalized trace element patterns display peaks for Cs, Ba, U, Pb, Sr as well as for the HFSE (Fig. 7h).
Samples from the ophicalcite occurrences have pronounced concave upward REE patterns with Ce (0.38-1.66) and Eu (0.74-1.41) anomalies (Fig. 7i). Samples 8140 and 8146 are less depleted in their REE compositions than the other samples. Concave upward patterns are also observed in PM normalized trace element diagrams (Fig. 7j). All samples show positive peaks in Cs, U, Pb and Sr as well as negative anomalies in HFSE.

Bulk rock composition of meta-gabbros
For comparison with similar rocks from Cape Steno, six meta-gabbro samples from the NW coast of Tinos (Agios Theodoros, Gavalas; Fig. 1c) were selected for bulk rock geochemistry. Analytical data is summarized in Online Resource 5 and shown in Fig. 8e-f. REE patterns are relatively flat with depletion towards LREE and small positive Eu anomalies. Trace element patterns normalized to PM values show depletions in HFSE (Zr, Hf, Nb, Th) but positive anomalies in Pb and Sr. Additionally, three samples show variable enrichments in Cs, Rb and Ba.

U-Pb and Rb-Sr geochronology
Zircon grains of sample 8170 have subhedral to variably rounded shapes (Fig. 9). CL imaging indicates a high number of grains with oscillatory zoning. Zircons with homogeneous or weak internal structures are also common. Some grains show patchy, banded, or sector-zoned CL patterns while a small number of zircons have low CL intensity. The filtered U-Pb data (181 out of 190 analyses) yielded an age range from 80 Ma to 2.7 Ga ( Fig. 9; Online Resource 3, ESM Fig. 3; and Online Resource 6). The data show a broad maximum at 275-375 Ma with a peak at 315-360 Ma. Two small peaks show up at ca. 250 Ma (n = 6) and ca. 470 Ma (n = 7). A broad age cluster occurs at 500-700 Ma. Fifteen analyses yielded ages in the interval of 1.0-2.7 Ga. The youngest age occurs at ca. 80 Ma and is constrained by two grains with 238 U/ 206 Pb ages of 83 ± 2 Ma and 85 ± 3 Ma (2σ). Three zircons are characterized by Late Jurassic ages (147 ± 4 Ma, 164 ± 2 Ma, 165 ± 3 Ma). A well-constrained coherent age group occurs at ca. 290 Ma (Isoplot TuffZirc age: 294 + 6/-3 Ma, 98.4% confidence, n = 7; weighted average: 294 ± 3 Ma, MSWD = 0.53, probability = 0.78). The prominent Triassic age cluster recognized in detrital zircon populations of CBU samples in other parts of Andros, Tinos and Syros is here only represented by a small peak at ca. 250 Ma (Fig. 10). The age peaks at 315-360 Ma and 290-300 Ma are also common in siliciclastic rocks from the MU .
Rb-Sr analytical data and the isochron diagram of the calcschist sample 8123 are shown in Online Resource 3. Alignment of all datapoints representing sized fractions of phengite (5x), epidote (1x) and calcite (2x) indicate a Rb-Sr date of 28.5 ± 0.7 Ma (MSWD = 18). Exclusion of  Hinsken et al. (2017) Sun and McDonough (1989) ◂ the smallest mica grain size fraction from the straight-line fit results in a similar apparent age of 28.7 ± 0.5 Ma but a lower MSWD (6.7).

Discussion
The meta-ophiolitic rocks occurring in block-in-matrix sequences of the study area yielded Jurassic (Andros) and Late Cretaceous (Tinos) U-Pb zircon ages (Bröcker and Pidgeon 2007;Bulle et al. 2010). This age difference either indicates the existence of a single mélange containing rock fragments with different protolith ages, or the existence of two mélanges belonging to distinct tectonic subunits. Recently reported Jurassic protolith ages of meta-ophiolitic rocks from the Tsiknias Ophiolite on Tinos (Lamont et al. 2020a) imply that the Cape Steno mélange may represent a previously misinterpreted occurrence of the UCU. A correlative relationship to the undated serpentinite belt stretching across NW Andros (Fig. 1b) cannot be ruled out yet either, especially as the geological map of Papanikolaou (1978a, b) indicates that both occurrences have a similar structural position above the uppermost marble horizon (m4) of the LU.

