The Secret Life of Lobsters

by Trevor Corson

  The dominant male waited in his shelter, peeing out the door of his apartment at the females who came calling. A female would poke her head in and pee back at her prospective mate, a love potion in her urine suppressing his bellicosity and putting him in the mood for courtship. He would stand on tiptoe and pulse his swimmerets, drawing her urine in and fanning it appreciatively about the boudoir.

  The scents of his dominant masculinity and her seductive femininity would mingle, Jelle supposed, and waft out the back door of his shelter, like an olfactory billboard posted in his backyard. With this billowing advertisement to the females of the neighborhood that his love nest was active, it hardly seemed surprising that a dominant male might develop a sense of confidence.

  14

  Against the Wind

  For all of Jelle Atema’s efforts, he had yet to truly enter the lobster’s Umwelt—to sense the world as a lobster did. The problem was that chemical signals were a messy business. Eyes detect light that mostly travels in straight lines at constant velocity. Ears detect acoustic waves that disperse through a medium according to well-established rules. But smell requires the detection of patches of molecules carried by ever-changing eddies inside chaotic plumes of moving air or water.

  The first time Jelle had ever dropped a chunk of fish into a lobster’s tank, he’d been amazed by the speed with which lobsters could pinpoint an odor source. At the first sign of an attractive scent they flicked their antennules and began to move, walking slowly at first and adjusting their heading with an accuracy that was nearly instantaneous. After a few seconds they were jogging toward the origin of the odor, whether it was a tasty morsel of food or an alluring lobster of the opposite sex. How did they do it?

  The lobster’s nose—its pair of antennules—is perhaps a less remarkable organ than its eyes, but the antennules are far more useful underwater. Attached to them are hundreds of sensory hairs with permeable walls. Inside each hair is a dense tuft of some four hundred grasslike neural cells that are attuned to particular combinations of molecules. If a lobster is walking into a current, the animal aims the antennules straight up like rabbit ears, so that water washes against the full length of the sensory-hair array. But if clear reception is interrupted by confused currents or imperfect orientation, the lobster flicks its antennules downward in swift strokes to obtain a stronger signal. Thus the analogy to sniffing.

  Detecting waterborne chemicals is one thing; tracking them to their source is quite another. The question of how lobsters use their antennules to locate the origin of seemingly random splotches of whirling odor had dogged Jelle for decades. He’d spent hours crouched over a drainage gutter in the floor of his basement lab, squirting dye into the flow of seawater in a vain attempt to detect patterns in the plume of swirls. It was an impossible task because he couldn’t see the plume from a lobster’s point of view.

  That changed when Jelle learned of a research project under way in Colorado. A neuroscientist in Denver was working on ways to repair the neurotransmitters of patients with Parkinson’s disease. He had developed a miniature electrode for detecting the neurotransmitter dopamine in rat brains. The electrodes were about the same size as the receptors on a lobster’s antennules. Jelle guessed that a pair of these detectors might be able to “see” a water current containing dopamine the same way a lobster “saw” a current containing an odor.

  Jelle began to collaborate with the neuroscientist and ran an initial trial in his flume tank. He injected a jet of dopamine into the current from a nozzle at the head of the tank. Then he positioned one of the electrodes at sixty successive locations across a grid downstream, recording the fluctuations in dopamine concentration at each location. As water flowed through the tank, it dispersed the dopamine in an increasingly turbulent plume of cascading eddies. But this time Jelle wasn’t watching the plume with human eyes. Instead, he was seeing what a lobster standing at different locations in the tank would sense with its antennules—a series of changes in chemical concentration.

  Neurological studies in Jelle’s lab had determined the detection frequency, response speed, and acclimation rates of the chemoreceptor cells in a lobster’s antennules. Using this data, Jelle calibrated the recordings from the dopamine detector to match the lobster’s own sensory abilities and graphed the results. The effect was stunning. The odor landscape inside a turbulent plume looked to a lobster something like a mountain range of chemical peaks, each following the next in time. Near the source, those peaks would be tall and steep because the patches of chemicals passing across the lobster’s antennules were dense and had clearly defined edges. Farther away from the source, or off to the side, the peaks the lobster would perceive were by degrees shorter and gentler in slope, and spaced farther apart, because the patches of chemicals had become fuzzier and more diffuse.

  Dye swirling in a current might appear chaotic to the human eye, but after several hundred million years of tracking odors underwater, a lobster inside a turbulent odor plume surely felt right at home. Most of Jelle’s colleagues thought the idea was crazy, but Jelle believed a lobster might be capable of pinpointing its own location in relation to the distant source of the scent simply by the look of the chemical slopes and hills in its immediate vicinity.

  The fact that lobsters had not one but two antennules, spaced a body width apart, was also essential to tracking odors. That was all too clear when Jelle snipped one antennule off a lobster and it started walking in circles. One of Jelle’s students set about constructing a lobster backpack that contained a submersible amplifier and two dopamine electrodes. When the backpack was attached to a lobster, one electrode sat directly behind each antennule. The lobster was blindfolded and lowered into the downstream end of the flume tank, and a brew of dopamine and squid extract was squirted from a jet upstream. A cable from the lobster’s backpack supplied a computer with the electrode readings while another cable supplied the computer with a video feed of the lobster’s movements, filmed by an overhead camera. As the lobster tracked the scent of squid up the tank, the computer synchronized the chemical and visual data. The scientists could see the pulses of odor the lobster was experiencing on the right and left sides of its head while it was deciding which direction to turn. As expected, the lobster turned toward the antennule that detected steeper and higher hills of odor before the other antennule did, enabling the lobster to “climb” the mountain range of the plume to the source, the animal’s trajectory growing ever more accurate as the steepness of the peaks increased.

  Perhaps it was inevitable that the next experiment would involve mounting two odor-release nozzles on the lobster itself, one pointing at each antennule. Now that Jelle understood how a lobster perceived an odor landscape, it was a simple matter to generate lobster virtual reality. A hungry lobster could be steered through an empty tank at will by squirting bursts of fish-flavored seawater at one antennule or the other.

  And after lobster virtual reality, perhaps it was inevitable that lobster artificial intelligence would be the next hurdle. When the diminutive automaton was complete, poised at the downstream end of Jelle’s flume tank, it had 256K of RAM and its name was RoboLobster.

  The hurricane warning crackled over Bruce Fernald’s radio aboard the Double Trouble two days before the end of August 1996. The storm was a monster and approaching Little Cranberry Island quickly. Of Bruce’s eight hundred traps, a third were still sitting in shallow water near shore, where a gale could beat them into tangles of wire and twine.

  Already, one of those traps had nearly killed him. Bruce had been breaking in a new sternman aboard the Double Trouble that year—a young man who’d never worked on a boat before. Two weeks before Bruce’s forty-fifth birthday, the fellow had lost his grip on a brick-laden trap at low tide and let it fall fifteen feet from the wharf to the boat. The trap had missed hitting Bruce by a yard. Bruce was glad to have escaped injury, not least because catches were still on the rise and he couldn’t afford to sit out the height of the trapping season. But with
a hurricane on the make, Bruce would have to remove his gear from shallow water to ride out the storm.

  The next two days passed in a blur of coiled rope and stacked traps. Aboard the Double Trouble Bruce and his sternman, like worshipers of some Pharaoh of lobsters, built pyramid after pyramid of wire-mesh rectangles as they hauled the dripping traps from the water and lugged them in boatloads to land. Little Cranberry’s fleet of old pickup trucks came and went from the wharf in a chorus of squeaky springs as the fishermen carted their traps up the road.

  The day the hurricane was to arrive a stiff breeze raced across the harbor. The men spent the afternoon hauling small boats out of the water and battening down equipment on the co-op wharf—electric scales, hundreds of wooden lobster crates, dragnets, and plastic bait trays. After checking the mooring lines to the lobster pens, the lobstermen rowed their skiffs into the harbor and double-checked the mooring lines to their lobster boats, which were too big to pull onto land on short notice.

  When the rain came crashing down in leaden sheets across the harbor and the ocean frothed white out of the west, the fishermen knew they were inside the leading edge of the hurricane, and there was nothing more they could do.

  Still in their rain slickers and dripping wet, they congregated in the bar at the end of the restaurant wharf to watch the storm come. It was the final day of the restaurant’s summer season, when the owners held their customary closing night for the islanders—no tourists allowed. Leftover beer would flow for free until the kegs ran dry. Clutching pints of Harpoon ale and Budweiser, the fishermen sat with their backs to the bar, gazing out through the windows while the rising tempest buffeted the wharf on its pilings and pulled their pitching boats tight on their mooring lines.

  One of their fellow fishermen had been running last-minute errands on the mainland, and they wondered whether he would attempt to return to the island. Not long afterward, through the gray dusk they noticed his boat approaching from the north, plumes of spray flying. The fishermen stood up from their barstools and strode to the rattling windows, their pints of beer forgotten, while a part of them, a descendant of the great cod fisherman Sam Hadlock and one-fifteenth of the whole that made up the island fleet, plunged laboriously toward home. He was half a mile away in jagged peaks of surf.

  Suddenly the boat turned east and headed out to sea. They knew what the man was doing. He was afraid the boat might roll on its beam and swamp, so he would try to surf the waves downwind on an angle, find a trough where he could gun the engine and spin the boat back to the southwest, and pound upwind into the oncoming walls of water, splitting them with his bow and making a slow, zigzag kind of progress until he found the harbor. It took him half an hour to cover that half mile, but he managed to reach his mooring and put the boat on its tether. When he walked into the bar he was soaked to the skin, but he had a big grin on his face. A cheer went up and a beer made its way into his hand.

  Little known to the general public, in a nondescript building on the northern fringe of the campus of the Massachusetts Institute of Technology is a small laboratory where research has been funded in part by the U.S. Navy. Inside, scientists have constructed torpedoes that can all but think for themselves. They are called AUVs, or autonomous underwater vehicles, and they disappear into the sea and carry out missions without remote control. Oceanographic research and oil exploration are among the civilian applications of these devices, but close cousins of these machines now assist the U.S. military in amphibious assault operations as well. During Operation Iraqi Freedom several were dropped overboard in the port of Umm Qasr. On dives that could last twelve hours or more, they swam free on their own recognizance, hunting antiship mines.

  Thomas Consi, a member of the AUV group of researchers at MIT, was a biologist by training but he liked having little robots walking across his desk. After a day designing intelligent torpedoes Tom would stop by a toy store and hunt for inanimate objects he could bring to life. When MIT hosted its first Artificial-Intelligence Olympics, Tom fielded a toy army tank that he’d taught to follow a beam of light.

  The term for this sort of work, and for the projects under way in the AUV lab, is biomimetics—the mimicking of natural physical and behavioral structures using artificial devices and algorithms. The goal isn’t only to create useful robots but also to gain insights into the biology of living organisms. Some of Tom’s colleagues were constructing a beast called RoboTuna. The purpose of the project was to understand the complex hydrodynamics involved in how a real tuna swims.

  When Tom heard about Jelle Atema’s plume-tracking project with lobsters, he invited Jelle to MIT, and the concept for RoboLobster was born. Working alongside the intelligent torpedoes in the AUV lab, Tom and a colleague machined, assembled, and programmed the device, and soon the lobster-sized robot had made its way into Jelle’s nine-foot flume tank in Woods Hole.

  RoboLobster sat poised like a jet fighter on a runway, ready to attack the oncoming current. His watertight body was a shiny cylindrical hull that housed an onboard computer with a 20-megabyte hard drive and sixteen AA batteries. Like an airplane, RoboLobster’s fuselage was topped with blinking red lights. A pair of direct-current motors powered his little rubber wheels and provided steering. Protruding straight up from RoboLobster’s head were two stainless-steel wires—a pair of metallic antennules that served as conductivity electrodes. Fresh water was running through the flume instead of salt water, and instead of dopamine or fish juice, RoboLobster would be tracking the scent of a salt-and-ethanol mixture injected into the current.

  The brain of a real lobster consists of several nerve ganglions strung together, and nearly half of their volume is dedicated to processing the signals collected by the animal’s sense of smell. RoboLobster’s brain was far simpler, and the only thing he had to process was smell. He was programmed to turn right or left depending on which electrode detected a higher concentration of salt in the water, to travel in a straight line if the concentration was similar on both sides, and to move backward if he lost the scent.

  Inside the flume RoboLobster managed to track down the source of the salt about 25 percent of the time. Compared to the swift efficiency of a lobster, the paths RoboLobster traveled were torturous. Still, RoboLobster was a start, and Tom was proud of him. RoboLobster was one of the first biomimetic automatons to function successfully underwater. And what was perhaps most startling was how “biological” RoboLobster’s performances looked. When plotted on paper, the robot’s tracks reflected more the fluid complexity of nature than the programmatic code inside RoboLobster’s microchip head. The implication was clear. A real lobster didn’t have to be very smart to find its way around inside the currents in the ocean. It just had to be equipped with the proper detectors.

  What the experiments suggested to Jelle was that a robot equipped with additional sensors for detecting the direction and speed of water flow might nearly match the skills of a real lobster. No actual animal tracked odor without reference to the movement of the medium containing that odor, be it air or water. The hairs on many parts of the lobster’s body detected touch or motion, so water flow was clearly a constant part of the animal’s sensory experience. Jelle guessed that a lobster’s antennules were involved in flow detection as well.

  Jelle and a student snipped the antennule off a live lobster and clamped the antennule upright inside a sort of wind tunnel constructed from a section of clear plastic pipe. They flooded the pipe with water and, while taking high-speed film of the antennule through a microscope, ran oscillating water currents through the pipe at different frequencies. The scientists discovered that different parts of the antennule itself resonated with different frequencies, like a guitar string of varying pitch. Jelle had been an avid flutist for decades—once he’d even played a tune on an old lobster claw for an audience at a meeting of the American Association for the Advancement of Science. For Jelle, the notion that lobsters might be not just smelling but, in a sense, listening to the symphony of the ocean’s current
s was entrancing.

  Elsewhere the quest for a robotic lobster had taken a more sinister turn. The U.S. Navy was now considering plans for a beachhead assault that would begin with thousands of biomimetic lobsters dropped offshore from low-flying aircraft. Clambering over rocks and sniffing their way through currents toward shore, the lobster robots would search out mines and blow themselves up on command. Soon the Pentagon was funding robotic-lobster research to the tune of several million dollars.

  The work was carried out in a lab at the end of a narrow peninsula on the northern coast of Massachusetts, accessible only by a two-lane causeway. There, a bank of computers belonging to Northeastern University’s Marine Science Center analyzed video feeds of lobsters walking on treadmills in glass tanks. Once lobster motion was translated into computer code it could be downloaded to microprocessors and fed via electrodes to artificial leg muscles made of nickel-titanium alloy. Fleets of thousands of robotic lobsters scurrying across the seafloor could have civilian applications as well as military uses, such as patrolling for pollution.

  And who knew—if the natural lobster population was being overharvested and the fishery had to be shut down, perhaps the federal government could task New England’s lobstermen with catching and disarming explosive automatons over which the navy had lost control. It wouldn’t be that much more dangerous than what they already did for a living.

  PART SIX

  Brooding

  15

  Gathering the Flock

  The old ship, a thousand tons of steel seesawing slowly on the undulating sea, groaned with the weight of the trawl. The twin electrohydraulic winches hummed with the strain of six thousand feet of metal cable. Scientists from the National Marine Fisheries Service waited under the yellow crane in the stern, bundled in orange overalls and rubber boots. They weren’t sure what the net would haul up. The aft deck of the R/V Albatross was a thousand square feet, and the trawl could pile it ankle-deep with spiny redfish. Or it could splatter out the occasional cod, huge flat halibut, or hideous hake with its stomach popping out of its mouth from the change in pressure. The Albatross had once caught a streamlined gray object that looked like a shark. It was the fuel tank of an F-16.