The Repository Abyss Report: Into the Dark

The Type Specimen of This Species Is a Video File, and the Animal Swam Away

August 26, 2026 · Runtime 17:04 · Watch on YouTube

Almost every named animal on Earth has one physical individual set aside in a museum, in a jar, as the standard that decides what the name means. This one has a video file instead — catalogued at the Smithsonian as USNM 1607331. The animal it shows was alive the last time anyone saw it: 10 April 2015, about 3,900 metres down off Puerto Rico. Nobody has seen it since.

Chapters

Transcript

Every named animal has one of these 0:00

There is a drawer in a museum in Washington with a number on it. Inside, in jars, are the animals that decide what words mean. Not a typical example. Not a good example. THE example. When a biologist names a species, one physical individual is set aside and becomes the standard, and every argument about what that name covers is settled by going back to that jar. Almost every named animal on Earth has one of these.

This one is a video file 0:29

This one does not. The standard for this species is a video file, catalogued and accessioned like any other specimen, with the number U S N M one six zero seven three three one. And the animal it shows was alive the last time anyone laid eyes on it. That was the tenth of April, two thousand fifteen, three thousand nine hundred metres down off the coast of Puerto Rico. Nobody has seen it since. It is still down there, and it is still the standard.

Guajataca Canyon, off Puerto Rico 1:00

Start with the place, because the place is on the record. Eighteen degrees, fifty-one point seven minutes north. Sixty-six degrees, forty-eight point nine minutes west. That is inside the Guajataca Canyon, in a feature the survey team called the Arecibo Amphitheater, cut into the slope north-northwest of Puerto Rico. It is about two hours of ship time from a beach with hotels on it.

3,900 metres, four hundred atmospheres 1:27

Now the depth, and this is where the numbers stop being decoration. The animals were about three thousand nine hundred metres down. The dive itself went deeper, to a maximum of four thousand and sixty-two point eight metres. At that depth the water above you weighs about four hundred times the atmosphere you are sitting under right now. There is no daylight, and there has been none for three and a half kilometres.

EX1502L3 dive 1, hour by hour 1:54

The dive has a name and a clock. Expedition E X one five zero two, Leg three, Dive one. In the water at twelve twenty-one and forty-eight seconds, Coordinated Universal Time. On the bottom at fourteen thirty-one. Off the bottom at twenty-one twenty-two. Eleven hours and seven minutes in the water, six hours and fifty-one minutes of it on the seafloor. Every one of those numbers comes from the dive summary, which is a public document.

Deep Discoverer, and the platform that films it 2:25

Two machines went down. The one on the bottom is a remotely operated vehicle called Deep Discoverer, run from the ship by a pilot with a joystick and a screen. Above it, on the same tether, hangs a second platform called Seirios, which lights the first one from behind and films it. That is why there is footage of the vehicle itself. Everything you are about to see was watched live, in a control room, by people who did not know what they were looking at.

15:29:36 — the first one 2:54

Fifteen twenty-nine and thirty-six seconds. Just under an hour into the bottom time, something small crosses the frame, well out in the blue, and the camera goes after it. It is a few centimetres across and it is a long way off. The annotation the team typed at that moment, in the file name itself, is three letters. C T E. Ctenophore. Which is not the same word as jellyfish, and the difference is the reason this episode exists.

A comb jelly is not a jellyfish 3:25

A jellyfish is a cnidarian. It stings. Its stinging cells fire like harpoons. A ctenophore does not sting. It catches food on sticky cells instead, and it swims by beating eight rows of fused hairs called comb rows. Those rows scatter light into a running band of colour, which is what you see flickering along the body. The two groups are not close relatives with a family resemblance. They are separate phyla, at the same level of the tree as us and starfish.

No shell, no skeleton, no fossil 3:58

And ctenophores sit unusually deep in that tree. Genome work over the last two decades keeps returning the same uncomfortable answer: that this group may branch off before sponges do, which would make it the sister lineage to every other animal alive. That is still argued about. What is not argued about is that these are soft, unmineralised animals that leave almost nothing behind. A ctenophore has no shell, no skeleton, and essentially no fossil record.

