The earlier MO question, "[Length spectrum of spheres][1]," asked if the length spectrum of closed
geodesics determines the metric on $S^2$, and the answer was a clear **No** due to [Zoll surfaces][2],
all of whose geodesics are simple, closed, and of the same length.  This made me wonder about the
same question for metrics for surfaces of revolution, and I found a paper by
Steve Zelditch,
"Inverse Spectral Problem for Surfaces of Revolution"
[arXiv:math-ph/0002012v1][3], which seems to answer the question **Yes**:

> Thus, if $(S^2,g)$ and $(S^2,h)$ are isospectral surfaces of revolution in $\cal{R}^*$,
then $g$ is isometric to $h$.

The class $\cal{R}^*$ are rotationally invariant metrics of "simple type" and which
"satisfy some generic non-degeneracy conditions."
However, there are Zoll surfaces which are analytic surfaces of revolution, as explicitly stated,
e.g., in Dan Jane's "The Ricci flow does not preserve the set of Zoll metrics" 
[arXiv:0809.2722v1][4].
So I've hit an apparent contradiction.

I don't fully understand Zelditch's conditions, and perhaps the contradiction disappears
in that his conditions exclude Zoll surfaces.  Or I may have some other elementary misunderstanding.
If anyone can clarify the situation for me, I would be grateful.  Thanks!


  [1]: http://mathoverflow.net/questions/89882/
  [2]: http://mathoverflow.net/questions/28622
  [3]: http://arxiv.org/abs/math-ph/0002012
  [4]: http://arxiv.org/abs/0809.2722v1