For any of you who haven't seen it, here is a stereo image of Ultima Thule, alternating between two images which help if you don’t have 3D Anaglyph glasses. The interesting thing about it is that there is no sign that the two lobes are more than slightly squashed by touching each other. They seem to have formed like this, in a gentle process, as separate objects that gradually grew by gravitational attraction of smaller pieces. Then eventually, the system lost energy probably due to influence of some third moon, and as they came together gently touched each other.

With comets and asteroids such as the two lobe 67p which the European Rosetta mission studied, then it's hard to know if they were always like that or got that shape by collisions or by outgassing of ice and gases. But this is a pristine object that has never been near the sun and has no signs of a collision.

Anyway, the answer is, yes, a larger object such as a planet can also be approximately that snowman shape, but it would have to be spinning fast enough to be in gravitational equilibrium. Ultima Thule isn't quite in equilibrium - it is spinning four or five times too slowly for that. So, let's look at this a bit more closely.

Dubbed “the snowman”

Pluto and Charon are not far off this, two round objects that rotate around their common barycenter, and keep the same face towards each other all the time

Click to watch on YouTube

Bring them closer together, which could happen if there was some other object in the Pluto system, such as a large extra moon, to take away their angular momentum, and you can have them almost touching. This is gravitationally stable, to have two planets or dwarf planets touching each other. As they spin up, Charon would orbit Pluto fast enough to keep separate from Charon, and Pluto would spin up as well to keep facing Charon.

They can even touch and overlap. The result is still stable so long as it spins fast enough.

If you look theoretically at faster and faster spinning planets, first they become flattened at the poles like Earth, broader at its equator. Then they become rugby ball shaped like Haumea

Then if it spins fast enough, it will split into two or more lobes like this (a single object is no longer stable at that speed):

(Image NASA) 'Double Earths' Could Be Fun Exoplanets To Hunt For -- If They Exist see: Can binary terrestrial planets exist?

In theory anyway, a planet or star can turn into three or even four spheres joined together to make an "overcontact ternary" or "overcontact quarternary" planet or star, though so far we haven't found anything like this:

See Viewpoint: The many shapes of spinning drops

As you can see it can even in principle be donut shaped with a hole in the middle when spinning so fast it is on the point of tearing itself apart.

There are many other possibilities not shown in that diagram, including the rounded triangles, squares and pentagons, cross shaped, and also three, four or five bodies in contact with each other in a circle with hole in the middle, like this.

See page 494 of this paper.

Those would be very unstable, slight change in rotation speed and they can no longer keep that shape. But they are theoretically possible.

However a double planet is definitely possible. I think Pluto and all the ice dwarfs and Ceres should be called planets myself, and define the term as anything rounded under gravity. If you take that broader definition of a planet we might well find a large Pluto sized or larger double planet as we explore the Kuiper belt and Oort cloud.

Ultima Thule itself is spinning too slowly to be in gravitational equilibrium. 15 or 16 hours, should be more like 3 hours (depending on the density).

It seems to have been formed by two objects that slowly accreted together and then gradually they spiraled in together and then bumped together at a few miles an hour. Meanwhile there was some other third object, or several objects, a moon or moons, that interacted and took away the spin so that the resulting double object gradually slowed down. Meanwhile the moon probably spiraled outwards. It may still be there, or it may have been lost to the system.

The New Horizon team are searching to see if they can find this object, or objects. They think there is a chance they may find it.

Meanwhile because it isn’t in gravitational equilibrium, then the neck is actually like a valley, objects would roll down into it, which may be why it is whiter than the rest of the body.

Here is a gravity gradient diagram:

Correction: I was misreading the plot. The combination of gravity and centripetal forces on #MU69 would actually pull the material *into* the trench where the spheres meet. Shown here is a rough model of what direction "down" would seem to be around the object (figure is my own). pic.twitter.com/ftkqAzjCvh

— Max Fagin (@MaxFagin) January 3, 2019

Notice, that the top of the snowman’s “head” actually is spinning fast enough so that any loose material there should have been thrown off into space.

Yes. In fact, that's probably why M2 is the size it is, and no bigger. Any more material that accreted on the tip of M2 would have fallen off. If you walked from the neck to the tip of M2, the ground would seem to slope more and more downward until you would just fall into space. pic.twitter.com/QkfsgWBTRY

— Max Fagin (@MaxFagin) January 3, 2019

This is the anaglyph version, needs those red-green specs.

And this is one you can see in 3D if you look at it cross eyed - hold a finger in front of the image and when it merges into one then the two images will join together

This is their press conference

Some of the images from: Finally, New Horizons' First Photos of Snowman-Shaped Ultima Thule

So - if you imagine the same process as formed Ultima Thule operating at a much larger scale - why not have one that is in gravitational equilibrium?