16:40 — the close approach 4:30

Sixteen forty and twelve seconds. Roughly an hour later. The vehicle is working close to a rock face and the pilot puts up a cloud of sediment, which he apologises for on the open microphone. When it clears, the animal is there again. This is the encounter that produced the footage the description would eventually be built on, and it runs, on and off, for twenty minutes.

Two arms, and where the name comes from 4:56

Here it is close, and the shape is the whole point. The body is rectangular seen one way and rounded seen the other, about six centimetres long, with the comb rows running down it. And out of the front come two tentacle arms, thick, held apart, several times longer than the body itself. Two arms. Hold on to that, because in five years it becomes the animal's name.

It was not swimming. It was anchored 5:22

But watch what it is doing, because it is not swimming. It is holding station. The two tentacles run down and away from the body, taut, anchored to the seafloor, and the animal sits above them, steady, in one place. The people watching it live reached for the same comparison the paper would use later: a balloon on two strings. That behaviour is not in a jar. There is no way to preserve it. It only exists because a camera was running.

Who was watching, from shore 5:52

One more thing about that control room, and it matters more than it sounds. These dives are streamed live, and scientists on shore dial in and talk to the ship while it happens. On this dive, one of the voices on the line belongs to a biologist at N O A A Fisheries, sitting at a desk in Silver Spring, Maryland, watching the same feed you are about to watch.

He asks for the measuring tool 6:16

His name is Michael Ford. Five years later he will be the first author on the paper that names this species. Right now he is on a phone line, looking at a screen, and what he asks the ship for is not a better angle and not a sample. He asks them to turn on the measuring tool.

The dive audio, unedited 6:35

This is the dive audio, unedited. Nothing has been added to it.

Two red dots, ten centimetres apart 6:54

What comes on are two red dots, projected onto whatever the camera is pointed at. They are parallel laser beams, mounted a fixed distance apart on the vehicle, and that distance is the ruler. Here is the ship explaining it, live, while the dots are on screen.

Six centimetres, read off the frame 7:24

Ten centimetres between the dots. Put them next to the animal and you can read its length straight off the frame, which is how the control room arrives at about six to seven centimetres for the body while it is still watching. The published description settles on roughly six. This is the part people do not expect about deep-sea work. The measuring stick is two red dots, and it is the only one you get.

They did not collect it 7:50

So they had it in frame, lit, measured, holding still. And they did not collect it. The manipulator arm stayed where it was. For a scientist that is the expensive decision in the whole story, and it is worth being clear that it was not an oversight.

Why the jar was never an option 8:08

A ctenophore is about ninety-five per cent water, with no hard parts and no skin worth the name. Push one into a suction sampler and it shears. Bring it up through four hundred atmospheres and hours of warming water and what arrives at the surface is a bag of mucus. The thing you wanted to measure, the shape and the arms and the way it held itself, is exactly the thing that does not survive the trip. The jar was never an option.

17:36:58 — the third one 8:37

Seventeen thirty-six and fifty-eight seconds. Late in the dive, on the way out, a third one. And listen to how the control room talks about it now, because by this point they have stopped being surprised and started counting.

Observed three times now 8:52

Three individuals, one dive, six hours apart at most.

Footage, and nothing else 9:09

Then the vehicle came up, the expedition moved on, and what they had was footage. High definition video of three individuals of an animal that nobody could match to a described species. No tissue. No D N A. No specimen. In most of the history of taxonomy that is where the story ends: an interesting sighting, filed, unnamed.

Five years and seven months 9:34

It took five years and seven months. The dive was the tenth of April, two thousand fifteen. The description was published on the eighteenth of November, two thousand twenty, in the journal Plankton and Benthos Research, by Michael Ford, Nicholas Bezio and Allen Collins. Five years to describe a few minutes of footage, which tells you something about how carefully it had to be done.

