* Franziska Reinhard is a Postdoctoral Researcher at the Humboldt Centre for Philosophy and the Humanities at Ulm University. She recently completed a PhD in philosophy at the University of Vienna, with a thesis titled “Knowledge-Making in Origins of Life Research,” which combined aspects of general philosophy of science, philosophy of biology, and philosophy of the historical sciences. She writes…
In July of last year, NASA’s Perseverance Rover discovered something intriguing on Mars. While exploring an ancient Martian riverbed called Cheyava Falls, it spotted some bright patches of rock surrounded by dark ring-like patterns – quickly dubbed “Leopard Spots” – that might, just might, be signs of ancient life. On Earth, such patterns can form when chemical reactions involving hematite bleach red rocks to paler tones. And these reactions not only release iron and phosphate, the source of the dark rings, but also tend to be indicative of microbial metabolism. Adding to the intrigue, Perseverance’s SHERLOC instruments (short for ‘Scanning Habitable Environments with Raman and Luminescence for Organics and Chemicals’) identified some organic molecules within the Leopard spots (Hurowitz et al. 2025).
“Leopard spots” from the Cheyava Falls region of Mars
This is exciting. But it remains an open question whether any of this actually signals the presence of past Martian life. This uncertainty isn’t because we haven’t returned any samples to Earth yet (a plan currently facing logistical and financial difficulties). Even on Earth, identifying early evidence of life is notoriously difficult. Putative traces of early life are usually scarce, and are frequently ambiguous or misleading. Not everything that looks biological is the result of past life activity.
So, how do we distinguish genuine traces of life from geological imposters? This post explores that question through one iconic example: stromatolites. The post is an adapted version of a forthcoming paper of mine, “Direct and Circumstantial Traces,” in which I suggest that insights from criminal investigations– concerning how evidence is classified– can help conceptualize how historical scientists navigate particularly ambiguous cases.
Traces of life and how to find them
When it comes to the earliest evidence for life on Earth, there are two main types of traces: chemical and morphological. Chemical traces are molecules probably left behind by biological activity, particularly complex molecules such as lipids or pigments (“molecular fossils”), but also the isotopic composition of elements like carbon.* Morphological traces are things like sedimentary structures or putative microfossils, which are taken to hint at life due to their distinctive shapes and organization. Both kinds of traces have been found on Earth, suggesting that the record of life might stretch back over 3.7 billion years. But both can also be faked, so to speak, by abiotic processes. And genuine life traces can easily be transformed or contaminated over time. Adding to the complexity, Earth’s rock record only starts at about four billion years ago; but Earth may have been habitable as early as 4.3 billion years ago. So, we’re missing the earliest chapters (Javaux 2019).
[* Life on Earth favors the lighter carbon isotope, ¹²C, so a strong local presence of ¹²C can suggest biological processing.]
Due to these difficulties, some have argued that when dealing with putative traces of early life, we should start from the ‘null hypothesis’ that any given trace isn't from life. So, how do you reject the null? A big part of the answer lies in the context of the putative life trace. Finding traces of life in ancient lava flows would be suspect, but finding one in sediments from an ancient river or sea makes more sense. Scientists look for traces preserved in geologically plausible environments, using a range of indicators that tell us about past conditions, like signs of water, certain minerals, or chemical patterns (Cockell 2015).
Identifying traces of life is just as much about where and how it's found as about what is found. Let’s look at two cases that illustrate this.
Stromatolites from Australia to Greenland
Stromatolites are layered rock structures that formed through the activity of organisms. The term refers to the rock itself— not the organisms that created it– but it generally denotes sedimentary structures shaped by biological processes, particularly by microbes trapping and binding sediments. The resulting formations have a distinct, banded appearance, a bit like puff pastry. However, not everything that looks like a stromatolite is necessarily biogenic (Awramik and Grey 2005). Stromatolite-like structures can also form through purely abiotic processes, obscuring their interpretation. Schopf (2006) suggests that “it is clearly difficult, and is perhaps impossible, to prove beyond question that the vast majority of reported stromatolites… are assuredly biogenic.” There are even experimental studies demonstrating how similar structures can arise without life (Criado-Reyes et al. 2023; McLoughlin, Wilson, and Brasier 2008).
