Diagenetic and Biological Overprints in Geochemical Signatures of the Gigantoproductus Tertiary Layer (Brachiopoda): Assessing the Paleoclimatic Interpretation

Variations in the geochemical signatures of fossil brachiopod shells may be due to diagenesis and/or biological processes (i.e., ‘vital effects’). It is critical to characterise them in order to identify reliable shell areas suitable for paleoclimate studies. This investigation contributes to an in-depth understanding of geochemical variations in Gigantoproductus sp. shells (SW Spain, Serpukhovian age), throwing light onto the Late Paleozoic Ice Age interpretation. Microstructural, crystallographic, cathodoluminescence and geochemical (minor and trace elements, δ18O, δ13C, and strontium isotopes) characterisations have been performed on the tertiary layer of the ventral valve, to assess the preservation state. Poorly preserved areas exhibit microstructural and geochemical changes such as recrystallisation, fracturing and higher Mn and Fe enrichment. Moreover, these areas have a higher dispersion of ⁸⁶Sr, ⁸⁷Sr, δ18O and δ13C than well-preserved areas. Three structural regions have been identified in well-preserved areas of the ventral valve by differences in valve curvature and thickness, such as the umbonal and thick and thin regions. These regions have different proportions of Mg, S, Na, δ18O, and δ13C, which are interpreted as ‘vital effects’ and probably related to growth-rate differences during shell growth. The Gigantoproductus tertiary layer seems the most suitable for paleoclimate studies, because it retains the original microstructure and geochemical composition.


Introduction
The Late Paleozoic Ice Age (LPIA), elapsed from the Viséan to the Capitanian (~335 to 260 Ma) [1], was a dynamic climatic phenomenon that involved multiple glacial and interglacial events driven by the continental Pangea assembly, and low global atmospheric carbon dioxide concentrations (pCO 2 ). These constraints, in turn, produced glacio eustatic variations, changes in ocean circulation, and finally, a turnover in marine flora and faunas [1][2][3][4][5][6][7][8]. The Mid-Carboniferous glacial peak represented an important change in seawater chemistry and ocean circulation [9] and led to one of the most important and selective crises in the Phanerozoic Era [10]. Low speciation and high extinction rates of marine faunas are the outstanding features during this period, particularly in the late Serpukhovian [11].

Studied Material
Eleven specimens of Gigantoproductus sp. ( Figure S1 in Supplementary Material) have been selected for this study, from the 2-4 horizon of the San Antonio section (Figure 1). The material (DMP-A301-1014-1 to DMP-A301-1014-11) is housed in the Paleontological Collections of the Paleontology Area (Complutense University of Madrid, UCM). The  were sectioned in two halves longitudinally, along the sagittal plane from the umbo to the commissure, when possible, and each slab was polished using sandpaper and 1 µm and 0.3 µm alumina. Polished samples were scanned in order to cross reference digitally the macroscopic features of the shells. One slab of each of the 11 specimens has been used for this study, and the other 11 slabs have been reserved for future studies. A total of twenty thin sections and five ultra-thin sections were prepared (sensu Coronado et al. [66]).

Microscopy Methods
The specimens were studied under a petrographic microscope, using scanning electron microscopy (SEM), computer-integrated polarization microscopy (CIP) and cathodoluminescence microscopy (CL), in order to select well-preserved areas for subsequent drilling.
A petrographic microscope, LM Leica DMLP, with a coupled camera Leica DC 300, was used for the purpose of characterizing the microstructure of the brachiopod shell.
Samples were extracted by natural breakage from the ventral valves of Gigantoproductus sp. between the muscle scars and the commissure, perpendicular and parallel to the valve growth. Moreover, a complete shell section was polished and etched in a 5% HCl solution for 20-25 s. These samples were coated with gold and analyzed by scanning electron microscopy (SEM), using a model JEOL JSM-820 working at 20 kV, located in the research facility of Geological Techniques of Complutense University of Madrid (UCM, Spain), in order to complete the microstructural characterization.
Thin sections were photographed under CL working at 16 kV and a current of 0.5 mA, in order to assess the intensity and distribution of luminescence. A cold cathodoluminescence probe model 8200 MK4 of Cambridge Image Technology Ltd. attached to a petrographic microscope model Eclipse E400 POL, with a camera, was used for this purpose. This microscope is located in the Geological Storage Division, Hydro-geochemical Group Ciemat (Madrid, Spain).
In order to evaluate shell preservation, the crystallographic organization of brachiopod shells was studied using computer-integrated polarization microscopy (CIP, [67]). This method for texture analysis and optical-orientation imaging has been applied in biomineralization studies of fossil bio-calcite with relevant crystallographic and structural results, highlighting it as a robust tool in diagenetic characterization [68][69][70]. This method determines the c-axis orientations of uniaxial minerals (such as calcite and quartz) from optical micrographs, displaying the results in pole figures and orientation images, using an RGB color-code that represents each orientation.
Seven CIP analyses in different samples and regions of the shell were performed, using an experimental petrographic Zeiss microscope with an automated rotation, and tilting system, made with the Arduino UNO microcontroller (open-source hardware) and controlled with software implemented in the LabVIEW environment (details in Coronado and Rodríguez [71]). The lowest ratio of pixel-µm reached with this setting in the micrographs is 1:0.05. Finally, the micrographs were processed using Image SXM software [72].
Crystal sizes were measured via micrographs using the plug-in ObjectJ 1.03w [73] of the free and open-source ImageJ 1.47v image-processing software [74].