ICZN Article 73.1.4 10:01

Because the obvious objection is that you cannot name a species without a specimen. And the rulebook says otherwise, in writing, and has for a long time. The International Code of Zoological Nomenclature, Article seventy-three point one point four. Designating an illustration of a single specimen as a holotype counts as designating the specimen illustrated. And then the sentence that decides this episode: the fact that the specimen no longer exists, or cannot be traced, does not by itself invalidate the designation.

The video is not the specimen 10:35

Read that carefully, because it is easy to get backwards. The video is not the specimen. The video is the illustration, and the specimen is the animal in it. The type specimen of this species is an individual comb jelly that was last observed alive on the tenth of April, two thousand fifteen, and was never touched by anyone. The museum holds its portrait, not its body.

What a jar keeps, what a video keeps 11:02

And it is worth being honest about the trade, in both directions. A jar keeps the tissue. You can section it, stain it, sequence it, and go back in fifty years with a technique nobody has invented yet. A video keeps none of that. There is no D N A here and there never will be. But a jar does not keep colour in life, or how the comb rows fired, or the fact that the animal was tethered to the bottom by two tentacles and holding position. That behaviour is data, and it dies in the sampler.

USNM 1607331, and two paratypes 11:36

What actually went into the collection is three video files. The holotype is U S N M one six zero seven three three one. The two paratypes are one six zero seven three three two, and one six zero seven three three three. They are accessioned in the invertebrate zoology collection of the National Museum of Natural History, and they are public. There is a catalogue number, and behind it there is a file.

It was not the first, and it is contested 12:04

This is where the episode has to be careful, because there is a claim that gets made about this animal that is not true. It was not the first species described from an image alone. A bee fly, Marleyimyia xylocopae, was described from two photographs in twenty fifteen, and there are others. And the practice has serious opposition. In twenty sixteen, a position paper in Zootaxa argued that photography-based taxonomy is inadequate, unnecessary and potentially harmful, and it was published with a list of four hundred and ninety-three taxonomists standing behind it.

What is actually unusual here 12:41

What is unusual here is the combination. A deep-sea animal, filmed in high definition by a public research vehicle, described from that footage by the people who were watching it live, with the video deposited in a national museum under a catalogue number. The evidence, the citation and the specimen are the same object. That is not a workaround. That is the case for this kind of footage being treated as a record at all.

The white balance card 13:09

And the record is honest about its own limits, which is the part that convinced me this dive was worth an episode. Twelve minutes after reaching the bottom, the vehicle holds up a colour chart in front of the camera. That is a white balance card. It is how you tell a camera what white looks like down here, where the water has already eaten the red end of the spectrum.

The dive log records its own fault 13:33

It did not work. The dive summary for this dive says so in plain language: the main high definition camera on Deep Discoverer could not be colour balanced properly. That is written down, in the same document that gives the depth and the coordinates. And it is why every frame in this episode leans green. The type material of a species was shot on a camera with a fault the expedition recorded at the time.

incertae sedis 14:00

So where does the animal sit? The paper is unusually blunt about this. The title itself carries the words incertae sedis, which is Latin for we do not know where this goes. The features point towards the order Cydippida, but those thick tentacle arms look like the bottom-dwelling platyctenids, and unlike most of those, this one keeps its comb rows as an adult. A name, a number, and an open question.

What one dive left behind 14:28

One last thing about the dive, since the whole dive is public too. Dive one of that leg produced two hundred and four video segments and three hundred and fifty-nine still frames, all of it in the public domain, all of it carrying the dive number, the date, the depth and the position. The animal that ended up in a museum catalogue occupies a few minutes of that. The rest is sponges, corals, holothurians and rock, and nobody has gone through most of it.

The index nobody publishes 14:59

Which raises a fair question: how do you find one animal inside two hundred and four video files? Not by searching the text, because the text is thin. You find it because the watch team typed a code into the file name at the moment it happened. C T E for ctenophore. S P O for sponge. C O R for coral. R C K for rock. W H T E underscore B A L for the white balance card. That code is primary metadata. The person who typed it was looking at the screen.

The rest of the dive 15:33

And once you can read the index, the rest of the dive opens up. Holothurians on the sediment. A white sponge on a rock face. Bamboo coral. A red shrimp sitting in the open. None of that was the reason anyone remembers this dive, and all of it is sitting there in the public record, with a depth and a position attached, in the same resolution as the animal that got a museum number. The archive is not the highlight reel. The archive is everything the camera saw.

Still down there, still the standard 16:06

The drawer in Washington still exists. There is still a number, and the number still points at one individual, and every future argument about what this name covers will still be settled by going back to it. The only difference is that this one was never opened, never preserved, and never killed. It is somewhere in the Guajataca Canyon, three thousand nine hundred metres down, or its remains are, and it is the standard.

Next: the Bloop 16:33

This episode worked because a camera was pointed at the right thing and the recording survived. Next time the recording is all there is, and there is no camera at all. In nineteen ninety-seven, hydrophones spread across the equatorial Pacific picked up a sound loud enough to be heard on sensors five thousand kilometres apart. It got a nickname, and the nickname got away from the data. Next episode: what the spectrogram of the Bloop actually shows.

Description and sources

On one dive — EX1502L3 dive 1, in the Guajataca Canyon north of Puerto Rico — the ROV Deep Discoverer filmed three individuals of a comb jelly nobody could name. It was never collected, and that was the method, not an oversight: a ctenophore is about 95% water and arrives at the surface as mucus. Five years and seven months later it was described as Duobrachium sparksae, with high-definition video designated as the type material — which the Code allows, because Article 73.1.4 says designating an illustration designates the specimen illustrated, and that the specimen no longer existing does not by itself invalidate the designation. This episode reads the dive: the log, the clock, the two laser dots that are the only ruler down there, the white balance card that failed, and the ship's own audio.

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

Ford, M., Bezio, N. and Collins, A. — Duobrachium sparksae (incertae sedis Ctenophora Tentaculata Cydippida). Plankton and Benthos Research 15(4): 296-305, 18 November 2020. https://www.jstage.jst.go.jp/article/pbr/15/4/15_P150401/_article
ICZN, Article 73 — name-bearing types fixed in the original publication. https://code.iczn.org/types-in-the-species-group/article-73-name-bearing-types-fixed-in-the-original-publication-holotypes-and-syntypes/
NOAA Ocean Exploration — a new species of ctenophore, described from ROV video. https://oceanexplorer.noaa.gov/news/oer-updates/2020/ctenophore.html
NOAA Ocean Exploration — ROV Deep Discoverer, including the paired lasers 10 cm apart. https://oceanexplorer.noaa.gov/technology/subs-deep-discoverer/
Marshall, S. A. and Evenhuis, N. L. — New species without dead bodies. ZooKeys 525: 117-127, 2015. https://doi.org/10.3897/zookeys.525.6143
Ceriaco, L. M. P. et al. — Photography-based taxonomy is inadequate, unnecessary, and potentially harmful. Zootaxa 4196(3): 435-445, 2016. https://doi.org/10.11646/zootaxa.4196.3.9
WoRMS — Duobrachium sparksae. https://www.marinespecies.org/aphia.php?p=taxdetails&id=1468001
Video, frame grabs and dive audio: NOAA Ocean Exploration, EX1502 Leg 3, dive 1, 10 April 2015, NCEI accession 0127387. Public domain.

ARCHIVE PLATES: all 23 are credited on screen in the video and listed in this project's manifest.

The 3 generated plates were made for this channel with FLUX.1-schnell. No person is depicted in any of them.

Abyss Report reads the deep ocean through the dive that recorded it — ship, dive number, date, depth and position on screen, and the primary sources listed above. The footage is the evidence, not decoration.

About the images. All video, all frame grabs and the unedited dive audio are NOAA Ocean Exploration material from EX1502L3 dive 1 (10 April 2015), public domain, credited on screen. The diagrams are drawn in code from the published numbers. Three plates — the specimen drawer, the drawer with a file in it, and the canyon from far off — are AI-generated illustrations of things no camera recorded, drawn as ink on dark paper so they cannot be mistaken for the archive footage beside them.

#DeepSea #NOAA #MarineBiology #NewSpecies #OceanExploration