Are there supportive arguments for a correlation between the Cape Steno rock suite and the serpentinites of NW Andros or Tinos?
A discontinuous belt of mappable serpentinite bodies (up to several hundred meter in size) extends through NW Andros (Papanikolaou 1978a, b). Huyskens and Bröcker (2014) described this belt as a suitable marker of the tectonic contact between the MU and LU but avoided a clear allocation of the ultramafic rocks to one of the two nappes. Based on lithostratigraphic observations, Papanikolaou (1978b) interpreted the ultramafic rocks as an olistostromatic horizon in the upper parts of the LU, placing the tectonic contact above the serpentinite belt. In contrast, Shin (2014) suggested that the serpentinites belong to the MU, inferring a tectonic contact at some distance below the ultramafic rocks. Based on new mapping, Gerogiannis et al. (2019) concluded that the serpentinite bodies are exposed at different structural levels within both tectonic units and thus cannot be used for delineating the nappe contact. Gerogiannis et al. (2019) used lithological contrasts between both nappes and the presence of mylonitic rocks for demarcation of the MU-LU boundary.
In contrast to Cape Steno, serpentinites of NW Andros are not found together with other meta-ophiolitic rocks in the same outcrop, but a meta-gabbro block with well-preserved igneous texture and relics of sodic amphibole was described from nearby schists of the MU (Huyskens and Bröcker 2014).
Using the geological map of Gerogiannis et al. (2019) as reference, we have systematically studied ultramafic rocks collected on both sides of the inferred tectonic contact. However, systematic differences in the chromian spinel or bulk rock geochemistry that correlate with the tectonic assignment were not recognized. Rather, these rocks represent a largely homogeneous sample suite, whose original composition has been modified by alteration processes. Bulk rock Al 2 O 3 /SiO 2 and MgO/SiO 2 ratios plot at the refractory end of the "terrestrial array" (Jagoutz et al. 1979;Hart and Zindler 1986) indicating that these samples were derived from ultramafic precursors which had experienced moderate to high degrees of partial melting (Fig. 6). The low Al 2 O 3 contents suggest harzburgitic protoliths. There are no bulk rock compositional features that would allow discriminating between serpentinites of the MU and LU. This observation also applies to the mineral chemistry of chromian spinel. Most samples ascribed to the MU have primary Cr-spinel/ chromite compositions suggesting supra-subduction zone and forearc affinities (Fig. 4a, b). We tentatively interpret the existence of distinct compositional groups of Cr-spinel/chromite as an expression of different degrees of partial melting. In samples assigned to the LU, the original chromite and/or chromian spinels were completely erased by superimposed alteration. The serpentinites of NW Andros either belong to different tectonic units but have identical geochemical compositions or occur within the same tectonic unit. Since the serpentinites are mainly exposed in the vicinity of the tectonic contact, their distribution could indicate the existence of a wider fault zone, possibly due to folding of the tectonic contact zone during exhumation (Gerogiannis et al. 2019).
The ultramafic rocks from NW Andros show little or no compositional similarities with the serpentinites from Cape Steno which lack HFSE depletions and show overall flatter REE patterns than serpentinites from the northern part of the island (Fig. 7a, b). Geochemical characteristics of ultramafic rocks do not provide clear indications for a correlative relationship between the meta-ophiolites of NW and SE Andros, but this could result from a stronger metasomatic overprint of the Cape Steno serpentinites. With the exception of sample 8115, bulk rock Al 2 O 3 /SiO 2 and MgO/SiO 2 ratios of Cape Steno ultramafic rocks deviate considerably from the melting trend, indicating severe alteration that is expressed by almost complete chloritization. It should also be noted that ultramafic rocks from the Upper Unit of Tinos show differences in the mineral and bulk rock geochemistry between different outcrop areas which, however, all belong to the same tectonic unit  Mavrogonatos et al. (2021) and this study. Normalizing values after Sun and McDonough (1989) ◂ 1 3 ( Fig. 7c-j). Likewise, the meta-gabbro blocks in the LU on Tinos represent a geochemically heterogenous group (Fig. 8 e-h) within the same tectonic unit.
