$SO(3)$ group is defined as linear transformations that preserve euclidean norm. Now if I identify $\phi$ (angle in $xy$ plane whose axis of rotation is $z$-axis) $0$ to $2\pi\beta$ where $\beta$ can vary from $0$ to $1$. What happens to our spherical symmetry? Will $SO(3)$ group still preserve the euclidean norm?

According to me as soon as I picked the $xy$ plane ($z$-axis of rotation) the spherical symmetry is broken symmetry due to both polar angle as well as azimuthal angle is lost. Is my reasoning right?

Edit: My terminology might be non-standard so what I meant by angle defect(or deficit) is $2\pi(1-\beta)$

like this

In my case, it will look like


with the open section stuck together. Also since all the line element for above case is $$ds^2=d\theta ^2+\beta^2\sin^2(\theta)d\phi^2$$. It's killing vector, which forms the generator of $SO(3)$ will be same as that of $S^2$ with $\phi$ replaced by $\beta\phi$ i.e.

$$K_1=\cos(\beta\phi)\frac{\partial}{\partial \theta}-\cot\theta\sin(\beta \phi)\frac{\partial}{\partial(\beta\phi)}$$ $$K_2=-\sin(\beta\phi)\frac{\partial}{\partial(\theta)}-\cot(\theta)\cos(\beta\phi)\frac{\partial}{\partial(\beta\phi)}$$ $$K_3=\frac{1}{\beta}\frac{\partial}{\partial \phi}$$

And oddly enough they do satisy their usual commutation relation $$[K_i,K_j]=\epsilon_{ijk}K_k$$ So there seems to be no issue with spherical symmetry since commutation relation is smoking gun to fire away any issue of coordinate representation. But clearly looking at the would-be distorted image of $\beta$ sphere is a clear no to spherical symmetry? What is going on? Is our intuition of spherical symmetry failing in the above case?

  • $\begingroup$ My terminology may be non-standard so to remove any ambiguity I have added as much info as possible in the question. $\endgroup$
    – aitfel
    Jul 15, 2020 at 16:35
  • $\begingroup$ The symmetry of a prolate spheroid (Aussie football) is not SO(3). $\endgroup$ Jul 15, 2020 at 17:56

1 Answer 1


I reckon that there is a gap in your argument. The Killing vectors of S² are correct, but two questions:

  • These vectors, in this representation, are they valid to the coordinates that you chose? Change the domain is change the system of coordinates. The step φ-> βφ is not so clear to me.

  • I don't know if this object is a 2-sphere. You cut a part. Even if the topology is the same, Killing vectors in this context are related with geometrical symmetries.

Summarizing, I believe that the problem is how you applied the Killing vectors to study your manifold. I would guess that changing to the correct Killing vectors, you would find that the relations change near the missing slice.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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