Geological Setting and Stratigraphy
The main material was collected near a cliff at the top of the Binolen section (“C-Layers” after Löw et al., 2022; GPS: 51°22’12’’N, 7°51’27’’E) within the Hönne Valley in north-western Sauerland, Germany. The Binolen section is located in the northern Rhenish Massif at the eastern edge of the Remscheid-Altena Anticline, which is surrounded by carbonate platform deposits of the Hagen Balve Reef. In terms of stratigraphy, the base of the Binolen section was positioned into the lower Givetian (likely within the timorensis conodont zone), defining the lower boundary of the basal part of the Hagen-Balve Formation (Binolen Member)64. However, the cliff at the top of the Binolen Member falls within the lower/middle Givetian boundary interval64,81.
During the Givetian, the Hagen Balve Reef was developed as an elongated carbonate platform surrounding a local submarine high on the Rhenish shelf, at the southern tip of Laurussia (Fig. 1). The onset of reef formation started approximately isochronous during the lower Givetian1. The depositional history of the initial reef formation of the Binolen Member was divided into several depophases1. The samples analyzed in this study were collected from strata within the upper part of Depophase VI (Beds 59 to 65 of “C-layers”)64 and stem from the initial reef platform of the Hagen Balve Reef. This part of the initial reef formation of the Binolen Member is characterized by coral-stromatoporoid frame-rudstones and coral-stromatoporoid-dominated float-bafflestones, representing a semi-open carbonate platform with argillaceous sediment input64.
Samples from the Eifel were provided by the Senckenberg Research Institute and Natural History Museum Frankfurt. The limestone synclines of the Eifel are located between the Lower Rhine Bay in the north and the Trier Bay in the south. Geologically, the region is part of the Rhenish Massif and consists of Devonian slates, sandstones, and limestones, interspersed with bioclasts, which were deposited in a coastal setting south of Laurussia (Fig. 1)82,83.
The Sötenich Syncline is characterized by changing assemblages of thin layered marly mudstones and thick layers of gastropod-coral-trilobite wackestones to floatstones which merge into stromatoporoid -coral-rudstones in the uppermost profiles. The coral associations are indicative of a low energy regime within a shallow marine lagoon. Faunal composition and facies types in the upper profiles suggest elevated sediment input and elevated nutrient supply83.
The Dollendorf Syncline yields a rich macrofauna characteristic of the Mid-Devonian. Local limestones are mainly composed of calcisphere-ostracod-wackestones or packestones and amphiporoid-floatstones, indicating a shallow marine lagoonal setting with restricted, low-energy water flow. Interspersed amphiporoid-rudstones suggest periods of high-energy regimes with potentially greater influence of open-marine conditions82.
Thin Section Analyses and Sample Storage
For taxonomic identification of collected coral samples from Binolen, fossil-rich rock samples were cut systematically for longitudinal- and cross-sections of individual coral skeletons. Thin sections were prepared with a thickness of 50 mm. Microphotographs were made under transmitted light with a Keyence VHX digital microscope to identify the analyzed tabulate and rugose corals84–86.
Nine thin sections from the initial Hagen Balve Reef at Binolen will be stored in the Geomuseum of the Westfälische Wilhelms University in Münster (GMM) under the inventory numbers GMM B2C.59-1 to GMM B2C.59-9.
Eifel samples were provided by the Senckenberg Research Institute and Natural History Museum Frankfurt, Germany, with Roemerolites brevis brevis from the Sötenich Syncline: SMF 40159; Temnophyllum cf. ornatum (= T. latum) from the Sötenich Syncline: SMF 40367/2 and Mesophyllum vesiculosum from the Dollendorf Syncline: SMF 73856.