In the absence of clear superimposed alteration trends (Fig. 6), the compositional differences among the Tinos serpentinites are interpreted as primary characteristics, possibly related to different degrees of partial melting. MgO/SiO 2 and Al 2 O 3 /SiO 2 ratios and PM like trace element compositions document a more primitive character of the Tsik01 samples than for the other serpentinites from this island (Figs. 6, 7). REE and trace element characteristics of the Cape Steno serpentinites are in good agreement with some of the Tsiknias samples (Tsik02), and with samples from the Marlas and listvenite areas (Fig. 7c-f). A possible co-genetic relationship to the meta-ophiolitic rocks of the Tsiknias area is further indicated by whole rock geochemical similarities of metagabbros from both occurrences (Fig. 8a, b), and their corresponding Jurassic protolith ages (Bröcker and Pidgeon 2007;Lamont et al 2020a). However, unambiguous evidence of a HP/LT metamorphic history leaves no doubt that the Cape Steno mélange is not part of the UU.

Evidence of a high-pressure metamorphic overprint of the Cap Steno rock suite
Andros and Tinos are only separated by a narrow sea channel (ca. 1500 m) and on both islands the metamorphic succession comprises marbles, schists and meta-ophiolitic block-in-matrix sequences Bulle et al. 2010;Bröcker et al. 2016). The Cape Steno mélange was originally interpreted as a meta-olistostrome (Mukhin 1996), and considered to represent the topmost part of the LU (Papanikolaou 1978a, b;Bröcker and Pidgeon 2007;Bulle et al. 2010), suggesting a HP/LT metamorphic history. This interpretation is based on (1) field relationships and petrographic similarities with the metamorphic succession of NW Tinos where various meta-igneous blocks, partly with HP/LT mineral assemblages, occur in meta-sedimentary host rocks with a clear blueschist-facies record, and (2) the presence of jadeitite in the Cape Steno block assemblage leading to the assumption that formation conditions correspond to those of similar rocks in the HP/LT serpentinite mélange on Syros Bulle et al. 2010). Other explanations include the possibility that the Cape Steno rock suite represents a tectonic unit with a different metamorphic record than the A relationship of jadeitite to ancient subduction complexes is clearly established (e.g. Harlow et al. 2015 and references therein) but formation of this rock type is not limited to peak blueschist-or eclogite-facies conditions. Temporal discrepancies between jadeitite formation and peak HP metamorphism recorded in other blocks of the same mélange were described from several occurrences (Tsujimori and Harlow 2012), indicating that jadeitite can already form in the overlying mantle wedge at T = 200-400 °C and P = 0.6-1.2 GPa (Tsujimori and Harlow 2012;Harlow et al. 2015). Jadeitite formation was also described from epidote amphibolite-facies P-T regimes (Tsujimori and Harlow 2012;Harlow et al. 2015).