Trace Elements and Isotope Analyses
Seven elements (Ca, Mg, Sr, S, Na, Mn and Fe) were analyzed using an electron microprobe analyzer (EMPA), with the analysis conducted on eleven thin sections (364 punctual analyses, 20 or 25 points per thin section), using a JEOL Superprobe, JXA-8900M with five wavelength-dispersive spectrometers, located in the Spanish National Centre for Electron Microscopy of the Complutense University of Madrid. Each point was analyzed with an accelerating voltage of 15 kV, a beam current of 10 nA and a spot size of 5 µm. Each analysis took an acquisition time of 45 s; because five elements can be analyzed simultaneously, the total time for analysis of the seven elements was 90 s. The following standards and detection limits were used in the EMPA: Ca: 201. 5  and S: 235.25 ppm (galena). In addition, the same seven elements were mapped. The EMPA mapping enables simultaneous analysis of different elements and the generation of distribution images for each element with 1 µm resolution. An accelerating voltage of 20 kV with a beam current of 100 nA and a spot size and step interval of 1 µm diameter (dwell time = 25 ms) were used.
Samples were powdered using a microdrill with an x-y micrometric plate attached to a binocular and a 500 µm tungsten carbide drill in selected areas, sampling for stable isotope analysis (δ 18 O, δ 13 C) and 87 Sr/ 86 Sr measurements. Stable isotope sampling requires only one drill hole to obtain enough material, whereas 87 Sr/ 86 Sr requires 30 to 40 drill holes.
Thermal ionization mass spectrometry (TIMS) was used in order to date the samples of study from the 87 Sr/ 86 Sr ratio. A total of 7 mg was obtained by the drilling of 4 samples of well-preserved zones and 2 samples of poorly preserved zones. The material was processed to obtain a Sr concentrate residue [75], which was loaded onto a Re filament by adding 1 µL of H 3 PO 4 1M 4 and 2 µL of Ta 2 O 5 . The Sr isotopic ratios were analyzed on a TIMS-Sector 54 ® Mass Spectrometer, located in the Laboratory of Geochronology and Isotope Geochemistry (UCM), following a dynamic multicollection data-acquisition method for 10 blocks of 16 cycles each, with beam intensity mass of 88 Sr at 3 V. The Sr analyses were corrected, to avoid possible interferences of 87 Rb. The 87 Sr/ 86 Sr ratios were normalized with respect to the measured value of the ratio 86 Sr/ 88 Sr = 0.1194, in order to correct the possible fractionation of masses that the sample might have undergone during the filament loading and/or instrumental analysis. During the analysis of the samples, the isotopic standard of Sr (NBS 987) was analyzed repeatedly, and the following values were obtained: 0.710239 ± 1.7 × 10 −5 (n = 8). These values were used to correct the analyses, taking into consideration the plausible drift referred to as the standard, and the certified value of the standard.
Stable isotope analyses (δ 18 O and δ 13 C) were undertaken using a triple-collector isotope-ratio mass spectrometer, Finnigan MAT 253 of the Stable Isotope Laboratory of the Department of Geological Sciences of the University of Michigan (USA). Forty-four samples were drilled, with 0.1 mg of each obtained, 30 from well-preserved zones and 14 from zones with evidence of poor preservation, previously characterized using optical microscopy, cathodoluminescence and CIP, and weighted and digested using H 3 PO 4 at 77 • ± 1 • C for 8 min in a Finnigan MAT Kiel IV. The analyses were calibrated using a betterfit regression line defined by two international standards, the NBS 18 (National Bureau of Standards; δ 13 C = −5.014‰ and δ 18 O = −23.2‰) and the 19 (National Bureau of Standards; δ 13 C = 1.95‰ and δ 18 O = −2.20‰). The data are given in ‰ notation, relative to the VPDB (Vienna Pee Dee Belemnite). The accuracy of the data was monitored through daily analyses, using a variety of carbonate-powder standards. At least 4 carbonate standards were reacted, analyzed and measured with an accuracy below ±0.1‰ for δ 13 C and δ 18 O.

Results
Collected specimens from the San Antonio section belong to horizons 2-4, which make them unsuitable for correlating geochemical variations with environmental factors (e.g., paleoseasonality). In order to differentiate diagenetic and biological overprints in Gigantoproductus shells, results have been divided into three main sections: shell microstructure, fossil preservation and geochemical characterization.

Shell Microstructure
Gigantoproductids have a concave-convex shell composed of a dorsal and ventral valve. The dorsal valve is thinner than the ventral valve. It has a mostly laminar microstructure, with lath crystals (sensu [76]),~1-2 µm in width, grouped in laminae (sensu [34,76,77]) 10 to 20 µm thick ( Figure 2a); a few columnar layers, with smaller crystals than the ventral valve, appear within the laminar microstructure. The crystal size decreases towards the inner part of shell, as was recognized in the ventral valve. The dorsal valve ( Figure 2a) shows a similar thickness (~2-3 mm), which decreases towards the commissure, whereas the ventral valve has a noticeable thickness difference. The shell shape can be divided into three regions of the ventral valve based on the thickness differences: the umbonal region (U-region), the thick region (Tk-region) and the thin region (T-region) ( Figure 2). The first region is more incurved than the rest of the shell, and is of ~8-10 mm in thickness, the second one is a thickened area, ~13-22 mm, following the U-region, and the third one is more planar than the U and the Tk region, with a ~5-7 mm thickness, which decreases from the thick region to the commissure. In the ventral valve, two layers were distinguished: (i) the secondary or laminar layer and (ii) the tertiary or columnar layer. The laminar is formed by lath crystals, ~1 μm in width, arranged in packages of thin sheets ~10 to 40 μm in width ( Figure 2f). The crystals are mostly tabular in appearance, with sharp contacts between the crystals (Figure 3e); the tertiary or columnar layer is formed of large columns (sensu [78]), with sizes ranging from The dorsal valve ( Figure 2a) shows a similar thickness (~2-3 mm), which decreases towards the commissure, whereas the ventral valve has a noticeable thickness difference. The shell shape can be divided into three regions of the ventral valve based on the thickness differences: the umbonal region (U-region), the thick region (Tk-region) and the thin region (T-region) ( Figure 2). The first region is more incurved than the rest of the shell, and is of~8-10 mm in thickness, the second one is a thickened area,~13-22 mm, following the U-region, and the third one is more planar than the U and the Tk region, with a~5-7 mm thickness, which decreases from the thick region to the commissure.
In the ventral valve, two layers were distinguished: (i) the secondary or laminar layer and (ii) the tertiary or columnar layer. The laminar is formed by lath crystals,~1 µm in width, arranged in packages of thin sheets~10 to 40 µm in width (Figure 2f). The crystals are mostly tabular in appearance, with sharp contacts between the crystals (Figure 3e); the tertiary or columnar layer is formed of large columns (sensu [78]), with sizes ranging from 100 to 1700 µm in length and 40 to 150 µm in width ( Figure 2e). Occasionally, the tertiary layer shows intercalation of laminar growth lines with a similar microstructure to the secondary layer embedded between the columns. The columnar crystals are composed  Likewise, two types of microstructural changes parallel to the shell growth were identified in the tertiary layer, called growth lines: (i) a dashed and diffuse aspect (Figure 2(b2)) with several dark dots (i.e., opaque under transmitted light and in BSE images), passing through the columnar crystals,~50 to 100 µm of thickness; (ii) a laminar microstructure (Figure 2(b1)), densely packed and with~30 to 80 µm of thickness. These growth lines cut off the columnar crystals. A variation of the laminar microstructure with a granular appearance is observed in these growth lines.

Fossil-Shell Preservation
Fossil preservation has been addressed herein by the combination of cathodoluminescence, textural changes under both optical microscopy and SEM, the crystallographic arrangement of microstructures and the geochemical signatures (i.e., minor and trace elements and Sr isotopes). This set of techniques was applied to distinguish between wellpreserved and poorly preserved shell regions, with the aim of comparing the luminescence and geochemical differences between the different shell regions and their relationship with the microstructure.

Textural Changes
Microtextural changes were recognized in some areas of the intra-shell surface, due to diagenetic processes. Fractures of the shell are usually oriented perpendicularly, or at a high angle to the shell surface, starting from the inner shell and moving to the outer shell, and are usually filled with micrite, sparite or iron oxides. On the other hand, thinner fractures, with micrite filling, parallel to the shell surface were observed in the center of some ventral valves. Some samples expose delamination processes near the shell surface.
Changes in the crystal size and the orientation of the tertiary layer are produced in adjacent zones of fractures or near to the edges of the shell. These areas contain a decreasing trend in crystal size towards the outer shell, with a mosaic appearance. Some areas exhibit micritization of the secondary layer and degrading neomorphism in the tertiary layer. These recrystallized zones are recognized by small and equigranular crystals, which totally obliterate the primary microstructure. Moreover, other zones contain partially dissolved crystals with a cloudy appearance, where the primary microstructure is roughly recognized. Rounded dark dots are opaque inclusions and are mostly concentrated in growth lines (Figure 2(b2)), whereas other scattered inclusions have sharp edges and different interference colors.

