Gravity inside a spherical shell in non-Euclidean space

It is well-known that the gravity from an isotropic spherical shell of matter in flat 3-space cancels throughout the interior volume, so an object stationary at some point in this volume is in free-fall and does not tend to "fall" either to the center of the shell's enclosed volume or to the nearest part of the shell surface.

What I am curious about is the limit for when these conditions hold. In particular, is it dependent on the force being inverse-square? If so, it should fail for, say, a four-dimensional analogue, or if the space is not Euclidean. Gravity would fall off as an inverse-cube in the former case, and in the latter case either slower than inverse square (with focusing back to a high intensity) or exponentially quickly (for a hyperbolic space).

In particular, what is the behavior of gravity in a spherical shell in a hyperbolic 3-space?

To ensure that this is well-defined, let's say that the hyperbolic 3-space is perfect vacuum except that at all of, and only, the points reached by starting at a given center-point and traveling along a geodesic for between a and b metres, where 0 < a < b; on this set of points the mass density is, say, that of silicate rock.

It's probably safe to say that gravity still cancels at the exact center of the sphere. The question is what happens at a point, say, halfway out to the shell, say along one of those geodesics at a distance of a/2.

Even for the Newtonian inverse-square gravity it isn't clear to me what will happen. The system is axially symmetric about a line throuh the center point and the a/2 point, so it seems likely the gravity is either null, directed straight in to the center, or directed straight out to the nearest point on the surface.

As for which it is, I've tried figuring it out. The tricky thing is that it's not clear how to translate between coordinate systems in such a space. The only coordinate system I can easily give it is a spherical polar one. Centering this about the sphere's center and assuming the point of interest to be halfway to the south pole leads to the question of just how wide a "circle of constant latitude" is at a given latitude, for example. Centering it about the point of interest leads to trying to figure out the sphere's translation to the coordinate system; or at least, how far the sphere surface is from the point in any given direction.

There may even be multiple tangential geodesics at a single point; I'm not completely certain. Certainly the lack of unique parallel lines renders anything as simple as an xyz Cartesian coordinate system moot.

Last but not least, there's no guarantee that the Newtonian answer will be the correct one in a curved space. So if anyone knows what GR has to say about such a shell in a hyperbolic 3-volume (of constant negative curvature) I'd like to know. (Assume the shell has no tension or pressure inside, only mass-density; a "dust" solution. A spherical shell of dust in an otherwise-empty expanding universe with no cosmological term, say; that would have a negative curvature, since it's approximately empty and therefore well below the critical density omega.)

Failing that, I'd at least like to know what happens to normally-inverse-square forces in spherical shells in non-Euclidean spaces, or higher-dimensional ones.
.

Relevant Pages

  • Re: Gravity: an alternative to "dark energy".
    ... and behave according the the usual laws of gravity, to wit, the ... net gravitational force on a body located inside a uniform ... you are closest to the spherical shell of matter surrounding the ... Most of the universe is not at the exact center, ...
    (sci.astro)
  • Re: Gravity: an alternative to "dark energy".
    ... Only if you are at the exact center. ... you will be influenced by more gravity from where you ... are closest to the spherical shell of matter surrounding the universe. ...
    (sci.astro)
  • Re: A newcomer here
    ... >> gravity, not greater gravity. ... > inverse square force and a spherical shell has the same field, ... > the outside, as a point mass of the same mass located at its center, ... > increasing surface gravity. ...
    (rec.arts.sf.composition)
  • Re: Gravity inside a spherical shell in non-Euclidean space
    ... or to the nearest part of the shell surface. ... Gravity would fall off as an inverse-cube in the former ... Even for the Newtonian inverse-square gravity it isn't clear to me what ... The only coordinate system I can easily give it is a spherical ...
    (sci.physics)