Tiny sea creatures make huge, fantastic houses to protect themselves and get food

Of all the posts I’ve written about the amazing things animals do, testifying to the power of natural selection, this is one of the most amazing. It concerns a very tiny animal, Bathochordaeus stygius, a “giant larvacean”, which is a free-swimming marine tunicate, a chordate in the same phylum as we humans.  It’s about 1.5 inches (4 cm) long, with a “trunk” where the organs reside, and a tail that helps it swim and, in this case, pumps water to help it eat. The tail contains the notochord (a stiff rod that we have as embryos; it develops into our spinal column), as well as muscles that are crucial in the activity described here.

Here’s a photo of the naked tuncate from Ocean:

The larvacean is rarely found “naked,” however, for it builds not one but two houses for itself out of mucopolysaccharides (mucus), a big net-like house about a meter across as well as a smaller, complex house (about 10 cm or 4 inches across) in which the animal resides. And both houses are built and discarded every day!

A new paper in Nature, below, tells how researchers used a new laser apparatus in free-swimming animals off Australia to dissect the structure of the inner house to reveal its workings. That inner house, known for a long time, serves not only to protect the animal (it even has an escape hatch that it uses when discarding the inner house or when something bumps it), but mainly to concentrate small organic food particles, which, after being moved through several chambers by the tail undulations, wind up caught in a net by the trunk, where the animal eats it.

The group largely worked from the Monterey Bay Aquarium in California, which produced this wonderful 4-minute video summarizing the paper’s results:


You can also read a short summary at the New York Times, but it’s not all that great, leaving out really interesting information (don’t worry; I’ll supply it):


You can get the article free by clicking on the screenshot below (you must have the legal Unpaywall app), or find the pdf here. The full reference is at the bottom. And don’t miss the five videos, here, especially the one in which they use dye to track the water flow through the house.

The outer net (“oh”) in the (a) bit below, presumably serves (as does the inner house) to deter predators like fish and jellyfish, but also to catch larger food particles that the larvacean couldn’t eat and would clog the filter. It surrounds the inner house, which is quite complicated and serves mainly as a place to filter organic debris and convey it to the mouth of the larvacean. “si” in the first picture is an abandoned house; after being used for just a day, these sink to the sea floor where they and their food-particle contents are consumed by other animals. There are two channels between the outer “net” house and the inner house, allowing food to be transported neatly to near the larvacean:

(From Nature): a, Inner and outer house structures of the mucus feeding structure. b–e, White-light (b, c) and laser-sheet (d, e) illumination of both the lateral (b, d) and dorsal (c, e) views of a midwater giant larvacean, B. stygius. fcf, food-concentrating filter; ih, inner house; ihw, internal house wall; oh, outer house; si, abandoned house or sinker; st, suspensory thread; ta, tail; tc, tail chamber; tr, trunk. Scale bars, 4 cm.

The structures of the houses, particularly the inner houses, were determined using a “laser-sheet” apparatus called “DeepPIV”, shown below. This was put in the mid-level depths where the larvae reside and laser scans revealed sections of the houses, which were then assembled by computer to regenerate the three-dimensional structures.

Here’s a photo from the New York Times showing the DeepPIV in action:

And the result of the scanning.  The larvacean itself is seen in (a) and (b), with (b) also showing the filters by the head (trunk) where the animal can snack on what’s caught. Water comes in the two inlet channels (e), and then is moved by the flapping of the tail through two other channels that move the food to the filters.


Here’s the flow of water through the house as determined by both the laser scanning and dye-injection experiments that track water flow. It’s hard to see how the water moves (but watch the movies); thewater, after traversing the chambers, winds up flowing through two filters beside the head, where the larvacean takes the food. The filters get clogged up after a day or so, and the larvacean discards both the inner and outer houses and builds the two structures anew. That’s got to take a lot of metabolic energy!

