An inversion formula for the Radon transform on trees (Q1106421)

An inversion formula for the Radon transform on trees is proved. For that Schwartz spaces on trees are introduced, and a Paley Wiener type theorem for the spherical Fourier transform is proved. Moreover a Plancherel formula for the Radon transform is giv \(\sigma (X)=X\}\), \({\mathfrak q}=\{X\in {\mathfrak g}|\) \(\sigma (X)=-X\}\). Let G denote the adjoint group of \({\mathfrak g}\) and H the analytic subgroup with Lie algebra \({\mathfrak h}\). H acts on \({\mathfrak q}\). By D(\({\mathfrak q})\) we denote the algebra of constant coefficient differential operators on \({\mathfrak q}\) and by D(\({\mathfrak q})^ H\) the H-invariants in D(\({\mathfrak q})\). Given \(X\in {\mathfrak q}\), call \({\mathcal O}_ X\) the H-orbit of X in \({\mathfrak q}\). \({\mathcal O}_ X\) is a regularly embedded subvariety of \({\mathfrak q}\). \({\mathcal O}_ X\) is closed in \({\mathfrak q}\) if and only if X is a semi-simple element of \({\mathfrak g}\). \(X\in {\mathfrak q}\) is called regular if the dimension of \({\mathcal O}_ X\) is maximal. One of the main results of the author is the following. If \(X\in {\mathfrak q}\) is a regular semi-simple element, then (a) \({\mathcal O}_ X\) admits an H-invariant measure \(\mu_ X\) (unique up to scalars). (b) \(\mu_ X\) is in a natural way a tempered Radon measure on \({\mathfrak q}.\) The author presents a new proof, suggested by M. Duflo. These results were earlier proved, with a different method, by G. van Dijk and the author. The Fourier transform of such a measure \(\mu_ X\) is an H- invariant eigendistribution for D(\({\mathfrak q})^ H\). It coincides with an analytic function, say \(J_ X\), on the set of regular semi-simple elements of \({\mathfrak q}.\) The author determines the asymptotics of \(J_ X\), applying the method of the stationary phase. By specializing to the `group case', where \({\mathfrak g}={\mathfrak g}_ 1\times {\mathfrak g}_ 1\), \({\mathfrak g}_ 1\) a real semi- simple Lie algebra, and \(\sigma\) is defined by \(\sigma (X,Y)=(Y,X)\), the author obtains a very nice byproduct. If one assumes in addition that \(G_ 1\) (the adjoint group of \({\mathfrak g}_ 1)\) admits a discrete series, a new proof of Rossmann's formula is obtained.