Die Turánsche Matrix als Produkt einer Euler-Knopp- und einer Taylor- Matrix. (The Turán matrix as a product of a Euler-Knopp and a Taylor matrix) (Q920231)

Two conformally equivalent infinite series \(\sum a_ n\) and \(\sum b_ n\) (with complex terms and partial sums \(s=\{s_ n\}\), \(t=\{t_ n\}\) respectively) do not necessarily possess the same convergence-behaviour. This surprising discovery goes back to Turán. He showed in a quite subtle way that in the relation \(t=As\) the matrix A has unbounded row- norms, and hence is not a regular summability matrix. The fact that \textit{P. Turán}'s result [Acad. Serbe Sci., Publ. Inst. Math. 12, 19-26 (1958; Zbl 0084.060)] is rather hidden makes it desirable to gain more knowledge about the transformation \(t=As\). We write it in series-to-series form, i.e. in the form \(b=Ka\), denoting \(K=K_ u\) as Turán's matrix (the parameter u characterizes the underlying conformal mapping). Then we exhibit a decomposition \(K_ u=E_ pT_ u\), where \(E_ p\) is a Euler-Knopp matrix and \(T_ u\) a Taylor matrix. A crucial point is the relation between the parameters u and p. The decomposition allows more insight into the structure of \(K_ u\) and its summability properties.