Probability-based preservational variations within the early Cambrian Chengjiang biota (China)

Abundance vs preservation data

Internal organ preservation (Fig. 1D) is not correlated to diversity (Fig. 1A) and abundances (Fig. 1B; see Spearman test for correlation in the Supplementary Material). For example, Euarthropods have the highest number of genera as well as the highest specimen abundance (Figs. 1A and 1B). However, they do not show the highest number of tissues per taxon (Fig. 1C) and they preserve one of the smallest proportions of internal organs (Fig. 1D). Chordates, on the other hand, are not the most diverse or abundant groups in Chengjiang (Figs. 1A and 1B). However, they preserve the highest proportions of internal organs (Fig. 1D). Thus, there is a certain group-selective preservation filter at place in Chengjiang, and this selective filter is not a matter of abundances as the most diverse and abundant taxa are not the best preserved in terms of internal organs (i.e., it is not simply a matter of luck aided by the large numbers of abundant and highly diverse taxa.). As a result, there must have been an interplay of both taphonomic and ecological conditions that dictated what and how animals got preserved in the Chengjiang biota.

Biostratinomy impact on fossil preservation

The internal organ preservation sequence between radiodonts (navy) and priapulids (pink) (Fig. 1D) reflects a transition from nektonic to endobenthic taxa. In the Walcott Quarry, it has been previously suggested that endobenthic taxa are easier to capture by sedimentary flows and are often buried alive, showing little evidence for degradation when compared to actively swimming taxa (Saleh et al., 2021a). Actively swimming taxa could escape to some extent transported sediments (Saleh et al., 2021a). In this sense, what is mainly preserved for these groups are carcasses that were already decaying on the seafloor and were passively transported by flows resulting in a lower degree of preservation than endobenthic taxa (Saleh et al., 2021a). Considering that fossils in the Chengjiang biota were also transported by sedimentary flows (Saleh et al., 2022), it is possible to suggest that groups containing endobenthic taxa (e.g., Brachiopoda; Priapulida) are the easiest to capture and are the best preserved (right side of the internal organ sequence; Fig. 1D). Radiodonta and Euarthropoda (left side of the internal organ sequence; Fig. 1D) contain many actively swimming genera that can escape the flow resulting in a low preservation fidelity (Fig. 1D). In this sense, the internal organ sequence “Radiodonta (nektonic/nektobenthic)—Euarthropoda (nektonic/nektobenthic/epibenthic)—Annelida (nektobenthic/epibenthic)—Lobopodia (epibenthic)—Brachiopoda (epibenthic/ endobenthic)—Priapulida (mainly endobenthic)” (Fig. 1D) represents a gradual preservational transition reflecting the interaction between sedimentary flows and the modes of life of organisms. This gradual transition could explain why no significant differences are observed when comparing consecutive animal groups one by one to each other, and significant differences are mainly observed between the extremes of this sequence (priapulids and radiodonts; Fig. 1D). Early in the Cambrian, the functional and ecological diversifications have not yet been unfolded, and each animal group was closely tied to a limited number of ecological roles (Bambach, Bush & Erwin, 2007). As a result, the phylogenetic signal in preservation variation is similar to that of the mode of life/ecological signal.

Hyolitha and Chordata are outliers to this sequence (Fig. 1D). For instance, hyoliths are epibenthic (maybe shallow endobenthic) but they are on the extreme left side of the internal organs sequence (Fig. 1D). Chordates are nektobenthic/nektonic but they plot on the extreme right side of the internal organs sequence (Fig. 1D). This pattern may therefore have been induced by favorable chemical conditions for chordate preservation at Chengjiang (Saleh et al., 2021a). These conditions can vary from those inhibiting the decay to those favoring the replication of chordate internal tissues in authigenic minerals or both (Parry et al., 2018; Purnell et al., 2018). On the contrary, it appears that hyoliths internal organs decayed faster than the rest, and were not preferentially stabilized through mineral precipitations at this site (Fig. 1D). This indicates that taphonomic processes acted differentially on the soft tissues across the major animal groups. This resulted in wildly variable preservation quality of functionally similar and homologically evolved bilaterian organs which nevertheless can be constructed of different materials depending on the animal group.

