The Chilling Mariana Station Mystery

I've spent 11 years mapping junk that doesn't belong on the ocean floor, and most of it turns out to be ordinary once you get close enough, a container that went off of freighter in a storm, a trawler that went down with nobody aboard, a wreck somebody remembered from a story their grandfather told them, which usually turns out to be half the size and twice as old as they said it was. My name's Dale Lasseter. I work on the Rhea, a private survey and recovery vessel that spends most of its time in places nobody sees. We find things on the seabed, sometimes people hire us because something valuable was lost, map the wreck, run sonar over the bottom, put the submersible down when there's something worth looking at, and bring back whatever were legally allowed to bring back. My job is mostly the part in between. I've spent enough time around sonar and submersibles to know what the ocean looks like when you're seeing it through a machine instead of your own eyes, and to know what bad data looks like, what loose sediment does to a return, what clean steel does, how a shape can fool you from the wrong angle. The ocean doesn't give you much for free. After a while you stop getting excited about a strange mark on a screen. That morning Cora Pickett got excited, and that's rarer than it sounds. She has the sonar, and she won't say a word about a return until she's run the line three times. I've never known her to break that. I was standing behind her at the console when she said that's weird and leaned in and moved her coffee out of the way to do it. We were surveying the edge of a deep canyon a little under 4,000 meters down, running parallel lines since before sunrise and building the bottom one pass at a time. The screen was mostly what you'd expect out there. Slopes of sediment, rock, scattered debris. Then there was the shape. It was large, too large to look natural from the sonar, with straight lines on it, hard corners, and irregularity the seabed doesn't have. I asked what I was looking at, and she said that was what she was trying to figure out. She brought up another display in the return sharpened. The thing sat on the bottom like a building, except it wasn't exactly a building, no roofline, no obvious windows or entrances, nothing familiar. It was symmetrical, and it was enormous. My first thought was a wreck, a very unusual wreck. I said ship, and she said it didn't look like one, and ran the scan again. The outside of the structure came back strong, clean, easy to read, the way steel comes back. Then Kora frowned and told me to look underneath it. Below the center of the thing there was nothing, no measured depth, no second surface, no bottom, no blank patch, maybe. I asked if it was a bad angle, and she said maybe. I asked about sediment, and she said it could be. She checked the raw feed and got the same thing. I've seen sonar do strange things before. You get a steep wall, and the return can stretch out into a smear. You get a pocket of soft sediment, and the computer loses the bottom entirely. That's why nobody down here trusts one pass. We trust repetition. Kora ran another line from a slightly different angle. The structure came back. She called the software a mess and brought up the depth estimate, and it was clean everywhere on that structure except directly beneath the center, where it gave us nothing at all. I asked if she could get the raw data. She said she already had. When I asked what it told her, she looked at me and said she still didn't know. That was enough to get my attention, not because I thought we'd found something impossible, because Kora knew that equipment better than I did. She zoomed the image out. The object was enormous. Harold Pruitt was already coming down from the wheelhouse. He asked what we had, and Kora told him a large artificial-looking structure on the bottom. He asked the depth, and I said just under 4,000. Then came his usual pair, historical or commercial, always in that order, and I told him it could be either. He rubbed his chin and told us to send finch. Roscoe stallings came down with me to get ready. He's been my partner on enough dives that we barely need to talk before getting into the sub, which doesn't stop him talking. He narrates every item of the pre-dive out loud, whether or not I'm answering. Batteries, lights, coms, manipulator arms, all the things we check every time he ate a sandwich while he did it. The descent takes better than an hour. At first there's still some color outside the viewport, and then the blue goes dark, and then there's nothing at all except the lights on finch. Roscoe watched the depth and called it out for 3,000, 3.5. Then the bottom came up the way it always does, not all at once, a little ridge of sediment first, then a slope, then rock, and I slowed her down and took her in toward the numbers Korra had given us. The sonar hadn't prepared me for the size of it. The beam from our lights went up into the dark above the thing and didn't come to anything, a wall, then another wall, a support as thick as a car, a whole section of metal buried almost entirely in sediment. Up close the surface was covered in years of