Discovery of giant and conventional magnetofossils bookending Cretaceous Oceanic Anoxic Event 2

The study investigates the distribution, morphology, and ecological implications of magnetofossils across the Cenomanian–Turonian boundary, focusing on sediments bracketing Oceanic Anoxic Event 2 (OAE2) in the Holland Park (Virginia, USA) core. Magnetotactic bacteria (MTB) biomineralize magnetite or greigite and play key roles in redox-sensitive biogeochemical cycles (iron, sulfur, nitrogen, silicon, phosphorus, carbon). Although molecular studies suggest MTB may have Archean origins, robust conventional magnetofossil records only extend back to the mid-Cretaceous. Giant magnetofossils (>1 μm) of uncertain biological origin (from giant iron-biomineralizing organisms, GIBO) were previously associated mainly with Paleocene–Eocene hyperthermal events, but recent reports indicate broader temporal distribution. The research questions include: (1) Do robust conventional and giant magnetofossils occur in mid-Cretaceous sediments associated with OAE2? (2) What are their morphologies, preservation states, and abundances across pre- and post-OAE2 intervals? (3) What ecological and environmental information do these assemblages record? The purpose is to document the oldest robust magnetofossil assemblages (~97 Ma), identify potentially new giant morphologies, and evaluate how magnetofossil assemblage changes relate to environmental and diagenetic processes across OAE2. The work is significant for reconstructing past iron cycle dynamics and for establishing magnetofossils as sensitive, widely applicable environmental proxies.

Prior putative conventional magnetofossils reported from Archean-Proterozoic rocks (2.7 Ga stromatolites; 2.0 Ga cherts) do not satisfy robust criteria requiring combined morphologic, crystallographic, and magnetic evidence. The oldest irrefutable conventional magnetofossils date to the mid-Cretaceous. Giant magnetofossils (unusually large magnetite crystals, >1 μm) were initially documented around Paleocene–Eocene hyperthermals and interpreted as proxies for environmental change associated with warming, stratification, deoxygenation, and increased iron and organic matter supply. More recent studies report giant magnetofossils from ~93 Ma to 1.8 ka, including intervals of global cooling, indicating broader ecological distribution. Widespread MTB in modern and ancient oceans, where iron is often limiting, and the proliferation of giant magnetofossils during major hydrological and biogeochemical shifts, suggest links between MTB/GIBO and the marine iron cycle. This context motivates systematic searches and robust multiproxy identification of magnetofossils through the geologic record.

