Note: I am aware of this question Could two planets be tidally locked to each other so close they share their atmosphere? My question is not about that. It is specifically about human travel.

By a cosmic accident, Slartibartfast's design, or through God's will there is a planet that is shaped like an old-fashioned dumbbell. enter image description here

Each lobe is Earth-sized and Earth-like in terms of their basic composition - at least on the surface. What lies under the surface may be assumed to be whatever is necessary to hold the structure together. You can assume a degree of hollowness if that helps with the plausibility.

The 'bar' joining the lobes is mainly a natural form of steel. It connects directly to the core of each lobe. Its proportions are exactly as shown in the picture. The bar is the same length as the diameter of the lobes.


Human life develops on one lobe, either by evolution or by being created.

Under the normal laws of physics:

Would the humans be able to walk from one lobe to the other or would it be like climbing and then descending an enormous mountain? If they can't simply walk - what stage of technology would they need in order to cross the bridge? How would they actually do it?


1. Please feel free to discuss factors like: shared atmosphere, the most stable axis of rotation, the effect on seasons, on tides and on day and night etc. However my prime interest is whether humans can cross the bridge and what level of technology it would take.

2. If you need to presume a particular internal structure for the planet to make it more plausible, then please do so and say how that will affect the answer to the travel problem.


The answers so far (29/07/15) are very useful. However I feel that maybe not enough attention has been given to the gravitational pull of the bridge itself. If this picture is accurate with regard to relative size of Earth and Moon, then I imagine the gravitation due to the bridge will be at least equal to that of the moon - probably a lot more. It's certainly not negligible. This would mean that, at the half-way point between lobes, humans could indeed walk around without floating off. Presumably they could colonise the area. Maybe they could mine it and simply throw the product down towards either side. I imagine they could use it as a platform for space-travel.

Wouldn't the proximity of the two lobes, combined with the gravitational pull of the bridge mean that it could maintain at least a tenuous atmosphere?

enter image description here


Thanks to comprehensive and convincing answers so far, I am now resigned to using large amounts of unobtanium in the actual construction of the planet.

The joining rod is coated in steel that has magnetic properties but the steel is coated with thick flaking rust. The surface underneath is pockmarked to some degree according to the weather at any give altitude. The bar is supported on the inside with unobtainium scaffolding. This is 'plated' with 1000 metres thickness of steel. In theory one could drill through to the hollow interior. There are holes in the unobtainium scaffolding that are big enough to allow very large items of machinery to be introduced once the steel has been broached (perhaps even as big as a moderate sized trawler).

My request now is to provide the earliest possible technology (by our Earth date) that would allow travel between lobes.

Clearly hot air balloons wouldn't come anywhere near, whereas rocket science would suffice. The question is, How far back in standard Earth history could we go and still be able to make the journey?


If you suggest welding, then please give evidence of when the necessary sort of welding was invented.

If you suggest magnetic vehicles, please say when magnets would have been strong enough for the purpose.

If you suggest breathing apparatus, please say when it was invented or at least could have been invented with materials available at the time.

If stone-age man could have made breathing apparatus out of flint (unlikely!) then that is okay.

The winner will be whoever comes back offers and substantiates the the oldest possible technology.

Please feel free to amend current answers or add new ones.

Good luck and thank you.

  • $\begingroup$ If dumbbell planet isn't orbiting a star the bridge will be squashed in no time, if orbiting a star the tug-of-war between star and the 2 planets will cause the bridge to collapse, so the question isn't about crossing the bridge is whether there is any bridge to cross😛 $\endgroup$
    – user6760
    Commented Jul 28, 2015 at 23:00
  • $\begingroup$ Cross-posted to Physics. Please don't do this. $\endgroup$
    – HDE 226868
    Commented Jul 29, 2015 at 13:55
  • $\begingroup$ It would have some unusual properties like moons that grow. hou.usra.edu/meetings/lpsc2014/pdf/2319.pdf $\endgroup$
    – King-Ink
    Commented Jan 27, 2016 at 7:57
  • $\begingroup$ cosmic accident, Slartibartfast's design, or through God's will I can't see a difference between each. $\endgroup$ Commented May 30, 2018 at 14:26

7 Answers 7


Climbing the bridge will be a job for the Vacuum Scouts.

