“…Nevertheless, if every orbit in the compact G-set Clos(U ) has the same orbit type, then there exists a uniform injectivity radius of orbits in Clos(U ). Therefore, the author add a technical assumption in [32] that M \ M reg is a submanifold with dimension no more than n − 2, which guarantees the closure of the small G-annulus around M \ M reg is contained in M reg , and thus the interpolation arguments in [26, ] would carry over with exp ⊥ G•p in place of exp p (see [32, Theorem 5.5, Appendix B]). Inspired by the work of Li-Zhou [19, Appendix B], we can now overcome this technical problem without any further assumptions on M \ M reg .…”