growth, barnacles, and pale organisms, and sediment packed into every seam, and whatever this was, it had been down there a long time. We moved along the exterior, and the further we went, the more obvious the construction became. Metal plates, vaulted sections, reinforced joints, the kind of heavy industrial work that gets done when the people paying for it expected to be standing in a hundred years. I stopped the sub beside a large panel, half covered in mud with something stamped into it, and Roscoe brought the light around. Numbers first, then lettering, and the lettering was Cyrillic. I couldn't read all of it, but I recognized enough to know what we were looking at, and so did he. Soviet. There was no sign anywhere on it that it had been destroyed, no collapsed section, no impact damage, no torn open hull. It hadn't been abandoned because something smashed it, it had been left there. Eventually we found an entrance, a circular docking opening set into the side, and the outer door was standing open. That was the first thing that felt wrong. The second was what was behind it, which was a corridor. The opening was large enough to take finch with room to spare, and the bay inside it was flooded, but at the back of that bay there was an inner hatch, and when we put the lights on it, it was sound. Doggid shut and holding on a station that had been sitting at 4,000 meters since before I could drive. And then the bay pumped itself down, which I didn't believe would work, and which took 19 minutes. At the end of it there was a dry steel room behind that hatch with air in it nobody had breathed since the 1980s. The difference was immediate. Outside everything was buried and covered with life. Inside it looked like time had stopped. The corridor was narrow and industrial, pipes overhead in a metal rail along the wall. Doors on either side of us, some standing open, and some closed. We moved slowly and let the lights find things one piece at a time. A room with bunks in it, a table, storage lockers, a work area with the tools still arranged along a wall, each one hung in its own painted outline, a mess room with cups and a rack. Everywhere there were signs of people having lived and worked in that place, and there were no people, no remains, nothing that explained where anybody had gone, and none of the equipment damaged. The facility had simply been left. The station went deeper than we'd expected and it was bigger too. Not a small military outpost, but a full installation, preserved in the cold, Soviet signage on the walls that Roscoe held his phone over as we passed. He keeps a translator on it that runs without a signal for port work, which is the only reason it was any use to us down there. And I kept coming back to the same thing. Why anybody would leave all of it? We found the internal power system in another section, most of it dormant, and the system itself intact. Roscoe went over the controls and said powers still here. I said looks like it. We followed it to a larger compartment with a heavy door at the far end, and behind that door was another level of machinery. The reactor sat behind its barrier, untouched. It wasn't damaged and it hadn't failed, and there was a bullet. open on the desk beside it, with a column of entries in it, and a signature at the bottom of the last one. And something went through me that wasn't fear yet. It was the understanding that we hadn't found an abandoned building. We'd found something somebody had built and used, and then deliberately turned off and left, sitting on the bottom of the ocean. Rosco asked why anybody would abandon this. I looked back at the dead panel and didn't have an answer for him, but it was the first moment since we'd found the structure that I was glad we'd come down with the cameras running, because whatever had happened here, somebody had gone to a lot of trouble to make sure it stayed in the dark. We spent the next few hours going through the reactor, and the first thing we established was that it could be started again. The systems around it were old, but the shutdown had been orderly. No evidence of a catastrophic failure, nothing that looked damaged. The safety systems still in place. The procedure had been completed rather than interrupted, right down to the last line in that book and the signature under it. Somebody had turned this off, and then they'd left it that way. Rosco stood at the panel looking over the gauges and said so they didn't lose power. They shut it down. We looked back at the shielding and asked why. There were still too many parts of that station we hadn't reached. Doors were locked, whole systems were dead, and the records were behind both. With eight hour sets on our backs and to place the size of a city block to understand, that's why we decided to restart it. Not because we knew what the reactor powered. We didn't. Rosco was against it. We wanted to go back up, tell the office, and come down in a week with people who knew reactors. What I had on my side was the state of the thing. I know the difference between a machine that has failed and a machine that has been parked. Eleven years of looking at the bottom of the ocean through a machine, and it had finally handed me something that wasn't a container or somebody's grandfather's story. The restart wasn't the business people imagine. I closed the breakers. After that, the station handled the start-up itself, stepping through its own checks one at a time. The first systems came back slowly. A light came on in the control room, and then another one. Emergency lighting appeared down the corridor outside, red and dim after so many years. Then, fitting after another, going away from us into the dark, pumps started somewhere deeper in the station. Ventilation began moving air through old ducts, and forty years of dust came out of the vents in a gray stream and hung there in the light of my helmet. For a few minutes, it felt almost ordinary, and abandoned station was becoming an operating one. Then the warning system activated. A speaker somewhere down the corridor crackled and a damaged recording started repeating, seven or eight seconds of a man's voice in Russian, and then a break, and then the same seven or eight seconds again. The tape was in bad condition. Some of the words were clear enough to identify as words, and the rest of it swam under the static. Rosco held his phone up to the grill for a long time. The app caught the word for personnel twice, and something that might have been "do not remain" and might have been "do not remain" below. And it never gave us the same version of it, twice running. He asked me if I could make any of it out. I said no. We left it running and started checking the newly powered sections. It was a dead man's voice on a loop, and after twenty minutes I'd stopped hearing it. That's when the vibration started. It came up through the floor underneath us, a low mechanical movement, strong enough that I felt it in the souls of my boots. Then it stopped. Then it came again. Rosco said "reactor," and I shook my head and said it didn't feel like a reactor, which was me talking about something I know nothing about. The third one was longer than the other two, and it wasn't coming from the room we were standing in. It was traveling through the structure from somewhere below it, so we followed it. That's the thing about power coming back, doors that had been dead all afternoon had green lights on them, and one of them opened onto a stair going down into a part of the station we hadn't known was there. The air got colder. The vibration got stronger the further down we went, until you could feel it in your teeth, and Roscoe stopped talking, which he doesn't do. The chamber at the bottom was bigger than everything above it put together. There was a shaft in the center of the floor with a rail around it and heavy machinery standing over the shaft, going up into the dark past where our lights reached, and when we got to the rail and looked down, the drilling mechanism went away into the earth further than we could see. Real structures all around the shaft, hydraulic equipment, the size of shipping containers along both sides, racks of pipes stacked against the bulkheads in their hundreds, thick lines and conduit running from the machinery back into the station. Roscoe stood there with his light on it and said, "This is a drill." We moved around it trying to understand the layout, and the scale made one thing obvious straight away. The installation hadn't been built with a drill in it. The installation had been built around the drill. Everything we'd walked through upstairs, the bunks, the mess room, the tools in their painted outlines, was accommodation for the men who ran this thing. I followed one of the big conduits back toward the main distribution board. It went to the reactor. The power was being routed straight into the drilling system. The reactor was feeding it, and the string was already turning. Roscoe went to the controls and said, "Shut it down." He worked through the controls, and the drill didn't stop. We found the local shutdown at the shaft head, then the emergency interrupt behind a plate on the wall, and then I got it into my head that I could beat it at the distribution board. Half an hour of pulling feeds took out half the engineering section and nothing else. The shaft has its own supply of the reactor, upstream of anything a man can reach. It wasn't waiting for an operator. It had come up when the reactor came up, run its own checks the same way the reactor had, and entered an automatic sequence. That was the first time I understood what we'd done. We hadn't restored power to an abandoned station. We'd restarted something. He said we needed the operating records and we found them in a cabinet by the control position, which is where anybody would put them. Most of it was technical. Depth measurements, operating reports. Drilling stages, sheet after sheet of one man's handwriting and neat columns. He started following the records down through the history of the project. The drill had reached a specific depth, then it stopped. That number turned up more than once. It was on the last sheet. It was on the panel in front of us, and it was written on a card taped to the frame of the control position where the operator would see it every time he sat down. 3,150 meters below the sea bed. The Soviet operation had gone down to that point and no further. Roscoe found the specification for the machine and the same cabinet. He looked at the specification, then at the recorded depth, then back at me, and he said it could go farther. I looked at the diagram. He was right. The string and the racks around us, the rams, the whole installation was built and rated to take a hole down to 9,000 meters. Nearly three times what they'd done. They'd stopped with two-thirds of it still in their hands. I asked why anybody would build all this and stop there. Roscoe shrugged at me and went back into the cabinet. There had to be a reason somewhere in it. Eventually we found another entry in the drilling records. It wasn't a maintenance instruction. It was a warning, and the same line was on the card taped to the control frame. And Roscoe read the translation out twice because the first time neither of us believed it. Do not continue below the established mark. The wording was absolute and there was nothing attached to it. No explanation, no mechanical reason, no note about structural instability, nothing about the ground, just the instruction. Roscoe read it again to himself and asked if that was it, and I said it looks like it. Everything else in that station was written by engineers full of tolerances and permissions and cool and cool. agreements. That one line had all of it stripped off. He looked toward the shaft. The vibration was still coming up through the floor of the room we were standing in. They hadn't stopped because they couldn't go farther. We went through the rest of the material and found no explanation. Whatever had happened to the Soviet project had happened below the level where they stopped drilling. Something made those men draw a line and then they walked away and left the machine sitting there fully capable of crossing it. Roscoe was back at the panel. He said we're still going down. The depth display said he was right. The automatic cycle was running. The bit wasn't cutting anything yet. It was tracking back down the hole those men had left through ground that had already been drilled at about 90 meters an hour. And the number on the panel was climbing steadily toward the one written on the card. 2,410. 2,415. 2,420. Roscoe did the arithmetic out loud. 8 hours near enough and then it would be at the bottom of their hole with nothing underneath it but the ground they wouldn't touch. The drill didn't stop and by then I'd given up expecting it to. Whatever the Soviet system had been built to do when the power came back, it was doing it now and it was going to go on doing it whether we stood there and watched or not. So we went back to the records. Neither of us wanted to stand at that rail and let it go down without knowing what those men had found. There wasn't enough desk space at the control position so the technical documents ended up spread across the floor of the room in Rose and we worked through them one at a time, checking the depth measurements against the operating reports. The deeper the records went, the stranger they got. At first there was nothing unusual in them at all. Depth, pressure, drilling speed, composition of the rock coming up. Roscoe read a column of it over my shoulder and said it was the tightest set of drilling records he'd ever seen and he's worked rigs so that counts for something. Then the measurements changed. The later sheets had a second column of readings that didn't fit anything geological and Roscoe stopped over one of the reports and asked what that was. The line he was pointing at carried a set of figures beside the recorded depth. I didn't understand it. Neither did he. We found the same readings on another report than another and they turned up every time the drill got into the deeper section. Something had been detected below them. It wasn't described as rock. It wasn't described as a cavity or a void or a fault. The word those men used for it was signal. The records were clear about what it wasn't. It wasn't radio in any normal sense and it wasn't a seismic event and it wasn't the ground ringing. One man had spent two pages ruling out his own equipment before he'd write a single word about anything else. The signal was simply there and it was coming from below and as the drilling got closer to its source the signal got stronger. We went back through the sheets and the pattern held everywhere. The nearer the drill the stronger the transmission the farther away the weaker it was. Roscoe read one of the reports twice and said they knew. The language in the earlier reports was careful, unknown source, unknown transmission, possible instrumentation anomaly, further testing required. Then farther in that language changed because they'd isolated the thing and there was nothing left to be careful about. In the reports from the deepest stage of the project they referred to an object, not a geological formation, not a machine they'd built. The app gave us the word as artifact, no theory about where it came from, nothing anywhere in it that so much is reached for an explanation. They established only what those men could actually determine that it was artificial, that it was not of human construction and that it was transmitting. Just the object buried far below them sending something outward. I've read a lot of survey reports and I've never seen anyone work that hard to write down only what they could prove. We went through the measurements again and that's where the last piece of it was. The surrounding geological material was interfering with the transmission. The farther the signal traveled through the earth the weaker it became. That was why the drill mattered. The scientists hadn't simply been digging toward the artifact. At some point in the project they'd worked out that every meter of hole changed the strength of it. They never found out what the signal carried or who it was for or what would happen if it got any louder. The later reports stopped being about discovery at all, containment, non-interference, preservation of existing conditions, no further drilling and then the same line again at the end of it, the one off the card. The machine was running the whole time we read that. The reactor was feeding it, the number on the panel was climbing and we were doing the one thing those men had spent their last months on the job telling somebody not to do. We came off the sets at about four in the morning because we ran out of the other option. They're rated eight hours and we'd been in there longer than that. I sat with the air meter we'd carried in and watched it make up its mind. Oxygen 20.4%, carbon dioxide higher than you'd want and nothing like dangerous, the rest of it ordinary. The ventilation had come back on with everything else and the plant had been scrubbing for hours. Somewhere in the hours after that the place stopped being comfortable to be in. It started with a pressure behind both eyes, low down the kind you get from a long day at a screen, then a taste in my mouth like I'd been holding a coin in it. The meter kept telling me the air was fine and I kept checking it three times in an hour and then I'd go and find something to do with my hands. What I wanted was the decimator. We'd carried one in from the start because of the reactor and if that thing had started ticking we'd have had a reason for the headaches, a reason to leave, and a story that ended with two men getting out because the radiation alarm went off. It read normal all night. I wanted one instrument in that station to tell us to leave, not one of them would. At 10 past nine the bit reached 3,150 and stopped running and started cutting. The sound of it changed. Everything up to then had been the string going down through open hole and now it was a machine with its weight on the ground and it came up through the structure into the floor and the rail and the deck plates. The number on the panel held for a moment at 3,150. I thought it might stop there. It didn't. 3,151. 3,152. Nothing happened. No explosion. No collapse. No failure. No judgment out of the dark. The machine put its teeth into the ground. Those men wouldn't touch and went on cutting. It cut for nine hours at about four meters an hour. The station shook harder, the deeper the whole got. Nothing broke. No alarm sounded anywhere in the place. At some point in the afternoon, Rosco wiped his nose on his cover all and went on reading and when I said something about it, he said it was the dry air and asked me to hold the light steadier. Then at eight minutes past six in the evening, with the panel reading 3,186, the vibration changed. For a second it felt as though the whole station had gone still around us. The drill broke through. The drill went through and the vibration stopped and my whole body kept moving for another second after it did. I had one hand flat against the console and I could feel the muscles in my arm still braced for a shake that wasn't there anymore. My teeth were clenched. I didn't notice that until my jaw ached. Rosco was breathing hard beside me and didn't say anything and the room had gone so quiet that I could hear the little fan inside the monitor housing. Then he asked what happened. The depth reading had stopped, so had the drill. Not because anything had broken and not because anything had hit it. The machine had simply come to the end of its sequence and stood still 36 meters below the line those men had drawn. Rosco brought up the camera on the drill head. The screen came on and I expected rock and instead there was black. For a second I thought the camera had failed. Then the image sharpened and I could see everything I should have been able to see. The broken rock around the opening, the edges of the hole, the housing itself. wall of the last few meters we'd cut through. And past that, nothing. The light from the camera went down into the opening and disappeared. It didn't land on anything, it just kept going. I told him to switch cameras, and he did, and we got the same thing. He moved the head slightly, and the rock moved with it, and the black state exactly where it was. He said the cameras fine, my mouth had gone dry enough that I had to work at it before I could get a word out. We turned the work lights up, all of them, and the beam went into it, and nothing came back at all. Roscoe said it's not a cave, I said no. He looked at me, and then back at the screen, and asked what it was, then I couldn't tell him. He started checking the other readings. I watched his hand while he did it. It was shaking just slightly, so was mine. I put it flat on the console, the metal was cold, and I left it there. He checked the depth and got nothing. He said come on under his breath, and try it again, and it gave him the same. All of it worked perfectly well everywhere around the opening. That's what frightened me, and it wasn't the dark. Dark was normal. We worked in dark all the time. Dark meant there was something there, and you couldn't see it, and everything I'd done for 11 years was about putting an instrument into the dark and getting a number back. There was nothing for the instruments to find. I thought about the sonar on the first day. That blank patch underneath the center of the structure, the one Cora had called an artifact of the processing, and I'd agreed was probably sediment. I'd looked at it for 40 seconds and stopped thinking about it. The sonar had seen this. It hadn't been failing, and it hadn't been confused. It had run its lines across that piece of seabed and reported correctly that there was no bottom there. Roscoe said look. I followed his finger to the screen and didn't see it. And then I did a tiny point of it. So far away that my first thought was a dead pixel or a fault in the housing, and I blinked and it was still there. I switched cameras myself, which meant a different sensor and a different feed, and the light was still there. There was no way to tell how far away it was, nothing to judge it against, no other lights anywhere, no sky, one point on its own. Roscoe turned the head slightly, the light didn't move. He turned it again and it stayed exactly where it had been, and he said that's not ours. Then he shut off the work lights, all of them, and the screen went almost completely black, and the point was still sitting in the middle of it. I stared at it until my eyes hurt. He was breathing through his mouth by then, and he had one hand on the edge of the desk, and when he did answer he said there's nothing around it. I looked again, and he was right. Something around it tells you there's distance or dust, or the black between them that still reads his space. There was none of that, just black, and the light. I made myself look away from the monitor and got down among the Soviet records. The pages were still spread across the floor where we'd left them, and I got down and ran my finger along the line I'd read four times already. All that rock between the thing, doing the transmitting, and whatever lay in the direction it was transmitting toward, and the Soviets had stopped before they took too much of it out. We hadn't. I wiped my palm on my leg, and said maybe that was what they had been afraid of. Roscoe looked at me and didn't say anything. He looked back at the black. Neither of us knew whether that was true. The records don't say it, but we were looking through the place where the ground had been, and there was something there, that little light. I wanted to back away from the monitor. I leaned closer instead. The black looked endless. It couldn't be endless because there was a light somewhere inside it, and I couldn't decide which of those was worse, that it went on forever, or that it didn't. Roscoe said quietly that we should get out. I said yeah. Neither of us moved, the drill was still sitting at the opening. The reactor was still running, the lights were still on. Everything on our side of that boundary was doing exactly what it had done all day. My own breathing. The cold of the console under my hand. The pressure holding at one atmosphere in a steel room four kilometers down, the way it had held for forty years. All of it real, all of it measurable, all of it doing what it was supposed to do. On the screen there was no rock, no water, no space we could put a number on, just a boundary, and something passed it that we couldn't get a single instrument to touch. I don't know how to say this next part so that it sounds like anything other than a man talking nonsense in a hole in the ground. I had the thought and didn't say it, and I think he had it at about the same time. He was standing there staring at the same screen, and it stopped pretending to check readings. Maybe we'd found the end of our world. Like whoever built this universe never planned on anyone digging this deep. A boundary line. We were out of bounds, staring into a place that was never meant to exist. And if the world had an edge, did that mean none of this was real? That we were living in a simulation? A cold, sick shiver went right through me, just letting the thought into my head. And we stood there in that room with the lights of a Soviet drilling station burning around us and the fan going in the monitor housing, looking at one point of light that had nowhere to be. We stayed there longer than we should have. One minutes maybe, an hour. Time went strange in that room because I kept waiting for the black to change, and it didn't. Roscoe stood up eventually and walked away from the monitor. He got as far as the edge of the room and stopped, and I watched him put both hands on his knees and breath. I asked if he was all right. He nodded. Then he said no. I was in no better shape. My shoulders ached from however long I'd been holding them up around my ears. My jaw was sore and my right hand had gone numb from keeping it flat on the console. I pulled it off and flexed the fingers and they were shaking. He came back and we checked the opening again. Nothing had changed. I've spent 11 years working in water you can't see through, and you get used to the ocean being empty in a way that isn't really empty. You can't see the bottom, but the bottom is there. You drop something and you know it's falling toward something and however deep it is, there's a floor at the end of it. You can always picture a place underneath you. There wasn't an underneath. There was an edge. Roscoe put his hand out over the rail then and dropped a piece of scrap down the hole. It was just a broken section of liner off the deck about the size of my hand and we watched it on the camera. It went through the opening and out passed the rock into the black and it went on going. It should have got smaller. I've watched a thousand things go away from that camera and every one of them got smaller. That one stayed the same size. It got further away and it stayed exactly the same size until the camera lost it and it never arrived anywhere. Roscoe didn't say anything about it. I didn't have anything to say either. He asked if the light had moved. It hadn't. We went back through the Soviet records one more time and found nothing in them. We hadn't found the first time. Those men stopped before the hole reached it. He'd done the opposite of that inside a day, restored the reactor, restarted the drill, crossed their mark. Now we were standing in the place they'd refused to reach and we still didn't know a single thing more than they had. The first time I looked at it I thought there was nothing there. Now I knew there was something and it was on the other side. I looked around the chamber after a while. The station felt small, the walls, the pipes, the lights, the rail, the drill standing over its shaft, all of it parts of one sealed little place and I'd been walking around inside it for a day and a half thinking it was the world. I put my hand against the bulkhead. It was cold and solid and it didn't move when I pushed on it and I pushed harder than I needed to. I don't know why I needed that, I wanted something I could trust probably. Roscoe had gone back to the gauges. He said everything on this side is normal, pressure normal, temperature normal, all of it. We were both quiet for a while after that. That was when the leaning started. We were both at the rail by then and I was standing with my weight over my left foot, correcting the way you do on a deck in a swell. Except the rear was 4 km above us and the station hadn't moved since the drill stopped. I straightened up, a minute later I was doing it again. Roscoe was doing it too in the other direction, which is why I said something. He stood up straight and then he took a nut off the top of one of the pipe racks and crouched down and set it on the deck plate, careful, and took his fingers off it. It didn't roll. It sat there on the plate where he put it. We both looked at it for a while and then we both straightened up again. We were both still leaning. Roscoe said we need to leave. He looked at the drill. It was still operational and nothing we could see was stopping us going back the way we'd come. We should have been relieved. I wasn't. The thought of going back up through the station turned something over in me. The surface had stopped meaning what it meant that morning. Above us was the ocean. Above that, the sky. Above that, everything I'd spent my whole life assuming was there, Mars, distance, space, the whole of it, and I'd just seen a place where there wasn't any more of it. The point of light was where it had been all evening, small on its own. I couldn't stop wondering what it was, not what it meant, or what it proved, or what anybody would do about it, just what it was. I wanted to see it closer and I caught myself working out how much line we had on the winch before I'd finished having the thought. My legs went when I did that. I'd been standing with my knees locked and I hadn't noticed and I had to bend them and wait for it to pass. Roscoe saw it. He looked at me for a second and turned back to the controls. We started getting ready to leave. There wasn't much left to say. The station was still around us. The black was still down there under all of it. And then the simple question came to me. It would have been meaningless that morning. In that room, it took me a moment to work out why I was having trouble getting it out. How do we get back up? Roscoe was quiet for a moment. He was looking at the black beyond the drill. Then he said, "We don't. Why?" Because I don't think there is enough anymore. We've reached the end. I'm Rex. I know time is the most valuable thing you have and the fact that you chose to spend an hour of it here in the dark with me. I don't take that for granted. If you want to be here for the next one, make sure you're subscribed. Until then, stay safe. Take care.