But first, we need to assume that the two lobes orbit each other in such a manner that the connecting rod is not under tension or compression. To do otherwise requires real unobtanium to handle the stresses involved. If the OP is edited to provide a different outcome, fine. Until then, orbits it is.

Given Earth-like lobes, the connecting rod will be 8000 miles long. Let's ignore the gravitational component of the rod. The two centers of the two lobes will be separated by 4 R (Earth radii). The COM (Center of Mass) will be at the center of the rod, with the center of each lobe 2 R apart. The attraction between the lobes will be $$F_G = \frac{GMM}{(4R)^2}$$ and the centrifugal force on one lobe from the COM is $$F_C = M\omega ^2(2R)$$ so $$F_G = F_C$$ and $$\frac{GMM}{16R^2} = 2MR\omega ^2$$ and $$\omega ^2 = \frac{GM}{32R^3}$$ Then $$\omega = \sqrt {\frac{6.73\times 10^{-11} \times 5.97\times 10^{24}}{32(6.37\times 10^6)^3}} = .0022$$ Finally, the orbital period T is $$T=\frac{2\pi }{\omega} = 2855 seconds$$ or about 48 minutes. This is going to make for a pretty short day.

Not quite what you had in mind, you say? You want a 24-hour day and #%#! the unobtanium? Can do. I live but to serve. The rod is going to be under humongous compression, and the crusts of the two planets will have to be massively reinforced with more unobtanium in order to keep them from failing under the stress, but that's all in the budget, I presume.

So now it's time for the Vacuum Scouts. The upward gravitational attraction from one lobe will be felt at the surface of the other, but since the distance from the surface to the center of the other lobe is 3 times the distance to the center of the lobe you"re standing on, the upward pull will be 1/9 the downward pull, and surface gravity at the base of the rod will be about .89 standard. This will reduce air pressure by about 10%, to the equivalent of about 2500 feet above sea level if the rod is anchored at nominal sea level. Not a problem. However, just as the pressure is not much affected much, neither is the atmospheric gradient. The Death Zone will be a bit lower, but not much.

The Vacuum Scouts are outfitted, basically, with space suits and arc welders, and carry with them a bunch of ladder rungs which they weld to the rod about every foot or so. Each suit is equipped with a harness which attaches to the ladder rungs. They also carry a pulley with a rope which is twice as long as is needed to reach the ground. A ground crew is camped at the base of the rod with about 16,000 miles of rope and 84 million (16,000 times 5280) ladder rungs. At regular intervals the ground crew splices on some more rope, and hoists more rungs upward to replenish the welders' supply.

Once the welders are about 5 or 6 miles up (the equivalent of Everest) they have to be suited up at all times, and by the time they've climbed 100 km or so they're in vacuum. As they climb, gravitational pull will gradually decrease, and will reach zero at the midpoint. It will then increase in the other direction as they approach the other lobe. The rope will, nonetheless, need to be very strong and light.

Oh yes, and as the climbers climb, and the rope gets longer, and the amount of weight carried by the rope increases, it will get harder and harder for the climbers to reposition their pulley upwards, as they essentially have to lift the entire load to do so. This is left as an exercise to the reader, although setting up staged intermediate pulley systems would seem the way to go. A lot of intermediate pulley systems, each with a crew to transfer rungs and other supplies to the next system in line.

Assuming the welders install 1 rung per minute and work 12-hour days, it will take them awhile to do the job. How long? It the time is T, $$T = \frac {8000\times 5280}{60\times 12} = 117,000 \text{ days} = 321 \text{ years}$$

But of course, nobody would be dumb enough to join the Vacuum Scouts, so another approach needs to be tried. Since the connecting rod is steel (well, steel-clad unobtanium), an obvious approach presents itself.

This is a self-contained, vacuum-proof climbing vehicle which uses a fission or fusion reactor for power, has large rubber (or some vacuum-qualified synthetic) tires, and a very powerful electromagnet on its belly. The electromagnet is powerful enough at the surface to produce a pull on the rod stronger than the weight of the crawler. Then, with high-traction tires the crawler can climb straight up. Not only that, but as the crawler gets higher the weight of the vehicle diminishes and the electromagnets can be dialed down, saving energy.

