Catching a fish is a solved problem. You drop a net in the water, the fish swims in, you haul it up, and whatever lands on deck still looks like a fish. Deep-sea biology was built on that exact move, and for animals with bones and shells it works fine.<\/p>\n
But try it on a jellyfish living half a mile down. An animal that is mostly water, arranged into delicate structure, does not survive being dragged through a few thousand feet of ocean in a mesh bag. What reaches the surface tends to be soup, and for roughly 150 years scientists have been describing new deep-sea species from that soup, or giving up on describing them at all.<\/p>\n
Which is what makes an expedition off Brazil worth your attention. An international team working from the research vessel Falkor (too) sent a 7,000-pound tethered robot called SuBastian into the open water of the tropical South Atlantic, fitted with laser imaging gear that can reconstruct a gelatinous animal in 3D right where it lives. In about two weeks, the team confirmed 31 species science had never described, according to the Schmidt Ocean Institute<\/a>, which operates both the ship and the robot and announced the results in early June.<\/p>\n Most of those animals would not have made it to the deck in one recognizable piece. So the robot went to them instead.<\/p>\n The nets go back to HMS Challenger, the British ship that spent the mid-1870s dragging them through every ocean it could reach and more or less founded oceanography in the process. The method stuck because it works, as long as the thing you catch is sturdy.<\/p>\n The problem is where the sturdy things aren’t. Between the sunlit surface layer and the seafloor sits the midwater, the single biggest living space on the planet and the least explored one. It is packed with jellyfish, siphonophores, comb jellies and other gelatinous drifters whose entire body plan assumes nothing will ever squeeze them.<\/p>\n A net squeezes them. So the animals that dominate the largest habitat on Earth are the exact animals the standard tool destroys on the way up. Formally describing a new species from damaged material can take decades, when it happens at all.<\/p>\n The April expedition, led by Dr. Karen Osborn of the Smithsonian National Museum of Natural History, was designed to attack that problem directly. Instead of bringing the animals to the lab, it bolted the lab onto a robot.<\/p>\n SuBastian is a remotely operated vehicle rated to 4,500 meters, or nearly 15,000 feet of depth. Per the official spec sheet<\/a>, it weighs 3,200 kilograms in air, about 7,000 pounds, packed into a frame 2.7 meters long. That is roughly the footprint of a Smart car carrying the weight of a heavy-duty pickup, and it tops out at 3 knots, so figure walking pace.<\/p>\n It carries 4K cameras, sonar, sensors and a 440-pound payload allowance for whatever instruments a science team wants to strap on. For this cruise, the key additions were two imaging systems built by the Bioinspiration Lab at MBARI, the Monterey Bay Aquarium Research Institute: DeepPIV and EyeRIS.<\/p>\n DeepPIV projects a laser sheet into the water and uses it to measure how fluid moves and to capture the three-dimensional shape of transparent animals, slice by slice, without touching them. Paired with EyeRIS, the setup let the team build detailed 3D reconstructions of animals at depth, alive, doing whatever they were doing when the robot found them.<\/p>\n When the scientists did want physical material, they had gentler options than a net. The robot carried insulated sample boxes for fragile animals and a folding, twelve-sided sampler that closes around a creature to take a small tissue sample. Back on the ship, a genetics team led by Dr. Cheryl Ames of Tohoku University and Dr. John Burns of Bigelow Laboratory sequenced genomes from collected specimens on board, which is how identifications that normally crawl through years of academic review got confirmed in days.<\/p>\n One more first happened at the microscope bench. Using a confocal scope developed by Dr. Manu Prakash’s lab at Stanford, the team observed the living 3D cellular structure of a microbe while still at sea, something the institute says had never been done on a ship before.<\/p>\n The 31 confirmed species break down like this: one amphipod, a small crustacean cousin of crabs and lobsters; one gossamer worm from the genus Tomopteris that swims faster than its body shape says it should; nine jellyfish; seven siphonophores, which are colonies of specialized units operating as one animal; seven comb jellies; four larvaceans; and two giant rhizarians, single-celled organisms big enough to spot with the naked eye.<\/p>\n The larvaceans deserve a second look. They are tadpole-shaped animals that build disposable filter houses out of mucus, and they are more closely related to you than to the jellies drifting past them. Four of the new ones were labeled by Russ Hopcroft, a professor of oceanography at the University of Alaska Fairbanks, and per UAF<\/a> they measured between about half a centimeter and two centimeters.<\/p>\n Run the arithmetic and 31 species in roughly 14 days comes out to a new one about every 11 hours, around the clock, for the entire cruise. The team also recorded plenty it wasn’t naming, including glass squid and a female seven-arm octopus, Haliphron atlanticus, wrapped around a bright red jellyfish and eating it at a depth of 800 meters, about 2,600 feet.<\/p>\n A larval fish turned up on camera at 966 meters. A jelly collected at 1,157 meters was photographed on board with a prototype macro camera system. None of this is seafloor work, either. It all happened in open water, the part of the ocean where there is nothing to stand on and nowhere to hide.<\/p>\n The midwater is “filled with incredible animals we are only just starting to understand,” Osborn, the expedition’s chief scientist, said in the institute’s announcement. This was the third time the broader project had taken its midwater tools to sea, after expeditions in 2019 and 2021 that first tested the technology in open water. Two Ocean Shot research grants from the Sasakawa Peace Foundation, through programs at the University of Western Australia and Bigelow Laboratory, funded the current push, and Bigelow’s own writeup<\/a> is worth a read if you want the full instrument rundown.<\/p>\n The distinction matters. SuBastian works on a tether, a cable that feeds it power and carries live video up to a control room on Falkor (too), where human pilots fly it in real time. That cable is why scientists on the ship could watch, live, while an octopus ate a jellyfish half a mile down.<\/p>\n A true underwater drone runs untethered on batteries and its own judgment, like the Orpheus AUV that surveyed the abyssal plain near the Mariana Trench<\/a> or the wave-powered gliders that cross entire oceans with no engine at all<\/a>. Saab builds a machine called Sabertooth that can operate either way<\/a>, tethered or free. Different tools, different jobs.<\/p>\n Schmidt Ocean Institute now owns both kinds. In late 2025 it added a Kongsberg Hugin Superior autonomous vehicle rated to 6,000 meters that can stay down for as long as 72 hours, and that unit ran its first science mission in May and June of 2026, mapping a fracture zone in the middle of the Atlantic. Its name is The Childlike Empress, because this organization names its hardware after characters from The NeverEnding Story: the ship Falkor (too), the robot SuBastian, and now the Empress. Somebody over there is having a wonderful time, and honestly, good for them.<\/p>\n As of mid-August, Falkor (too) is in the Caribbean on a monthlong study of an ocean-mixing quirk called salt fingers, running through September 2. The 2026 schedule<\/a> then sends the ship across the equatorial Atlantic from Brazil toward Ghana, exploring a fracture zone under a UN-endorsed effort called Challenger 150, named for the anniversary of the very ship that started the net era. In December it heads back to seamounts off northern Brazil, with the calendar running into January 2027.<\/p>\n Nets are not going anywhere, and for a lot of ocean science they remain the right tool. But describing the deep sea’s most fragile animals from whatever survives the ride to the surface now has a working alternative, and it came back from Brazil with 31 entries in the record to prove it.<\/p>\n","protected":false},"excerpt":{"rendered":" Catching a fish is a solved problem. You drop a net in the water, the fish swims in, you haul … <\/p>\nDeep-sea biology has had a net problem since the 1870s<\/h2>\n
A 7,000-pound robot with a laser scanner, flown from a ship on a cable<\/h2>\n
So what did two weeks in the water actually buy?<\/h2>\n
One thing SuBastian is not: an underwater drone<\/h2>\n
The robot is booked into 2027<\/h2>\n