Site and core: The Holland Park core (Suffolk, Virginia; N36° 40' 55.6", W76° 46' 50.1", elevation 23.9 m) was drilled by the Virginia Department of Environmental Quality in 2012. Approximately 139 m of core was recovered; this study focuses on 126.0–81.4 m spanning the Cenomanian–Turonian transition and OAE2. Lithostratigraphy indicates shoreline-proximal shelf deposition, with dark laminated clays (OAE2) transitioning to micaceous quartz sands up-section; unconformities occur near 124.7 m, 86.9 m, and 82 m. Biostratigraphy: 67 samples (138.4–83.4 m) were analyzed for calcareous nannofossils using established smear-slide techniques for sandy and TOC-rich samples. Zonation followed Sissingh and Burnett and was correlated to Gradstein et al. Key markers include Eiffelithus turriseiffelii, Gartnerago obliquum, Helenia chiastia (upper CC9/UC2), first occurrence of Lithraphidites acutus acutus (base CC10a; UC3–UC4 undifferentiated), placement of UC5 (115.5–102.8 m) by absence of L. acutus acutus and H. chiastia, and first occurrences of Eprolithus moratus and E. octopetalus (base Turonian; CC10b/UC6b). UC7 marker Quadrum gartneri not recorded; CC12 (UC8) at 86.9 m; Danian markers at 82 m indicate a ~25.5 My hiatus. TOC and carbon isotopes: 83 samples were measured for total organic carbon and δ13Corg following USGS protocols. TOC increases to 0.5–2% and a ~+2‰ δ13Corg excursion (−23 to −21‰) from 117.9–100.5 m delineate OAE2. Rock magnetism: 103 specimens (126.0–81.4 m) had hysteresis loops and backfield curves measured to obtain Ms, Mr, Bc, and ratios (Mr/Ms, Bcr/Bc). 26 specimens were selected for first-order reversal curve (FORC) measurements: 15 in pre-OAE2 Cenomanian (124.4–118.0 m), 2 within OAE2 (115.1 m, 107.0 m), 7 in post-OAE2 Turonian (95.8–92.9 m), and 1 each at 125.2 m and 90.0 m. Measurements used an 8604 Lake Shore VSM; standard processing and mass normalization applied. Magnetic extraction and electron microscopy: Eight horizons were targeted for magnetic extraction and EM: pre-OAE2 (HP40825, 124.4 m; HP40625, 123.8 m; HP39705, 121.0 m; HP38715, 118.0 m) and post-OAE2 (HP31435, 95.8 m; HP31375, 95.6 m; HP30900, 94.2 m; HP30465, 92.9 m). Two additional horizons outside the magnetofossil intervals (HP35110, 107.02 m; HP29530, 90.01 m) were processed to test for absence. Magnetic extracts followed a modified protocol (Strehlau et al.; Wagner et al.). SEM sample prep involved depositing extracts on carbon-taped stubs, air drying, and 15 nm carbon coating; ~15 random SEM images per sample at ~69,000× were acquired to quantify morphology proportions and estimate magnetite volume. EDS spectra were acquired (Thermo-FEI Quattro SEM; dual Thermo EDS detectors, 15 kV). TEM grids were prepared for HP40825, HP40625, HP31435, HP31375; TEM imaging and SAED used JEOL F200 Cold FEG TEM and TECNAI TF30 STEM (200–300 keV). Gatan GMS 3 was used for particle measurements and SAED analysis; CrystalMaker/CrystalDiffract aided crystallographic identification. Dimensional analyses compared Cretaceous giant magnetofossils with published Paleocene–Eocene datasets.

• Oldest robust conventional and giant magnetofossils documented at ~97 Ma in Holland Park (Cenomanian–Turonian), bracketing OAE2.
• Rock magnetic results: Pre- and post-OAE2 intervals (124.4–118.0 m; 95.8–92.9 m) exhibit FORC features consistent with single-domain biogenic magnetite (central ridges, negative regions, coercivity peaks ~20–50 mT). OAE2 horizons lack clear biogenic signatures.
• Electron microscopy: All eight targeted horizons yielded conventional (cuboctahedra, bullets—small/medium/large, prismatic) and giant (giant bullets, spindles, needles, spearheads) magnetofossils, except HP31375 lacked giant bullets and spearheads. SAED and lattice spacings confirm magnetite in HP40825 and HP31375; EDS indicates iron oxide consistent with magnetite in all imaged morphologies. Two out-of-interval horizons (HP35110, HP29530) lacked magnetofossils.
• Three potentially new giant magnetofossil morphologies documented: seeds (~1350×900 nm; hexoctahedral, elongated), squash (~2570×1300 nm; oblate ellipsoids with blunt stalk), and spades (~430×320 nm; spearhead-like but smoother, lacking [220] segmentation planes).
• Dimensional analyses show Cretaceous giant magnetofossils fall within size ranges of Paleocene–Eocene counterparts; seed morphologies have a tight length–width relationship (R²≈0.89).
• Abundance changes across OAE2: total conventional magnetofossils decreased by ~80% post-OAE2 (from n=15,716 to n=3,140); each conventional category dropped by >60%. Total giant magnetofossils increased slightly (~3.5%; pre n=85, post n=88). Giant magnetofossils constitute <4.3% of total particles but contribute disproportionately to magnetite volume; post-OAE2 percent magnetite volume from giants rose markedly (reported as ~96.1–98.1%; Tables S7–S8).
• Conventional morphologies: cuboctahedra decreased by ~11.5% and small bullets increased by ~8% after OAE2; other categories relatively stable.
• Giant morphologies (relative abundances among giants): giant bullets −20.5%, spearheads −7.5%, spindles +6.7%, needles +7.5%, seeds +9.9%, squash +3.6%, spades +0.3% post-OAE2.
• Preservation: SEM indicates films on post-OAE2 extracts and sparser conventional magnetofossils; giants are well preserved in both intervals, implying greater diagenetic resistance, likely due to size.
• Environmental interpretation: Changes in conventional assemblages (notably increased bullets) are consistent with increased seasonal stratification and nutrient/organic matter supply; lithologic changes support altered depositional conditions.

The findings extend the robust record of both conventional and giant magnetofossils back to the mid-Cretaceous, demonstrating the resilience and long-term persistence of iron-biomineralizing organisms. Consistency in morphologies and size ranges with Paleocene–Eocene counterparts suggests similar ecological niches and biomineralization controls through time. Differences in abundance and preservation across OAE2 likely reflect significant changes in sediment and/or water chemistry. Two non-exclusive scenarios are proposed: (1) baseline environmental shift after OAE2 that stressed MTB, reducing conventional magnetofossil production and favoring bullet producers, while GIBO remained less affected; (2) enhanced diagenesis post-OAE2 preferentially diminished conventional magnetofossil preservation, while larger giants were more resistant, and surface films potentially impeded magnetic extraction. Regardless, giants appear better preserved when present. Coordinated changes in giant and conventional morphologies (e.g., decreased giant bullets with increased conventional bullets) imply that GIBO, like MTB, exhibit ecological preferences and may modulate particle morphologies with environmental conditions or depth habitats. The documentation of three additional giant morphologies (seeds, squash, spades) expands known morphological disparity; preliminary observations (e.g., hexoctahedral habit of seeds; lack of [220] segmentation planes in spades) suggest distinct biogenic pathways or functions from spearheads and other forms. Broader occurrence of giant magnetofossils beyond hyperthermal intervals, together with their strong contribution to magnetic mineral volume even at low counts, argues for a widespread but historically under-sampled distribution in the rock record and underscores the utility of integrated magnetic and EM approaches to detect them.

This study reports the oldest robust conventional and giant magnetofossil assemblages (~97 Ma) from Cenomanian–Turonian outer neritic sediments in the Holland Park core, bracketing OAE2. It confirms abundant, highly disparate magnetofossil assemblages with FORC signatures indicative of single-domain magnetite, documents three potentially new giant morphologies (seeds, squash, spades), and quantifies pronounced shifts in conventional versus giant abundances and preservation across OAE2. The results support the interpretation that both MTB and GIBO have distinct ecological preferences and that giants may be more diagenetically robust, enhancing their preservation potential. These observations expand the temporal and environmental range of giant magnetofossils, strengthening their value as proxies for redox and iron cycle dynamics. Future work should include: (1) comprehensive crystallographic, micromagnetic, and isotopic analyses to confirm biogenicity of seeds, squash, and spades; (2) targeted, high-resolution sampling and EM in sediments with biogenic single-domain signatures to improve detection; (3) cross-archive comparisons of robust giant assemblages to identify environmental drivers of morphological disparity; and (4) studies constraining diagenetic effects on conventional versus giant magnetofossil preservation.

• Lack of a modern analogue for GIBO limits direct biological interpretation of giant magnetofossil function and ecology.
• Potential diagenetic overprint, especially in post-OAE2 intervals, complicates separation of ecological signals from preservation effects; surface films may have inhibited magnetic extraction.
• Giant magnetofossils occur at low counts relative to conventional forms, increasing sampling uncertainty and susceptibility to under-sampling.
• Biogenicity of newly described morphologies (seeds, squash, spades) is not fully demonstrated; further crystallographic and isotopic evidence is needed.
• Stratigraphic gaps and variable nannofossil preservation (hiatus near 82 m; sporadic occurrences) add chronological uncertainty in parts of the section.