As with Ultima Thule then there would be some third object that would take away the angular momentum of the system in order to get them close enough together to touch. Then the main problem would be that this third object would gradually slow them down even further meanwhile spinning out further itself. As it did that then the double lobe shape would no longer be stable and it would flow into a shape more like Haumea, ellipsoid in shape.

So - it would need that third moon to be lost from the system early on, which maybe it could be if it was far enough away initially. Either that or the process is so slow that the two main objects touched each other only recently.

Still there are large numbers of ice dwarfs out there and we already have the very ellipsoid Haumea, which is strikingly different from the inner solar system planets, dwarf planets or moons. Though there are many elongated smaller objects none of the rounded ones are as elongated as that in the inner solar system.

So, I wonder if there’s a decent chance that some day we find a two lobed double dwarf planet looking just like Ultima Thule but more rounded under gravity?

Maybe even larger Roche-world planets as large as Mars out there, or even Earth sized ones or larger?

Here is Mike Brown's video where he predicts using a statistical argument from the orbit of Sedna, that we are likely to find planets beyond Neptune that are Mercury, Mars or even Earth sized. This is a transcript, he describes Sedna’s orbit then goes on to say:

"What is strange about this one, we happened to find this one almost at the closest point it ever gets to the sun. Not by coincidence. Because there is only about a 200 period shown in red here when we could have seen it. ... So 200 years out of 12,000 years means 1 in 60 chance of finding it. So either we are very lucky, which astronomers don't like to think of themselves as lucky, or scientists in general, but instead what we like to think is that if we found something and we only had a 1 in 60 chance of finding it, probably there are 60 of them and we just found the one that happens to be close. ... Now maybe it's not 60. Maybe it's 30 and we got a little bit lucky. Maybe it's 90 and we got a little bit unlucky. But there are a lot of objects in this very distant region where we never knew of anything before. This is the first time we found anything in this region out here."

""Now the fun thing to think about is, if there are 60 of these, and Sedna is about 3/4 times the size of Pluto, if there are 60 objects 3/4 the size of Pluto there are probably, oh, 30 objects the size of Pluto, that's a lot of objects the size of Pluto. There are probably 10 objects that are twice the size of Pluto and there are probably two or three objects that are three and four and maybe even five times the size of Pluto. in this region here. It's a little bit vague, since we have only found one object, to be able to extrapolate to these things, but it is not that vague. There must be some of these very large objects out there. …

"I am willing to go out on a limb there and say, we will find something like that, the size of Mars, in this region of space ...."

He goes on to say that if / when this happens then probably the IAU will probably re-open the big debate about whether Pluto is a planet, since these Mars and Earth sized objects will most likely not be big enough to count as planets according to their definition (according to the current IAU definition objects have to be much larger to count as “planets” the further they are from the sun).

• Does 'The Goblin' Make Neptune Sized Planet 9 More Likely? ... And Are Mars Or Earth Sized 'Goblin's Possible?

Anyway - that’s a bit of an aside for this answer.

But it does seem likely there are vast numbers of objects out there rounded under gravity. Some may be very large. Given that we have Haumea already, there may be other rugby ball shaped ones.

So just to make a fun guess here, perhaps the equivalent of Kuiper belts and Oort clouds, well away from all the action in the inner star systems, are the best places to search for Roche worlds?

Incidentally, it wouldn't matter of the dwarf planets are too small to resolve. Alan Stern makes an interesting point in the press conference. They got the shape of it amazingly accurately by the occultation profile.

This is when a group of astronomers using telescopes observe the shadow of Ultima Thule as it passes between Earth and a very distant star. The shadow is the same size as Ultima Thule itself.

The white lines show the occultation observations and the blue outline is from the New Horizons photograph.

They could do the same with other distant objects. It needs lots of enthusiastic astronomers and precision timing. In this way they could also find the exact shape of any of the dwarf planets if it happens to pass in front of a distant star.

If that happens, it would be direct evidence, not just interpreting a light curve, that there are Roche world dwarf planets out there, if there are any.

They think there may well be more double lobed snowman type objects at least. So long as you see it side on rather than edge on, you get a dip as the second lobe hides behind the first one a bit like an occultation.

Perhaps some day in the future, as we find more and more dwarf planets, we might spot potential Roche world dwarf planets in the same way? And then check them out with occultation studies?

This is extending that idea of more snowman like Kuiper objects to planets. A system like the Pluto system but with a large extra moon that removes the angular momentum until they are touching? After that as it slowed down further the central double object would become an ellipsoid like Haumea but if something stopped the process before then e.g. lost the extra moon, we could end up with a Roche-world type double planet.

Perhaps the outer reaches of a star system, where planet formation is more gentle, is the best place to search for Roche-world objects?

Just a fun thought.

See also my

• So You Thought You Knew What Spinning Planets Look Like? ... Surprising Shapes Of Rapidly Spinning Planets

This is my answer to an interesting question someone asked on Quora recently.

The press conference images are here