At this point, the petrogenesis of the Cape Steno jadeitite remains unclear, but this rock type clearly differs from the jadeitites of the Cycladic HP/LT mélanges in terms of bulkrock composition, age complexity of the zircon population and Fig. 10 Probability distribution diagrams of detrital zircon from a Cape Steno (this study), b Makrotantalon Unit, c Lower Unit of Andros ) and d Lower Unit of Tinos (Bulle et al. 2010;Hinsken et al. 2016) 1 3 protolith age. The jadeitites of the CBU are characterized by homogeneous zircon populations which only yielded a single Late Cretaceous U-Pb age group Enders 1999, 2001;Bulle et al. 2010). In contrast, the Cape Steno jadeitite is characterized by a more complex zircon population with Jurassic overgrowths (163.1 ± 3.9 Ma and 174.3 ± 2.0 Ma) on Middle Proterozoic (ca. 1126 Ma and ca. 1421 Ma) and Permian (ca. 273 Ma and ca. 281 Ma) grains (Bulle et al. 2010). The Jurassic ages of the overgrowths broadly correspond to the protolith ages of meta-gabbros and meta-plagiogranitic gneisses (156.2 ± 2.3 Ma, 160.0 ± 2.0 Ma) from the same occurrence (Bröcker and Pidgeon 2007). Furthermore, the REE and trace element patterns of jadeitites from the CBU and Andros are different Enders 1999, 2001;Bulle et al. 2010). In the case of Andros, the REE patterns of both jadeitite and cpx-free gneisses are characterized by a strong enrichment in LREE (La N /Yb N = 29-44), whereas jadeitites and omphacitites from Tinos and Syros often show sinusoidal REE variations, concave REE patterns, weak LREE enrichment (La N /Yb N = 1.6-2.2) or a continuously increasing REE distribution from La to Lu (Online Resource 3, ESM Fig. 4; Bröcker and Enders 2001;Bulle et al. 2010). We therefore consider it unlikely that the Cape Steno jadeitite represents an exotic fragment of the CBU but instead assume a different origin and P-T history.
In meta-gabbros and felsic gneisses, relics of blue amphibole were not recognized but high-Si values of phengitic mica in the gneisses (3.5-3.6 per formula unit; Fig. 5; Online Resource 4) clearly document that these rocks were affected by HP/LT metamorphism. The lack of HP/LT assemblages in the meta-gabbros is not at variance to a high-pressure history. Similar observations were also reported from some meta-gabbros of the CBU on Tinos (Bulle et al. 2010). Besides bulk rock compositional differences, short duration of metamorphism, lack of deformation, and limited availability of fluids or infiltration paths may explain the lack of equilibrium mineral assemblages.

Is the Cape Steno mélange a previously misinterpreted outcrop of the Makrotantalon Unit?
Field observations, geochemical and geochronological data can be reconciled by two scenarios: (1) Two distinct mélanges are exposed in SE Andros and NW Tinos that belong to different tectonic units with different tectono-metamorphic histories, both containing HP/LT rocks related to the same or different metamorphic events. The Cape Steno rock suite is correlative with the MU of NW Andros and is separated by a tectonic contact from the meta-olistostromatic sequences of the LU. The mélange exposed near Aghios Theodoros, NW Tinos, could be a lateral equivalent of the Cape Steno mélange. (2) The Cape Steno mélange is a litho-or tectonostratigraphic equivalent of the metamorphic succession exposed in NW Tinos which can be clearly assigned to the LU. The different jadeitites were formed almost simultaneously from precursor rocks with different protolith ages. In this case, the existence of Jurassic meta-igneous blocks in mélanges of the CBU would be documented for the first time.
We consider the first alternative as the more likely explanation and suggest a correlative relationship of the Cape Steno mélange with the Makrotantalon Unit of NW Andros. The MU records a more complex polymetamorphic history than other tectonic subunits of the CBU, including Early Cretaceous and Eocene blueschist-facies events (Huyskens and Bröcker 2014;Huet et al. 2015;Gerogiannis et al. 2019). Unequivocal evidence of Cretaceous HP/LT metamorphism in the CBU has not yet been documented (for a discussion of contrasting views see Fu et al. 2010Fu et al. , 2012Bulle et al. 2010), but occurs in the Pelagonian zone of mainland Greece (e.g. Schermer et al. 1990;Lips et al. 1998). Accordingly, Huet et al. (2015) suggested a Pelagonian affinity for the MU. The coexistence of both Cretaceous and Eocene HP/LT rocks led to the interpretation that the MU was incorporated into the nappe stack of the CBU at deep subduction levels during the Eocene or somewhat earlier, resulting in a common metamorphic history since that time (Gerogiannis et al. 2019). New Rb-Sr dates of blueschist-facies rocks of the MU (Bröcker et al., unpublished data) further substantiate interpretations suggesting that the MU was affected by both Cretaceous and Eocene blueschist-facies events.
The combination of Pelagonian meta-ophiolitic rocks and HP/LT metamorphism suggests a correlative relationship between the Cape Steno rock suite and the MU which both occur in a similar litho-or tectonostratigraphic position overlying the topmost part of the LU (Papanikolaou 1978a, b). Jurassic protolith ages are unknown from the HP/LT mélanges of the CBU. So far, all dated meta-ophiolitic rock fragments only yielded Late Cretaceous U-Pb zircon ages (Keay 1998;Tomaschek et al. 2003;Bulle et al. 2010;Bröcker and Keasling 2006;Bröcker et al. 2014). Furthermore, Triassic protolith ages (249-240 Ma) were reported for gneisses of the Syros mélange (Bröcker and Keasling 2006) and felsic meta-igneous rocks from Andros considered to represent either olistoliths of a meta-olistostrome or fragments related to large-scale boudinage (Bröcker and Pidgeon 2007).

Is it justified to correlate the schists below the Cape Steno mélange with similar rock sequences in NW Tinos?
Lithological characteristics and similarities in the tectono-metamorphic history indicate that the marble-schist sequences of SE Andros and NW Tinos represent lateral equivalents. In NW Tinos, meta-ophiolitic blocks (meta-gabbros, glaucophanites, serpentinites) are more widespread than in southern Andros but generally are rather rare constituents in siliciclastic host rocks Bulle et al. 2010). Mineral assemblages of the LU on both islands mostly indicate greenschist-facies conditions, but mineralogical evidence of earlier HP/LT metamorphism was not completely erased. Relics of sodic amphiboles are often found.
Similar maximum depositional ages of ca. 80 Ma provide further evidence of a correlative relationship between schist sequences on both islands (Bulle et al. 2010;Shin 2014;this study). Such ages are a typical feature of detrital zircon populations of the CBU on Tinos and Syros (Bulle et al. 2010;Löwen et al. 2015;Hinsken et al. 2016) and were also recognized in the newly studied sample 8170, collected directly below the Cape Steno mélange, interpreted to represent the topmost part of the CBU in SE Andros.
The Rb-Sr date of a calcschist (28.7 ± 0.5 Ma; Online Resource 3; ESM Fig. 4) from Cape Steno, collected close to the tectonic contact at Aghios Stephanos, corresponds to similar Rb-Sr dates (~ 29-25 Ma) from other parts of Andros and Tinos (Huyskens and Bröcker 2014), lending support to the assumption that this date could indicate a distinct P-T-D stage during exhumation affecting both islands.

Summary and conclusions
This study focused at unravelling the status of the Cape Steno mélange within the structural architecture of the Cyclades. The Cape Steno occurrence is different to the HP/LT mélanges of the CBU on the neighboring islands of Syros and Tinos (e.g. Dixon and Ridley 1987;Bröcker and Enders 2001;Bröcker and Keasling 2006;Bulle et al. 2010;Gyomlai et al. 2021). The combination of Jurassic meta-ophiolites and HP/LT metamorphism is unusual and a unique feature. Judging from field observations, geochronological and bulk rock geochemical data, we consider a relationship to the Makrotantalon Unit of NW Andros to be very likely. The MU represents a distinct tectonic subunit of Pelagonian derivation in the nappe stack of the CBU that is mainly exposed in NW Andros, but possibly correlates with the Ochi nappe on Evia (Papanikolaou, 2013;Gerogiannis et al. 2019). The MU includes serpentinites and rare meta-gabbro in a marble-schist sequence that was affected by both Cretaceous and Eocene HP/LT episodes (e.g. Huet et al. 2015;Gerogiannis et al. 2019). The meta-gabbros and serpentinites exposed on the NW coast of Tinos could also represent remnants of this Pelagonian nappe, but so far there is insufficient evidence to support this assumption. Lithological and mineralogical criteria, maximum depositional ages and similarities in the tectono-metamorphic evolution suggest that the schist sequence below the Cape Steno mélange corresponds to the CBU on Tinos which was only affected by Eocene HP/LT metamorphism.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.