Suppose we have a distance cutoff $r_0$.

Let's say that any two stars in the galaxy with a relative distance below $r_0$ have a path between them.

Would the path-connected Milky Way look like a Web in which every star is potentially connected to all of its neighbors or would we form a type of branched Tree structure with junctions and bottlenecks?

I understand that for a high enough $r_0$, it will always look like a web. But my doubt is whether the standard-deviation of interstellar distances in the Milky way is such that there exists a value of $r_0$ for which the topology is unequivocally a Tree.


The process where you link random points within distance $r_0$ into a graph is known as continuum percolation. As $r_0$ increases from 0 at first there are just isolated clusters (binaries, randomly very close stars) that gradually merge. At a critical distance of a few parsecs these clusters mostly merge into one major galaxy-spanning cluster, and above that distance this cluster just gets more densely connected. The big network at the critical range has a scale-free structure of local clusters, bridges, and trees: neither a proper web nor a tree.

Some illustrations from my book manuscript:

enter image description here enter image description here

It should be noted that these pictures assume a constant star density. In reality the stellar density varies, so for a given $r_0$ there is a part of the disk and bulge that are connected and parts of the outer galaxy and halo that are not connected.

enter image description here

  • $\begingroup$ Very interesting! I did not know graph theory was used like this. What is it then useful for in understanding these systems/images? Also, the last question of the OP, "But my doubt is whether the standard-deviation of interstellar distances in the Milky way is such that there exists a value of r0 for which the topology is unequivocally a Tree." seems to perhaps require some graph theoretic techniques to answer? $\endgroup$ – Daddy Kropotkin May 20 at 23:07
  • $\begingroup$ @DaddyKropotkin - The main application of continuum percolation is in nucleation of crystals, but it shows up in many domains. I have seen people apply it to the cosmic web, trying to analyse its topological properties by turning it into a graph. As for tree structure, I doubt it happens since it would require a very particular distance distribution to avoid creating self-loops. When looking at actual star distance data it is always clusters merging to a web. $\endgroup$ – Anders Sandberg May 21 at 1:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.