Earliest organisms: 3.7 billion years old?

There’s a new paper in Nature that has everyone excited, for it reports what is said to be the earliest evidence for microbial life—”microbial structures” dated 3.7 billion years ago. The paper, by Allen P. Nutman et al. (reference and free link at bottom), describes what are said to be ancient traces of stromatolites—layered colonies of cyanobacteria that trap sediments and are thus fossilized—from a part of southwest Greenland that harbors old rocks.

The earliest previous evidence for microbial life are microfossils dated at 3.4-3.5 billion years old, coming from the Strelley Pool formation of West Australia. (Wacey et al., Nature Geoscience 4:698-702). The Nutman et al. finding, if true, pushes back the known existence of cells by 200-300 million years, no small chunk of time. (There is some evidence, though not very convincing, for carbon of biological origin dating back 4.1 billion years.)

What is the new evidence for 3.7 billion-year-old life? It’s largely structures in dated rocks that Nutman et al. interpret as stromatolites, structures like those shown below (“strom” means “stromatolite”). The pointy structures are identified as the remains of ancient stromatolites, though I wonder why the middle one isn’t labeled “strom”:


Image is inverted because layering is overturned in a fold. b, Interpretation of a, with isolated stromatolite (strom) and aggregate of stromatolites (stroms). Locally, lamination is preserved in the stromatolites (blue lines). Layering in the overlying sediment (red lines) onlaps onto the stromatolite sides. A weak tectonic foliation is indicated (green lines). c, Asymmetrical stromatolite and d, linked domical stromatolites from the Palaeoproterozoic28 Wooly Dolomite, Western Australia. The lens cap is 4 cm in diameter. Image c is left-right-reversed for comparison with panels a, b.

Nutman et al. give other evidence too, including isotopic data, the presence of minerals that said to be biogenic, and the presence of layers (“lamellae”) in the stromatolite-looking bits. But the most touted (and convincing to others) evidence are pictures like those above.

I checked with some well known paleontologists and sedimentologists, however, and they don’t find even the “fossil” data very convincing. (I’ll withhold their names for the time being.) The pointy bits above, they say, could be “flame structures“: simple deformation of clay or mud that occurs when it’s pushed up by heavier overlying layers of sand. This could produce (and has produced) the kind of structures seen in the photo above, but without any presence of life. Further, the layers in the structures might not represent layers of ancient microbes, but simply layers in the underlying mud that, after all, could be produced by successive sedimentation events.

The rest of the evidence, I’m told, may be suggestive of life but hardly convincing. The paper is tough going, which you’ll see if you read it, so all I want to do is note that the evidence for life given in this paper is questioned by some experts.

Nevertheless, we still have pretty good evidence for bacterial cells existing 3.4-3.5 billion years ago, and such cells are pretty complex. That means that life got started pretty soon after the Earth cooled down, roughly 4.3 billion years ago. These cells, after all, had to have undergone a very long period of evolution from the initial replicating molecule (or whatever it was) that constituted the first “life”.

So take this 3.7 billion year date with a grain of NaCl.  That doesn’t mean it’s wrong, just that there are formidable problems with finding solid evidence for life in ancient rocks. Not many of those rocks exist on Earth any more, and those that do could have been changed or deformed in a way that would make life hard to detect. Further, the best evidence for life are microfossils like those shown below, but even these are somewhat controversial. Proving that such structures are fossil bacteria rather than inclusions or artifacts is often hard to do.

Nevertheless, the photos below, and other data from the Wacey et al. paper, have convinced most paleontologists that there were microbial cells around 3.4-3.5 billion years ago.

Why do paleontologists fight bitterly about the “first” cells if it’s only a mere matter of 300 million years (!)? Well, there’s cachet to be gained by finding the earliest good evidence for life, but, beyond that, finding complex cells soon after Earth cooled down gives us a good time scale for how long it takes to go from simple chemicals to “life” (I see the origin of “life” as a somewhat subjective point, as it varies depending on your definition). Pushing dates of cells further back tells us that that transition could be even faster than we once envisioned.