And here is the BIG MYSTERY about this whole thing: how do they build the outer net and, especially, that complex inner house? After all, what we have here is essentially a tadpole without limbs, and yet somehow it’s able to construct two complex structures out of mucus, one inside the other. And I’ve bolded what really knocks me over from the paper’s summary:

The greatest remaining mysteries of larvacean houses concern how they are produced. Whereas a spider builds a complicated web one silky strand at a time, the house of a larvacean is extruded all at once as a rudiment and is then inflated. This leads to the question of how a bank of mucus-producing cells can create such an intricate form within a small, tightly packed bubble. Given their remarkable architecture, it seems almost implausible that these complex marvels should be built to last only a day or two. Future observational tools and vehicles will enable us to observe the construction of giant larvacean houses in their entirety, and to precisely document the frequency with which they are built.

It’s constructed compactly and then inflated! How on earth can this tiny creature do that? Well, we have no idea, so there’s lots of work to be done.  What is clear is that the houses, inner and outer, are examples of Dawkinsian “extended phenotypes,” structures that aren’t part of the animal’s body itself but can be conceived of as extensions of the animal (like a termite mound or a beaver dam) The behaviors for making the houses must clearly reside in the larvacean’s genome, as the animal doesn’t learn to do this. What a tangled house we build!

This is from Encyclopedia.com:

Larvaceans move their tail inside their house to make a current that filters food particles and moves the house through the water. If the filters become clogged or something bumps the house, the larvacean leaves the house through a trap door. The beginnings of a new house lie on the trunk of the animal’s body, and the larvacean inflates the new house and flips inside.

If you don’t find that stunning, you need to buff up your capacity for wonder!

Finally, just to show that this is by no means the largest invertebrate “extended phenotype” in the sea, here’s a video of a giant siphonophore (a class within the phylum Cnidaria, which includes jellyfish, corals, and sea anemones ) estimated to be 150 feet (46 meters) long. How the individual animals (“zooids”) work as a team, and the advantage of such a length, is yet to be determined.

This commentary is from SciTechDaily:

The discovery of the massive gelatinous string siphonophore — a floating colony of tiny individual zooids that clone themselves thousands of times into specialized bodies that string together to work as a team — was just one of the unique finds among some of the deepest fish and marine invertebrates ever recorded for Western Australia. Scientists from the Western Australian Museum, led by Chief Scientist Dr. Nerida Wilson, were joined by researchers from Curtin University, Geoscience Australia and Scripps Institution of Oceanography in exploring the Ningaloo Canyons in the Indian Ocean. Using an underwater robot, ROV SuBastian, they completed 20 dives at depths of up to 4,500 meters over 181 hours of exploration.

This all reminds me of biologist J. B. S. Haldane’s comment about the cosmos: “My own suspicion is that the Universe is not only queerer than we suppose, but queerer than we can suppose.” The “can” says everything about the limitations of our imagination. Nobody could ever have predicted or guessed that a larvacean like this could exist. Nor a frog, nor almost any other organism!


Katija, K., Troni, G., Daniels, J. et al. 2020. Revealing enigmatic mucus structures in the deep sea using DeepPIV. Nature (2020). https://doi.org/10.1038/s41586-020-2345-2


  1. Posted June 17, 2020 at 10:18 am | Permalink

    I wonder if these creatures are also counterexamples to the idea that only humans represent internally the goal of an activity before starting it.

    • Posted June 17, 2020 at 10:19 am | Permalink

      I doubt it; they are just instantiating a neural program, or so I think.

      • Posted June 17, 2020 at 10:28 am | Permalink

        I agree; of course I am no expert.

        I do find it interesting trying to figure out when this ability evolved – and it seems to have happened twice, perhaps even three times. (At least)

        I’m thinking of in mammals, in birds and in cephalopods, even. Not sure about the latter, but when I see those “octopus opens jar” things, I have to wonder …

      • eric
        Posted June 17, 2020 at 10:36 am | Permalink

        Yep. Genes build their body, which has mucus extrusion organs. Those organs just do what they do, evidently producing goo in the right conformation which is then expanded.

        It would be interesting to look at the variation in house shapes and sizes. Since the entire thing is extruded at once and inflated, it would seem to me that the shape would be more dependent on phylogeny than environment. So the variation in houses may reflect variation in genetics or development, rather than environment. But that’s just a guess.

  2. Mark Reaume
    Posted June 17, 2020 at 10:18 am | Permalink


  3. ThyroidPlanet
    Posted June 17, 2020 at 10:20 am | Permalink


  4. jezgrove
    Posted June 17, 2020 at 10:29 am | Permalink

    Incredible and fascinating – thanks for posting it.

  5. merilee
    Posted June 17, 2020 at 10:32 am | Permalink


  6. Posted June 17, 2020 at 10:55 am | Permalink

    Great videos.

    Just makes me want to go back to the ocean. Have had to move my plans for Hawaii back twice. I would welcome a dozen jelly stings if I could just sit in the waves above the coral.

  7. Posted June 17, 2020 at 11:22 am | Permalink

    They are also a player in carbon sequestration.

  8. rickflick
    Posted June 17, 2020 at 12:00 pm | Permalink

    I hope there are plans to try to capture the house building using time-laps photography. It would be nice to watch it unfold. There must be something about the contours and movement of the extruding orifices that shapes the house as it emerges. Like blowing soap bubbles. Very cool.

  9. Posted June 17, 2020 at 12:02 pm | Permalink

    I just wanted to congratulate on the wonderful sentence, “What a tangled house we build!”

  10. Mike
    Posted June 17, 2020 at 12:49 pm | Permalink

    The larvaceans are really spectacular. I did not know about these giants – our local species are only a few mm in length. I don’t think it’s quite right to call the house an extended phenotype. I tell my students to think of the larvacean house as just a very large and very complex cuticle: it’s homologous with the outer extracellular layers on other animal bodies, it’s just made mostly of mucopolysaccharide instead of protein or chitin or carbonate. Shedding the larvacean house is like moulting by an arthropod (at least the event is like moulting but it’s not under the same hormonal control). I think if the house was built of found objects (like a termite mound or a beaver dam) I’d agree with the extended phenotype label.

  11. Steve Pollard
    Posted June 17, 2020 at 12:49 pm | Permalink

    This is just mind-blowing, and I hope there is much more of it to come.

    The mucal extrusions might be constrained to take their specific configurations by the pH and maybe dissolved mineral content of the ambient seawater. I wonder if they undergo a chemical reaction with the water as they expand after extrusion.

  12. darrelle
    Posted June 17, 2020 at 1:29 pm | Permalink

    My theory, which is mine, is that this is what happens when a larvacean sneezes.

    More seriously, nature never ceases to amaze. I can imagine all sorts of things figuring this puzzle out could inform. Such as how to design and build tiny compact “seed” structures that can then easily be expanded into large complex structures.

  13. Posted June 17, 2020 at 1:47 pm | Permalink

    Well, that is way cool. Its as if the inner house in particular is an extension of the larvaceans’ body.

  14. stevenintokyo
    Posted June 17, 2020 at 4:13 pm | Permalink

    That is astonishing. Here I was imagining that any such formation of mucus would be completely random in structure.
    This sense of wonder is so much greater than anything I ever felt about religion, even when I was my young innocent self.

  15. Posted June 17, 2020 at 4:16 pm | Permalink

    Yeah, Hollywood doesn’t have the tiniest fraction of the imagination that it would take to imitate nature’s variety and weirdness.

  16. gscott
    Posted June 17, 2020 at 4:39 pm | Permalink

    Now I’ll be trying to find a way to work the phrase ‘enigmatic mucus structures’ into casual conversation.

  17. chrism
    Posted June 17, 2020 at 5:25 pm | Permalink

    Deep PIV? Oh dear, I’m such a child.

  18. Stephen Wilson
    Posted June 18, 2020 at 10:46 am | Permalink

    What pops into my mind is a tiny spider having a web within a web. The outer web being to catch the mantises that would try to eat it, and the inner web to catch the yummy fruit flies and gnats that make through the strands of the outer web.

  19. Mark R.
    Posted June 18, 2020 at 9:30 pm | Permalink

    And I imagine there is even more to learn from these critters. On and on nature goes and quite beautiful it all is; for some of us humans anyway.

%d bloggers like this: