Metamorphism obscures primary taphonomic pathways in the early Cambrian Sirius Passet Lagerstätte, North Greenland

Burgess Shale–type (BST) biotas provide critical insight into the function of Cambrian marine ecosystems and into the soft-part anatomy of diverse animal stem lineages (e.g., Daley and Edgecombe, 2014). Much progress has been made toward resolving the depositional controls on their occurrence (see Gaines, 2014) and the resulting biases in the view they provide. They are primarily preserved as carbonaceous cuticular compressions (Butterfield, 1990; Gaines et al., 2012), locally augmented by early diagenetic pyrite coatings (e.g., Gabbott et al., 2004). This commonality in preservation has promoted the view that BST biotas form a coherent taphonomic grouping, the consequence of a complex trade-off between decay, organic stabilization, and early diagenetic mineralization (Schiffbauer et al., 2014; Anderson et al., 2020; Saleh et al., 2021). However, little is known about the impact of deeper burial and very low- to low-grade metamorphism (i.e., anchizone to epizone) on the taphonomic outcome of these biotas. These processes have predictable consequences for the maturation of the carbonaceous fossils (Butterfield, 1990; Topper et al., 2018) and for the mineralogy and texture of their host sediments (e.g., Powell, 2003; Strang et al., 2016b; Lerosey-Aubril et al., 2018), but the extent to which they overprint primary taphonomic signals or introduce artifacts is unclear.

We combine geochemical and petrographic analyses from scanning electron microscopy and stable silicon isotopes (see the Supplemental Material¹) to resolve the impact of these processes on the early Cambrian (Series 2, Stage 3) Sirius Passet Lagerstätte (North Greenland), one of the oldest and least well understood BST biotas (Harper et al., 2019). It, like most other BST biotas, is dominated by arthropods but is distinguished by having experienced an unusually high grade of metamorphism, reaching a peak temperature of 409 ± 50 °C (lower greenschist facies) during the Devonian Ellesmerian orogeny (Soper and Higgins, 1987); by comparison, the Burgess Shale (British Columbia, Canada) reached a peak temperature of 335 ± 50 °C (Topper et al., 2018). The occurrence of silicified muscles in this biota (Fig. 1) has led to the hypothesis that an Ediacaran silicification window (Tarhan et al., 2016) continued into the Cambrian, resulting in a unique style of BST preservation (Strang et al., 2016b). However, we show instead that this is a consequence of the Lagerstätte's burial history, which has profoundly altered original fossil preservation. This new understanding provides both a context for interpreting this biota and for recognizing modified primary taphonomic signals in others.

The mineralized labile soft tissues of panarthropods are preserved in a diversity of silicate and non-silicate minerals (Figs. 1C and 2A), most shared with the host sediment. Cross-cutting relationships reveal the general paragenetic sequence: apatite > quartz > chlorite ± muscovite > chloritoid > xenotime.

Apatite [Ca₅(CO₃,PO₄)₃(OH,F)] is invariably the first phase, consistent with its early preservation of soft tissues in other biotas (Briggs et al., 1993). It is extensively replaced by subsequent phases, except in the guts (Fig. 2B) where it frequently remains important. Quartz [SiO₂] is the dominant phase (Fig. 2A) and formed in multiple generations, including after late-stage chloritoid (see below; Fig. 2G). However, it everywhere succeeds apatite, as evidenced by abundant relict (<1 µm) apatite inclusions (Fig. 2C). Silicification is focused at the margins of the fossils (Figs. 2B and 2D), where interlocking crystals of quartz (~5–20 µm) growing inwards from the sediment may obliterate fine morphological detail (Fig. 2E; Fig. S1A in the Supplemental Material). Elsewhere, the quartz is microcrystalline and faithfully replicates subcellular details originally captured by apatite, such as muscle myofibrils (Figs. 2C and 2G). Silicified muscle δ³⁰Si values range between −0.76‰ and −0.99‰ (Fig. 3; Table S3), and are notably closer to that of the matrix (−0.7‰) than to other potential contemporary early diagenetic and biogenic low-temperature sources (see Geilert et al., 2014), consistent with a metamorphic origin.

Chlorite [(Mg,Fe)₆AlSi₃O₁₀(OH)₈] and muscovite [KAl₂(AlSi₃O₁₀)(OH)₂] typically occur as similarly sized (<20 µm long) lath-shaped crystals; locally, they are intergrown, indicating cogenesis. Whereas chlorite is widespread (Fig. 1F) and may faithfully pseudomorph silicified soft tissues (Fig. 2D; Fig. S1B), muscovite is generally confined to discrete domains (Figs. 2A and 2D; Fig. S1C) and does not replicate ultrastructural details. Chloritoid [(Fe,Mg,Mn)₂Al₄ Si₂O₁₀(OH)₄] principally occurs in the sediment (Fig. 2A) along with pyrite [FeS₂] porphyroblasts (generally 50–90 μm), but it locally overprints silicified muscle and gut diverticula (Fig. S1C), particularly near the fossil margins or adjacent to sediment inclusions. The chloritoid forms large (as much as 320 µm long) lath-shaped euhedra that traverse multiple muscle fibers and destroy all original morphology (Fig. 2G). Their long axes exhibit a degree of preferred orientation (Fig. S1D), implying growth under stress.

Cuticles typically appear as thin kerogen films (carbon compressions), commonly accompanied by a patchy coating of (now oxidized) pyrite (Figs. 1B and 1G; Fig. S3). However, they may also be preserved by interlocking muscovite and quartz, indistinguishable from the sediment (Fig. 2H).

Accessory phosphate minerals occur in close association with the fossils and overprint the mineralized soft tissues, largely destroying their fine morphology. Xenotime [YPO₄] forms rosettes within the digestive tract, may crudely preserve muscle tissue (Fig. S2D), and locally grows outwards (Figs. S2A–S2C) and inwards from the walls of fossils (Fig. 2H; Figs. S2B and S2C). Similarly, subhedral to euhedral apatite (as large as ~20 µm) (Fig. S2G) occurs both inside and outside the fossils, in the latter case as a diffuse corona extending ~400 µm away (Figs. S2E and S2F). By contrast, monazite [(REE)PO₄, where REE indicates rare earth elements] is generally confined to the adjacent sediment (Fig. 2F), where it may be concentrated on only one side of fossils (Figs. S2E and S2F), suggesting growth linked to fluid movement (cf. Evans et al., 2002). In all cases, it is poikiloblastic and comparatively coarse (30–380 µm), locally replacing accessory apatite (Fig. S2H) and at least partially overlapping syntectonic chloritoid formation (Fig. S2I) (see Wilby et al., 2007). Lastly, an amorphous P, Ca, Fe, Al-rich phase, interpreted to be a weathering product of apatite and pyrite, fills interstices in the fossils (Fig. S2J).

The bulk rock is enriched in aluminum and depleted in calcium compared to the Burgess Shale, but its major element composition otherwise overlaps (for details, see the Supplemental Material; Table S4, Fig. S4). Powell (2003) considered the Burgess Shale to be unremarkable for a pelite, and though the mineralogy of the Sirius Passet protolith cannot be ascertained with certainty (especially the starting clay composition), it likely passed through a typical prograde sequence of mineral reactions for an aluminum-rich pelite (e.g., see Bucher and Grapes, 2011), leading to the observed succession of soft-tissue replacements (Fig. 4).

Silicified soft-bodied biotas are scarce in the Phanerozoic; they rarely preserve labile soft tissues (e.g., muscle) and are confined to exotic non-marine settings (e.g., Trewin et al., 2003). By contrast, early diagenetic silicification was active in diverse marine environments in the Ediacaran and has been implicated in the preservation of several of its soft-bodied biotas (Muscente et al., 2015; Tarhan et al., 2016). This dichotomy has led to the suggestion that a silicification taphonomic window may have persisted into the Cambrian and been responsible for silicifying soft tissues in the Sirius Passet Lagerstätte (Strang et al., 2016a, 2016b; Topper et al., 2018). Our petrographic and isotopic data refute this idea and reveal instead that silicification was a product of very low- to low-grade metamorphism, consistent with its late-stage formation in other BST Lagerstätten (e.g., Powell, 2003; Lerosey-Aubril et al., 2018). Evidence for multiple episodes of silicification in the Sirius Passet Lagerstätte (e.g., before and after chloritoid formation) is consistent with the release of silica from the host sediment during successive clay mineral transformations (van de Kamp, 2008; Fig. 4).

Textural relationships indicate that preserved muscle and digestive systems were initially stabilized by early diagenetic apatite (cf. Wilby et al., 1996; Butterfield, 2002) prior to being progressively replaced by other minerals. Overall, the mineralogy, isotopic composition, and texture of the soft tissues have converged on those of the host sediment, and the primary preservation mode of the soft tissues has been irrevocably altered. Likewise, the originally phosphatized guts have been replaced by quartz and clays, supporting an alternate explanation for apparently sediment-filled guts (Butterfield, 2002), which have been used elsewhere to invoke deposit-feeding habits in certain arthropods (cf. Hou and Bergström, 1997; Lerosey-Aubril et al., 2012). Each new replacing mineral retains a particular level of detail and is focused in discrete anatomical regions (Figs. 1C and 2A), imposing significant biases on the ultimate survival and fidelity of preservation of different tissues. Muscovite and chloritoid growth were especially deleterious, suggesting that aluminum-rich protoliths, such as the Burgess Shale (Powell, 2003), have the least potential for conserving labile soft tissues during very low- to low-grade metamorphism. By contrast, pyrite is unaffected, meaning that this taphonomic pathway at least is reliably conserved (now weathered).

Phosphorus released during successive mineral transformations was redistributed into new accessory phosphates. Accessory monazite is widely reported in other BST biotas (Conway Morris, 1990; Moore and Lieberman, 2009; Broce and Schiffbauer, 2017), implying that phosphatized soft tissues may have been lost from these too and have been a more important component of BST preservation than their present distribution suggests (Daley and Edgecombe, 2014; Paterson et al., 2015). The sum of evidence, including other examples of late-stage overprint (e.g., Gaines et al. 2019), demonstrates the potential for deep burial and metamorphism to modify primary taphonomic signals and biological information, and emphasizes the need for site-specific characterization of the BST preservation.

Silicified soft tissues in the Sirius Passet Lagerstätte are a product of progressive alteration of originally phosphatized soft tissues during very low- to low-grade metamorphism. Contrary to previous assertions, they do not record a novel mode of primary BST preservation, but rather an extreme example of a spectrum of late-stage processes that operated in other BST Lagerstätten. The mineralogy and chemistry of Sirius Passet fossils have converged on those of the host sediment, preferentially destroying certain tissues and inducing a reduction in resolution of others. Growth of muscovite and chloritoid was especially destructive, suggesting that protolith composition has a bearing on the ultimate fate of soft tissues during metamorphism. Accessory monazite formed in response to the replacement of the originally phosphatized soft tissues and may serve as a useful proxy for the former presence of such tissues elsewhere.

We thank Stuart Kearns (University of Bristol, UK) and Sebastian Næsby Malkki (Geological Survey of Denmark and Greenland) for assistance with electron microscopy, and three anonymous reviewers for helpful suggestions. M.L. Nielsen was supported by a UK Natural Environment Research Council GW4+ Ph.D. studentship (grant NE/L002434/1) with CASE-partner British Geological Survey and a Korea Polar Research Institute (KOPRI) Arctic Science Fellowship Program 2018 grant. J. Rushton and P. Wilby publish with the permission of the Executive Director, British Geological Survey. K. Hendry and H. Ng are supported by a European Research Council grant ICY-LAB (678371). This work was funded by a Korea Polar Research Institute (KOPRI) project “Advancement into unexplored areas of North Greenland through paleoenvironment and animal evolution research (PE21060)”.