a,b,e, Clusters of cells, some showing cell wall rupturing (arrows in a,b), folding or invagination (arrow in e). c,d,h, Chains of cells with cellular divisions (arrows). f,i–j, Cells attached to detrital quartz grains, exhibiting cell wall rupturing and putative escape of cell contents (arrow in f), preferred alignment of cells parallel to the surface of the quartz grain (arrows in i), and constriction or folding between two compartments (arrow in j). g, Large cellular compartment with folded walls (arrows).

h/t: Latha Menon


Nutman, A. P., V. C. Bennett, C. R. L. Friend, M. J. Van Kranendonk, and A. R. Chiva. 2016. Rapid emergence of life shown by discovery of 3,700-million-year-old microbial structures. Nature, published August 2016, doi:10.1038/nature19355


  1. Griff
    Posted September 2, 2016 at 10:08 am | Permalink

    If the earliest replicators had generations of only minutes, I could imagine cells evolving very quickly.

    • gravelinspector-Aidan
      Posted September 2, 2016 at 11:05 am | Permalink

      On the other hand, I’d be surprised to see the earliest replicators having generation times as fast as replicators that have had, literally billions of years and possibly hundreds of billions of generations.
      The oft- quoted “2 minute generation” of bacteria is a “fastest-observed, today”, not even an “average, today” (and we won’t even tart on he mean / median/ mode question).
      OTOH, even with a generation time in the days or weeks, evolution can proceed at a pace which is blistering to us metazoans with our generation times in the decades. I was seeing reports recently of increasing worries about the development of fungicide-resistant fungi (not renowned for their speedy reproduction) in addition to the better-known cases of antibiotic-resistant bacteria.

      • Posted September 2, 2016 at 11:40 am | Permalink

        I thought it was 20 minutes.

        • gravelinspector-Aidan
          Posted September 2, 2016 at 1:28 pm | Permalink

          Typo. This keyboard is noisy but not very positive. I doubt I’ll be buying another Lenovo.
          20 minutes is the figure I’d heard (and always been decidedly dubious of.)

          • Torbjörn Larsson
            Posted September 2, 2016 at 1:50 pm | Permalink

            It is derived from the model organism E. coli. [ http://textbookofbacteriology.net/growth_3.html ]

            But notably it is a modern organism, which can initiate new replication forks on daughter strands when it has optimum environment. I.e. when cleaving into daughter cells, the latter have already themselves forked substantially. [I cant find the picture right now, but they have caught the chromosomes in the act.]

            How rapid replication early cells had is an open question. And how that mapped to evolution rates is yet another question, I think. Early simple strand products – whether polynucleotides or polypeptides – could have been more promiscuous in function.

            • gravelinspector-Aidan
              Posted September 2, 2016 at 5:15 pm | Permalink

              If only we knew what systems DNA and RNA replaced.

              • Torbjörn Larsson
                Posted September 4, 2016 at 11:52 am | Permalink

                RNA seems to be primordial, supplied by the environment. Certainly the LUCA presented by Martin et al didn’t do much of basic nucleotide metabolism.

              • gravelinspector-Aidan
                Posted September 4, 2016 at 6:45 pm | Permalink

                RNA seems very clear to pre-date DNA. “Primordial” in the sense of “present from the earliest times” is a lot steeper claim.
                The presence of a succession of “information substrates” as “life” evolved, and the replacement of one substrate by another of faster metabolism and lower energetic cost isn’t an axiom in O(s)OL research, but is there in a lot of thinking. If nothing else, the selection of 20-odd amino acids from a field of hundreds, and their encoding onto a space of 64 words with a very non-uniform degree of redundancy, all hints at a deeper history. Possibly back behind the use of nucleic acids at all.

  2. Posted September 2, 2016 at 10:15 am | Permalink

    Very interesting, thanks for this.

    We have a ~1.2 byo stromatolite fossil at home (origin: Australia). I love looking at it and trying to conceive of that depth of time.

    • gravelinspector-Aidan
      Posted September 2, 2016 at 11:08 am | Permalink

      The “deep time” moment is when you irrevocably warp from being a normal human being into being a geologist.
      The beard, check shirt, and liking for beer are secondary characteristics compared to the appreciation of time, but they’re easier to spot at a distance so convenient in the field. Or in the office.

      • Kevin
        Posted September 2, 2016 at 1:32 pm | Permalink

        There are plenty of ‘deep time’ feelings.

        An Al+ ion clock with ~ 6/10^18. It misses only one second every >3.5 billion years. Or the photons from the cosmic background microwave radiation…those photons help constraint the age of our known universe. Or all starlight, for that matter.

        Or watch every YouTube as many times as they have already been seen…that would be more than 10^10 years…that’s a lot of kitties and scantily clad humans.

        • gravelinspector-Aidan
          Posted September 2, 2016 at 5:13 pm | Permalink

          I’m not talking about technical measures of time. I’m talking about the visceral realisation that the landscape and rocks that you’re looking at or holding in your hands represent hundreds or thousands of millions of years. I gave my Dad a specimen a couple of months ago which encompassed a 1200 million year gap in sedimentary history. Or mapping your way up though a monotonous section of mudrocks with sub-millimetre bedding, which you think are annual layers. A kilometre of that is barely a million years. That realisation is “deep time”.

  3. busterggi
    Posted September 2, 2016 at 10:43 am | Permalink

    If only that bacteria hadn’t eaten the microbial apple!

  4. Damien McLeod
    Posted September 2, 2016 at 10:44 am | Permalink

    If I’m gonna do the wishful thinking and the I wanna believe dance I prefer this one: https://whyevolutionistrue.wordpress.com/2016/04/29/oldest-traces-of-possible-life-on-earth-4-1-billion-years-ago/—-thank you very much.

    Hope all the kitty’s are doing well Dr.Coyne.
    I enjoy your email’s quite a lot.

    • gravelinspector-Aidan
      Posted September 2, 2016 at 11:20 am | Permalink

      Since the Jack Hills and Acasta data on very early zircons having been formed with water which had been through a meteoritic fractionation, few geologists have a problem with the *idea* of life having evolved so early [SELF : lifts the “define life” file out of storage, with a forklift truck]. Whether the evidence presented is sufficient to persuade us that this really did happen is the sticking point. All of the rocks in question have been absolutely through the mill, big style, and all the evidence (to date) is quite delicate geochemical data. (I’m still chewing on the current story – but it’s interesting because it involves macroscopic evidence of the sort that I’ve seen myself in rocks a mere 1.1- 1.2 billion years old.)
      You might even be tempted to say that the presence of dolomite masses in a sediment-like form actually moves these rocks from the class of “fubarite” to being definite metasediments – that in itself has been a question over some claims for very early rocks. (“furbarite” is an acronym attributed to various surveys, all working 3 days beyond the death of the last camel and standing for variations on “fouled up beyond all recognition”-ite.)

      • Posted September 2, 2016 at 11:42 am | Permalink

        Life forms should have had the decency to evolve after the rocks largely settled.

        • gravelinspector-Aidan
          Posted September 2, 2016 at 1:29 pm | Permalink

          Which rocks? The ones on Earth, or the ones in the sky?

          • Posted September 2, 2016 at 2:50 pm | Permalink

            The ones on Earth. Imagine that life evolved entirely on Earth but some of its chemistry occurred independently on other places. Then the fact that the oldest extant rocks are meteorites will lead to an impression that some key organic compound have been imported from space.

            • gravelinspector-Aidan
              Posted September 2, 2016 at 5:35 pm | Permalink

              Hmm, I see the sampling problem. It is exacerbated by the suggested propensity of large impacts to potentially deliver enough energy to the hydrosphere and atmosphere to boil the oceans. Which is presumed to be bad for (proto-) life in the oceans. Less bad for life in the seabed and hydrothermal systems, which treads a spectrum from Russell’s entirely reasonable interpretation of Wachterhauser’s iron-sulphur world hypothesis, to the much more dubious propositions of Tom Gold.

      • Torbjörn Larsson
        Posted September 2, 2016 at 1:58 pm | Permalink

        What does the claim of being formed with water that had been through meteoritic fractionation mean? That is not something I have seen in the papers. They implicitly assume that the formed oceans had essentially the same isotope fractionation values as today, since that date roughly coincides with the end of bulk volatile delivery.

        • gravelinspector-Aidan
          Posted September 2, 2016 at 5:18 pm | Permalink

          Well, if we knew when bulk volatile delivery ended …
          I’d have to go back and re-read the papers, but I thought the oxygen stable isotope ratios were different from mantle ratios, indicating that the water had been through a meteoritic fractionation cycle.
          But like I said, I’d have to go and read the papers again.