Crystallographic Arrangement
The observation of crystallographic patterns is a precise approach for the evaluation of diagenesis of polycrystalline skeletons formed by microcrystals, with preferred crystallographic orientations and a notable arrangement of supra-specialized structures [29]. This criterion has been previously applied to brachiopod shells, achieving excellent results in experimental diagenesis and in the evaluation of fossil shells [79][80][81]. In addition, crystallographic assessments of other calcium carbonate skeletons have been successful [65,67,68]. This approach is based on the principle that controlled mineralized shells have constrained crystallographic arrangements (i.e., preferred orientations), which, after diagenesis, can be compromised and substituted by random orientations and/or loss of textural features.
Seven areas of five ultra-thin sections of Gigantoproductus sp. ventral valves were studied after sampling for geochemical analysis ( Figure 4). The CIP method is limited to assessing the crystallographic organization based on azimuth and inclination of the c-axis. Criteria to identify diagenetic alterations are based on abrupt changes in the c-axis orientation (highly disoriented), and textural changes with or without subtle c-axis variations.  (7), showing a better constrained c-axis orientation. (10) Corresponds to the complete area (g), showing a pole maximum, with c-axis which is strictly oriented perpendicular to the growth direction of the Gigantoproductus shell. (11)(12) Corresponds to selected areas. (11) Prismatic area with random c-axis orientations, due to recrystallisation. (12) Equivalent area to (11), showing a well-constrained c-axis orientation. Pole figures were calculated as an orientation-distribution function and are provided in multiples of uniform distribution intervals of 0.5 for c-axis orientations of the complete area. Red points correspond to the punctual c-axis orientation of each pixel of selected areas.  (7), showing a better constrained c-axis orientation. (10) Corresponds to the complete area (g), showing a pole maximum, with c-axis which is strictly oriented perpendicular to the growth direction of the Gigantoproductus shell. (11)(12) Corresponds to selected areas. (11) Prismatic area with random c-axis orientations, due to recrystallisation. (12) Equivalent area to (11), showing a well-constrained c-axis orientation. Pole figures were calculated as an orientation-distribution function and are provided in multiples of uniform distribution intervals of 0.5 for c-axis orientations of the complete area. Red points correspond to the punctual c-axis orientation of each pixel of selected areas. The c-axis is strictly oriented perpendicular to the growth direction of Gigantoproductus shells (i.e., parallel to the radial axis of the valve), regardless of microstructure, shell region and texture, although some variations have been recognized. Columnar and laminar microstructures exhibit different c-axis orientations: parallel to the elongation axis of the crystals in the case of the columnar microcrystals, and perpendicular in the case of the laminar crystals (Figures 4 and S2). In those areas where the laminar microstructure changes gradually to columnar, and vice versa (e.g., in a growth line), the c-axis is kept constant between the interconnected crystals (Figure 4b), similar to kinked crystals (sensu [82]).
All the studied areas, which include all the shell regions identified, demonstrate preferred orientations (based on pole figures), with quite narrow azimuthal dispersion. However, some distinctive arrangements have been observed in some shell regions. Usually, c-axis orientations in the tertiary layer show two narrow pole maxima, with an azimuthal dispersion between them of ca. 55 • in the umbonal region ( Figure S2e Evidence of diagenetic alteration based on crystallographic orientation and textural changes occurs in some areas. In those areas, where the original microstructure is apparently well-preserved and lamellar twinning is notable, they are characterized by random c-axis orientations in twinned crystals and the surrounding areas (with hundreds of microns, Figure 4b,h). These areas correspond mostly to thin fractures without displacement, favored by microstructural changes (e.g., columnar to laminar or contact between columnar crystals, Figure 4h). In addition, in some of these areas the original microstructure has been completely obliterated by recrystallisation processes (Figure S2b), and these areas exhibit random azimuthal orientations, forming large patches with similar characteristics. On the other hand, in some areas the surrounding fractures are characterized by a degrading neomorphism with co-oriented microcrystals, compared to the main crystallographic orientation of the shell (Figure 4e), although the azimuthal dispersion of orientations in this region is larger than in individual crystals (i.e., columnar, and laminar crystals). In Figure 4g,h the diagenetic alteration is confined by a growth line, and involves only a part of the underlying crystals. This area is recrystallized because the original texture is lost, and random crystallographic orientation occurs.

Geochemical Characterisation
Geochemical analyses, minor and trace-element (MTE) contents and isotope compositions, have been conducted in well-and poorly preserved areas, taking into The luminescence in the columnar crystals decreases from the edges to the inside ( Figure 5), as well as in degrading neomorphism areas (Figure 5i). The luminescence of the growth lines increases in those areas affected by fractures, in comparison with the surrounding area (Figure 5d,g).

Geochemical Characterisation
Geochemical analyses, minor and trace-element (MTE) contents and isotope compositions, have been conducted in well-and poorly preserved areas, taking into consideration the luminescence under CL (Table 1, Figure 5), and the three shell regions identified (i.e., umbonal, thick and thin, Tables 2, S2, and 3).    The Gigantoproductus ventral valve shows different trace-element concentrations related to diagenetic alteration, layers, and shell regions ( Figure S3, Tables 1 and 2).
The proportion of trace elements varies with the function of the preservation state ( Figure S3, Table 1). In general, well-preserved areas of the ventral valve show lower mean standard deviation of the Ca, Mg and S values than the poorly preserved areas. The NL well-preserved areas contain more Ca, less Mg and less dispersion of all elements, except Sr, than the NL poorly preserved areas. The SL well-preserved areas contain more Mn and Sr, less S, and exhibit lower mean standard deviation of Ca and Mg than the SL poorly preserved areas. The highest mean of Mn occurs in the luminescent areas ( Figure S3, Table 1).
The mol% of MgCO 3 /CaCO 3 of the NL and SL well-preserved areas, for each shell region, represents a linear correlation (r = −0. Statistics per element (Table 2) of the different brachiopod-shell regions demonstrate a higher average concentration of Ca and Mn in the umbonal region than in the rest of the regions. The thick region (Tk) exhibits less proportion of Mg and Mn than the U-and T-regions, but more Ca. The T-region contains slightly more Mg, Sr, and S than the U-and Tk-regions, but lower Na. Mg and Sr average values are higher in the secondary layer than in the tertiary layer, whereas Ca is lower in the secondary layer.
Ca, Mg, S and Na of all well-preserved areas are shown in boxplots in Figure 6. Mg is higher in the U-and T-regions, whereas Ca is higher in the Tk-region. Na is lower in the T-region; however, S is higher. The Mg/Ca, S/Ca and Na/Ca transects measured in well-preserved areas of three individuals varies, from a higher concentration near the shell edge to a decreasing one towards the shell interior (Figure 7). In addition, the concentration of Mg, S and Na covaries in the measured transects. In poorly preserved areas (e.g., recrystallized areas) a decreasing-metal/Ca trend is not observed, and each element is not positively correlated ( Figure  7c).  The Mg/Ca, S/Ca and Na/Ca transects measured in well-preserved areas of three individuals varies, from a higher concentration near the shell edge to a decreasing one towards the shell interior (Figure 7). In addition, the concentration of Mg, S and Na covaries in the measured transects. In poorly preserved areas (e.g., recrystallized areas) a decreasing-metal/Ca trend is not observed, and each element is not positively correlated (Figure 7c).

Isotope Geochemistry
The values of 87 Sr/ 86 Sr obtained from the TIMS range from 0.707830 to 0.707860 in the NL well-preserved areas (Table 3), while SL in the poorly preserved areas deviated from these values by ±0.00003. One sample is enriched in 87 Sr, while another is depleted, in the cases of the well-preserved samples.
The δ 13 C-δ 18 O relationship can be subdivided according to the dispersion degree into poorly preserved and well-preserved areas and the consistency within the identified shell regions (Figure 7; Table 3). Dispersed values, with a variation of ~4.6‰ for δ 18 O and of ~3‰ for δ 13 C, occur in poorly preserved areas, whereas ~1.7‰ for the δ 18 O and ~1.3‰ δ 13 C occur in well-preserved areas. Poorly preserved areas show different alteration signatures: some samples have similar δ 13 C but are depleted in δ 18 O, others are depleted in δ 13 C and δ 18 O, and others are δ 13 C-and δ 18 O-enriched (Figure 8; Table 3).
A correlation with each identified shell region is recognized in well-preserved areas (i.e., the same drilled regions contain rather similar values; Figure S5). Thus, the δ 18 O val-

Isotope Geochemistry
The values of 87 Sr/ 86 Sr obtained from the TIMS range from 0.707830 to 0.707860 in the NL well-preserved areas (Table 3), while SL in the poorly preserved areas deviated from these values by ±0.00003. One sample is enriched in 87 Sr, while another is depleted, in the cases of the well-preserved samples.
The δ 13 C-δ 18 O relationship can be subdivided according to the dispersion degree into poorly preserved and well-preserved areas and the consistency within the identified shell regions (Figure 7; Table 3). Dispersed values, with a variation of~4.6‰ for δ 18 O and of~3‰ for δ 13 C, occur in poorly preserved areas, whereas~1.7‰ for the δ 18 O and 1.3‰ δ 13 C occur in well-preserved areas. Poorly preserved areas show different alteration signatures: some samples have similar δ 13 C but are depleted in δ 18 O, others are depleted in δ 13 C and δ 18 O, and others are δ 13 C-and δ 18 O-enriched ( Figure 8; Table 3).

Discussion
In this section, the preservation of the tertiary layer of Gigantoproductus ventral valves and the influence of diagenesis on the geochemical signatures are discussed. Moreover, the kinetic and biological effects that overprint the geochemical signatures suitable for paleoclimatic purposes, are assessed.

Fossil Preservation
The Gigantoproductus ventral valve can retain its original structural features at macroand at microscale. The secondary and tertiary layers are well-defined by their microstructure, with a controlled preferred-crystallographic arrangement (the c-axis perpendicular to the shell edge). The microstructure is interpreted as unaltered because it retains its original morphological features, similar to those described in some extant brachiopods [82,83]. Likewise, submicrometric laminae have been recognized (Figure 3d) in columnar crystals of the tertiary layer, such as those described by Schmahl et al. [83] in the columnar crystals of Terebratulina septentrionalis. Furthermore, preserved nanocrystals, with a granular texture, were identified in some columnar microcrystals from natural breakage (Figure 3f,h). This finding suggests good preservation (i.e., the preservation of biogenic features) in fossil carbonate biominerals [29,84], because the nanocrystals are also characteristic of extant brachiopods [85].
The secondary and tertiary layers in Gigantoproductus shells are characterized by different microstructure and geochemical compositions, as was recognized in other extant and fossil brachiopods [36,43,86]. The secondary layer, in the Gigantoproductus of this study, is characterized by a laminar microstructure, with the elongate axis of crystals oriented parallel to the shell surface and with a small crystal size in comparison with the A correlation with each identified shell region is recognized in well-preserved areas (i.e., the same drilled regions contain rather similar values; Figure S5). Thus, the δ 18 O values of well-preserved areas range from −3.8 to −2‰ and from 2.1 to 3.4‰ for δ 13 C. The U-region is where the data shows the lowest δ 18 O, with averages of −3.4‰ for δ 18 O and 2.6‰ for δ 13 C ( Table 3). The Tk-region exhibits similar δ 13 C data to that of the U-region, but lower δ 18 O, with averages of −2.6‰ for δ 18 O and 2.6‰ for δ 13 C. Higher δ 13 C values are recognized in the T-region compared to the U-region, with averages of −3.2‰ for δ 13 C and 2.7‰ for δ 18 O.

Discussion
In this section, the preservation of the tertiary layer of Gigantoproductus ventral valves and the influence of diagenesis on the geochemical signatures are discussed. Moreover, the kinetic and biological effects that overprint the geochemical signatures suitable for paleoclimatic purposes, are assessed.

Fossil Preservation
The Gigantoproductus ventral valve can retain its original structural features at macroand at microscale. The secondary and tertiary layers are well-defined by their microstructure, with a controlled preferred-crystallographic arrangement (the c-axis perpendicular to the shell edge). The microstructure is interpreted as unaltered because it retains its original morphological features, similar to those described in some extant brachiopods [82,83]. Likewise, submicrometric laminae have been recognized (Figure 3d) in columnar crystals of the tertiary layer, such as those described by Schmahl et al. [83] in the columnar crystals of Terebratulina septentrionalis. Furthermore, preserved nanocrystals, with a granular texture, were identified in some columnar microcrystals from natural breakage (Figure 3f,h). This finding suggests good preservation (i.e., the preservation of biogenic features) in fossil carbonate biominerals [29,84], because the nanocrystals are also characteristic of extant brachiopods [85].
The secondary and tertiary layers in Gigantoproductus shells are characterized by different microstructure and geochemical compositions, as was recognized in other extant and fossil brachiopods [36,43,86]. The secondary layer, in the Gigantoproductus of this study, is characterized by a laminar microstructure, with the elongate axis of crystals oriented parallel to the shell surface and with a small crystal size in comparison with the columnar microstructure of the tertiary layer, where the elongation and c-axis are oriented perpendicular to the shell surface and with larger crystals. Therefore, the morphological features of the secondary layer make it more prone to fracture, delamination, and recrystallisation. Within the tertiary layer, the thick region (Tk) of the valve is less affected by recrystallisation and fractures than the thin and umbonal regions (T, U). This may be due to their large crystal size and/or central shell position. On the other hand, the thin dorsal valve contains more fractures than the ventral valve, which makes it more susceptible to alteration and to more micritized patches (Table 4). The secondary layer is highlighted by more intense luminescence under CL than the tertiary layer of the ventral valve ( Figure 3). Similar differences in CL-luminescence intensity between the secondary and the tertiary layers had been recognized in Composita subtilita and Neospirifer pattersoni [87]. Although classically Mn and Fe are indicative of diagenetic alteration in calcite [88][89][90][91], not all luminescence, or its absence, can be correlated with diagenesis [65,78,79,92]. Biogenic carbonates can incorporate small amounts of Mn and Fe during their metabolic activity because of physicochemical variations of the environment in which they develop [92][93][94].
The luminescence differences in Gigantoproductus might be related to variations in the chemical composition; thus, the secondary layer is slightly enriched in Mn and Mg (Table 2), possibly related to more organic-matrix remains (Figure 3h). In addition, some fractures with a higher amount of Fe are non-luminescent under CL, and the growth lines with an apparently higher Mn concentration are luminescent under CL. This is partly due to the fact that Mn is a luminescence activator and Fe is an inhibitor of CL [91]. High Sr and Fe concentrations in some areas of the secondary layer suggest alteration by fluids favored by delaminated areas.
Slightly luminescent areas with a well-constrained crystallographic arrangement (considered here as well-preserved) have been identified in Gigantoproductus. These areas exhibit a luminescence pattern under CL that corresponds to the growth lines, and, in turn, show a slightly higher Mn content. This luminescence pattern under CL seems to be the result of organic matter and/or Mn [95] incorporated during shell growth. These areas also contain a higher Mg and S content, but no extensive Mn changes. The determination of small amounts of organic inclusions in growth lines (Figure 3h), and higher concentrations of Mg and Mn across the growth lines, suggest a combination of both sources, probably induced during the biocrystallization process. Similar luminescence patterns under CL were reported in the growth lines of the extant brachiopod Megerlia truncata [78], in fossil brachiopods [92,95] and in belemnite rostra [96]. Higher organic concentrations are recognized when the secretory regime of calcite is reduced [92]. This fact has been interpreted as growth cessation during environmental-stress events, produced by mantle anaerobiosis or by the acidification of calcification fluid by the closure of the valve [95,97,98]. Evaluation of these hypotheses deserves further study.
As described previously, not all non-luminescence areas of Gigantoproductus shells correspond to well-preserved zones. Crystals with disordered c-axis orientations have been identified in non-luminescent areas under petrographic microscopy. This may be due to slow recrystallisation processes, which keep the original microstructure but change the orientation of the c-axis. Examples of this process can be observed in coral skeletons [67,70]. Therefore, luminescent areas are not always indicative of diagenetic alteration [92] and non-luminescent areas are not always indicative of good preservation [29,37]. Banner and Kaufman [18] and Barbin and Gaspard [92] documented non-luminescent areas with altered 87 Sr/ 86 Sr and δ 18 O values in fossil brachiopod shells and slightly luminescent areas without evidence of alteration in an extant brachiopod.
The combination of CL and geochemical analyses helps to evaluate diagenetic alteration [65], but it is necessary to fully understand the original brachiopod-shell chemistry before using geochemical features such as a diagenetic indicator, in order to consider those biotic and abiotic factors that could control the original shell chemistry [36]. The chemical composition of the Gigantoproductus shells in this study is equivalent to that in the gigantoproductid reported by Popp et al. [17], Bruckschen et al. [20], Armendáriz et al. [9], Angiolini et al. [34,35] and Nolan [32]. The luminescence pattern under CL in well-preserved areas is similar to those described by Angiolini et al. [34,35]; additionally, a similar luminescence of diagenetically altered areas was described by Armendáriz et al. [9] and Nolan [32].
Minor and trace elements have been traditionally used to evaluate brachiopod-shell preservation by comparison between individuals and genera and with extant, unaltered brachiopod shells [17,28,36]. The NL and SL poorly preserved areas of the tertiary layer of the Gigantoproductus ( Figure S3, Table 1) have~800 ppm more Mg than the well-preserved areas. Moreover, the NL poorly preserved areas have~800 ppm more Fe and~120 ppm more Mn than the equivalent well-preserved areas. These values for the poorly preserved areas probably reflect the influence of Mg-, Fe-, and Mn-rich fluids during diagenesis in different stages (burial and meteoric waters). These data agree with the observations made by Grossman et al. [36] regarding altered brachiopod fossil shells, which are Fe-and Mn-enriched. Additionally, the diagenetic alteration of valves modifies the isotopic record of δ 18 O, δ 13 C and 87 Sr/ 86 Sr. Poorly preserved areas exhibit δ 18 O and δ 13 C with larger standard deviations than those of the well-preserved areas (Table 3). Five samples of poorly preserved areas ( Figure 8) are depleted in δ 18 O, with a similar depletion for δ 13 C; two samples are depleted in δ 18 O and δ 13 C; three samples are enriched in δ 18 O and δ 13 C; and two have similar amounts of δ 18 O, but show a δ 13 C enrichment with respect to the well-preserved areas.
In relation to Sr isotopes, well-preserved areas contain homogenous 87 Sr/ 86 Sr ratios, in contrast with poorly preserved areas (with one 87 Sr/ 86 Sr-enriched sample and one 87 Sr/ 86 Sr-depleted sample). Sr-isotope values from the Gigantoproductus (from 0.707830 to 0.707860) are equivalent to those shown by Bruckschen et al. (1999) in brachiopods from the Pendleian substage (mostly gigantoproductids shells, from 0.707828 to 0.707879), which validate the good preservation of the samples, except for two samples.
Different signatures of stable isotopes (δ 18 O and δ 13 C), trace elements (Mg, Mn, and Fe), and Sr isotopes suggest the co-existence of different diagenetic processes and stages acting on the Gigantoproductus shells, a fact which is also supported by the showcasing of microstructural and crystallographic changes.
On the other hand, the δ 18 O and δ 13 C trends in well-preserved areas can be explained by other mechanisms, such as kinetic and biological variations in ionic and isotopic values from the equilibrium of seawater ('vital effects').

Biological Overprint of Geochemical Signatures
'Vital effects' are geochemical deviations from the thermodynamic equilibrium of seawater produced by the vital processes of organisms, such as kinetic effects during ontogeny. Curry [99] observed three growth phases with different growth rates in Terebratulina retusa: a fast growth rate during the first stages of growth (in the umbonal region), followed by a constant growth period of three years and a progressively decreasing growth rate with brachiopod ageing. Variations in Mg across the shell have been widely studied in the literature as indicative of growth rates. Buening and Carlson [56] observed a higher Mg amount near the umbo of extant brachiopods, which decreases during brachiopod ageing.
The tertiary layer of the Gigantoproductus ventral valve can be divided into structural regions (umbonal, thick and thin regions) and growth stages (younger and older), because during ageing the shell grows in length (from umbo to commissure) and thickens (towards the interior of the valve), as a coupled process. Therefore, the U-, Tk-and T-regions must include younger (the outermost parts of the shell) and older (the innermost parts of the shell) areas, which may disguise some kinetic fractionation, considering the mean values of each area. For instance, structural regions exhibit differences in averaged Mg: the umbonal region has more Mg than the thick region (Tk), and the thin region (T) has the highest amount of Mg. On the other hand, transects analyzed for the function of distance exhibit a Mg depletion towards the interior from the outermost part of the ventral valve. Tk-region shows an exponential variation of Mg in contrast to T-region, which varies linearly. Different Mg trends have been identified in extant brachiopod shells: an Mg-enrichment with ageing in the extant brachiopod shell Magellania venosa and a parabolic Mg trend with ageing in the extant brachiopod shells Liothyrella neozelanica and Gryphus vitreus [51]. Variations in Mg concentration of brachiopod shells seems to be species-specific [54,100], with different distribution trends among species [51] or within the same species [29,54]. Moreover, Rollion-Bard et al. [43] showed differences in the Mg/Ca incorporation across the shell in different extant brachiopod species, emphasizing the fact that Mg does not incorporate by a simple pathway during ontogeny. These differences may be related to different growth rates [56], crystallographic features [45], different proportions of organic components in the shell, seawater temperature and pH variations [42], and/or Mg exclusion from the calcification fluid [43]. Besides the Mg variation, S and Na vary from the outer edge of the ventral valve of Gigantoproductus towards the interior. The microstructure of the tertiary layer shows abundant smaller crystals located in the interphase between the secondary and tertiary layers (the initial biocrystallization event of the tertiary layer, Figure 2d), where nanograins and organic inclusions are frequent. Remarkably, these areas with small crystals are Mg-S-and Na-enriched, and during ageing the shell crystals enlarge and these elements decrease towards the interior. Na in Gigantoproductus seems positively correlated with Mg and S (Figure 7) in the well-preserved areas. Na concentration in carbonates can be influenced by different processes that might explain this trend. For example, Na incorporation in abiogenic calcite and biogenic calcite may be favored by crystalline defects and by higher precipitation rates [101,102]. Moreover, Na can occupy an interstitial position in the carbonate-crystal lattice or can form part of the carbonate lattice by an altervalent substitution, creating a vacancy in the lattice [103,104]. On the other hand, Na can form part of the organic matter in biogenic carbonates; for instance, Rollion-Bard and Blamart [102] described a positive correlation between Na/Ca and Mg/Ca with the organic matter in bioaragonite in corals. Finally, a composition of carbonates can be strongly influenced by diagenesis, more than by environmental processes [105].
Mii et al. [106] showed a positive correlation between Mg, S and Na in fossil samples of Neospirifer dunbari, similar to the Gigantoproductus of this study, supporting the hypothesis that these trends are original signatures, indicative of well-preserved areas in fossils.
Mg, Na and S incorporation into a crystal lattice have been associated with crystal growth rates in crystal-growth experiments [101,107]. Moreover, growth rates associated with trace-element partitioning have been observed in both extant and fossil brachiopod shells [36,56,106]. Mii et al. [106] and Grossman et al. [36] related the high concentrations of these elements to seasonal growth rates, increasing during summer and decreasing during winter. Intra-shell small variations in Mg, Na and S in Gigantoproductus-shell transects ( Figure 6) may be explained by this interpretation, despite the overprinted depletion trend from the shell edge towards the interior caused by brachiopod ageing.
All specimens of Gigantoproductus exhibit a depletion of δ 13 C and δ 18 O from older regions (the Tk-and T-regions) to younger regions (the U-region). It is remarkable that data from the thick region (Tk) are split into two clusters ( Figure 8): one near the shell edge (younger), more depleted in δ 13 C, and another close to the shell interior (older), enriched in δ 13 C. Fractionation of δ 13 C and δ 18 O in Gigantoproductus shells exhibits an inverse trend to that observed in Mg, S and Na partitioning. Younger regions with higher growth rates are Mg-, S-, and Na-enriched and δ 13 C-and δ 18 O-depleted, whereas older zones are Mg-, S-, and Na-impoverished and δ 13 C-and δ 18 O-enriched. Preferential isotope fractionation has been previously described in extant and fossil brachiopods: for instance, Auclair et al. [49] described δ 13 C and δ 18 O depletion from the outermost (younger) to the innermost (older) part in the secondary layer of Terebratalia transversa. Similarly, Batt et al. [24] recognized δ 13 C depletion in the secondary layer of a fossil productid and δ 13 C and δ 18 O depletion in the tertiary layer from the outermost (younger) to the innermost (older) part of the genus Composita shell. Isotope fractionation was related to growth rate in brachiopod shells [24,49,50]. δ 13 C and δ 18 O direct-and inverse-covariations trends have been reported in extant and fossil brachiopod shells [24,42]. Moreover, the depletion trend of high growth-rate areas is not only linear, due to the parabolic trend, and inverse parabolic trends have been described in other brachiopod shells [51]. These variations in the partitioning and fractionation of ions and isotopes observed in different species and genera are called 'vital effects' and have been related to specific growth rates or metabolic changes in the brachiopods. Some authors were concerned about their influence on the shell geochemistry (e.g., Mg/Ca, δ 13 C and δ 18 O) and the possible paleoclimatic misinterpretations [50]. According to McConnaughey [108], 'vital effects' represent a noteworthy problem for isotopic geochemistry and paleoclimatology studies. The current results on Gigantoproductus geochemical variation underline the fact that they can be extended to the fossil record and, more specifically, to the interpretation of the Late Paleozoic Ice Age (LPIA). This requires the identification of 'ideal' sampling regions, in order to avoid differences in ion partitioning and isotope fractionation related to the differences in the growth rates (kinetic effects).
The Gigantoproductus regions less affected by "vital effects" (the sensu "equilibrium zone" or "plateau zone" identified by Perez-Huerta et al. [55] and Rollion-Bard et al. [43]) are the T-zone and the inner zone of the Tk-regions (Figures 7 and 8), because of the low dispersion of the data. The umbonal region and external parts of the Tk-region of the Gigantoproductus ventral valves should be discarded, to avoid paleoclimatic misinterpretations (up to~1.05‰ of δ 18 O and 14 mmol/mol of Mg/Ca). Herein, the Gigantoproductus tertiary layer offers great potential for paleoclimate studies, due to (i) a higher volume of shell substance compared with the secondary layer; (ii) growth-rate variability, which allows the characterization of the 'vital effects' of the shell; (iii) being less prone to diagenetic alteration than the secondary layer; and (iv) higher growth-line spacing, which allows for the avoidance of growth lines.

Conclusions
A comprehensive study of the microstructure, crystallography, and geochemistry (trace elements, δ 18 O, δ 13 C and 87 Sr/ 86 Sr) of the tertiary layer of the Gigantoproductus sp. ventral valves, has allowed the differentiation of areas with signs of diagenesis, from the primary biogenic features. In addition to fossil preservation, a special emphasis has been given to the characterization of the geochemical 'vital effects'.
The Gigantoproductus shells consists of three layers, but only two are still preserved: the secondary layer with a laminar microstructure, and the tertiary layer with a characteristic columnar microstructure.
Evidence of diagenesis has been recognized in some poorly preserved areas, because the original crystallo-chemical features have been obliterated. These areas, regardless of luminescence under CL (i.e., NL or SL), show neomorphism, a random crystallographic distribution, and a disperse geochemical composition. It should be highlighted that the luminescence under fluorescence microscopy, identified in some well-preserved areas as growth lines, corresponds mainly to preserved organic-matrix remains.
Well-preserved areas of secondary and tertiary layers exhibit a well-constrained crystallographic arrangement, with the c-axis oriented parallel to the elongation axis of the prisms and perpendicular to the laminae. Three structural regions were identified in the ventral valve, based on the curvature and thickness changes in the ventral valve: the umbonal (U), thick (Tk) and thin (T) regions.
Furthermore, the geochemical composition of these areas is well-constrained in MTE (Ca, Mg, Sr, Na, S, Mn, and Fe) as well as in isotope composition (δ 13 C, δ 18 O and 87 Sr/ 86 Sr), with a characteristic variability possibly related to shell growth-rate differences, these being identified as geochemical 'vital effects'. Younger regions with higher growth rates (e.g., umbonal and the secondary-tertiary-layer interphase) are Mg-, S-and Na-enriched and δ 13 C-and δ 18 O-depleted, whereas older zones (the inner zones of the Tk-and T-regions) are Mg-, S-and Na-impoverished and δ 13 C-and δ 18 O-enriched. These latter regions are the most suitable areas for geochemical analysis of the Gigantoproductus.
The 'vital effects' caused by kinetic effects have a direct influence on the paleotemperature estimation; up until~1.05‰ of δ 18 O and 14 mmol/mol of Mg/Ca, intra-shell variation was detected, which can modify the ulterior palaeoclimatological interpretation, underestimating or overestimating the environmental conditions of the LPIA.