The operational difficulty with this approach is controlling the magnet pull. Too much and the magnet contacts the rod and the vehicle screeches to a halt. This may be alleviated by some clever suspension design, but since at close range the pull of a magnet follows a 1/R cubed law, this is very tricky. And, of course, if the magnet does not pull hard enough gravity will triumph once again as the crawler falls off the rod. Near the center point, magnet failure combined with suspension rebound will simply leave the crawler stranded in space, slowly receding from the rod until one or the other lobes captures it. It is, after all, not orbiting either lobe, so it will eventually reenter.

In order to minimize the magnet control problem the crawler will have to maintain a low speed even near the center. Assuming this to be on the order of 1 mph, it will take 8000 hours to cross the bridge, or 330 days. Best bring a lunch.

  • 1
    $\begingroup$ If your magnets are equally spaced on a wheel, then the forces are symmetrical, allowing free rolling. It's an interesting solution. Though the magnets will need to be particularly strong depending on the composition of the steel. Especially if it's stainless :) $\endgroup$
    – Samuel
    Commented Jul 29, 2015 at 4:44
  • 2
    $\begingroup$ This means that for modern day technology, the most feasible and cheapest (but still very expensive) method would be to just launch a spacecraft, change orbits, and re-enter the other sphere. However, technological development might be accelerated by the desire of exploring the other sphere. Reaching the Moon was not such a strong drive in itself to lead us to discovering rocketry, and only after the WWII provided us with rocket technology, did the space exploration start. On such a world, there might be a motivation for discovering rocket science even in the absence of warfare. $\endgroup$
    – vsz
    Commented Jul 29, 2015 at 6:13
  • $\begingroup$ Only after a colony is established on the other side and the amount of materials and people to be transported rises enough, would a system using the rod be feasible. For example, installing some sort of rails onto the connecting rod, so a vehicle can connect to it. If it can grab a rail from both sides, an electromagnet is much more feasible, just like with real-life maglev trains. However, this requires economies of scale. Just like in real life, a space elevator would be only cheaper in launching things into orbit if we launched a lot more things into orbit. $\endgroup$
    – vsz
    Commented Jul 29, 2015 at 6:17
  • $\begingroup$ @vsz He didn't mention a vehicle with rails because I already said it. It was a neat alternate though. $\endgroup$
    – Samuel
    Commented Jul 29, 2015 at 6:36
  • $\begingroup$ My point was not the exact vehicle type used, but that it would cost so much that it would only pay out if used for a lot of cargo. Otherwise spacecraft would be more practical for initial exploration with probes. This raises another question, and can provide for interesting stories: flying to the Moon was easier because the it has less gravity and the lander could take off again. Taking off again from an Earth-sized planet would be difficult, you would need a lot of fuel, and even more fuel to transport it... so initial landings would be one-way trips until a large enough colony is built. $\endgroup$
    – vsz
    Commented Jul 29, 2015 at 14:58

Under normal laws of physics this entire thing would have long collapsed into itself to form a spherical planet. It doesn't matter what it's made of it won't be stable.

If you want to handwave that away then it can be more interesting.

Would the humans be able to walk from one lobe to the other or would it be like climbing and then descending an enormous mountain?

No. They could not walk across it. They could possibly climb part of the way up depending on how smooth the walls of the connecting bar are. The cross section of the bar is as big as Brazil, it'll have appreciable gravity, but not enough to maintain an atmosphere. People would suffocate trying to travel the 12,742 km (7,917.5 mi) across the bar. See my answer regarding a cube world on how the atmosphere of such megastructures would behave (hint: it tries to be as spherical as possible).

If they can't simply walk - what stage of technology would they need in order to cross the bridge? How would they actually do it?

Late industrial. They would need air-tight suits and some way to climb the bar. If the material can be worked then this will be easier, because they can mount platforms to it and a track for a ratcheted vehicle. The vehicle would be required to bring the tools and food required to lay more track and make the long trek to the down side.

