Suppose you have a stationary circle, and inside that circle is a second smaller circle, touching the larger circle from within. Pick an arbitrary point on the inner circle, call it . Imagine pushing the inner circle against the outer circle, so that the inner circle revolves around its own center, as it moves inside the larger circle. What path does describe?
If you saw my earlier blog post on cats on ladders, you would recognize it to be an astroid!
Try setting in the slider to see what happens when the radius of the inner circle is a third of the larger circle as well, although the rounding error becomes more visible. (It’s a deltoid)
Now this is the interesting part: what happens if the diameter of the inner circle is exactly the radius of the larger circle? Make a guess!
Here is a more detailed version, emphasizing that the angles at all times.
Did you guess that it would be a straight line? The fact that the path described is a straight line is sometimes called Copernicus’s theorem. However, mechanical devices exploiting this fact have been made long before Copernicus: see Tusi couple.
In general, shapes traced by small circles rolling inside of larger ones are called hypocycloids. For a nice exposition of this, as well as some related phenomena (some quite a bit deeper), see this post on John Baez’s blog. For more problems like this, see the book Lines and Curves.