The Repository Abyss Report: Into the Dark

The Squid Nobody Has Ever Caught

August 27, 2026 · Runtime 16:23 · Watch on YouTube

In 2021 a NOAA robot spent four minutes filming a five-metre squid at 2,382 metres in the Gulf of Mexico. No intact adult of this animal has ever been collected - everything science knows about it comes from video like this. Which makes what the ship's log did with the name a bigger problem than it looks.

Chapters

  1. 0:00 Two white lines cross the frame
  2. 0:19 An animal nobody has ever caught
  3. 0:46 EX2107, dive 10, Gulf of Mexico
  4. 1:16 Two vehicles, and a plan
  5. 1:38 An hour and forty-four minutes down
  6. 1:59 A metre a minute, up a ridge
  7. 2:31 Fourteen specimens, none of them this
  8. 3:01 17:56:34 - "Bigfin squid!"
  9. 3:44 "We followed the tentacles"
  10. 4:11 The fins, the mantle, the arms
  11. 4:30 The bend no other squid has
  12. 5:16 What an ordinary squid looks like
  13. 5:43 About five metres, mostly string
  14. 6:07 Guessing on an open microphone
  15. 6:38 The man who named the family
  16. 7:26 Damaged and/or juvenile
  17. 8:09 Why the arm and the hose cannot take it
  18. 8:43 About a dozen sightings, worldwide
  19. 9:26 4,735 metres, and a submersible in 1988
  20. 9:57 6,212 metres: the first hadal squid
  21. 10:34 "As far as your logging is concerned"
  22. 10:57 The same animal, logged twice
  23. 11:44 Twenty-four of twenty-five
  24. 12:20 What the archive returns
  25. 12:51 The file called BIGFIN_SQUID
  26. 13:28 The index errs both ways
  27. 14:05 There is no jar
  28. 14:43 The right name in the wrong field
  29. 15:17 Four minutes, and back to the coral
  30. 15:51 Next: the dive that forgot the biologist

Transcript

Two white lines cross the frame 0:00

Two thousand three hundred and eighty metres down, off the west coast of Florida, a robot is photographing black coral on a ridge when something crosses the top of the frame. Two thin white lines, almost vertical, moving. Somebody in the control room sees them before anybody else does.

An animal nobody has ever caught 0:19

What they are looking at is an animal that has never been caught. Not rarely caught. Never. There is no intact adult of this thing in any museum on Earth, and everything science knows about it, it knows because a camera happened to be pointing in the right direction. This is a recording of one of the times that happened. It is also a story about what the archive did with it afterwards, which is where this gets uncomfortable.

EX2107, dive 10, Gulf of Mexico 0:46

The dive is E X two one zero seven, dive ten, the ninth of November twenty twenty-one. The expedition was called Windows to the Deep. The ship is N O A A Ship Okeanos Explorer, the vehicle is remotely operated vehicle Deep Discoverer, and the target was a steep ridge on the West Florida Escarpment that the dive plan calls Shark Fin. Twenty-four degrees north, eighty-four degrees west, in the eastern Gulf of Mexico.

Two vehicles, and a plan 1:16

Before any of that, there is a conversation. Two vehicles go down on this system: Deep Discoverer does the looking, and a second platform called Seirios hangs above it on the tether and lights it from behind. How far above is a decision, and it gets made out loud, in the recording, hours before anything interesting happens.

An hour and forty-four minutes down 1:38

The vehicle went into the water at thirteen thirty-one and reached the bottom at fifteen fifteen. An hour and forty-four minutes of falling, which is normal, and which nobody films. Maximum depth on the day was two thousand four hundred and fifty-six metres. Bottom time, four hours and fifty-five minutes.

A metre a minute, up a ridge 1:59

And the dive was not about squid. It was about the ridge. The science team spent those five hours climbing a spine of sediment and rock, cataloguing what grows on it: bamboo corals, black corals, glass sponges standing on spicules driven into the sand, a precious coral colony three metres tall. In five hours the vehicle covered three hundred and twenty-three metres of ground. That is about a metre a minute. This is what deep-sea exploration is actually like.

Fourteen specimens, none of them this 2:31

They also collected, because that is what the arm and the sample drawers are for. Fourteen specimens came up from this dive: two corals, a glass sponge, a loose rock, a polychaete worm, an amphipod, a brittle star riding on a coral, and five bottles of water for environmental D N A. All of it between two thousand three hundred and fifty-seven and two thousand four hundred and sixty-one metres. All of it, in other words, from the bottom.

17:56:34 - "Bigfin squid!" 3:01

At seventeen fifty-six and thirty-four seconds Coordinated Universal Time, somebody watching the feed types two words into the ship's annotation log. Bigfin squid, exclamation mark. Hold that timestamp. We are coming back to it, and the exact wording is going to matter more than anybody in that room could have known. What you are about to hear is the control room, live, from the dive recording. It is public domain, and as far as I can tell nobody has ever transcribed it.

"We followed the tentacles" 3:44

The dive summary, written up afterwards by the science team, describes what happens next in one flat sentence. The tips of the tentacles of a large squid were just visible by the vehicle's cameras. We followed the tentacles. That is the whole trick. Somebody had to decide, in a few seconds, to turn a two-tonne machine away from the rock it was surveying and go up into open water after two white threads.

The fins, the mantle, the arms 4:11

And this is what they found. The two enormous fins at the top are where the name comes from: Magnapinna, big fin. Underneath them is the mantle, the body, and it is small. Then eight arms and two tentacles, and this is the part that has no parallel anywhere else in the group.

The bend no other squid has 4:30

Every other squid you have ever seen holds its arms roughly along the line of its body, trailing behind it like a rocket. This one holds them out sideways, and then they bend. The thick part of each arm goes out at close to a right angle, and the thin part drops straight down from the bend, hanging, for metres. The published description of the posture calls it exactly that: arms spread outwards, filaments dangling at almost ninety degrees. Nobody knows what it is for. The best guess is that the filaments are a trap, and that guess is about twenty years old.

What an ordinary squid looks like 5:16

For comparison, here is an ordinary deep-sea squid, filmed by the same ship, on the same expedition, the following afternoon, a thousand metres shallower. This is Illex, the shortfin squid, and it is what a squid is supposed to look like: a torpedo with the arms held together at the front, pointed the way it is going. The archive has a hundred and nineteen video segments of Illex. Remember that number.

About five metres, mostly string 5:43

Size, because a screen destroys scale. The people on the ship put this individual at about five metres. The largest one ever measured properly, from Australian video, came to one thousand eight hundred and thirty millimetres in total, of which the body was one hundred and forty-nine. So more than nine tenths of the animal is filament. It is mostly string.

Guessing on an open microphone 6:07

Telepresence means a dive like this is watched live by people who are not on the ship, and one of the voices you can hear is calling in from shore. Somebody in the room is working out the anatomy on the open microphone, and then does the thing people do when they are guessing in front of someone who does not have to.

The man who named the family 6:38

The person who answers is Michael Vecchione. In nineteen ninety-eight, he and Richard Young erected this family in the South African Journal of Marine Science, and it is worth knowing what they had to work with. Three specimens. Two juveniles, one of which was recovered from the stomach of a lancetfish. And a paralarva nineteen millimetres long. Not one of the three had grown the arms this family is famous for. Twenty-three years later he is on a phone line, watching a live feed from the Gulf of Mexico, and he says it himself.

Damaged and/or juvenile 7:26

So here is the state of the evidence, and it has barely moved since. The published record on this animal is video, plus a small number of specimens that the literature describes, in a phrase worth reading twice, as damaged and slash or juvenile. Torn fragments and babies. Nothing that would let you put an adult on a table and measure it. Which means the defining feature of the family - the bend, and the filaments hanging off it - has never once been in a jar. It has only ever been on film.

Why the arm and the hose cannot take it 8:09

It is not for lack of trying, and it is not carelessness. It is physics. The tools on the front of that vehicle are a hydraulic arm and a suction hose, and both of them are built for things that stay still: a coral fixed to a rock, a sponge in the sand, a shell. This animal is neutrally buoyant in open water, it is faster than the vehicle, and it is mostly filament of a thickness you can see through. Grab it and you get a piece. Suck it up and you get a smear. The only instrument that works on it is the camera.

About a dozen sightings, worldwide 8:43

Which is why the counting is done in sightings, not specimens. The Australian survey that Vecchione mentions on the recording went through the whole published record and found, before their own work, only about a dozen in situ sightings globally. They then added five, all from the Great Australian Bight, which roughly doubled the tally for the southern hemisphere. Their deepest was three thousand and sixty metres.

4,735 metres, and a submersible in 1988 9:26

The depth record is older than any of that, and it is in the two thousand and one paper, the one that pulled eight of these sightings together from three different oceans. First of September, nineteen eighty-eight. The French submersible Nautile, in the western Atlantic, ten degrees north and forty-one west, in open ocean a long way from any coast. Four thousand seven hundred and thirty-five metres. That number stood as the deepest squid ever seen, of any kind, for thirty-three years.

6,212 metres: the first hadal squid 9:57

Then in twenty twenty-one, in the Philippine Trench, a manned submersible passed a small one swimming near the bottom at six thousand two hundred and twelve metres. That is the hadal zone, below six kilometres, which no squid of any family had ever been recorded in. It extended the depth range for the entire order by one thousand four hundred and seventy-seven metres, about thirty per cent, in a single observation. Do that subtraction and you land exactly on the nineteen eighty-eight record. Same animal, same author, twenty years apart.

"As far as your logging is concerned" 10:34

Back on the Okeanos, two hundred and five seconds into the encounter, Vecchione stops describing and starts dictating. He is not talking to the audience here. He is talking to whoever is keeping the log.

The same animal, logged twice 10:57

And now go back to that timestamp. There are two entries for this animal in the dive log, eighty-six seconds apart, and they are not the same kind of entry. The first, at seventeen fifty-six thirty-four, is from the outreach watch. Vocabulary: education. Comment: bigfin squid, exclamation mark. Taxon field: organism. The second, at seventeen fifty-eight, is from the science lead. Vocabulary: the World Register of Marine Species. Taxon field: Magnapinna, with the full path from Biota down through Mollusca, Cephalopoda, Oegopsida, Magnapinnidae. Both are true. Only one of them is a searchable fact.

Twenty-four of twenty-five 11:44

And that is not a slip on one line. It is how the log works. Two hundred and twenty-five annotations were made on this dive. A hundred and forty-eight use the scientific register. Twenty-five are outreach entries, and of those twenty-five, twenty-four carry the taxon organism and nothing else. Anemone, sea fan, coral, glass sponge, crinoid: every plain-English name on the dive lives in a free-text comment box. Two parallel records of the same afternoon. One for people, one for the index.

What the archive returns 12:20

So ask the archive. The NOAA Ocean Exploration video catalogue is searched by keyword matching on the annotation, and here is what the terms return across the whole collection. Cephalopoda, one thousand and thirty-two segments. Illex, the ordinary squid this same ship filmed the following day, one hundred and nineteen. Opisthoteuthis, the dumbo octopus, eighteen. Magnapinna, six. Bigfin: zero. Bigfin squid: zero.

The file called BIGFIN_SQUID 12:51

There is one more turn of the screw. NOAA cut a ninety-three second highlight of this encounter and put it in the archive under a filename that says, in plain letters, E X two one zero seven, video, dive ten, bigfin squid. Its keywords are highlight and trailer. Nothing else. The one file in the collection actually named after this animal cannot be found by searching for the animal's name, including the name that is in its own filename. It is not lost. It is unfindable, which is a different problem with the same consequences.

The index errs both ways 13:28

And the index gets it wrong in the other direction too, in the same week, from the same ship. The following day, dive eleven at Key West Scarp, a thousand metres shallower. Two of its segments are indexed under Magnapinna. The ship's own annotation log for that dive records no Magnapinna at all. It records an Illex, a dumbo octopus, a gulper shark, and a plastic bag. I sampled both cameras across those five minutes and the animal is not in them. A false negative on Tuesday and a false positive on Wednesday, in the same download.

There is no jar 14:05

For most things in a museum, an indexing error is an inconvenience. You can go back to the shelf. The jar is still there, the label is wrong, somebody re-reads the animal and fixes it. This animal has no shelf. There is no type specimen you can put under a microscope, no drawer, no accession number that leads to a body. There is a video file with a timestamp, a depth, and a coordinate. The archive is the specimen. Which means the index is not administration. It is the only thing standing between recorded and lost.

The right name in the wrong field 14:43

So the rule this dive leaves behind is narrower than it sounds, and harder to act on. Nobody was wrong here. Nobody wrote down a guess, the way the golden orb dive did. The right name was said out loud by the right person, typed into the log within seconds, and it is sitting in the catalogue correctly spelled today. The failure is that the other name, the one every human being who wants this footage would actually type, went into a field that nothing reads. The right name in the wrong field is the same as no name at all.

Four minutes, and back to the coral 15:17

Four minutes of video, from one afternoon in twenty twenty-one, and they are a meaningful fraction of everything our species has ever seen this animal do. The people on that dive knew it while it was happening. You can hear somebody say it near the end, once the animal is gone and the vehicle is being lined up for the next twenty-metre move. They got a data point. And then they went back to the coral.

Next: the dive that forgot the biologist 15:51

Next time: the seventeenth of February, nineteen seventy-seven, the Galapagos Rift, two thousand five hundred metres. The submersible Alvin goes down to look at a temperature anomaly, finds hydrothermal vents, and finds an entire ecosystem living around them that does not run on sunlight. There was not one biologist on board, because nobody expected anything to be alive down there. They had to preserve the first specimens in vodka.

Description and sources

On 9 November 2021, dive 10 of NOAA's Windows to the Deep expedition was surveying corals on a ridge called Shark Fin, on the West Florida Escarpment, when the tips of two tentacles crossed the top of the frame. The pilot turned the vehicle and followed them, and for the next four minutes two cameras recorded a bigfin squid, genus Magnapinna - one of roughly a dozen confirmed in-situ sightings in the world. The published specimens are, in the literature's own words, damaged and/or juvenile. Michael Vecchione, who erected the family in 1998 from two juveniles and a paralarva, was on the phone line, identified it live, and dictated how it should be logged. The dive's own control-room audio runs in this episode, transcribed here for the first time I can find. It also follows what happened to the name: the animal was logged twice, 86 seconds apart, once as "Bigfin squid!" with no taxon, and once as Magnapinna with the full path. Only the second is searchable. Search the catalogue for BIGFIN and you get zero - including the one file named BIGFIN_SQUID. For an animal that exists scientifically as video, the archive is the specimen, which makes the index the only thing between recorded and lost.

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

Abyss Report reads the deep ocean through the record that resolved it - instrument, date, position and the primary sources above. The evidence is the recording, not decoration.

About the material. Every moving shot is NOAA Ocean Exploration video from EX2107 dives 10 and 11, 9-10 November 2021, public domain, credited on screen. The eight passages of control-room audio are the dive's own soundtrack, unedited. Every diagram is drawn in code from the numbers in the sources. There is no generated image in this video at all - not one frame.

A note on position: the catalogue's granule records for this cruise return the longitude without its minus sign. The coordinates on screen come from the ROV dive summary, checked against the published dive track.

#BigfinSquid #Magnapinna #NOAA #DeepSea #OceanExploration