Conodont Alteration Index (CAI)
The assessment of the textural alteration of conodonts as a proxy for the maturation of rocks has long been known87. The first systematic approach on quantifying temperature regimes a rock has experienced during diagenesis with the CAI dates back to 197788,89. Generally, conodont elements are composed of calcium phosphate (frankolite)87. During the growing phase of the conodont animal, frankolite lamellae are separated by thin organic layers. This organic matter can alter as a consequence of a carbonization reaction and change color in a characteristic way known as the CAI (Extended Data Fig. 6, Supplementary Table 1). Since then, many authors successfully applied the CAI to assess and quantify the maturation of regional rock formations and basins90–96.
The CAI has been used within the Rhenish Massif93,97. Helsen and Königshof (1994) yielded a useful map of CAI isoclines for the region. We used 30 conodonts from Binolen to determine the temperature-induced diagenetic overprint of the limestones which have been collected from slightly older strata only a few meters away and narrowed values down to 4.0–4.5 (which correspond to maximum temperatures of 190–300°C)98 for most of the Middle Devonian strata of the Rhenish Massif. The CAIs of the different synclines of the Eifel Hills yield nearly homogenous values between 1.5–2.0 (which correspond to maximum temperatures of 50–95°C)60,93,99.
Analysis of Coral-Bound Nitrogen Isotopes
The CB-δ15N measurements were performed in the Martínez-García Lab at the Max-Planck Institute for Chemistry in Mainz (MPIC). We used the persulfate oxidation-denitrifier method74,100, first applied to corals by Wang et al. (2014, 2015)33,101, with the analytical modifications described by Moretti et al.102.
The collected samples of fossil-rich carbonate rocks were cut into smaller handpieces using a stationary rock saw. Sample material was carefully extracted from the handpieces with a mm-drill bit attached to a hand-Dremel®. Only those specimens were considered that were located on the edge of a handpiece to ensure that the different phases of material (coral skeleton, secondary sparite and surrounding carbonate sediment) and their respective dimensions were visible (Fig. 2 and Extended Data Fig. 1). Each phase was collected exclusively from the center of mass to minimize contamination of adjacent material (Extended Data Fig. 1). Subsequent samples were sieved to separate coarse (250–63 µm) and fine (63–5 µm) aliquots. The coarse fraction was used for N-isotope analysis, while the fine fraction was further prepared for C- and O-isotope analysis.
Twenty mg (±2 mg) of unclean, coarse powder was weighed into a 12 ml Wheaton® tube. Subsequently, for the removal of clay particles, 10 ml of a 2% Na-Polyphosphate solution was added, left on a shaker at 120 rpm for five minutes and then put in an ultrasonic bath for one minute. Afterwards, the tubes were taken out and the supernatant was decanted. 8–10 ml of Milli-Q® water was added and samples were centrifuged at 300 rpm for two minutes before being removed. The procedure was repeated three times.
To remove potential Fe-Mn oxides, 5 ml of pH-adjusted dithionite-citric acid (pH 8) was added to each sample tube, which was placed in an 80°C DI-water bath for 30-40 min. The samples were removed and centrifuged, the supernatant was decanted, and the sample was rinsed three times with Milli-Q® (see steps above). Afterwards, sample material was transferred to a previously muffled 4 ml VWR® borosilicate glass vial and 3 ml of a potassium peroxydisulfate oxidative reactant (POR) solution (pH>12) was added. Samples were then autoclaved at 121°C for 65 min for the oxidation of non-bound organic matter. Finally, the supernatant was removed with a muffled pipette attached to a vacuum line set at 500 mbar, and the sample was rinsed at least three times with Milli-Q® until the supernatant reached pH 7. Cleaned samples were stored in a drying oven at 60°C overnight.
Once the powder had dried fully, 15±5 mg of cleaned powder was weighed inside an in-house clean room to minimize the opportunity for contamination. Thereafter, skeletal organic matter was released by dissolving the cleaned coarse powder with 4N hydrochloric acid (HCl). This leads to a solution of calcium chloride (CaCl2) at pH < 2. The amount of 4N acid used had been calculated based on the sample weight. We used the stoichiometric calculation of the reaction (CaCO3 + 2HCl ⇌ CaCl2 + H2O + CO2), which translates to 5µl 4N HCl per 1 mg of clean carbonate powder. We added an additional 20 µl 4N HCl to each sample to ensure very low pH and thus complete dissolution.
Concurrently, a new POR solution was prepared inside the cleanroom with a 4 ml spike of 6.25N NaOH to reach high pH. 1 ml of POR solution was pipetted to each dissolved sample and at least 10 empty cleaned vials (blanks), and the batch of vials was placed in a custom-built sample rack that was tightly sealed with a PTFE sheet before being autoclaved at 121°C for 65 min. The basic POR pH conditions convert the sample N to nitrate within the closed vial. After the autoclave run, supernatant was tested for pH to make sure every sample is basic (pH > 10). Eventually, each sample was balanced with the same HCl aliquot previously used for dissolution to reach a pH near 7. From the resulting aliquot, nitrate concentration was measured for each sample by quantitative conversion to nitric oxide and subsequent chemiluminescence detection103, and these measurements were used to calculate the sample volume yielding 5 nmol N of nitrate.
A volume of 1 ml of concentrated denitrifying bacteria (Pseudomonas chlororaphis) was injected into 800 ml of growth media and left 4–6 days to grow in the dark at room temperature on a shaking rack. Once the bacteria have grown sufficiently, the medium was transferred to autoclaved PE bottles and centrifuged at 7600 rpm for 10 minutes. The supernatant was then discarded and the remaining bacteria pellet was resuspended with a buffered (pH 6.3) resuspension medium. From this, 3 ml was pipetted into muffled 12 ml glass vials, capped with a septum and tightly sealed were placed upside-down on a needle rack with a small extra needle for pressure release. The needle rack supplied a continuous flow of N2 for at least three hours to replace the internal atmosphere and dissolved gases with pure N2. Bacteria vials were removed from the rack, and ~0.8 ml of the oxidized sample was injected into each bacteria vial. Once all samples have been injected, bacteria vials were placed in the dark for 2–3 hours to ensure quantitative transformation of nitrate to nitrous oxide before being frozen at -21°C.
On the day of analysis, bacteria were thawed, lysed with several drops of 10N NaOH and finally placed on a mass spectrometer for isotope analysis. The δ15N of the N2O was determined by a purpose-built inlet system coupled to a Thermo MAT253 Plus stable isotope ratio mass spectrometer100,104. Long-term precision was determined by running internal carbonate standards with each sample batch, which yielded an average carbonate standard reproducibility of ±0.2‰. Average reproducibility for replicate Devonian coral measurements was 0.22‰ (n=45) and 0.68‰ (n=20) for clean and unclean samples, respectively.
Modern samples of Tubastraea spp. and Porites spp. were taken from four different collections in the Senckenberg Research Institute and Natural History Museum Frankfurt. Subsequent samples were drilled with a hand-held Dremel® and powder was transferred into 4ml borosilicate glasses using aluminum foil. Each sample was then sieved into coarse (250–63 µm) and fine (63–5 µm) aliquots, where 6 mg coarse and 100–200 µg fine powder was used for N isotope and C/O isotope analysis, respectively.
Before analysis of modern coral samples, 8 mg of clean coarse powder was weighed into a 4 ml VWR® borosilicate glass vial and filled with 4.25 ml of 2% Na-hypochlorite before being left on a shaking table at 120 rpm for at least 24 hours. Afterwards, supernatant was removed with a muffled pipette attached to a vacuum line set at 500 mbar and was further treated the same as described with the Paleozoic samples.
Coral Oxygen and Carbon Isotopes
Amounts of 100–200 µg of coral carbonate sample material was analyzed for δ18O in the inorganic stable isotope laboratory at the Max-Planck-Institute for Chemistry in Mainz. Within a run of 55 samples, one international carbonate standard (IAEA-603, n=10) and one internal carbonate standard (VICS, n=11) were used to calibrate the analyses to the Vienna Pee Dee Belemnite (VPDB) scale. Samples were measured with an isotope ratio mass spectrometer (IRMS) (Delta V Advantage, Thermo Scientific, Bremen, Germany) which is connected to a GasBench II unit (Thermo Scientific). Each sample was placed in a 12ml exetainer vial (Part no. 9RK8W; Labco, Lampeter, UK). Samples and standards were then put into a 70°C-heated hot block. Firstly, vials were flushed with He to remove atmospheric CO2. Then, 5-10 drops of >99% H3PO4 were added and the sample was left to dissolve for 1.5 hours. Finally, the sample was transferred in He carrier gas to the GasBench II, where water and contaminant gases were removed before subsequent isotope analysis in the IRMS. Average analytical precision, based on the reproducibility of IAEA-603, was 0.11‰ (1SD, n=42) and 0.09‰ (1SD, n=42) for oxygen and carbon isotopes, respectively.
The δ18O and δ13C values for our Paleozoic samples show mean δ18O values ranging from -5.57‰ for secondary sparite to -6.64‰ for sediment samples (Supplementary Table 4). Mean δ13C values of all samples cluster around 1.95‰, where lowest δ13C values are recorded for dendroid rugose corals (1.17‰), whereas sediment and secondary sparite show higher δ13C values (2.39‰ and 2.29‰, respectively). Notably, solitary rugose coral samples cluster within narrow δ18O and δ13C values (-5.72‰ and 1.52‰, respectively).
Skeletal δ18O and δ13C values of modern samples are relatively widespread ranging from -7.39‰ to 3.57‰ and -10.27‰ to 1.60‰, respectively (Supplementary Table 5). Symbiont-bearing and symbiont-barren coral species do not show a distinct offset in either δ18O or δ13C. However, modern symbiont-bearing and -barren species can be distinguished in a cross-plot of δ18O vs δ13C44 (Extended Data Fig. 3).
The original oxygen and carbon isotopic composition of corals can be altered by partial dissolution of aragonite, precipitation of secondary carbonates or by recrystallisation of metastable aragonite to calcite (McGregor and Gagan, 2003; Müller et al., 2001; Sayani et al., 2011; Swart, 2015). While secondary carbonates (sparite) are predominantly observed in submarine environments107, partial dissolution or recrystallisation is more common for subaerial settings (Hendy et al., 2007; McGregor and Gagan, 2003). According to previous studies on the Hagen Balve Reef and the Eifel region, samples have probably experienced both, submarine and subaerial alteration64,82,83.
δ18O and δ13C values of our samples from Binolen and the Eifel cluster within the range previously discussed for marine limestones111. The narrow range of δ18O and δ13C values would suggest photosymbiosis across species24,44,112 and thus, would stand in direct contrast to our result from CB-δ15N measurements, and previous suggestions, based on morphological analysis, that no Paleozoic corals harbored symbionts4,5 (Extended Data Fig. 3). Previous studies have highlighted that diagenetic processes and geochemical comparisons of polymorphs (i.e., calcitic skeleton for Paleozoic coral samples and aragonitic skeleton for modern Scleractinia) can bias the interpretation of δ18O and δ13C values and thus, are thought to be less robust proxies for fossil reef settings49,106,109. In addition, increasing temperatures and recrystallization can bias carbonate oxygen isotope values towards more negative δ18O values113,114. Thus, it is possible that diagenetic alteration of the coral carbonate δ18O can bias the interpretation towards more symbiotic associations.
Statistical Analysis
Standard deviations are given as ±1SD and statistical significance tests were conducted either by a Welch’s t-test given a similar sample size and a heterogenous variance or an individual t-test for similar sample sizes and variances115,116. All analyses were conducted using Python3 on a Jupyter Notebook® (version 5.7.4). Data were imported using the Pandas library and plotted with Seaborn or Matplotlib libraries.