A note on the C. Neumann problem (Q1207187)

The motion of a point on the sphere \(S^{n-1}\) under the influence of a quadric potential \(U(x)=\langle Ax,x\rangle\), \(x\in\mathbb{R}^ n\), \(A=\text{diag}(a_ 1,\dots,a_ n)\) is called the C. Neumann problem. It is well-known that this system is completely integrable when \(A\) has all entries distinct. In this case \(n-1\) independent integrals in involution may be given by means of a family of confocal quadrics [\textit{J. Moser}, Prog. Math. 8, 233-290 (1980; Zbl 0468.58011)] or Ratiu's method in which the problem is fitted into the framework of Euler-Poisson equation on a coadjoint orbit of Lie algebras [\textit{T. Ratiu}, Trans. Am. Math. Soc. 264, 321-329 (1981; Zbl 0475.58006)] and others [\textit{M. Adler} and \textit{P. van Moerbeke}, Adv. Math. 38, 267-317 (1980; Zbl 0455.58017) and \textit{R. L. Devaney}, Am. J. Math. 100, 631-642 (1978; Zbl 0406.58019)]. In the case where \(A\) has multiple eigenvalues the integrals obtained by Ratiu are also in involution but not independent. In this paper we give \(n-1\) independent integrals in involution for the case where \(A\) has multiple eigenvalues.