In response to your edit:
You're right, the connecting steel rod would have appreciable gravity. Enough to stand on the rod in the middle but not enough to hold and atmosphere. I approximated the gravity in the middle of the rod by assuming it to be an infinite line of mass, a fair assumption considering the diameter and distance from the end. In that case, for the exact center where the gravity of the attached planets is mostly cancelled, the gravity would be approximately 0.57g; half earth gravity. I used a cylinder with the cross-sectional area of Brazil and a linear mass density ($\lambda$) of $6.854575*10^{16}{{kg}\over{m}}$

$$g = {{2G\lambda}\over{r}} = {{2G(6.854575*10^{16} {{kg}\over{m}})}\over{1.646*10^6m}} = 5.558 {{m}\over{s^2}} = 0.5668g $$

As they travel away from the center the gravity from the planet below would increase until they felt like they were sliding down a very very steep slope. That is, everywhere away from the center would be downhill. The atmosphere from each planet would not even get close to the center.

  • 1
    $\begingroup$ If the planet was built by Slartibartfast himself, could it be made sufficiently hollow and with sufficient internal cross bracing to resist collapse or would anything that size collapse regardless of how clever the structure? What about if it was a mostly hollow honeycomb structure? Could centripetal force keep the two lobes from falling towards each other? $\endgroup$ Commented Jul 28, 2015 at 23:38
  • 3
    $\begingroup$ @chaslyfromUK You can't make a hollow planet structurally stable, not matter the internal structure. If you want it to have the gravity to support the humans and an atmosphere then the planets will collapse into smaller solid spheres. Which will then collapse into one larger sphere. $\endgroup$
    – Samuel
    Commented Jul 28, 2015 at 23:43
  • $\begingroup$ Why would it collapse? Wouldn't each ball be gravitationally (and internally) stable separately? Admittedly, I can't see it ever forming, but assuming it was artificially created and set in motion, what would cause the collapse? $\endgroup$
    – Bobson
    Commented Jul 29, 2015 at 19:38
  • 1
    $\begingroup$ @Bobson Because the spheres would be gravitationally drawn to each other. See WhatRoughBeast's answer for a rapidly spinning configuration, which also won't work. The connecting rod itself would still collapse into a sphere. $\endgroup$
    – Samuel
    Commented Jul 29, 2015 at 19:44
  • $\begingroup$ @Samuel - Oh, I see. I missed the obvious there. $\endgroup$
    – Bobson
    Commented Jul 29, 2015 at 20:11

Assuming that the planetary configuration was possible, and stable, most of the atmosphere would be confined to the planets due to their having most of the gravity and also centrifugal force. This would extend far off planet, into the arm, but as a tenuous, unbreathable atmosphere. This is for earth sized planets. For smaller planets the atmosphere could extend even further, beyond the typical range in a massive lobe, or tide between them. As far as climbing this dumbbell, at first it would be like trying to scale a cliff, not quite straight up and down, but close. This would continue well off the surface of the planet, past the breathable atmosphere. But as you start to near the center, gravity would transition your ascent into a steep uphill. Then your uphill battle would gradually lessen until you reached the gravity center where gravity would be light, but still present as you reached "flat ground." Then the process would reverse slowly transitioning from a nice downhill trek, to a free-fall onto the other planet. In order to do this a team would need some way to breathe in space like a spacesuit, and either elevator built in or some way to trek nearly vertical terrain over extremely long distances. You could use a tether: https://en.wikipedia.org/wiki/Space_elevator but this would require at least the technology we currently have to make it work. As far as seasons, that depends on the tilt of both planets and their eccentricity of their orbit around the sun. What would be common though if their center of orbit was lined up, would be eclipses. Practically every day or every other day the sun would be blocked by the other planet for up to several hours. But if the orbit was not lined up this way, and instead of spinning around the sun like a top, rolled like a wheel around the sun (like Uranus) seasons and day night cycles would be vastly different. One side of the planet would receive light for one half the year, and the other would be in perpetual night. The day side would be scorching hot, and the nightside would be completely frozen. Tides would never change. There would always be a MASSIVE high-tide on the side facing the other planet, and on the opposite side. The sun would raise tides a little, but there would always be a mound of water several thousand feet higher at the start of the arm.

  • $\begingroup$ Would it be possible that the planet spun on its lengthwise axis as well as like a wheel - thus providing day and night? Or would the two types of rotation be be physically incompatible? $\endgroup$ Commented Jul 28, 2015 at 23:31
  • $\begingroup$ It is possible, but I have yet to see any examples of this in our solar system beyond asteroids and other very small bodies. Perhaps a physics law unknown to me acts to stabilize the rotation onto one axis. $\endgroup$ Commented Jul 29, 2015 at 6:21
  • $\begingroup$ You have mentioned the gravity of the arm. I've edited the question to highlight the importance of that. Thanks. $\endgroup$ Commented Jul 29, 2015 at 9:07
  • $\begingroup$ Thanks! And your question about a planet sized object rotating on two axis (or axies?) I will ask as a separate question. I have always wondered that myself. $\endgroup$ Commented Jul 29, 2015 at 19:22
  • $\begingroup$ Without spending a lot of time on it, I doubt you'd get a useful result. If the object is spinning on its long axis it will be gyroscopically stabilized, and in order to get it to spin around the center would require a net torque on the long axis (which will, of course, provide rotation at 90 degrees to the applied force). $\endgroup$ Commented Jul 29, 2015 at 21:03

I would imagine more of an hourglass shape would be realistic, but we can assume there is a scalable gradient and not simply a shear, perpendicular intersection. The Hike would become easier as you ascend. Gravity should be neutral at the exact center of the bar, except for what is produced by the bar itself.

As such, there would be considerably less atmospheric pressure. We could assume the formation of the bar could have evolved from two moons colliding, or some other aggregate of material caught in the Lagrange point. It would probably be not very dense but we can go ahead and assume the middle third of the the journey is at roughly 1/6 g. I'm not sure exactly what effect this has on atmosphere. I know that if you lessen the pressure of a breathing mixture, you have to increase the percentage of oxygen to keep form asphyxiating. I also know that at low enough pressure, your own body's moisture will boil away. But I can imagine a situation where the low pressure at the bar 'draws' the atmosphere from the bells. This might inflate the atmosphere enough to allow passage. It's hard for me to find exact numbers for this - I think living at .25 bar would be impossible. I think .33 is do-able. But you have to account for the less strain the weaker gravity is going to have on your system. You might not need a pressure tight suit if it's above a 10th bar, but again it's hard to find sources for this. It might be possible to live and breath in a 'meditative' state similar to freedivers here on earth, except to do so might prohibit strenuous activities, such as WALKING. A mere re-breather might not go the distance. Perhaps an air-compressor (a hand pump might suffice) and an oxygen tank. those are fairly recent inventions. Perhaps a really really long hose (which would be heavy). You only need that for that stretch in the middle. You probably need to wait until someone invents a something akin to a stillsuit.

It would appear the distance across the bar is roughly proportional to the diameter of one of the bells, ~ 13,000km. I once hiked nearly continuously for a tenth that distance in the mountains of Wyoming and Montana with one of these, one of these, and averaged 40 km a day (if you don't count the days I slacked off completely). During that time I came upon two grocers and was able to keep well stocked with food. I also ate about a lb of foraged flora (fireweed/glacier lily/salsify/pennygrass) a day, but I was bing conservative in my harvesting. I estimate I could go 3 weeks before previsions ran out, provided I could always find fresh sources of water. I can imagine someone who was a more serious trekker could average 50 km a day and we might alleviate nutritional requirements if they where more active and better at foraging. But traveling this way often requires a couple of days to stop and stockpile, maybe one day out of every 10 or 15. If we're talking about an established trail, there could be catches of food intended to help those attempting the journey. Native Americans used to plant food along trails to appease travelers between settlements. I imagine something similar would occur at the bar. If so, your trekker might actually pack very little and 100km a day would be possible.

If you ignore the disparages in gravity, the trip might take as many as 260 days. We should cut out a third of this and reduce it by a sixth, a very rough approximate reflecting the ease of traversing the middle section. The latter half of the journey would be downhill and your pack would be lighter, so we might half the number of days for this leg, but than we'd probably go ahead and double the number of days for the first leg for the exact opposite reasons, so that balances out. That brings us to 180 days of travel, and maybe 10 days of rest and food collection/preparation. 190 days as a rough reckoning, if they travel with a full pack like me. Less than a hundred if we have a truly wily coyote who is packing light. around 60 days if I use the figures supplied by these folks. This sorta sounds do-able at even a very primitive stage, provided you can breath the whole way. You might have to wait for the industrial revolution or early enlightenment era otherwise. I image the tantalizing prospect of crossing the bar, though, would push humans to develop the means much earlier in history. Glider/balloon assistance might actually be a perfectly feasible means by which one would accomplish the feat in much less time. Again provided we can fathom a reasonably dense intermural atmosphere.

  • $\begingroup$ *instead of a stillsuit as Herbert depicted, conserving water, our planet's stillsuit would use the body's locomotion to draw in and compress air. That would be a fairly late-stage Tech . $\endgroup$ Commented Jul 29, 2015 at 14:46
  • $\begingroup$ I like the idea of vegetation growing on the bar. I can imagine a whole ecosystem developing. $\endgroup$ Commented Jul 29, 2015 at 16:42

They can't walk across it because it would be vertical with relation to gravity. (despite the fact that gravity is significantly lower on that side of the planet.)

Maybe they could mine into it and build a spiral stair case up a shaft. Once they reach the middle they would be in a weightless environment because they are between two planets. The can then continue mining down the other side.

I don't believe that atmosphere would be a huge problem because the idea of a tidally locked planet is that the atmosphere bridges across. I'm not sure how to estimate the drop in atmospheric pressure.

Ascending this staircase would be fairly easy going because gravity keeps reducing the higher you get.


There are several factors which will make this arrangement unlikely and have interesting effects.

First off, the two planets need to be orbiting around each other and be at a distance greater than the Roche limit, otherwise the planets will destroy each other in an orgy of mutual gravitational attraction (technically, the tidal forces will tear the planets apart). A simple equation to derive the limit is: Roche Limit = 2.4 x (Radius of Larger Object) x (Density of the Larger Object/Density of the Smaller Object)^1/3.

Assuming the two planets are whirling around each other very close but beyond the Roche limit, we will see the two planets should be tidally locked, so the same sides face each other. There will be a noticeable lessening of gravity when the other planet is directly overhead, but not enough to allow you to jump to the other planet or anything silly like that. A compression tower may be possible, as gravitational force will lessen as the tower approaches the centre of rotation, but then the difficulty will be the building material will be pulled towards the other planet, putting tension on the structure until the other end is grounded. The engineering of this would be a very tricky piece of work. On the other hand, a relatively small rocket should be able to blast off from one planet and coast to the other with a great deal less energy that what we need to get to orbit. (Note the rocket will hit the other planet at the same speed needed to take off, unless you use some sort of system to shed the excess velocity).

Robert L Forward wrote a SF novel called "Flight of the Dragonfly" which pushed the idea to the limit (the planets were close enough that they had deformed over the ages into an egg shape) and at the climax of the novel, a sort of tidal alignment with the two planets and their Sun triggered an event where water was able to "flow" over the Roche limit of one planet onto the other. I'm fairly certain Forward would have been able to calculate if this was possible or not, but suggest that this is a very extreme case, and I think that over the aeons it would become unstable and cause the planets to either crash into each other or fly apart.

  • $\begingroup$ You say they are tidally locked. I'm assuming they are also structurally locked. I presume if they weren't then the tides (and also continental drift) would wear away the base of the bridge pretty quickly. What do you think? $\endgroup$ Commented Jul 29, 2015 at 10:15
  • $\begingroup$ Although they are orbiting bodies and would be tidally locked, other forces such as perturbations of the Sun's gravity and any giant planets will perturb the system slightly. A rigid bridge would collapse as the planets shifted slightly during their orbits around each other and the Sun. Even a tension cable would need to be actively managed. $\endgroup$
    – Thucydides
    Commented Jul 29, 2015 at 23:33

According to physics, you won't float off the bar at any point in the journey, and if you were worried about that, you're walking on a giant steel bar. Bring same magnets with you :) In fact, you might end up wanting to use them to help you speed up the journey depending on the level of tech the humans have.

Since you mentioned these are Earth-like planets, let's take a look at Earth's gravity well.

In order to escape Earth's gravity well at 1 MPH, you would need to be 26,000 astronomical units (1 AU == the distance between the Earth and the Sun) away.

So if they are Earth-like planets, they're more than close enough to have overlapping gravity wells.

In regards to the more nuanced problems like you know, surviving, it would be analagous to any high-altitude mountaineering expedition:

  1. Bring food and water
  2. Bring sufficient garments, gear, and shelter to protect yourself from killing cold and oxygen deprivation

Bonus points for packing some lead-lined underoos to make sure the people who make it to the other side can still reproduce after being bombarded by cosmic rays and other fun stuff :)


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .