Possible life found in sediments between 3.8 and 4.3 billion years old

March 3, 2017 • 11:15 am

The Earth is about 4.54 billion years old, and the oldest undisputed life on our planet appears as bacterial “microfossils” 3.5 billion years ago. But because bacteria are already quite complicated organisms, it’s a good bet that life (however you define it), began well before that. But how long? The seas weren’t around much before about 4 billion years (the Earth was too hot), and there was no oxygen. Life, if it existed about then, was probably adapted to extreme temperatures and was anaerobic (not requiring oxygen).

A new paper in Nature by Matthew Dodd et al. (free access, reference below) has reported what may be traces of life (iron-containing filaments and tiny tubes) that resemble the kind of life found in modern hydrothermal vents, as well as in undisputed microfossils. The age of the sediments, which are from Hudson Bay in Quebec, Canada, spans a range between 3.77 and 4.28 billion years. They can’t narrow it down much more than this large range, and of course the press is concentrating on the 4.28-billion-year date, because that’s about the earliest life could have formed given the state of the Earth then.

I won’t go into detail about the paper: it’s a hard slog even for an evolutionary biologist, for it’s largely geology and paleobiology. But one expert I asked said that the results are very interesting but not conclusive, and that the age range of course is quite large. Here are a few photographs of what may be the remnants of ancient bacteria. First are the filaments (click to enlarge pictures):

(From paper): a, Filaments from the NSB attached to a terminal knob (arrow) coated with nanoscopic haematite. b, Filaments from the Løkken jaspers coated with nanoscopic haematite and attached to terminal knobs (red arrows) and branching (orange arrows). Inset, multiple filaments attached to a terminal knob. c, Filaments from the NSB in quartz band with haematite rosettes (green arrow). Inset, branching filament (orange arrow). Green box defines d. d, Filament from the NSB enveloped in haematite (inset, same image in cross polars).

And the tubes:

(From the paper): a–f, Tubes from the NSB. a, Tubes associated with iron oxide band. b, Depth reconstruction of tubes with haematite filament (arrow). Inset, image of tubes at the surface. c, Tube showing a twisted filament (red arrow) and walls (black arrow). d, Strongly deformed tubes. e, Depth reconstruction of tubes. f, Two tubes attached to terminal knob (arrows); lower image taken in false colour. g, h, Tubes from the Løkken jaspers. g, Tube showing filament (red arrow) and walls (black arrow). h, Aligned tubes (green arrows).

These of course are not microfossils themselves, which are the fossilized remains of ancient bacteria, but simply traces of what may be ancient bacteria.

Carl Zimmer’s article about the find in Tuesday’s New York Times also shows that some experts are dubious. Several seem to think the find represents real organisms, while others think they’re artifacts. Here’s a bit of Zimmer’s piece:

But many experts in the field were skeptical of the new study — or downright unconvinced.

Martin J. Van Kranendonk, a geologist at the University of New South Wales, called the patterns in the rocks “dubiofossils” — fossil-like structures, perhaps, but without clear proof that they started out as something alive.

. . . And if these are fossils 4.2 billion years old, then scientists will have evidence that life began quickly on Earth, not long after the oceans formed.

Yet Frances Westall, the director of research at the CNRS-Centre de Biophysique Moléculaire in Orléans, France, isn’t convinced these are fossils at all. “I am frankly dubious,” she said.

For one thing, she has argued, the filaments in the Nuvvuagittuq rocks are too big. She and her colleagues have found filaments formed by bacteria in rock dating back 3.3 billion years, and these are far smaller.

On the early Earth, bacteria were forced to stay small, Dr. Westall said, because the atmosphere did not yet have enough oxygen to fuel their growth.

From someone more enthusiastic:

“I think the authors have done a good job,” said David Wacey, who researches the origins and evolution of life at the University of Western Australia. With the new evidence, he said, “One comes up with a pretty convincing biological scenario” for the origins of the mysterious rock features.

Dr. Wacey was not surprised that the new work had drawn criticism. “It may be many years before a consensus is reached,” he said. “But this is how science progresses.”

I think the last sentence is the operative one. This is by no means evidence for early life, or even for its age, but it was certainly worth publishing and will doubtlessly lead to more work. If life really did exist 4.3 billion years ago, then it means that it didn’t take long after Earth’s conditions were “right” for carbon-based and water-requiring life to begin proliferating.


Dodd, M. S., D. Papineau, T. Grenne, J. F. Slack, M. Rittner, F. Pirajno, J. O’Neil, and C. T. S. Little. 2017. Evidence for early life in Earth’s oldest hydrothermal vent precipitates. Nature 543:60-64.


42 thoughts on “Possible life found in sediments between 3.8 and 4.3 billion years old

  1. Perfect for my high school earth science and middle school bio classes: photos, debating scientists, anaerobic life . . . they will actually like this.

  2. What are the alternative hypotheses? More than just “not life”, but what specifically? For example, I never did read in any popularizations at least what is known (now) about the Martian soil that created the false positive on one of the life-tests.

      1. The consensus (heavily criticized by Levin, who constructed the experiment) is that any organics would be broken down by oxidants. They are produced from unblocked UV interacting with mainly the CO2 atmosphere, and detected at the Mars surface by modern landers.

        Especially the polar perchlorates detected by Phoenix would have a “slow fuse” and start to react in the clement or hot conditions of chemical experiments of landers. But the peroxides that are believed to be at least partially relevant for equatorial sites like the Vikings and Curiosity seems to have behaved much the same. (That is an analysis problem that Curiosity’s construction did not avoid, and it has been frustrating to try to follow the hunt for organics.)

        And that is likely all what the Viking labeled carbon (and more) experiments detected. For one reason or other (optimism? not recognizing the potential UV problem?) the experiments did not account for such confounds.

        1. So I remember from organic chemistry class that one can photodissociate bromine to then produce peroxides. Is that the idea? (I can’t imagine it is actually bromine, of course) Is there really that much free oxygen?

    1. Interesting coincidence; it looks from the Maven satellite around Mars that the atmosphere there lasted probably about a billion years. Meaning that if these results are true, the origin of life on Earth overlapped quite well with the time at which Mars had oceans and an atmosphere and was being bombarded by the same stuff as earth (i.e. possibly the same organic stuff).

      I’m not implying Mars seeded Earth (or vice versa), but an early start to life on Earth IMO ups the probability that we might find the remains of life on Mars.

  3. From my non-expert perspective being that it is undisputed that there are 3.5 billion year old bacterial fossils, it already seems to be perfectly accurate to claim that life started very quickly on planet Earth. Granted, 4.28 billion is even quicker.

    It seems to me that the more we have learned over the years it seems to have become more and more probable that life is relatively common in the Universe. It sure doesn’t look like it was difficult, or improbable, to get the ball rolling on Earth.

    Forget my flying car, I want my warp drive starship so I can get some answers.

    1. We need to send many, many more Philae probes to comets [& other targets] pronto rather than this silly, expensive-on-intellect-better-used elsewhere preparations/studies for humans on Mars [which I often think is a marketing exercise to preserve/increase NASA budgets]

      The comets are particularly attractive because it’s supposed they originate from the hypothesized Oort cloud [200 to 200,000 AU perhaps?] – see if there’s ‘tool kits’ for life in those vastly ancient & fairly unblemished lumps of ice/rock. I’m thinking the element phosphorus which seems to be essential for ATP processes & is a pretty rare element. Also perhaps some organic chemistry molecules or perhaps little alien pre-life/life from other star systems. The Ooort is so far out that it must exchange materials with other passing star systems – perhaps a mechanism by which ‘life’ arrived in our locale even before our sun lit up.

      1. perhaps a mechanism by which ‘life’ arrived in our locale even before our sun lit up.

        I find the concept of panspermia thoroughly uninteresting. At absolute best, all that is does is take exactly the same problem (turning non-living chemistry into life), and move it to a place that we don’t know, at a time that we don’t know, under conditions that we don’t know, and which we’re never likely to find out.
        At least by looking at the abiogenesis problem on Earth, we’ve got quite strong constraints on the environment in which it happened, and some handle on the chemical conditions.
        I don’t even find the idea of “wooo, you cn form amino acids on ice grains in a vacuum, given a million years of sun-like UV” to b terribly interesting, because when you drop this mix of chemicals into a plausible terrestrial ocean, you’re still going to seriously dilute any chemicals you inherit from space. So you’re still going to have to have some concentration mechanism – which are typically proposed to sticking chemicals to mineral surfaces (a mineralogy problem) and/ or pumping fluids past fixed (“metabolising”) objects – which is a geological problem.
        Panspermia is a plausible idea, but I simply don’t see it as being a useful contribution to the study of abiogenesis. Kind of like arguing over whether the shape of Utnapishtim’s ship (“boat”-shape or “guffa”-shaped) is important to the story of Noah’s Ark.

        1. Well, Gravel, why are you mincing your words sir? 🙂

          Yes. It’s my hopeful, wishful face I’m showing in that comment of mine you’ve replied to. I would *like* to think, against all reason, that our galactic [Milky Way] environment is old enough for there to be some sort of seed ecosystem that existed before Sol which plies interstellar space like dandelion seeds writ large. This would probably have to be via an ID mechanism such as aliens unable to leave a doomed solar system a vast age ago choose to be remembered by some clever bio mechanism that’s helped on its interstellar travels by the supernovae that kills them. [bits of Superman origin legend in there]

          Or like the way a dead whale carcass food source drops onto an impoverished ocean floor & suddenly all sorts of scavengers seem to appear from nowhere.

          Or as an analogy – the way those mid-ocean ridge hydrothermal vent ecologies perhaps [I speculate] can invade fresh vents many miles down the spine of the ridge.

          Just speculations…

          I accept all you say, BUT I still want a load of Philae probes – they’ll find stuff we never thought of & as usual generate more Qs than As [which is always nice].

          My serious face position is more Loren Russell – her two comments at 6. below.

          1. I would *like* to think, against all reason, that our galactic [Milky Way] environment is old enough for there to be some sort of seed ecosystem that existed before Sol

            Ah, there are fairly strong arguments on that point about the increasing metallicities of stars with time. I know people are looking at Kepler data to try to confirm (or refute) the idea by looking at some actual data, but I don’t recall a consensus result, as yet.
            I feel the temptation of “great galactic civilisation” ideas too. Then the ghost of Richard Feynman looms up at me, rattling frozen O-rings like an engineering Marley, and moaning “Don’t fool yourself!” and “You are the easiest person to fool!”

            1. Of course Gravel, but the Vorlons & the Bushmills Black Bush in a tea cup next to my right front primate paw think differently!

    2. Yes. Even the most skeptical about this claim must admit that life started here very early. That alone should orient us to expect that life starts often. What I mean is to say that life should be pretty common out ‘yonder.

    3. From my non-expert perspective being that it is undisputed that there are 3.5 billion year old bacterial fossils

      It is disputed. More precisely, the specific evidence that Schopf et al brought forward definitely hasn’t convinced everyone that these structures are microfossils ; that evidence may not have convinced a majority of people in the field. I’m very much sitting on the fence myself – as an interested amateur. I’m perfectly happy to accept that there was life present at 3500 million years, but I’m much less convinced about Schopf et al having pointed at genuine microfossils in their samples.

      Granted, 4.28 billion is even quicker.

      That’s an oldest-possible age for the rocks, not a most-likely age. The range of ages given is longer than the Phanerozoic Era (“Era of Evident life”). What gets people strongly concerned is that the amount of time (integrated with volume of abiogenetically interesting environments) to go from an biotic system to a biotic system is getting uncomfortably tight.

      It sure doesn’t look like it was difficult, or improbable, to get the ball rolling on Earth.

      That’s certainly the way that most people who seriously follow the science, but the strict evidence is somewhere behind that opinion. We know of the existence of a number of interesting chemistries, and a number of very interesting and geologically plausible environments ; but the full story hasn’t been put together and demonstrated yet. Plus there’s the unavoidable problem that all the evidence has been steam-rollered at least once in geological history.
      Even at the earliest age given for the Nuvvuagittuq supracrustal belt (NSB, the rock unit under examination here) would mean that a little over a billion years after they had formed, they got steam rollered (“lower amphibolite grade metamorphism – think of a steam roller 20-odd km thick, hitting temperature at a dull to medium red heat for longer than the time since the mammals originated). Looking at the UK’s geology, we’ve got several hundred square miles of sedimentary rocks that are a little over a billion years old (and my personal rock pile includes stromatolite fossils from them), but of the order of 5 times more rock of the same age, but “steam rollered” to the point of partial melting (upper amphibolite grade metamorphism, 30-odd km and medium-bright red heat). In fact, the steam rollering was so bad that we’re still not sure of it’s 5 or 10 times as much area as the corresponding unmetamorphosed (well, 5-10 km and not more than 200°C) sediments.
      What I find moderately surprising is that the NSB only got the one steam rollering in around 4 billion years. Our 2.7 billion year old rocks in Scotland have fitted 4 steam rollerings into their puny short lives, which haven’t been particularly busy.

      1. @gravel Very interesting. Thank you.

        Perhaps the best place to look for fairly unmolested early-Earth rock is in the Sun-Earth Lagrange points – where it may have ended up after being ‘launched’ by asteroid bombardment of Earth’s surface?

        1. The Lagrange points are relatively stable on the multi-million year timescale, but not so much on the billion year timescale.
          People have been looking for Trojans and Greeks (in the Earth’s leading/ trailing L5, L4 points) since at least the early 1990s ; TTBOMK, no results yet.
          The Jovian Trojan/ Greek points are “deeper” (because Jupiter is bigger) and less-perturbed (because further out), and they do have a population of asteroids. But I think they’re transients too.

    1. The link to Valley 2015 I provide in my longer comment (which I think is hold for approval; I cannot see it) is an update which largely confirms that but illustrates quite nicely how a habitable global ocean gradually formed > 4.3 billion years ago [http://www.geology.wisc.edu/facilities/wiscsims/pdfs/Valley_AM2015.pdf , figure 17]. Since O’Neill is on the new paper I wonder if he has walked back his claim of the Nuvvuagittuq rocks as > 4.4 billion years old. He showed in a NASA seminar how it is the best fit to radiological data, and would explain the absence of zircons by forming before any subduction had occurred! In any case such an old crust seems reasonable from Valley’s zircon data as well as seen in figure 17, making for a consistent set of observations. On consistency one could also add Cavoiser data on the early mantle which fits a primordial unmixed Nuvuuagittuq source material to a tee [his observation, presented at AGU Fall 2016, as I remember it].

  4. The idea that sea floor hydrothermal vents were the sites for abiogenesis has been in the air for some time — and such vents must have been ubiquitous and chemically diverse in the earliest oceans.
    As we go back, at some point, we may be talking about metabolism, rather than bacterial cells, so the presence of oxidized iron and sulfur minerals, as well as isotopically-sorted carbon is very intriguing, whether the tubules, etc, pan out.

    Furtherm the comment of one critic that lack of atmospheric oxygen would limit cell size is not persuasive if there are habitats where hematite can form.

      1. Yes, though not necessarily identical to any vents observed on earth today. At this age, plate tectonics hadn’t gotten organized; the oceans were not very salty, and there was no long history of biological alteration of land, sea and air.

        And the dynamics of the early earth and oceans must have contributed to an astounding range of chemistry and thermal profiles in these vents. Undoubtedly there were hundreds of millions of these experimental systems cycling though hundreds of millions of years.

        It would be much more exciting [especially if these rocks date from the earlier, 4+ bya, range, if these are signs of life-like chemistry rather than simply older bacteria!

        1. The non-cellular stage was virtually non-existent in main hydrothermal vent theory since vents provide pores that had metabolic cell function by constriction and inorganic or organic (lipid) membranes. On the other hand the non-life stage was persistent since the best analysis method to date has LUCA as still half-alive [http://microbialcell.com/researcharticles/physiology-phylogeny-and-luca/ ]. Whether or not you call that half-alive cell “half-prokaryotic” is perhaps a philosophic question.

        2. Yes, though not necessarily identical to any vents observed on earth today.

          Identical to what degree?
          Water would still be water. Basalt, as a 5-20% partial melt of dunite-peridotite of previously un-melted mantle material, would still be basalt. The chemistry of hydrothermally circulating water wouldn’t be much different between then and now. IMO.

          At this age, plate tectonics hadn’t gotten organized;

          [SHRUG] Even if the deep convection of modern plate tectonics, with the baked remnants of plates accumulating at the top of the outer core, hadn’t started, most of the surface chemistry would have been pretty similar. Sure – no significant granite/ granitoid/ tonalite massifs had accumulated (we call them “continents”), but there aren’t actually that many oceanic hydrothermal systems rooted in granitoid rocks (more on continents, but they’re not in the frame for the origin of life).

          the oceans were not very salty

          Ah, now that might make a significant difference. But … I can’t off the top of my head think of any strong evidence (e.g. fluid inclusions in authigenic minerals very old sediments) that the oceans are a lot saltier today than they were 1, 2, 3, 4 billion years ago. I know that people hypothesised a steady accumulation of salt in the oceans, and that this might be a potential dating system. In itself, that dates the idea back to before ~1910, when the idea of radiometric dating was gaining traction (it took decades to get to worthwhile measurements – I’ve got a reprint of one of Holmes’ papers giving a 3+ billion year estimate for the age of the Earth from the early 1950s, signed by the man himself). There’s a little bell tinkling in my third braincell from the left saying that one of the Darwin offpring (fils or grand-fils, I don’t remember) tried getting an age for the Earth from the method. But since then, it’s only contribution to science has been to give creationists something to flap their gums over.
          The hole in the idea is that it stops working if there are processes to remove “salt” from the oceans. As a geologist who has drilled holes in the thousands of cubic kilometres of “salt” in the Zechstein Basin, and in the thousands of cubic kilometres of salt in the Mediterranean Basin (found a geopolitically important province on that one!), and in the thousands of cubic kilometres of salt in the Arabian Shelf … there have been effective mechanisms for removing salt from the oceans for at least a billion years, and probably a good deal longer (salt belts don’t last long under metamorphism – bit on the soluble side of insoluble).

          It would be much more exciting [especially if these rocks date from the earlier, 4+ bya, range, if these are signs of life-like chemistry rather than simply older bacteria!

          It would certainly be very exciting. And while I like this work, the rocks do look pretty “steam rollered” to me, and I’m not screamingly convinced that they’ve made their case. I’m going to let the braincell stew over it for the night, and re-read the paper tomorrow. But this geologist remains unconvinced. Despite wanting to be convinced.

          1. “The chemistry of hydrothermally circulating water wouldn’t be much different between then and now. IMO.”
            But if the atmosphere above the ocean were significantly different then from what it is now, I’d think the chemistry of the water could be quite different. However, I’m not a geochemist.

            1. Oxygen content – yes. As it goes into the system; by the time it has been circulating through hot (say, 450-500K) rock with ferrous minerals and some metal sulphide minerals … not so much oxygen. That’s the water that comes up from depth, to form the black, white or grey “smokers”.
              Ammonia content – that’ll have changed ; but the pH is going to be pretty strongly buffered by bicarbonate, even if there’s a moderate amount off ammonia. Atmospheric CO2 – same buffering against serpentine and carbonate minerals.
              Hydrogen? If hydrogen formed a significant atmospheric component which survived through the down-leg of the hydrothermal systems, then it would drag down the oxygen content – but would itself be buffered against sulphide/ sulphate species. And the presence of authigenic (formed in situ) haematite – iron III oxide – argues against significant hydrogen being present.

  5. “On the early Earth, bacteria were forced to stay small, Dr. Westall said, because the atmosphere did not yet have enough oxygen to fuel their growth.”

    Yeah, couldn’t have been different kinds of life ever/

  6. It is easy to forget how long a million years is, when talking about things 4 billion years ago. It’s only 0.001 billion years.

    If the earth were completely sterilized tomorrow, do you think bacteria could start growing again a million years in the future? I do. It is a very long time, and this earth is ripe with potential.

  7. Such discoveries are completely thrilling. I have to imagine the scientists working on this must be having a really good time. While, now having to defend their conclusions, I suppose there’s more pressure. Still I envy them.

  8. Fossils or not, they are a new discovery of something incredibly ancient. Close to the dawn of time on earth. Science is a wonder.

  9. The bible doesn’t say so, but may we assume that all life forms are/were created in god’s image? If so, bacteria and viruses must be too. The many faces and bodies of god.

    Seriously, this is a wonderful scientific discovery however it turns out. More progress is being made in the genealogy of life on Earth. Very exciting.

  10. A new paper in Nature by Matthew Dodd et al. (free access, reference below)

    I was getting a “cough up dosh” page following PCCE’s link, but I got the full paper from Sci-Hub. And I’m only getting time to have a read of it just now, the thick end of 2 days after hearing about it.

  11. yes, I have been following this best I can. We shall see how these proposed fossils hold up over time. It might garner support; it might lose support.

  12. There’s another thing that’s troubling me about this as a claim of “the earliest fossils known” : they’re outlined with authigenic co-depositional haematite, which implies a significant oxygen fugacity at the time of deposition, and therefore a fair chance of at least some localised production of oxygen. So, even if they’re fossils (but they do look disturbingly like some corrosion features in other rocks – which makes me uncomfortable with this interpretation), there is evidence pointing towards earlier lifeforms with detectable geochemical effects.
    I think I’ll fence-sit on this one. But the evidence does seem to be pushing OOL back to the very deepest of time.

  13. I’ll have to peruse the article better. I’m on the fence. I do think life would be that early, but at first sight I do not find these fossil trace convincing (not really unconvincing either). That is the difficulty if one is not an expert in the field.
    The question is at what stage we’d call something ‘life’, only at the already very complex bacterial level, or some earlier complex biochemistry? I don’t think there is a discrete boundary between ‘organics’ and ‘anorganics’.

Leave a Reply