Some differences might exist between different outcrops at Chengjiang, likely reflecting the various depositional settings represented within this Lagerstätte (Saleh et al., 2022). However, considering that most flows transported organisms from their habitat to environments favorable for their preservation (Saleh et al., 2022), the impact of any of these flows will remain more pronounced on the endobenthic community in comparison to epibenthic, nektobenthic, and nektonic taxa (regardless of the precise sedimentary sub-environment investigated and the proportion of deposits associated with a particular flow in a given locality). The only exception within the sedimentary environment of the Chengjiang biota is hemipelagic deposits where no flows are recorded (Saleh et al., 2022). However, this setting is not accounted for herein, because it yielded a low diversity of fossils mainly consisting of sponges (Qi et al., 2018). Sponges are not included in this study, because they are not formed of the same types of tissues, and body plans, as eumetazoans. In this context, while specific outcrop data is unlikely to alter the main modes of life hypothesis of this paper, further evaluation must be done in the future. For example, chordates that are outliers to the mode of life hypothesis presented earlier differ in their preservation between outcrops (Yang et al., 2021). Outcrop comparisons will likely lead to the refinement of the results presented herein and resolve some of these subtle preservational variations.

Comparisons with decay experiments

Decay experiments on modern taxa are a powerful tool that provides an alternative view of the patterns observed in the fossil record (Briggs & Kear, 1993a; Briggs & Kear, 1993b; Briggs & Kear, 1994; Sansom, Gabbott & Purnell, 2010; Sansom, Gabbott & Purnell, 2011; Sansom, 2014; Sansom, 2016; Murdock et al., 2014; Butler et al., 2015; Klompmaker, Portell & Frick, 2017; Hancy & Antcliffe, 2020). Decay experiments have been performed on many experimental models, such as euarthropods (Briggs & Kear, 1994; Butler et al., 2015; Klompmaker, Portell & Frick, 2017), onychophorans (close to lobopodians; Murdock et al., 2014), chordates (Briggs & Kear, 1993a; Briggs & Kear, 1993b; Sansom, Gabbott & Purnell, 2010; Sansom, Gabbott & Purnell, 2011), priapulids (Sansom, 2016), and cnidarians (Hancy & Antcliffe, 2020), providing a baseline for most of the major groups found in the fossil record of the Chengjiang biota. Most decay experiments were performed under limited environmental conditions (e.g., in the absence of sediments). Some decay experiments were performed with the presence of a substrate and indicated that animals could eventually get preferentially preserved in the presence of clays, in particular kaolinite (Wilson & Butterfield, 2014; Naimark et al., 2016a; Naimark et al., 2016b). Despite the limitations of decay experiments they can still be used to confirm a certain finding or to highlight where a gap in understanding exists (Purnell et al., 2018). For instance, it was experimentally shown that nervous tissues in numerous groups decay particularly fast in laboratory conditions (Fig. 2). However, numerous brains were discovered in many taxa from the Chengjiang biota (e.g., Ma et al., 2012; Ma et al., 2015). It was subsequently demonstrated that neural tissues in Chengjiang animals were replicated in authigenic minerals very rapidly following the death of the organism allowing for their preservation in the rock record to the exclusion of other tissues (Saleh et al., 2020a). This taphonomic model shows that information from decay experiments and the pattern of fossil preservation are complementary evidence for which explanatory processes can be discovered (Saleh et al., 2020a). As such, by investigating whether the increase in the proportion of internal organ preservation (Fig. 1D) reflects an increase in decay resistance (Fig. 2), it becomes possible to test some of the hypotheses that emerged in the previous sections. It is shown in decay experiments that the degradation of chordate internal organs is slower than in other groups such as Priapulida, Onychophora, and Euarthropoda (Fig. 2). This has been clearly seen in the fossil record of the Chengjiang biota (Fig. 1D). However, in decay experiments, it is observed that Priapulida decay faster than Onychophora, and Euarthropoda respectively (Fig. 2). An opposite pattern has been observed in the Chengjiang biota for these animal groups with Euarthropoda being faster to decay in comparison to Priapulida (Fig. 1D). Thus, there must have been other parameters influencing the sequence of decay at Chengjiang. The mode of life hypothesis suggested in this study can explain the contrast between the results of decay experiments and the fossil record. This hypothesis suggests that endobenthic groups such as Priapulida cannot escape sedimentary flows and are often buried alive (facing a very short decay time). Moreover, if dead prior to their transport, endobenthic taxa (e.g., Cambrian priapulids) are automatically buried at deeper levels within the sediment, slowing down their decay (Pratt & Kimmig, 2019). Epibenthic, nektobenthic, and nektonic taxa are able to escape, to some extent, the flow and only exposed carcasses of these animals are preserved (very long decay time). The interaction between modes of life and preservation is not limited to the Chengjiang biota and other Burgess Shale-type deposits. For example, it was long assumed that the rarity of ten-armed Decabrachia (squid and cuttlefish) in comparison to the relatively abundant phosphatized eight-armed Vampyropoda (octopuses and vampire squid) among cephalopod molluscs in phosphatic Mesozoic Lagerstätten is ecological. However, it was recently demonstrated that it is largely caused by differences in their chemical composition and buoyancy mechanism during life (Clements et al., 2017). Subsequent studies showed that Decabrachia and Vampyropoda overlapped in their habitats meaning that differences cannot be attributable to ecology sensu stricto (Košák et al., 2021). The first soft-tissue squid was also found in non-phosphatic sites further corroborating the interaction between preservation and the chemistry of the physiology (Mironenko et al., 2021).

Even though the slow decay of chordate tissues and the interaction between modes of life and sedimentary flows explain the patterns observed in the Chengjiang biota, it is important to acknowledge that there might be other explanations for this pattern. For instance, changing the conditions in decay experiments can often lead to contrasting results. It has been recently shown that some cnidarian tissues decay faster under anoxic conditions than in the presence of oxygen (Hancy & Antcliffe, 2020). Euarthropod tissues decay substantially slower under antimicrobial reducing conditions than in the presence of oxygen (Butler et al., 2015). In the future, it is essential to decay animal groups under various environmental conditions in order to build numerous experimental decay sequences that can be then compared to the patterns observed in the fossil record.

Other factors leading to preservation variation

The contrasting patterns of different groups at Chengjiang may not be resulting from differential preservation but rather from the ability of paleontologists to recognize a particular fossil. Euarthropoda and Radiodonta remains are generally easily recognizable in the rock record (Guo et al., 2019; Wu et al., 2021). For instance, in sites with exceptional fossil preservation, many radiodonts are known from their molting products, with complete carcasses being extremely rare. In the Walcott Quarry, only one complete specimen of Anomalocaris is known, but tens of fossil fragments can be confidently attributed to this genus based on key morphological features (Nanglu, Caron & Gaines, 2020). In the Fezouata Shale, complete carcasses of the radiodont Aegirocassis are similarly relatively rare (Saleh et al., 2020b), yet many fragments were easily assigned to this genus (Van Roy, Briggs & Gaines, 2015). The easy recognition of euarthropod and radiodont fossil fragments to existing or new genera (Zhu, Lerosey-Aubril & Ortega-Hernández, 2021) may account for the lower degree of preservation observed within these groups because most fossil fragments will be deprived of internal organs, resulting in data entry without the most labile parts (Figs. 3A and 3B). This does not arise for other major groups of animals. The designation of a specific fossil to the Chordata, for example, requires the preservation of very specific and labile characters. To avoid any ambiguity and uncertainty, paleontologists tend to define new species/genera based solely on almost complete specimens of chordates. Those complete specimens are often preserved with internal parts and thus, logically show a higher degree of preservation (Fig. 3C). Ctenophores and cnidarians are other examples of this dilemma (Figs. 3D–3G). Paleontologists often rely on complete specimens in order to say if a particular fossil belongs to cnidarians or ctenophores and even then, it is contentious (Daley & Antcliffe, 2019). It is almost impossible to recognize any fossil fragment belonging to these groups if it has not been described previously. Even when nearly complete specimens are discovered, it will be still challenging to assign those carcasses to either cnidarians or ctenophores because defining and recognizing characters can be ambiguous (Daley & Antcliffe, 2019; Zhao et al., 2019).

The possibility of paleontologists unconsciously and unwillingly inducing variations in a certain dataset is not limited to the ability to identify a certain fossil fragment. Preservational variations can also be induced by fossil collection, preparation, accession, and digitization (Whitaker & Kimmig, 2020). Variations can be induced following certain practices such as sampling from a preferred locality, or the background, interest, and specialization of the sampling team (Whitaker & Kimmig, 2020). However, within the context of this paper, we consider that induced preservational variations remain minimal. Sampling for the fossils that are currently housed at Yunnan University was always made by a group of international researchers covering numerous aspects of paleontological research. Future investigations can reveal different patterns in fossil collections at other institutions. Quantifying the divergence between institutional collections will be key to quantifying collection biases and understanding the impact we and other colleagues might unconsciously introduce in databases.