  5. ThyroidPlanet
    Posted September 2, 2016 at 10:52 am | Permalink


    • ThyroidPlanet
      Posted September 2, 2016 at 10:52 am | Permalink

      Sub for reelz

  6. Pyers Symon
    Posted September 2, 2016 at 11:12 am | Permalink

    Couldn’t they have used a better indicator of scale than a lens cap ( from, I assume, the camera that took the photos )? Someone surely had a ruler !

    • Posted September 2, 2016 at 11:12 am | Permalink

      That was exactly what I thought when I saw the photo!

    • Newish Gnu
      Posted September 2, 2016 at 12:30 pm | Permalink

      Or even a phone with a ruler app.

    • gravelinspector-Aidan
      Posted September 2, 2016 at 12:45 pm | Permalink

      I’ve been known to leave the photo-scale in the tent too. In the SI they resort to the traditional “geologist for scale.”
      Someone in the party – probably one of the younger members of the party – was also lugging a stone-saw with a 5cm+ cut (so probably 15cm disc). Field days with usable weather in Greenland are a rare commodity, even though this is quite well south and relatively balmy.
      Another oddity is that the SI gives the location coordinates “from Google Earth” ; I’m surprised that no-one had a GPS and set up to averaging a waypoint while sweating oer the saw.
      FYI, as closely as I can tell, the outcrop is sandwiched between two sets of moraines. I get the location at https://www.google.co.uk/maps/@65.1791667,-49.8041667,1149m while this link covers the wider outcrop of the “Isua Supercrustals” – with the eye of practice, you can see the wider folds where the Isua group are wrapped around (IIRC) a core of shallow-water lavas.
      They must have better aerial photography. And it’s a sufficiently remote location that even inveterate rock-botherers like me are unlikely to take dynamite up there to bag my own samples, so I can’t understand why the absence of properly controlled GPS coords.
      I do recall mention that one of the recent expeditions met unusually large amounts of bare (not snow-covered) rock, so possibly they were just surprised by the weather’s clemency.
      Incidentally, rulers are really poor for “object for scale”. The contrast is very low. Almost anything is better – including “geologist for scale”, “hammer for scale” and lens caps. Of course, a proper scale is better. My last-but-one rock hammer (last seen disappearing into the North Atlantic at 9.8m/s/s), I put saw-marks into the untempered parts of the head with contrasting colours of markings in 1 cm blocks. But the paint rubs off.

    • Kevin
      Posted September 2, 2016 at 1:37 pm | Permalink

      A lot of science is a mix with ‘that’s sufficient’ and ‘we must have real’.

      I would have measured the lens cap and substituted, digitally, a scale bar. Some scientists, whom I have worked with, would not have liked that. They would have thought it would be regarded as possibly doctored. This kind of insecurity and lack of trust is no good for science.

  7. Gregory Kusnick
    Posted September 2, 2016 at 11:19 am | Permalink

    I wonder why the middle one isn’t labeled “strom”

    The caption suggests they consider it part of the right-hand aggregate of “stroms”.

    • busterggi
      Posted September 2, 2016 at 11:52 am | Permalink

      Tehn they should have been labeled “strom”, “stromer” and “stromest”.

  8. rickflick
    Posted September 2, 2016 at 12:05 pm | Permalink

    I can’t remember the book title, but I read some speculations about the probability of life on other planets which was based on estimates of how hard it was to cook up life on Earth. They used what was then the estimate of the earliest life to help establish the amount of time between significant evolutionary phases – the first life, first prokaryotes, first eukaryotes, first multi-cell, first worm, etc., all the way up to first human. The parameters they came up with suggested its pretty damned rare. If the time for life to arise is set earlier, the probability for at least simple life elsewhere would seem to increase.

    • gravelinspector-Aidan
      Posted September 2, 2016 at 12:47 pm | Permalink

      That would be an important deduction from pushing the OOL date back.
      Which is one reason for being very cautions about accepting this.

    • Torbjörn Larsson
      Posted September 2, 2016 at 2:13 pm | Permalink

      There are many different models of deep time evolution and forms of selection bias (our observation, anthropic selection) inherent in those models. They have still to demonstrate usefulness, and they don’t really answer the interesting problem of emergence rates.

      Instead your observation is central, a fast result implies an easy process. But seriously, consequences should mean little in science. We would still play around with “impetus” if early scientists had bowed to the church at the time.

      • rickflick
        Posted September 2, 2016 at 3:37 pm | Permalink

        Well they are just speculation, but thoughtful speculation. To demonstrate usefulness we’d have to make some discoveries of real exobiotica. But from their analysis you can predict that an Earth-like planet around a star with a lifetime of only 1 billion years would be very unlikely to have intelligent or even multi-cellular life. That’s at least somewhat useful. The away team need not carry sidearms.

    • Gregory Kusnick
      Posted September 2, 2016 at 2:17 pm | Permalink

      I’m skeptical of attempts to infer probability purely from elapsed time. It takes a long time to fill a swimming pool from a garden hose, but once it’s full, you’ll very quickly have a large puddle on the patio if you’re not paying attention. Plotting the surface area of the water as a function of time would show a long period of stasis followed by a rapid increase late in the game, but it would be a mistake to infer anything about the probability of large puddles from that graph.

      Are there swimming-pool effects in the history of life on Earth? The oxygenation of the lithosphere, hydrosphere, and atmosphere is a plausible candidate. For all we know the mutations needed for aerobic metabolism were common during the first few billion years, but didn’t become adaptive until the oxygen level reached the puddle-forming stage.

      • rickflick
        Posted September 2, 2016 at 3:47 pm | Permalink

        I’m not here to defend the book (which I have forgotten the name of), but as I remember it they were estimating time scales based on historical events we already have some idea about. For example, oxygenation of the atmosphere was necessary for complex land animals to arise. So the appearance of complex land animals in the fossil record serves as a limiting marker for the oxygenation process. I think this was correlated with data on surface minerals, etc. There conclusion was that there are a great number of whoops to jump through on the way to humans, each whoop requiring considerable time, based on their historical review.
        Damn. I wish I could remember the book.

    • derekw
      Posted September 3, 2016 at 7:38 am | Permalink

      If the time for life to arise is set earlier, the probability for at least simple life elsewhere would seem to increase.That seems to be the conclusion most are making…though it also bolsters evidence for the miraculous ala panspermia or the creator god hypothesis(popular article at https://www.washingtonpost.com/news/speaking-of-science/wp/2016/08/31/3-7-billion-year-old-fossils-may-be-the-oldest-signs-of-life-on-earth/ )

      • rickflick
        Posted September 3, 2016 at 8:56 am | Permalink

        “though it also bolsters evidence for the miraculous ala panspermia or the creator god hypothesis.”

        I’m not sure how that would be true. Did you mean to link to info on that?

      • Torbjörn Larsson
        Posted September 4, 2016 at 12:04 pm | Permalink

        Having a faster process decrease the likelihood for competing processes such as transpermia. (Panspermia between stars is unlikely due to the times involved – no active repair mechanisms for genetic material – but a local process is possible.)

        Since magic is already rejected by thermodynamics it isn’t in the competition as an alternate theory. But if we assume it existed, that too would become less likely when having an easier natural process.

  9. Diana MacPherson
    Posted September 2, 2016 at 12:06 pm | Permalink

    Whenever I’ve had to suffer blue-green algae in my aquarium, I think about how old it is & how important it was for life on earth, even though I hate it.

    • infiniteimprobabilit
      Posted September 2, 2016 at 6:37 pm | Permalink

      Wassup? Your little avatar thingy seems to have changed colour. IIRC it used to be yellow.

      (Mine used to be mauve).

      WordPress is doing strange things to us.


      • Diana MacPherson
        Posted September 3, 2016 at 12:57 pm | Permalink

        Weird. Looks OK now.

        • infiniteimprobabilit
          Posted September 4, 2016 at 3:09 am | Permalink

          No, your little ‘avatar’ thingy is still a nice lime green one where it used to be chrome yellow (I’m sure my memory is accurate on that score).

          Actually – this *is* odd – you’re green in my Opera browser. And in Iceweasel. And Konqueror. But if I look at this same page in Chromium your avatar is the old yellow one, of a slightly different design.

          I have absolutely no idea why this should be.


          • Diana MacPherson
            Posted September 4, 2016 at 7:55 am | Permalink

            That’s weird. I’m using chrome and Safari on iOS and Mac and it’s yellow. Maybe your browser interprets that colour differently. I don’t know how the colour of the avatar is coded.

            • infiniteimprobabilit
              Posted September 4, 2016 at 9:08 am | Permalink

              Well, that’s the first thing that occurred to me – that Opera had somehow got its colour palette screwed. But, not so, it still displays colour photos correctly. Also, the ‘new’ green avatar is a different pattern from your old one – i.e. it’s a completely different one. And Konqueror and Iceweasel do the same.

              Has me baffled.

              (Reason I noticed it is, I was used to scrolling quickly down a long page and recognising the ‘Diana’ comments as they went past from the colour).


              • Diana MacPherson
                Posted September 4, 2016 at 10:09 am | Permalink

                That is odd. I know that the avatar is generated with math foo from your email address. Usually I have goofed my address when that happens.

                I wonder if the math foo is somehow a different math foo on those browsers. Very interesting anyway.

              • infiniteimprobabilit
                Posted September 4, 2016 at 7:44 pm | Permalink

                Yes, there’s obviously some deep juju going on.
                ‘View source’ is usually illuminating, but in this case both Opera (which shows the green avatar) and Chromium (the old yellow one) yield this: {img alt=” src=’https://2.gravatar.com/avatar/8b83f8fbcd1881ebc3d0e3ac9c4af808?s=32&d=identicon&r=G’ class=’avatar avatar-32′ height=’32’ width=’32’ /}

                But pasting that into a browser just yields Gravatar’s default icon.

                So whatever’s going on is way beyond my basic knowledge of HTML…


  10. µ
    Posted September 2, 2016 at 12:10 pm | Permalink

    Interesting! Didn’t know about Bell et al 2015 and the potentially biogenic carbon 4.1 billion-years old

    • gravelinspector-Aidan
      Posted September 2, 2016 at 12:48 pm | Permalink


      • Torbjörn Larsson
        Posted September 2, 2016 at 2:16 pm | Permalink

        I have the same grin, but probably for other reasons if I read correctly and take that as sardonic. The find is an exciting possibility, both as candidate and as (microscale) “locale”!

  11. Posted September 2, 2016 at 12:56 pm | Permalink

    Reblogged this on The Logical Place.

  12. Torbjörn Larsson
    Posted September 2, 2016 at 1:31 pm | Permalink

    This find places in a series pushing candidate and accepted fossils earlier, in no small chunks of time. Besides that earlier stromatolites – and Brasier et al microfossils – at 3.5 Ga has had a belated acceptance, and that Isua is metamorphic sediments, these specific candidates run up against the putative late bombardment posed between 4.1 and 3,8 Ga.

    Two weak series of evidence clash, and it is a historical quirk that the late bombardment came first. Notably early fossil candidates strengthens over time, while the late bombardment evidence weakens. (See e.g. the latest evidence that the Imbrium contamination of the Apollo samples were more serious than earlier estimates.)

    I haven’t read the paper, but the find appears to have the essential characteristics of context and repetition. (I.e. several stromatolites in this series, and stromatolites a dominating fossil in “near” time.)

    There is some evidence, though not very convincing, for carbon of biological origin dating back 4.1 billion years.

    My characterization is that the fossil candidate is such because it currently fails to be evidence. I.e. it lacks external context and repetition.

    As it is, the find should be internally convincing. The two organic carbon inclusions have narrow and consistent geometries and carbon isotope ratios. Incidentally they are better matches for photosynthesis bacteria metabolism than Isua organics. Finally geological dating of habitability at > 4.3 Ga – oceans – and biological dating of the bacteria/archaea split at > 4.2 Ga can not reject them.

    Hopefully the find can be repeated, there are more zircons out there. But if a convincing context can be established is unfortunately not a given. If not, the find(s) will remain curiosities.

  13. keith cook + / -
    Posted September 2, 2016 at 3:45 pm | Permalink

    You’re only as old as you feel.. heh heh.

One Trackback/Pingback

  1. […] nearest star & in its habitable zone: WOW! Early 3.7 billion y.o. life? Old stromatolites or just squashed mud? Pleasurable ponderings🙂 […]

%d bloggers like this: