The Repository Abyss Report: Into the Dark

The Predatory Tunicate: Six Filmed, One Possible Meal

November 1, 2026 · Runtime 22:36 · Watch on YouTube

The predatory tunicate is a sea squirt that stopped filtering: its intake siphon became a pair of lips that close into a hood. We went back to five NOAA dives, from Johnston Atoll to St. Croix, where six of them were filmed close up, measured them against the robot's lasers, timed how long each was on camera, and listened to what the scientists said live.

Transcript

The predatory tunicate 0:00

This is the predatory tunicate: six filmed by NOAA's robot, and one possible meal. A sea squirt is supposed to filter water. This one is called a hunter. So, nineteen hundred metres down, what is it actually eating?

Five dives, three oceans 0:16

We went back to every NOAA dive we could find where one was filmed close up. We measured them against the robot's lasers, timed how long each one was on camera, and listened to what the scientists said live. Every moving picture in this episode is NOAA footage from those dives. Five dives, between twenty fifteen and twenty twenty-two, in three oceans. The ship each time is Okeanos Explorer, and the robot is Deep Discoverer. Two dives off Johnston Atoll in the central Pacific. One in the Mariana Islands. One in the Gulf of Mexico. And one off St. Croix, in the Caribbean.

What a sea squirt is supposed to be 0:57

First, what a sea squirt is supposed to be. On that last dive, five hundred and twelve metres down off St. Croix, the camera finds a few ordinary ones on a rock. The room describes one, and listen for the word they use for its shape.

Two openings. That is the whole plan of the body. Water goes in through one siphon and into a basket called the pharynx, whose walls are full of slits lined with beating hairs, called cilia. A sheet of mucus traps the particles. The water leaves through the second siphon. And the pump is not small. In the shallow sea squirt Ciona, measured in the lab in nineteen seventy-four, the average was fifty millilitres of sea water an hour for every gram of animal. Sometimes much more, and sometimes it stops pumping altogether.

Why filtering fails deep 2:19

That works where the water is full of food. Most of the world's three thousand sea squirts live in the sunlit top of the ocean. But the plankton thins out fast with depth. At a thousand metres, about one percent of what is at the surface. At five thousand, about a tenth of one percent. A sieve that runs all day on water that empty is a bad deal. And the deep sea has an answer it keeps finding, in animals that are not related. A family of sponges did it too. They gave up filtering and started catching small crustaceans with hooks. The sea squirts did it with the mouth.

The sieve is gone 2:57

In the family called Octacnemidae, the intake siphon grew into two big lips, which close into a hood. And inside, the pharynx has no beating hairs on its slits. The sieve that every sea squirt is built around is gone. And DNA points to an ordinary family of filtering sea squirts as the likely ancestor. So the claim is simple. The lips wait open, something swims or drifts in, and they close. That is what the name says, and what the research institutes say. Now here is what the robot actually filmed, dive by dive, starting with the oldest.

Johnston Atoll, dive 04: the worm 3:36

Johnston Atoll, the eighteenth of September, twenty fifteen. Dive four, on a ridge of dark rock, one thousand three hundred and thirty- eight metres down. The camera is on a glass sponge when someone asks about a clear bubble at the bottom of the frame. Look at the left side of the hood. There is a worm on it, with little legs sticking out along its body. The dive report would later write it down as a polychaete worm inside the hood. On the feed, the first reaction is honest.

A predator with a worm in its mouth looks like a predator with a meal. But the room had seen this pairing before, more than once, and drew a different conclusion, right there on the day.

A lodger, not dinner 5:45

Ten years later, in twenty twenty-five, scientists at Scripps described seven new species of this family. In one of them, a worm of the same family was found in the mouth, and their verdict is the room's: likely a symbiont rather than prey. A lodger, not dinner. Two days later, dive seven, a volcanic cone south of Johnston Atoll, nineteen hundred and forty metres down. On the way up the cone, the robot stops on a sponge and a tall bamboo coral, and the room explains why so much here grows tall. It matters for everything that follows.

Johnston Atoll, dive 07: the shrimp 7:08

Food down here does not only fall from above. It travels sideways, on the current, and an animal fixed to a rock eats what the current carries past it. And then, behind the bamboo coral, the room spots something transparent, just to the right of the lasers. This is the animal in our opening. Through the wall of the hood you can see shapes, orange and red, folded inside. The room sees them too, and makes the only claim of a meal in all six encounters.

Maybe even two. That is the strongest evidence of hunting in all of this footage, and it is a guess, made through a wall of jelly. Nothing was collected. Nobody saw the shrimp go in, and nobody saw the lips close on it. One more line from that dive deserves a hearing, because it sounds wrong, and it is right.

Closer to us than to corals 9:24

Sea squirts are chordates, the group that includes fish, and us. The adult hides it. The larva does not. In Monterey, scientists raised the local species of this genus in the lab. Within two to four days, the eggs became swimming tadpoles, which could drift for up to three months before settling and growing a mouth.

Mariana Islands: the sock puppet 9:47

April twenty sixteen. The Mariana Islands, a volcanic caldera called Fina Nagu, two thousand seven hundred and sixty-eight metres down. The dive is almost over when the robot finds the clearest one of the six. The room is delighted.

Between the laser requests, and just before the robot has to leave, one of them says out loud what a lot of people feel about this animal, the first time they see one.

Then they ask how big it is, and the pilot puts the lasers on it. Deep Discoverer's two red dots are ten centimetres apart. Listen to the estimate that came back.

Nineteen centimetres, not ten 11:44

Slightly over ten centimetres. We measured the same seconds of video. With both dots on the front of the animal, the dots are a hundred and twenty-four pixels apart, and the body is about two hundred and forty wide. That is about nineteen centimetres. Nearly twice the guess. For comparison, the new species described in twenty twenty-five measure from about two centimetres to fourteen, from the base of the stalk to the outflow siphon. And the Monterey institute lists the genus at up to twenty-six centimetres tall.

April twenty eighteen, the Gulf of Mexico, on the Florida Escarpment, about two thousand eight hundred and ninety metres down. The rock here is stacked in thin, flat layers, and the geologist on the feed reads them out loud.

Gulf of Mexico: the misspelled family 12:18

Rock that formed in warm, shallow water, now almost three kilometres down. The camera moves on to a bamboo coral, and then someone spots a hooded tunicate and tries to remember the name of the family.

Octacnemidae, similar to the genus Megalodicopia. And whoever typed the annotation that day spelled the family with an extra letter. Search NOAA's video catalogue for the correct name, and this dive never comes up. We found it only through the genus. Here the lasers land right at its base, and it is out of focus, so our number is rougher. Eighty-five pixels for ten centimetres, and about a hundred and eighteen across the body. Roughly fourteen centimetres.

St. Croix: the first of the expedition 14:29

And August twenty twenty-two, off St. Croix, five hundred and ninety- three metres down. Just before the tunicate, the robot stops on a glass sponge, and the sponge researcher explains how it feeds. Listen, because this is filtering, at work.

The robot moves on, along the foot of a ledge. Under it, there is a pale bag, dusted with sediment. Somebody asks what that is. Something blobby. Someone on the call asks the sponge researcher what she thinks. The answer that comes back puts this animal next to sponges.

Sponges and sea squirts, two groups that are nowhere near each other on the tree of life, both turning from filtering to catching in the deep. And then she gives the reason in one phrase.

Capture larger prey, and be fed for longer, in a food-poor place. A minute later the lasers come on, at its base. Seventy-seven pixels for ten centimetres, and fifty-eight across the bag. This one is about seven and a half centimetres. The smallest of the six.

And the robot is not the only one nearby. Right above the tunicate, a crab is holding up a piece of glass sponge.

The geologists' tunicate 18:22

Two hours later, eighty metres shallower, a second one. By then the biologist is at lunch, and two geologists are watching the feed. They spot it in the corner, while arguing about two crinoids.

Two minutes later, with a good close-up on the screen, the two geologists get to the most important detail of all, on their own, with no biologist in the room.

How rare it is on NOAA's video 19:25

So, how rare is this animal on NOAA's video? We searched the whole catalogue for the family and the genus. They are attached to footage from about twenty dives, over nine years of expeditions. And on the very dive we just watched, the room called it the first one they had seen on that expedition. Compare that with Monterey Canyon, off California, where the Monterey Bay Aquarium Research Institute has been diving for decades. There, in a study published in twenty oh six, this genus turned up in seventy-one of a hundred and thirty-six survey dives. More than half.

Six animals, zero bites seen 20:03

Now add up our six. About ten minutes on the main camera in total. Between five hundred and twelve and two thousand eight hundred and ninety metres. Seven and a half to nineteen centimetres. Lips opening or closing while the camera watched: zero. Something inside: one, maybe.

What the papers actually show 20:24

That is not unusual. It is the state of the science. The study from Monterey says it directly: the shape of the mouth suggests that every genus in this family is a predator, although direct observations are few. The hunting is read from the anatomy. And when someone finally looked inside, the answer was mixed. In twenty oh two, a Japanese team opened animals collected in Toyama Bay. The guts held small crustaceans, yes. But also diatoms, which are single-celled algae, and detritus. Whatever drifted in.

Not a hunter, a trap 21:00

A second test, in twenty oh five, read the fats in the animal's tissue, which carry a signature of what it eats. Their conclusion: both a suspension feeder and a scavenger. And the twenty twenty-five paper sums it up: an emphasis on small particles, with occasional opportunistic big prey. So here is the answer to the question we started with. The predatory tunicate is not a hunter in the way the name sounds. It does not chase and it does not choose. It lost the sieve, and turned its mouth into a trap big enough for a shrimp, but it keeps whatever the current brings.

Exactly what is missing 21:40

And that is exactly why it works down here. A sieve only catches what is small, and the small things are almost gone. A pair of lips catches the small things and the rare big one. In water this empty, the animal that refuses nothing is the one that eats. What is still missing is simple to say. A clean video of the lips closing, timed frame by frame. No paper we found has one. And a look inside the stomach of one of these six, which no report we read describes. Until then, the shrimp from Johnston stays a maybe.

Next: the tar lily 22:16

Next time, the place. In twenty fourteen, in the Gulf of Mexico, NOAA sent this same robot to a target on the seafloor that looked like a sunken ship. Within minutes, it was clear it was not man-made. The scientists called it a tar lily. That is the next Abyss Report.

Description and sources

Six predatory tunicates (family Octacnemidae) filmed by ROV Deep Discoverer from NOAA Ship Okeanos Explorer: EX1504L4 dives 04 and 07 (Johnston Atoll, September 2015, 1,338 and 1,940 m), EX1605L1 dive 05 (Mariana Islands, April 2016, 2,768 m), EX1803 dive 10 (Gulf of Mexico, April 2018, about 2,890 m) and EX2206 dive 08 (St. Croix, August 2022, 593 and 512 m). Against the robot's 10 cm lasers we measure three of them at about 7.5, 14 and 19 cm. In about ten minutes on the main camera, no hood opens or closes; one, on Johnston dive 07, holds what the room read as one or two partly digested shrimp. The worm in the dive 04 hood was called a commensal on the day, as a 2025 paper concludes for a similar worm. Gut contents and fatty acids say mostly small particles, with occasional large prey.

PRINT-READY, FROM THIS CHANNEL

Read This Dive, Vol. 01 — the method, and five dives read with it

https://therepository.gumroad.com/l/read-this-dive

PRIMARY SOURCES

NOAA Ocean Exploration, ROV dive summaries: EX1504L4 dive 04 (https://repository.library.noaa.gov/view/noaa/19874) and dive 07 (https://repository.library.noaa.gov/view/noaa/19877); EX1605L1 dive 05 (https://repository.library.noaa.gov/view/noaa/21303); EX1803 dive 10 (https://repository.library.noaa.gov/view/noaa/21411); EX-22-06 dive 08 (https://repository.library.noaa.gov/view/noaa/49989)
Havenhand, J.N., Matsumoto, G.I. & Seidel, E. (2006). Megalodicopia hians in the Monterey submarine canyon: distribution, larval development, and culture. Deep-Sea Research I 53(2): 215-222. doi:10.1016/j.dsr.2005.11.005
Mandre, P. & Rouse, G.W. (2025). Molecular phylogeny of the deep-sea predatory Octacnemidae (Ascidiacea), with seven new species. Diversity 17(12): 859. doi:10.3390/d17120859
Okuyama, M. et al. (2002). Morphological studies on the bathyal ascidian, Megalodicopia hians Oka 1918, with remarks on feeding and tunic morphology. Zoological Science 19(10): 1181-1189. doi:10.2108/zsj.19.1181
Kurabayashi, A. et al. (2003). Zoological Science 20(10): 1243-1247. doi:10.2108/zsj.20.1243
Ogawa, A. et al. (2005). Inferring the feeding habit of the deep-sea 'big mouth' ascidian tunicate, Megalodicopia hians, by fatty acid analysis. Aquatic Ecosystem Health & Management 8: 185-193
Kustin, K. et al. (1974). Water transport rates of the tunicate Ciona intestinalis. Biological Bulletin 147(3): 608-617. doi:10.2307/1540744
Vacelet, J. & Boury-Esnault, N. (1995). Carnivorous sponges. Nature 373: 333-335. doi:10.1038/373333a0
MBARI, Predatory tunicate: https://www.mbari.org/animal/predatory-tunicate/
Next episode: NOAA Ocean Exploration, EX1402L3 dive 12 summary (WR0325, 24 April 2014): https://oer.hpc.msstate.edu/okeanos/ex1402l3/ex1402l3-DIVE12-SUMMARY-FINAL.pdf

ARCHIVE PLATES — public-domain documents, with author, date, archive and licence.

2015-09-18 | NOAA Ocean Exploration | public-domain
2018-04-27 | NOAA Ocean Exploration | public-domain

Abyss Report reads the deep ocean through the record that resolved it -- instrument, date, position and the primary sources above.

About the material. Every moving frame is NOAA Ocean Exploration video from the five dives named above, public domain, credited on screen; St. Croix depths come from the vehicle's 1 Hz track. Several cuts play the original control room audio; each window was re-transcribed in isolation twice and used only where both passes matched. The laser measurements, the time on camera and the count of dives in NOAA's video catalogue annotated with the family or genus are ours. Every diagram is drawn in code.

#DeepSea #NOAA #OceanExploration #Tunicate #Okeanos