Imagine the following star system: The star is very much like ours. Orbiting this star is one planet, very much like Earth. Let's call it the Orb. Orbiting the Orb are two moons: a large one, Mona, that is very much like ours and a small one, Wick, that's big enough to be round but small enough to not cause havoc on the Orb with it's rather close orbit. There is nothing else in the system, not even a mote of dust, and the system itself is in a very quiet neighborhood.

An Orb day, specifically a sidereal day, is very close to twenty-four Earth hours long but unlike an Earth day is always exactly the same length, every day. An Orb year, specifically a tropical year, is exactly 364 Orb days long. An Orb month, the length of Mona's phase cycle, is exactly one-twelfth of an Orb year (i.e., there are exactly twelve months in a year.) An Orb week, the length of Wick's phase cycle, is exactly 7 Orb days long, meaning there are exactly 52 weeks in a year.

Now obviously this system can't maintain this level of precision indefinitely. And in fact the system isn't natural, but was created and maintained by outside entities. For story reasons, these entities are to be permanently removed and nature is to take its course. My question is how quickly and how catastrophically would the system as described destabilize. My first guess is that Wick, in a 3:13 resonance with Mona, would start moving pretty fast, but I don't know which way or how soon.

Also: Assume Wick is sturdy enough that its Roche limit is within its orbit. If this simply isn't possible, although I doubt it isn't, please say so.

  • 1
    $\begingroup$ "Unlike an Earth day is always exactly the same length, every day": To a first order approximation his simply means that the Orb's orbit is a circle and Orb's axis has zero inclination. Strictly speaking this "always the same length" is not possible, because the tidal forces exerted on Orb by its primary and its satellites will slow down its rotation continuously, so every day will be a tiny weeny little bit longer than the previous day. For example, the our second is defined so that the mean solar day was exactly 24 hours on 1-Jan-1900; today it's about 24 hours plus 2 milliseconds. $\endgroup$
    – AlexP
    Sep 15, 2017 at 16:50
  • 2
    $\begingroup$ 3 body systems are not solved, we would have no way to know the stability of your specific 3 body system (ignore the star for now) without exact numerical simulations. If all 3 bodies become resonant and stable, then they will remain that way for a long time until the angular momentum loss from tidal forces moves them away from the resonant stability. $\endgroup$ Sep 15, 2017 at 17:00
  • $\begingroup$ @AlexP, I'm well aware of that, as are the entities in charge of maintenance. They're a little obsessed with harmony and precision, so they spin up the Orb ever so slightly. Or rather, they did, before they were removed. Also, I would like there to be axial seasons, which in my understanding only change the day:night ratio over the course of the year. Would the axial tilt cause its own problems once the stabilizing entities are removed? $\endgroup$
    – No Name
    Sep 15, 2017 at 17:23
  • $\begingroup$ If there is an axial tilt then the Sun will move across the sky on a circle inclined with respect to the celestial equator and the length of the solar day will vary during the year according to the equation of time. $\endgroup$
    – AlexP
    Sep 15, 2017 at 17:43
  • 1
    $\begingroup$ I am going to comment on what can change the rotation of earth. Large events, such as large earthquakes and tsunamis, have been shown to measurably change the length of an earth day. It is a very small change indeed, but has been documented. The change is analogous the the rate of spin of an ice skater who puts her arms in towards herself (spinning faster.) Sufficiently large events on Earth can change the mass distribution of the planet and thereby the spin and day length. $\endgroup$
    – SFWriter
    Sep 15, 2017 at 19:09

1 Answer 1


If Wick is big enough to be round, then it would have to be outside Roche's limit -- about 40K km for earth, or about 12% of the orbital radius of Luna. 1/8 the distance gives about 1/16 the orbital period. At 1/8 the distance it needs 1/8 the diameter of Luna to have the same apparent diameter. However a moon of 1/4 apparent diameter would still look like a disk. Luna is 2000 miles. 1/8 is already down to 250 miles, which is marginal for self rounding. This gives it 1/64 the mass of luna, so that it is small compared to Luna.

Now you have a set of possible values, you need to tweak the orbital periods to get something that is stable.

I have played with this on various simulators. To date I have been unable to give a planet two moons that are large enough to present disks to the surface that is stable.

In general multiple moons of significant mass and close spacing are a Bad Thing (tm)

Maybe your Powers are still involved...

  • $\begingroup$ Drat. Are we talking catastrophe on the order of days, months or years? Also, do you have cheap (read: free) simulators I could try to fine-tune my measurements with? It might help me fix the question. $\endgroup$
    – No Name
    Sep 22, 2017 at 3:58
  • $\begingroup$ Most of my systems failed within a few orbits, either with one crashing into the primary, or getting hurled into the outder darkness. I'm not convinced that the simulator wasn't at fault however. You can check the accuracy of a simulator by running it with known working planets. E.g. Mars with Deimos and Phobos. Instability can be rewarding. Larry Niven found that the Ringworld wasn't stable, and was able to get 3 sequels to the book as a result. Go further: Make it obviously artificial, and leave open whether it's stable. Later you can invoke aliens, ancient machinery, or gods to fix. $\endgroup$ Sep 22, 2017 at 19:03

You must log in to answer this question.

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