Taphonomy of fossilized embryos

There are these fossilized embryos from the Ediacaran, approximately 570 million years ago, that have been uncovered in the Doushantuo formation in China. I’ve mentioned them before, and as you can see below, they are genuinely spectacular.

Parapandorina raphospissa

But, you know, I work with comparable fresh embryos all the time, and I can tell you that they are incredibly fragile—it’s easy to damage them and watch them pop (that’s a 2.3MB Quicktime movie), and dead embryos die and decay with amazing speed, minutes to hours. Dead cells release enzymes that trigger a process called autolysis that digests the embryo from within, and any bacteria in the neighborhood—and there are always bacteria around—descend on the tasty corpse and can turn it into a puddle of goo in almost no time at all. It makes a fellow wonder how these fossils could have formed, and what kind of conditions protect the cells from complete destruction before they were mineralized. Another concern is what kinds of embryos are favored by whatever the process is—is there a bias in the preservation?

Now Raff et al. have done a study in experimental taphonomy, the study of the conditions and processes by which organisms are fossilized, and have come up with a couple of answers for me. Short version: the conditions for rapid preservation are fairly easy to generate, but there is a bias in which stages can be reliably preserved.

The experiments are straightforward. They worked primarily with embryos of the sea urchin Heliocidaris erythrogramma, as well as a number miscellaneous smaller species to test the effect of size, and killed them under various conditions and watched to see what would happen. There are two secrets to getting good embryo preservation: the animals have to be kept in anoxic, reducing conditions, and the fertilization envelope must be intact. The preserving conditions were generated by putting the embryos in a solution of sea water with 100 mM β-mercaptoethanol; this is analogous to immersing them in the H2S environment believed to have killed the Doushantuo specimens, without the nasty toxic risks of working with hydrogen sulfide in the lab. Photos A-C below show embryos kept in the β-ME solution for days to weeks, and while they are as dead as can be, the cells are still beautifully intact. D and E are embryos that were killed with β-ME, but then placed in pure seawater—the deterioration is obvious.

Cleavage-stage H. erythrogramma embryos under preserving and
nonpreserving conditions. (A and B) Embryos killed at the two-cell stage by
placing them in seawater containing 1% ammonium for 10 min, then trans-
ferring them to seawater containing 100 mM
after 2 (A) or 10 (B) days. (C) Embryo killed at the eight-cell stage by transfer
into seawater containing 100 mM
Embryos from the same two-cell stage culture as in A and B but returned to
normal seawater after killing, photographed after 2 days. Embryos have
undergone autolysis: cytoplasmic lipid and pigment have coalesced (arrows);
cleavage furrows have degraded (asterisks); and fertilization envelopes are
disintegrating (arrowhead). Autolysis is further advanced in the top embryo
than in the bottom embryo. In the bottom embryo, the process is further
advanced in the left-hand blastomere (arrow) than in the right-hand blas-
tomere. (E) Decaying surface of an embryo from the set shown in A and B,
returned to normal seawater after 4 days in reducing conditions, photo-
graphed 7 days later (total 11 days postdeath). Onset and progress of decay is
slower than autolysis in embryos never exposed to reducing conditions. The
fertilization envelope degrades and the cytoplasm of the embryo is then
exposed to external decay processes, including attack by protists (arrows).
(Scale bar: 200 µm A-D; 32 µm E.)

The reducing conditions of a β-ME solution would keep embryos intact for weeks, long enough for phosphatization to begin. The effect was independent of embryo size, too, with both small and large embryos being successfully preserved. That suggests that the relatively large size of the fossilized specimens may not be a preservation artifact, but may reflect the actual size distribution of Ediacaran embryos.

Unfortunately, there are serious biases in preservation. The fertilization membrane is a kind of spherical protein coat that surrounds newly fertilized embryos, forming a kind of shell within which the embryos develop. Without that membrane, the embryos disintegrate fairly quickly, even in the β-ME solution. That means that we’re unlikely to ever find post-hatching larva in these fossil layers—once they’ve emerged from that membrane, the conditions just don’t work to preserve their structure intact any more. In addition, the reducing conditions that work so well to preserve cell structure dissolve calcite skeletal elements, so skeletons and shells would also be lost. In addition, later stage embryos, where cell-cell adhesion is relatively weaker, are poorly preserved as the arrangements of the cells become scrambled.

Here is a summary of the results:

Consequences of death and postdeath conditions for
soft-bodied embryos
  Condition Outcome
Mode of death
  Fast Normal cell cleavage pattern
  Slow Abnormal cell cleavage pattern
Fertilization envelope present
  Normal seawater Rapid autolysis
  Reducing conditions Extended preservation
  Cleavage stages Cell arrangement retained
  Prehatching blastula Cell arrangement lost
Fertilization envelope absent
  Normal seawater Rapid decay
  Reducing conditions Rapid loss of morphology;
individual cells preserved

The positive answer from this work is that those Ediacaran fossil embryos probably are faithfully preserved, and represent a reliable picture of the distribution of embryonic characters over a broad range of phyla. The bad news, though, is that larvae wouldn’t have been preserved; the absence of feeding larvae in those formations doesn’t mean that they weren’t there, just that the conditions present wouldn’t have preserved them. From the Raffs’ other work, we know that one strong interest is in the evolution of direct- and indirect-developing forms, or life history evolution. We aren’t going to get the answers directly from fossil observations, I’m afraid, but they do hold out some hope that the presence of feeding larval stages can be inferred from indirect evidence, such as the size distribution of the embryos.

Raff EC, Villinski JT, Turner FR, Donoghu PCJ, Raff RA (2006) Experimental taphonomy shows the feasibility of fossil embryos. Proc Nat Acad Sci USA 103(15):5846-5851.


  1. says

    Are there any other conditions that are OK for fossilization that may be more conducive to preservation of older embryos and larvae? Was the H2S environment the only kind of environment at the time? How about later, post-Ediacara?

  2. Dustin says

    Wow. I didn’t think that something like that was even possible. That’s pretty amazing.

  3. says

    They didn’t say, and they couldn’t — maybe there is some special condition that would work. However, these experimental specimens did mimic the fossils well, and the conditions are similar to those proposed to have preserved the originals, so what this is saying is that the conditions that generated the phosphatized fossils of the Doushantuo are not compatible with preserving other stages.

  4. Flounder says

    The lack of scale bars sucks. But then again I don’t have time to reference the original paper.

  5. Chris Nedin says

    Firstly kudos to the National Academy of Sciences for allowing access to the complete paper on their web site.

    Secondly, that’s the Australian sea urchin Heliocidaris erythrogramma :-)

    And thirdly, its a good paper. They’ve shown that embryos can survive intact for the required period to allow phosphate preservation.

    However, I’d query the need for a “strong’ (p.5847) reducing environment. It’s not necessarily the strength of the reducing environment (= low pH, = lack of oxygen), but the integrity of the environment leading to the required duration of the low pH event. Provided the environment restricts the influx of oxygen and retains the low pH environment then the low pH event is sustained for the length of time necessary for phosphate precipitation. Integrity can happen for a couple of reasons:
    1) rapid burial in fine grained organic rich sediments (e.g.possibly Burgess Shale and Emu Bay Shale.
    2) Trapping on a low pH sea floor.

    The first is straightforward burial in organic rich fine grain sediments seals off the organism from oxygen, within a rapidly lowering pH environment.

    The second is even more likely! Low oxygen, low pH bottom waters are common where there is fine grained organic rich sediment and little current activity (a bit of a prerequisite for fine grained sediment accumulation). In this situation the decay by-products of the organic matter rapidly accumulate and strip the oxygen from the water column immediately above the sediment (water-sediment interface if you want to be fancy) faster than it can be replaced by diffusion from the rest of the water column. Oxygen is a slow diffuser in sea water. So the result is a low oxygen/low pH layer of sea water up to a few millimetres thick. Sure a few mms doesn’t sound like much, but remember the SIZE of the critters we are dealing with here, 300-500 micrometres (0.3-0.5 mm), easily small enough to get swallowed up by the layer. Not only that but death is rapid once in this layer.

    This is where a good membrane becomes important. Without a membrane, even a little decay rapidly increases the mushiness index, and the thing falls apart. Now, with a good membrane you can tolerate a small amount of initial decay as the membrane will hold everything together in the original layout. However, the most important role is as an environmental integrity device. The membrane holds in the initial decay products, allowing the low pH environment within the structure to be maintained, which stops further decay. Maintaining the low pH environment both inside and outside the structure will allow any stray carbonate (calcite) and phosphate to enter solution. Low pH solutions can carry more dissolved carbonate and phosphate that higher pH solutions.
    Studies have shown that it is the change in pH levels that initiate phosphate and calcite precipitation. as pH rises, the environment becomes supersaturated in phosphate and carbonate and precipitation (ppt.) results. Here the membrane also plays a part because it can act as a template for the formation of phosphate and calcite crystals, starting the ppt. process off (once started, precipitation is rapid) allowing ppt. on and around the tissues providing a mask of them in phosphate. Now if the low pH environment is too short we only get carbonate ppt. This may be that it takes time to build up sufficient concentrations of dissolved phosphate. However, when you do get dissolved phosphate, it acts to inhibit the ppt. of calcite (it blocks the sites of carbonate crystal growth) so you always get phosphate ppt. first and carbonate ppt second.

    So the presence of a membrane is essential both to hold everything together, maintain the low pH environment and act as a template to initiate ppt.

  6. Kai-Mikael Jää-Aro says

    So, it is likely that the autolytic process was present even in Ediacaran cells? I have understood autolysis as a part of the immune system, so that cells which probably are sick (considering dying to be a subset of that) kill themselves so as not to harm the rest of the organism. What is the current state of knowledge about Precambrian immune systems?

  7. Doug M. says

    [raises hand] So, these embryos show signs of a fertilization membrane?

    That’s sort of interesting, ennit? I mean, a fertilization membrane is no small thing. That suggests that embryological development was already more complicated than cell -> blob of cells -> mat of cells, even well back in the Precambrian.

    Are there any modern animals that /lack/ a fertilization membrane? Sponges?

    Doug M.

  8. says

    “the absence of feeding larvae in those formations doesn’t mean that they weren’t there, just that the conditions present wouldn’t have preserved them.”

    This is just another tired example of ad-hoc reasoning by appeal to naturalism-of-the-gaps. Did Eidicaran embryos develop into adults through a larval stage? The fossils STILL SAY NO! There are profound missing links between fossil embryos and adults.
    This is science, here, not atheism. We mustn’t dogmatically state as fact those theories that leave out the possibility of adulthood reached by means of an intelligent developer! Tee hee.

    Great article, I really like hearing about the new and creative ways that science studies the past. Despite claims to the contrary.

  9. Coragyps says

    Beta-mercaptoethanol. Wow. I’m probably the only one that visits here that uses 20,000 kg of it a month….and at 0.1 molar, I can imagine it preserving cells, even if it will make ’em stink!!!

  10. iGollum says

    Whoa, Coragyps, did I read that right, you use 20,000 kg of 2-ME a MONTH? What the hell do you DO? I hate the stuff and use it only microlitres at a time (for RNA extraction), I can’t imagine being around literally tons of it… Yuurgh.

  11. says

    I wonder if this discovery will help scientists look for the right indicators in the future to find more of these fossils.

    Also, is there a similar process, or would the same process help to preserve soft tissue?

    Neat stuff, thanks, PZ and those who did the research!


  12. Coragyps says

    BME goes into a couple of products that are used in the hydrochloric acid that gets pumped down oil wells – it reduces ferric iron to ferrous, and keeps the crude oil from turning to pasty goo after the acid contacts it. US Patent 6 060 435 if you’re interested.

    But we can’t get the stink out – and it smells much worse after it starts forming thiirane in warm acid.

  13. Elizabeth says

    Could you go into a bit more detail as to how we know these are indeed fossilized embryos rather than fossilized something else or some interesting rock formation?

    Not a creationist, I just remember the controversy with the Mars rocks with “signs of life”.

  14. CCP says

    “a fertilization membrane is no small thing. That suggests that embryological development was already more complicated than cell -> blob of cells -> mat of cells, even well back in the Precambrian.”

    I don’t know about sponges specifically–but the function of a fertilization membrane/envelope starts during fertilization itself–it’s the “slow block to polyspermy” that prevents entry of >1 sperm nucleus. As such I’d guess it would be useful to any sexually reproducing species, even the unicellular.

    Another bit of info we can glean from these embryos is the egg size of whatever multicellular animal laid them: initial development to at least the morula stage (maybe blastula? Dr. Myers?) occurs without growth.

  15. says

    Could you go into a bit more detail as to how we know these are indeed fossilized embryos rather than fossilized something else or some interesting rock formation?

    Size, similarity to extant embryos, regularity, uniformity of blastomere size, and the presence of multiple near-identical exemplars in the same samples. There are also anatomical details like the preservation of the fertilization membrane. The fossils are so well preserved that you can even zoom in with the SEM and see surface featurs. The Raff paper has SEMs of urchin embryos that are in many ways indistinguishable from the fossils.

    Just out of curiosity, what species are the sploding embryos in the video?

    Those are my zebrafish embryos. I do teratological experiments that sometimes go awry, and instead of abnormal embryos I get lots of mortality.

  16. Bruce Thompson says

    Any information on the environment? How did the embryos get into the reducing environment. Did these embryos drift into a local H2S feeder spring draining into the ocean? How widespread is the fossil deposit? Does the distribution of phosphatized microfossils correlate with any other mineral deposits?

  17. CanuckRob says

    It is great posts like this that won the Koufax for Best Expert Blog. This is truly astounding, embryos over half a billion years old. I feel sorry for creationists because they cannot see the absolute wonder (entirely natural) of life. There are none so blind as those who will not see.

    PS Is it possible to ge the delete cookies to post thing fixed?

  18. says

    But, the important thing, obviously, is that you can tell, even at this embryonic sttage, that this is….well, uh, it kind of looks pretty much like any other cluster of cells, doesn’t it?

  19. Coragyps says

    “How did the embryos get into the reducing environment. Did these embryos drift into a local H2S feeder spring draining into the ocean?”
    I’m betting that the eggs were laid/shed/released near the surface in oxygenated, “sweet” water but sank into a sour, anoxic layer nearer the bottom. The Black Sea is like this today: everything below a couple of hundred meters deep is oxygen-free and full of hydrogen sulfide.