Genuine stromatolites result from the activity of microbial communities, predominantly cyanobacteria, in shallow aquatic environments. As the cyanobacteria communities grow and photosynthesize, they trap sediments and minerals, creating thin layers that build up on top of each other over time. These microbes played a major role in Earth’s early history, gradually oxygenating the atmosphere and setting the stage for more complex forms of life. Stromatolite-forming microbes are still around today, but they are limited to a few locations – the most prominent ones being the stromatolites in Shark Bay, Western Australia.
The stromatolites of Shark Bay
Western Australia is also home to one of the most compelling candidates for Earth’s earliest life traces. The stromatolites of the Strelley Pool Formation (formerly the Strelley Pool Chert) in the Pilbara region of Western Australia are estimated to be around 3.4 billion years old. Thanks to its exceptionally old rock record, the Pilbara region is something of a hotspot for early life research. It is also home to the famous but now contested Apex Chert microfossils (Brasier et al. 2015; Wacey et al. 2018)– but that’s a story for another day.
The Strelley Pool Formation stromatolites first became the subject of intense research and debate in the 1980s. Upon first analyzing them, geologist Donald Lowe (1980) interpreted them as microbial in origin. However, he later reversed his opinion on the matter, suggesting instead that hydrothermal processes might have formed the structures without life being involved (Lowe 1994). At the time, this abiotic explanation fit what was known about the past environmental context of the Strelley Pool Formation.
The landscape that houses the Strelley Pool Formation, in the Pilbara region of western Australia
But the picture shifted again when further stromatolites in the Pilbara region were discovered and analyzed, revealing more diverse structures than in those studied by Lowe (Hofmann 2000). Importantly, new evidence constraining the past environmental context was discovered as well. Van Kranendonk and colleagues (2003) analyzed the trace-element geochemistry in the surrounding rocks, particularly the distribution of rare-earth elements (REEs). REE distribution patterns can serve as chemical fingerprints of sorts, revealing past environmental conditions. Modern shallow seawater environments– including those where presently living stromatolite-forming microbes still thrive– contain a particular REE distribution, and a similar distribution pattern was found in the sediments tested around the Strelley Pool stromatolites.
A Strelley Pool stromatolite (~12” long), showing characteristic fine layering
In addition, a large-scale study conducted by Allwood and colleagues (2006; 2007; 2009) sampled stromatolite-bearing rocks in the Pilbara region along a 10km line. They identified seven different stromatolite types with distinct morphology. Their analysis showed that stromatolites were most diverse and abundant in locations that used to have a reef structure, forming isolated platforms on the shoreline. This marine environment was evidenced by, among other things, the location’s contemporary geochemical signature. By contrast, stromatolites were not found, or were scarce or underdeveloped, in rocks associated with hydrothermal or volcanic activity, or in locations deeper in the ocean. This environmental setting, again, closely resembles where modern stromatolites are found – calm, shallow, often tidal zones with plenty of sunlight.
Here is how Allwood et al. (2007) summarize their approach themselves:
In the present study, we further demonstrate that the criterion of stromatolite morphology, if studied in spatial and temporal context in the palaeoenvironment, is a valid and powerful criterion for the analysis of stromatolite genesis. That is, if palaeoenvironmental aspects such as spatial relations, conditions and processes are woven into the analysis, then a range of stromatolite features including morphology take on greater validity and importance as criteria to use in the assessment of stromatolite genesis.
Another place that has garnered scientists’ attention in the search for Earth’s earliest life traces is Greenland. Greenland, specifically the Isua Greenstone Belt in its southern part, is home to some of the oldest rocks on the planet, estimated to be around 3.7 to 3.8 billion years old. In (1996), Mojzsis et al. reported an enrichment of the lighter carbon isotope ¹²C in rocks from the Isua Greenstone Belt, raising the possibility of extremely early microbial activity. This finding on its own, however, wasn’t considered definitive.
Scientists collecting from the Isua Greenstone Belt in Greenland
Generally, it was thought unlikely that structures like stromatolites or microfossils could be preserved in the Isua Greenstone Belt (Allwood 2016). That’s because most rocks in the area are metamorphic, having gone through several transformative episodes, likely obscuring and distorting any information about past life that there may have been.
However, that seemingly changed in 2016 when Nutman et al. reported the discovery of stromatolite-like structures in newly exposed outcrops, previously covered by snow (which had melted in response to climate change). If correct, this would push back the stromatolite record by over 200 million years. The structures were small (1-4 cm tall), roughly layered, and located close to dolomitic sediment (a form of carbonate), which can be an indicator of an early marine environment. The ‘classical’ stromatolite-shape was not as clear as it was in Western Australia– some of the characteristic shapes are not pointing in the direction one would expect a stromatolite to grow, namely upward toward the sunlight. However, Nutman et al. highlight that this could be explained by the subsequent deformations of the rocks in the Isua Greenstone Belt.
An alleged stromatolite from Greenland
However, Nutman et al.’s interpretation has been met with skepticism. Critics, including Allwood et al. (2018) and Zawaski et al. (2020), argue that the Isua Greenstone Belt structures are more plausibly explained by ‘mere’ geological deformation. The structures themselves lack key features seen in confirmed stromatolites, such as consistent upward growth and lamination. And while it is a possibility that these features were lost due to metamorphism, an abiotic explanation seems more likely. Allwood et al. (2018) agree that the stromatolite-like structures are found in what used to be a marine environment, but point out that there is no evidence that they formed in the shallow water environment modern stromatolite-forming microbes prefer.
One reason the stromatolite interpretation seemed plausible was the presence of carbonate in and around the Isua structures, but this is also one of the points of contention. Carbonate can signal the past presence of life, but can also be associated with non-biological processes. Critics of Nutman et al.’s interpretation have pointed out that there are actually carbonate bands all around the area where the putative stromatolites occur, but without the presence of stromatolite-like structures. And they are near deposits of magnesium silicate (‘talc’ in layman’s terms), which, when put under pressure or heat, releases carbon dioxide that can form carbonates – providing a plausible non-biological explanation for carbonate presence in the Isua Greenstone Belt (Kaufman 2019)
The stromatolite investigations, both in Australia and in Greenland, rely on a variety of traces. All these play a role in reconstructing the past, but they don’t all do so in the same way. Some traces seem to speak more directly to the presence of early life, others more indirectly. That is, not all traces relate to the research target (early microbial life) in the same manner. To make sense of this difference, I suggest we turn to a distinction familiar from legal contexts: the difference between direct and circumstantial evidence.
Direct and Circumstantial Evidence … and Traces
If you’ve ever watched a courtroom drama, you may have heard a lawyer say, “Objection! Circumstantial.” And indeed, in the courtroom, circumstantial evidence is regarded as less potent than direct evidence, its more conclusive counterpart. Still, this doesn’t mean it’s irrelevant. In fact, circumstantial evidence plays a crucial role in building a case, especially when direct evidence is lacking. Direct evidence in court might include a defendant’s confession or an eyewitness directly observing a crime. Circumstantial evidence, by contrast, supports an inference to a defendant’s guilt through an intermediate step, for example, by showing that the defendant had the opportunity or motive to commit the crime. (This cuts the other way, too. If the glove doesn’t fit…)
In my forthcoming paper, I suggest that we can apply a similar way of thinking to some historical reconstructions, and, in particular, to how we understand traces. Now, if you’ve been following the 'philosophy of the historical sciences' literature, maybe you're thinking, "Haven't we said enough about that? Surely we know what traces are?" And you would not be wrong. Broadly, there are three main views on traces: the causal view sees a trace as an effect of a past cause (Cleland 2002). The informational view sees it as something that carries information about a past event (Turner 2007). And the evidentiary holds that, for a piece of material evidence to count as a trace, (1) that piece of evidence must be directly causally downstream from a past event and (2) there needs to be sufficient background knowledge connecting trace and event, such that the former can serve as evidence for the latter (Currie 2018).
And as a reader of Extinct, you’re also likely aware that studying the past involves much more than just uncovering traces. Still, I want to suggest that thinking just a bit more about different kinds of traces and how these interact might help clarify even more of how scientists investigate the deep past.
The interesting thing about direct and circumstantial evidence is also how they connect in supporting a legal hypothesis. One way to illustrate this is through Bayesian networks (which I’ll here use in a purely qualitative way) that capture the structure of legal arguments (Lagnado 2011; Fenton, Neil, and Lagnado 2013). Look at the following network, for instance:
Let’s say we’re dealing with a legal investigation of whether a defendant robbed a shop (G). Two witness testimonies have been obtained. One witness (W1) claims to have seen the robbery take place. This is direct evidence; it links straight to the event in question. (In Bayesian terms, W1 probabilistically depends on G. The testimony W1 is more likely to obtain if G is the case: hence the arrow between G and W1.) The second witness (W2) only saw the defendant at the scene beforehand. This is circumstantial evidence: it supports the hypothesis of guilt (G) indirectly, by establishing that the defendant had the opportunity (O) to commit the crime. Now, O obtaining is probabilistically relevant to G. The defendant being at the shop makes it more likely (though not certain) that they committed the robbery. Also, neither O nor G is observable as such; they are hypotheses inferred through evidence. (O, in particular, is an auxiliary hypothesis that needs to be established for the circumstantial evidence, W2, to become relevant to G – that is the extra inferential step required when legal evidence is circumstantial rather than direct.)
Now, how does that help in characterizing traces? I want to suggest that, in analogy with the legal case, we can distinguish between direct and circumstantial traces as well. Direct traces are those that connect to a past event under investigation through a direct causal chain. That is what the philosophical discussion of traces has so far been about. In the stromatolite case, that might mean the discernible, layered structures. If those are traces of early microbial activity, they are direct traces of it. There is a direct causal chain from the past event to the present-day phenomenon studied.
Circumstantial traces work differently. Let’s look at a possible network for the stromatolite case for illustration:
In this network, S (the layered rock structures) is a direct trace of L (early microbe-activity). But C (a particular geochemical signature, like the distribution of REEs) is not directly linked to L. Instead, it indicates M (an ancient shallow marine environment), which in turn makes L more likely to have been the case. In my terminology, C is a circumstantial trace of L– it supports L (our research target) via an intermediate inferential step.
What kind of things qualify as circumstantial traces? I think two conditions must be met:
Causal: the trace shares a common cause with the target event in the past.
Epistemic: the trace supports an inference to the target event via an intermediate inferential step.
Both conditions together avoid letting in too much: not every contemporary remnant linked by a distant, and perhaps unknown, common cause is useful. It must also be made epistemically relevant to the target under investigation.
The distinction between direct and circumstantial traces offers a way to conceptualize how different types of traces function in practice, or so I suggest. Circumstantial traces can support or challenge hypotheses about the past. In the Strelley Pool Formation case, they supported the view that ancient microbes cause the formation of the observed stromatolites. Alongside analyzing the rock structures themselves, researchers studied the site’s paleoenvironment, including its geochemical signature. These are direct traces of an ancient marine setting (M), but function as circumstantial traces of microbial activity. In this case, both direct traces (S) and circumstantial traces (C) aligned – they pointed to the same conclusion. If direct and circumstantial traces align, they are both positively relevant for the historical research target in question, even though only one of them is directly causally downstream from the research target.
But such alignment isn’t automatic; it must be carefully established in the course of an investigation. In the above network, connecting C to M, and M to L, requires substantial background knowledge. For the Strelley Pool Formation investigation, considering the environmental setting of contemporary stromatolite-forming microbes was important to foster these links, for instance.
In the case of the putative Greenland stromatolites, this kind of alignment is much less clear. The team of researchers supporting the stromatolite-interpretation cites geological evidence obtained in the surrounding area as evidence suggesting an ancient marine environment – so they invoke circumstantial traces in my terminology. But critics question whether these traces do evidence an environment supporting early microbial life, arguing that the field evidence can be explained without invoking biology. In this case, circumstantial traces don't clearly support the presence of early microbes. They may even undercut it. To invoke the comparison to the legal case again, imagine an investigation in which the prosecution fails to establish opportunity. Perhaps the defense produced a witness willing to testify that they saw the defendant elsewhere. In a related way, stromatolite-forming microorganisms are less likely to have been active in past if they did not have the opportunity either.
Conclusion
You may have noticed that I started this post talking about possible life traces on Mars, only to spend most of the text discussing cases from Earth. While that ‘detour’ was mostly self-serving (my forthcoming paper is about stromatolites), it was not completely without intention. In fact, what I have proposed to call circumstantial traces may be even more important beyond Earth, where we lack familiar analogues in the form of known life forms and access to relevant sites is severely restricted and indirect. That makes it even harder to identify direct traces with confidence. Lessons from Earth, drawn from cases like the stromatolite investigations, may give examples of what the epistemic contribution of different types of traces is, and the conditions required from them to support hypotheses about the presence of life. How exactly these lessons apply to Mars – especially the putative life traces at Cheyava Falls – remains to be seen. After all, that investigation is really just beginning.
References
Allwood, A. C. 2016. Evidence of Life in Earth’s Oldest Rocks. Nature 537 (7621): 500–501. https://doi.org/10.1038/nature19429.
Allwood, A. C., J. P. Grotzinger, A. H. Knoll, I. W. Burch, M. S. Anderson, M. L. Coleman, and I. Kanik. 2009. Controls on Development and Diversity of Early Archean Stromatolites. Proceedings of the National Academy of Sciences of the United States of America 106 (24): 9548–55. https://doi.org/10.1073/pnas.0903323106.
Allwood, A. C., M. T. Rosing, D. T. Flannery, J. A. Hurowitz, and C. M. Heirwegh. 2018. Reassessing Evidence of Life in 3,700-Million-Year-Old Rocks of Greenland. Nature 563 (7730): 241–44. https://doi.org/10.1038/S41586-018-0610-4.
Allwood, A. C., M. R. Walter, I. W. Burch, and B. S. Kamber. 2007. 3.43 Billion-Year-Old Stromatolite Reef from the Pilbara Craton of Western Australia: Ecosystem-Scale Insights to Early Life on Earth. Precambrian Research 158 (3–4): 198–227. https://doi.org/10.1016/j.precamres.2007.04.013.
Allwood, A. C., M. R. Walter, Balz S. Kamber, C. P. Marshall, and I. W. Burch. 2006. Stromatolite Reef from the Early Archaean Era of Australia. Nature 2006 441:7094 441 (7094): 714–18. https://doi.org/10.1038/nature04764.
Awramik, S. M., and K. Grey. 2005. Stromatolites: Biogenicity, Biosignatures, and Bioconfusion. Proceedings of SPIE - The International Society for Optical Engineering 5906 (September):1–9. https://doi.org/10.1117/12.625556.
Brasier, M., J. Antcliffe, M. Saunders, and D. Wacey. 2015. Changing the Picture of Earth’s Earliest Fossils (3.5-1.9 Ga) with New Approaches and New Discoveries. Proceedings of the National Academy of Sciences of the United States of America 112 (16): 4859–64. https://doi.org/10.1073/pnas.1405338111.
Cleland, C. E. 2002. Methodological and Epistemic Differences between Historical Science and Experimental Science. Philosophy of Science 69 (3): 447–51. https://doi.org/10.1086/342455.
Cockell, C. S. 2015. Astrobiology: Understanding Life in the Universe. Wiley Blackwell.
Criado-Reyes, J., F. Otálora, À. Canals, Cristóbal Verdugo-Escamilla, and J. Manuel García-Ruiz. 2023. Mechanisms Shaping the Gypsum Stromatolite-like Structures in the Salar de Llamara (Atacama Desert, Chile). Scientific Reports 13 (678): 1–10. https://doi.org/10.1038/s41598-023-27666-5.
Currie, A. M. 2018. Rock, Bone, and Ruin: An Optimist’s Guide to the Historical Sciences. MIT Press.
Fenton, N., M. Neil, and D. A. Lagnado. 2013. A General Structure for Legal Arguments About Evidence Using Bayesian Networks. Cognitive Science 37 (1): 61–102. https://doi.org/10.1111/cogs.12004.
Hofmann, H. J. 2000. Archean Stromatolites as Microbial Archives. In Microbial Sediments, edited by R.E. Riding and S.M. Awranik, 315–27. Springer. https://doi.org/10.1007/978-3-662-04036-2_34.
Hurowitz, Joel A., Michael M. Tice, Abigail C. Allwood, Morgan L. Cable, Kevin P Hand, A. E. Murphy, K. Uckert, et al. 2025. The detection of a potential biosignature by the Perseverance Rover on Mars. In 56th Lunar and Planetary Science Conference, 2581. The Woodlands (Texas), United States: Lunar and Planetary Institute. https://hal.science/hal-04930754.
Javaux, E J. 2019. Challenges in Evidencing the Earliest Traces of Life. Nature 572 (7770): 451–60. https://doi.org/10.1038/s41586-019-1436-4.
Kaufman, M. 2019. On the Ground in Greenland: Astrobiology Program-Supported Researchers Travel to a Site of Disputed Evidence of Ancient Stromatolites. Astrobiology at NASA. 2019. https://astrobiology.nasa.gov/news/on-the-ground-in-greenland/.
Kranendonk, M. J. van, G. E. Webb, and B. S. Kamber. 2003. Geological and Trace Element Evidence for a Marine Sedimentary Environment of Deposition and Biogenicity of 3.45 Ga Stromatolitic Carbonates in the Pilbara Craton, and Support for a Reducing Archaean Ocean. Geobiology 1 (2): 91–108. https://doi.org/10.1046/j.1472-4669.2003.00014.x.
Lagnado, D. A. 2011. Thinking about Evidence. In Evidence, Inference and Enquiry, edited by Philip Dawid, William Twining, and Mimi Vasilaki, 183–223. Oxford University Press. https://doi.org/10.5871/BACAD/9780197264843.003.0007.
Lowe, D. R. 1980. Stromatolites 3,400-Myr Old from the Archean of Western Australia. Nature 284 (5755): 441–43. https://doi.org/10.1038/284441a0.
Lowe, D. R. 1994. Abiological Origin of Described Stromatolites Older than 3.2 Ga. Geology 22 (5): 387–90. https://doi.org/10.1130/0091-7613(1994)022<0387:aoodso>2.3.co;2.
McLoughlin, N., L. A. Wilson, and M. D. Brasier. 2008. Growth of Synthetic Stromatolites and Wrinkle Structures in the Absence of Microbes – Implications for the Early Fossil Record. Geobiology 6 (2): 95–105. https://doi.org/10.1111/J.1472-4669.2007.00141.X.
Mojzsis, S. J., G. Arrhenius, K.D. McKeegan, T.M. Harrison, A.P. Nutman, and C.R.L. Friend. 1996. Evidence for Life on Earth before 3,800 Million Years Ago. Nature 384.
Nutman, A. P., V. C. Bennett, R. L. Friend, M. J. van Kranendonk, and A. R. Chivas. 2016. Rapid Emergence of Life Shown by Discovery of 3,700-Million-Year-Old Microbial Structures. Nature 537 (7621): 535–38. https://doi.org/10.1038/nature19355.
Schopf, J. W. 2006. Fossil Evidence of Archaean Life. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 361 (1470): 869–85. https://doi.org/10.1098/RSTB.2006.1834.
Turner, D. 2007. Making Prehistory: Historical Science and the Scientific Realism Debate. Cambridge University Press.
Wacey, D., N. Noffke, M. Saunders, P. Guagliardo, and D. M. Pyle. 2018. Volcanogenic Pseudo-Fossils from the ∼3.48 Ga Dresser Formation, Pilbara, Western Australia. Astrobiology 18 (5): 539. https://doi.org/10.1089/AST.2017.1734.
Zawaski, M. J., N. M. Kelly, O. Felipe Orlandini, C. I. O. Nichols, A. C. Allwood, and S. J. Mojzsis. 2020. Reappraisal of Purported ca. 3.7 Ga Stromatolites from the Isua Supracrustal Belt (West Greenland) from Detailed Chemical and Structural Analysis. Earth and Planetary Science Letters 545 (September):116409. https://doi.org/10.1016/J.EPSL.2020.116409.